Selected Abstracts from the 2018 NCRI Cancer Conference of National Cancer Research Institute

Scottish Event Campus, Glasgow, Sunday 4 – Tuesday 6 November 2018 Attendees of the 2018 NCRI Cancer Conference will be able to contact corresponding authors through the Conference App and e-poster platform at the Conference. W: https://conference.ncri.org.uk: E: conference@ncri.org.uk Organised by the National Cancer Research Institute, which is a partnership of 19 cancer research funders, the NCRI Cancer Conference provides a platform for researchers, clinicians, people affected by cancer and industry representatives to come together to discuss, present and showcase high-quality research. Sponsorship Statement: Publication of this supplement was sponsored by Roche. All content was reviewed and approved by the NCRI Scientific Committee, which held full responsibility for the abstract selection.


1
Abertay University, 2 Macmillan Cancer Support, 3 Glasgow School of Art, 4 Digital Health and Care Institute, 5 Glasgow City Council Background: Improving the Cancer Journey (ICJ) was launched in 2014 by Glasgow City Council and Macmillan Cancer Support. As part of routine service, data is collected on ICJ users including demographic and health information, results from holistic needs assessments and quality of life scores as measured by EQ-5D health status. There is also data on the number and type of referrals made and feedback from users on the overall service. By applying artificial intelligence and interactive visualization technologies to this data, we seek to improve service provision and optimize resource allocation. Method: An unsupervised machine-learning algorithm was deployed to cluster the data. The classical k-means algorithm was extended with the k-modes technique for categorical data, and the gap heuristic automatically identified the number of clusters. The resulting clusters are used to summarize complex data sets and produce three-dimensional visualizations of the data landscape. Furthermore, the traits of new ICJ clients are predicted by approximately matching their details to the nearest existing cluster center. Results: Cross-validation showed the model's effectiveness over a wide range of traits. For example, the model can predict marital status, employment status and housing type with an accuracy between 2.4 to 4.8 times greater than random selection. One of the most interesting preliminary findings is that area deprivation (measured through Scottish Index of Multiple Deprivation-SIMD) is a better predictor of an ICJ client's needs than primary diagnosis (cancer type). Conclusion: A key strength of this system is its ability to rapidly ingest new data on its own and derive new predictions from those data. This means the model can guide service provision by forecasting demand based on actual or hypothesized data. The aim is to provide intelligent person-centered recommendations. The machine-learning model described here is part of a prototype software tool currently under development for use by the cancer support community.
Disclosure: Funded by Macmillan Cancer Support Corresponding author: Kean Lee Kang 5. Young colorectal cancer patients experience referral delays in primary care leading to emergency diagnoses Chanpreet Arhi 1 , Paul Ziprin 1 , Alex Bottle 1 , Elaine Burns 1 , Paul Aylin 1 , Ara Darzi 1 1 Imperial College London, UK Background: The incidence of colorectal cancer in young patients is increasing. In this population based case-control study, we hypothesise missed opportunities for diagnosis in primary care are leading to referral delays and emergency diagnoses in young patients. Method: We compared the interval before diagnosis, presenting symptom (red-flag based on NICE guidance 2005 vs non-specific), the odds ratio (OR) of an emergency diagnosis and stage at diagnosis for those under the age of 50 compared with older patients, sourced from the cancer registry with linkage to the Clinical Practice Research Datalink database of primary care records.
Results: 7315 patients were included of which 561 (7.7%) were under 50, 1287 (17.6%) 50-59, 2458 (33.6%) 60-69 and 3534 (48.3%) 70-79 years old. Young patients were more likely to present with abdominal pain (21.1%) and were diagnosed with later stage (48.0%) and via an emergency (29.1%) compared with all age groups. They experienced a longer interval between referral to diagnosis by 12.5 days compared with those aged 60-69, reflecting the higher proportion of referrals via the nonurgent pathway (33.3% vs 26.5%). The OR of an emergency diagnosis did not differ with age if a red-flag symptom was noted at presentation but increased significantly for young patients if the symptom was non-specific compared with all age groups (OR 0.62 95% CI 0.49-0.80 for the 60-69 age group). Conclusion: Young patients are more likely to present with symptoms that fall outside the referral guidelines. Referral criteria should be tailored according to the age group to reduce the risk of an emergency diagnosis due to missed opportunities in primary care. Disclosure: Funded by Imperial College London Corresponding author: Chanpreet Arhi 6. Delays in primary care are associated with a late stage and worse prognosis for colorectal cancer patients: A population based study Chanpreet Arhi 1 , Elaine Burns 1 , Alex Bottle 1 , George Bouras 1 , Paul Aylin 1 , Paul Ziprin 1 , Ara Darzi 1 1 Imperial College London UK Background: Delays in referral for colorectal cancer patients may occur if the presenting symptom is falsely attributed to a benign condition. We hypothesis primary care delays are associated with a later stage at diagnosis and worse prognosis. Method: Patients with a non-emergency colon or rectal cancer in the cancer registry with linkage to the Clinical Practice Research Datalink were included. All had a colorectal associated symptom defined as 'red flag' (rectal bleeding, abdominal mass, change in bowel habit, diarrhoea, anaemia) or 'non-specific' (abdominal pain, weight loss, constipation, other bowel function) and a relevant referral recorded in the year leading up to diagnosis. The primary care interval (presentation to referral), hospital interval (referral to diagnosis) and stage was compared for each symptom individually and after grouping as above. Cox modelling determined the hazard ratios (HR) of death for each symptom compared with rectal bleeding, and whether delays in referral were associated with a worse prognosis. Results: 4527(63.5%) colon and 2603 (36.5%) rectal cancer patients were included. 16.9% and 13.5% presenting with redflag symptoms respectively experienced a delay of over three months before referral, compared with 35.7% and 42.9% of those with non-specific symptoms. Delays in primary care did not increase the hospital interval. Each symptom was associated with reduced survival compared with rectal bleeding for both cancers. Patients referred after three months who had presented with red-flag symptoms (Colon: HR 1.53 (1.29-1.81) Rectal: HR 1.30 (1.06 -1.60)), demonstrated a significantly worse prognosis compared with referrals within two weeks with similar symptoms. This was not seen for patients with non-specific symptoms. Delays in referral significantly increased the proportion of late stage cancers. Conclusion: The first presentation to the GP provides a referral opportunity to identify the underlying cancer, which if missed for patients presenting with red-flag symptoms, is associated with a later stage in diagnosis and worse survival. Disclosure: Funded by Imperial College London Corresponding author: Chanpreet Arhi Method: Qualitative semi-structured interviews were undertaken in 2017 with 20 patients within 6 months of receiving their first treatment for UGI cancer referred through the urgent (2 week) GP route in the North East and Cumbria. Thematic analysis was undertaken.
Results: Key themes identified were organisation of care, diagnosis, support and expectations of the NHS. Some patients were offered immediate access to testing, others required specialist approval first. Patients' were often unclear about the purpose of visits and journey ahead. Patients contrasted the 'urgency' indicated by the requirements placed upon them (and therefore often upon family and friends) to attend numerous appointments, at short notice, in close succession, often across multiple hospital sites, with the seemingly long time taken to receive results, and the unsettling time spent waiting and 'not knowing' the prognosis. Patients' were grateful for the NHS and described how staff were time-pressured but working in their best interests. They did not want to be seen as 'complaining' and were often reluctant to express concerns or suggest improvements. Conclusion: Current pathways are complex, varied, can be poorly communicated to patients and do not always appear to be patient-centred. Clearer communication is necessary but insufficient without a review of the diagnostic pathway to reduce unwarranted variation and simplify the patients' journey. Patients' gratitude for the NHS and stoic acceptance of the status quo should not be seen as justification for no change. Disclosure: Funded by NHS England Specialised Commissioning (North East England hub) 1 Brighton and Sussex Medical School, 2 Royal Sussex County Hospital Background: The ageing population is providing new challenges for oncologists. Oncology trainees are underequipped to manage geriatric oncology patients with a recent survey showing 66% never receive any formal geriatric oncology trainingGeriatric experience prior to oncology training is mainly limited to inpatient settings, whereas the majority of oncology workload is outpatient. The Rapid Access Clinic for Older Persons (RACOP) is a potential environment in which oncology trainees could experience geriatric assessment techniques.
This project explores the quantity and quality of geriatric oncology training available within the RACOP clinic with a view to implementing a novel training opportunity. Method: Clinic lists for 3 months of RACOP clinics extracted. Notes from patients who had previous, current or new cancer diagnoses made during the clinic were examined. 2 oncology trainees attended RACOP clinics and provided structured reflection on the potential training value. Results: 220 patients were identified in total. 30% had a previous, existing or new cancer diagnosis made. Of the 30 patients (14%) that had a new cancer diagnosis made from their appointment only 2 received active treatment from oncologists due to either overwhelming comorbidities or patient choice. Approximately 75 other geriatric conditions were managed during this time period. Trainees are currently attending RACOP clinic and feedback will be available by the time of the NCRI conference. Conclusion: Geriatric oncology services are being developed in a small number of centres, usually run by geriatricians. The growing numbers of older people living with cancer is likely to outstrip the supply of geriatricians if this model were to be implemented nationally. There is a drive to empower non-geriatricians to perform basic geriatric assessments.
The RACOP clinic provides an excellent resource of an unselected geriatric population and new geriatric oncology patients. This experience can up skill clinical oncology trainees to improve confidence in assessing and managing older oncology patients.
Disclosure: None declared Corresponding author: Cressida Lorimer 10. The role of dual modality imaging in the detection of residual nodal disease following radical chemoradiation for locally advanced head and neck cancer Vanita Gandhi 1 , Shelly English 2 , Anna Thompson 2   1 Royal Free NHS Trust, 2 North Middlesex University Hospital Background: Adjuvant treatment following definitive chemoradiotherapy (CRT) for node-positive head and neck squamous cell carcinoma (HNSCC) has evolved from routine neck dissection (ND), towards risk-stratification with positron emission tomography with computed tomography (PET-CT). Patients with a complete response (CR) remain on clinical follow-up, and those with residual disease reported as an incomplete response (IR) have surgery. However, there remains no consensus for those with an equivocal response (ER). The sensitivity of PET-CT is reported at 100% with specificity being only 84%. Studies have reported that combining diffusionweighted magnetic resonance imaging (DW-MRI) with PET-CT improves specificity to 94%. Our centre uses both modalities at 3 months after completing CRT to assess treatment response. We retrospectively analysed the outcomes of these patients. Selected Abstracts from the 2018 NCRI Cancer Conference of National. . . Method: Data was gathered from radiological reports and specialist multidisciplinary discussions for 43 patients diagnosed with advanced node positive HNSCC treated with definitive CRT at our centre from December 2015 to January 2018. Results: CR, IR and ER were seen in 84% (n=36); 2% (n=1); and 14% (n=6) respectively on PET-CT. All DW-MRI scans reported those with residual lymphadenopathy as reactive or post-radiation change.
The patient with a radiological IR declined a ND and had a subsequent negative node biopsy. The node remains stable at one-year clinical follow-up. Of the 6 patients with ER: one underwent a ND and 0/39 nodes contained viable tumour cells. The remaining 5 patients are all in remission at a median follow up of 13.7 months. Conclusion: These results demonstrate that combination imaging with PET-CT and DW-MRI may improve the evaluation of persistent lymphadenopathy following radical CRT. This may reduce the number of 'equivocal' responses identified and so avoid the need for salvage surgery, which is associated with additional morbidity and cost. We recommend performing a randomised control trial to examine this strategy further.
Disclosure: None declared Corresponding author: Vanita Gandhi 11. Effects of a 12-week exercise and nutrition programme on cardiovascular fitness, body composition and psychological well-being in women at all stages of the breast cancer care continuum Joe Othen-Price 1 , Matthew Thompson 1 , Yvonne Rumble 1 , Sam Olden 1 , Ivor Cradock 1   1 Clinical Prevention and Rehabilitation, London, UK Background: It is well established that breast cancer treatment can have a deleterious effect on physical function, body composition and psychological well-being. It has been proposed that lifestyle factors, such as exercise and nutrition can have an important role in offsetting these negative changes, however, exercise programme guidelines remain generic and unvalidated. We propose that a 4 Pillars® framework provides an effective model for exercise and nutrition programmes for all breast cancer patients, from diagnosis through treatment and into survivorship.
Method: 24 females, aged 50 ± 10.6 years at various stages of the breast cancer care continuum completed a 12-week 1-2-1 programme at CP+R® based on 4 Pillars; twice-weekly resistance sessions, twice-weekly cardiovascular training within a prescribed heart rate zone, daily step-count monitoring and nutritional guidance. The programme was led by a clinical exercise specialist and overseen by a clinical nurse. All patients underwent cardiorespiratory fitness (VO 2 peak) testing, body composition analysis and psychological profiling pre and post programme.
Conclusion: A 1-2-1 exercise and nutrition programme incorporating a 4 Pillar model significantly improved a range of physiological and psychological factors. The 4 Pillar model could provide a blueprint for exercise and nutrition programmes for breast cancer patients at all stages of the cancer care continuum. Disclosure: None declared Corresponding author: Joe Othen-Price University Department of Neurosurgery, Edinburgh Background: Patients with "Headache Suspicious of Cancer" should be referred quickly for evaluation, especially if there is optic disc swelling or visual field loss. Previous work shows patients with tumour headaches often have asymptomatic cognitive findings. Based on this we developed an Edinburgh Protocol Based Referral (PBR) Pathway. Method: The Edinburgh PBR for "Headache suspicion of cancer" incorporates an expedited GP open access CT (OACT) within 24-48 hours, if headache is associated with personality or cognitive change but in addition requires completion of a semantic verbal fluency test (SVFT) -a simple fast screening test of cognition (number of animals thought of in one minute). Also, if GP cannot see the discs or is uncertain about visual field defect, local optometry services will check the patient within 24-48 hours. The Edinburgh protocol is integrated into the electronic GP referral system RefHelpwith supporting information.
Results: In year 1, 20% (169/623) of all, and 74% (169/228) urgent, OACT scans came through the PBR pathway. Many forms were incompletely submitted. In forms completed correctly (including the SVFT) median time to scan was 2 days (1-7); Incomplete PBR was 9 days (3-13) and routine 13 days (9-16). 16.5% of all PBR requests were abnormal (3 had tumours: glioma 2, meningioma 1). Where SVFT was completed there was an association between abnormal CT scan and poorer SVFT (abnormal -median SVFT 13 (IQR 10-17); normal -median SVFT 19 . The Optometry pathway is being promoted and is being considered for extension across Scotland in the new Scottish Cancer Referral Guideline (2018/19). Conclusion: The Edinburgh PBR for "Headache suspicion of cancer" has been implemented. Expedited scans are being done. More work needs to be done around GP compliance with SVFT as this could be a valuable "risk factor" for secondary headache. Scottish Optometry support and promote the pathway.
Disclosure: Funded by The Brain Tumour Charity Corresponding author: Robin Grant 13. Give us the tools... Ensuring consumer involvement adds value Richard Stephens 1 , Mat Baker 1 1 National Cancer Research Institute, UK Background: The cancer research community has long encouraged consumer involvement in the design, delivery and dissemination of research, including its oversight, regulation, funding and priority-setting. Recognition of the benefits that consumers bring to the research process has prompted the development of a number of toolkits that assist researchers in recruiting and using patients, carers and members of the public in cancer research. Less well developed are approaches to ensuring that consumer involvement is relevant and effective and that it adds the value that researchers, funders, patients and carers are seeking. Method: The NCRI Consumer Forum provides consumer involvement services to the NCRI and to NCRI partners. It seeks to quality assure the contributions of its members and to ensure that these contributions are supported by training and development that enable members to operate to best effect. These collectively constitute an expert practice model (EPM) that carries a new recruit from selection and appointment to a Clinical Studies Group or other NCRI role, through induction training and developmental learning, supporting them to make a useful contribution in their individual role, and to add value to the wider contributions of the Consumer Forum as whole.
Results: Data from the Value and Impact Measures (VIM) indicate that this structured approach to the development of NCRI consumers' expertise is enhancing the value and the positive Macmillan Cancer Support, 2 NHS Lothian Background: Primary Care has an important role in supporting people following a diagnosis of cancer. The 'Macmillan Cancer Care in Primary Care: a quality toolkit' was developed to help GP practices and GP clusters improve care for people living with cancer.
Method: By using case studies, reflective practice, data collection and critical analysis, practices identified opportunities for improvements in current systems and processes as well as content of Cancer Care Reviews (CCR). Support and advice was offered from local Macmillan GPs and health board cancer leads. Practices completed baseline and follow-up questionnaires, and a quantitative and qualitative evaluation was also undertaken. Results: 251 Scottish GP practices (26% of total) from 12 health boards completed the toolkit. 190 practices (76%) completed both data collection questionnaires.
At baseline, 45% of practices had a robust system in place to routinely contact people after a diagnosis of cancer, increasing to 78% of practices following completion of the Toolkit (figure 1).
The range of topics discussed at a CCR increased following completion of the Toolkit (figure 2). More practices reported discussing the provision of cancer information (increased from 28% to 58%), the benefits of physical activity (increased from 25% to 46%), and the financial impact of a diagnosis (increased from 32% to 61%) at follow-up.
Following completion of the Toolkit, 142 GP practices (76% of total) reported that they felt better equipped to support people living with cancer by either a moderate or large extent (figure 3). Conclusion: People with cancer often consult Primary Care professionals following their diagnosis. A significant proportion of Scottish GP Practices (26%) undertook the Macmillan Quality Toolkit. Following completion there is evidence of improved systems for contacting people following a diagnosis of cancer, a broader range of issues discussed at CCR, including further information regarding cancer, exercise, and the financial impact of diagnosis.
Disclosure: Funded by Macmillan Cancer Support Corresponding author: Paul Baughan 15. A systematic review update on cost-effectiveness of colorectal cancer screening: identification of an optimal strategy in Europe Chih-Yuan Cheng 1 , Tao Ran 1 , Michael Schlander 1   1 Division of Health Economics, German Cancer Research Center (DKFZ) Background: Biennial guaiac-based faecal occult blood test (gFOBT), biennial faecal immunochemical test (FIT) and 10-yearly colonoscopy are the most commonly adopted colorectal cancer (CRC) screening strategies in Europe. Evidence regarding the effectiveness and cost-effectiveness of those strategies have been substantiated, yet there is no consensus on the most cost-effective modality. We aimed to identify a cost-effectively optimal CRC screening strategy among the three by conducting a systematic review. Method: We searched PubMed, EMBASE, NHS EED, EconLit, and three other databases for European cost-effectiveness analyses evaluating any of the three strategies, published between January 2010 and December 2017. Two researchers independently conducted the review following the standard systematic review method. Net monetary benefit approach and incremental costeffectiveness ratio (ICER) were used to assess the optimal strategy. Costs were converted to 2016 USD. Results: We identified 17 European studies. Only two studies compared all three strategies-a French study measured the outcome with life-year gained (LYG) and a German study with quality-adjusted life-year (QALY). Both found biennial FIT cost-effectively optimal at and above the willingness-to-pay (WTP) threshold of $20,000 per LYG/ QALY. If focusing on stool-based strategies, ten studies in total compared biennial FIT and gFOBT. Compared with gFOBT, FIT was cost-saving and more effective in two English and one German study, and had ICERs ranging from $2,406 to $11,310 per LYG or $478 to $2,591 per QALY in the remaining seven studies. Conclusion: More evidence comparing all three strategies in Europe is needed, although two studies pointed at biennial FIT being costeffectively optimal among the three at various WTP levels, as opposed to biennial gFOBT and 10-yearly colonoscopy. Between the two stoolbased strategies, biennial FIT was shown to be highly cost-effective (at $50,000 WTP threshold), or even cost-saving. Our findings support the trend in Europe shifting from gFOBT to FIT in CRC screening programmes. Disclosure: None declared Corresponding author: Chih-Yuan Cheng 16. The development of an exercise focused self-management app (iExhale) for lung cancer survivors to improve their symptom control Catherine Henshall 1 1 Oxford Brookes University, UK Background: Lung cancer affects over 33,000 people per year in the UK and many of these have significant smoking related comorbidities, which can negatively impact on their activities of daily living and subsequent quality of life. Common symptoms experienced include breathlessness, fatigue and depression. Exercise practices have been shown to improve symptom control in lung cancer patients and exercise based self-management mobile applications have the potential to provide tailor made, individualised care to lung cancer survivors and decrease pressure on NHS services. The aim of this study was to design and develop an exercisefocused self-management mobile application for lung cancer survivors to improve their symptoms of breathlessness, fatigue and depression. Method: The study design consisted of three stages. Stage 1 was a systematic review to examine the impact of exercise interventions in improving breathlessness, fatigue and depression in lung cancer survivors. Stage 2 consisted of qualitative focus groups with lung cancer patients, carers and health professionals to explore their views on the usefulness of exercise in improving symptom control. Findings from Stages 1 and 2 informed Stage 3 which consisted of mobile application development and a usability study. Results: Stages 1 and 2 found evidence that exercise does improve symptom control for lung cancer survivors and that exercises need to be carefully tailored to meet the preferences and needs of individual lung cancer survivors. These findings were used to develop the content and format of the mobile application, which was further modified following a usability study with lung cancer survivors. Conclusion: A working, theoretically underpinned, patient centred exercise mobile application (iExhale) has now been developed to help improve the symptom control of lung cancer survivors. Plans Background: Exercise has been demonstrated to be beneficial to men with prostate cancer (PCa). However, maintaining these exercise behaviours remains a challenge. This study examined a 12-month person-centred physical activity behaviour change pathway, delivered through a London-based NHS cancer centre, and determined changes in self-reported physical activity (PA) and self-reported quality of life (QoL) over the course of 12 months. Method: 221 men diagnosed with PCa were given the option of exercising independently/in the community; or undergoing 8-12 weeks of supervised exercise, followed by independent/community exercise. All men were given an initial 1-hour consultation, subsequently followed up via telephone at 4, 7, and 12 months. All men were offered motivational interviewing, aimed at encouraging exercise behaviour, at all time points. Results: 68% (n=150) of men remained engaged with the pathway at 12 months. This retention rate was similar across treatment groups. Overall, there were increases in self-reported PA compared with baseline, which were significant at 4 and 7, but not 12 months. The trend for increases in PA amongst men on curative treatment showed significance at 4 months compared with baseline, but not at 7 or 12 months. Men undergoing systemic treatment reported increases in PA compared with baseline which were significant at all follow up time points. There was a trend toward increases in reported QoL over time, with significant increases at 4 and 12 months compared with baseline. Men undergoing curative treatment showed the most pronounced changes in QoL. Method: PBUK's counselling service was evaluated using the validated person-centred outcome measure MYCaW (Measure Yourself Concerns and Wellbeing), measuring the impact of cancer support services on the severity of people's cancer related concerns and wellbeing. Concerns were rated at the start of counselling and re-rated at the end of the last session. Qualitative data was captured on the MYCaW tool regarding other things going on in people's life affecting health and what was important about the service they received. Data was analysed using the accompanying qualitative coding framework. Data was analysed for all clients who had attended counselling appointments in January-December 2017 and provided full prepost-counselling MYCaW data. Results: 40 clients provided pre-post MYCaW data and attended an average of 6 sessions (range 1 to 18). Psychological and emotional concerns were top rated for concern 1 and 2 (81% and 78% respectively). Prior to counselling, concerns were rated fairly severely (5.1/6 for concern 1 and 2; 6 being the worst). After counselling concerns showed a statistically significant improvement (2.2/6 and 2/6 for concern 1 and 2 respectively, p<.000 for both). 90% of clients showed a clinically significant improvement in their concerns. Qualitative data indicated that counselling at PBUK gave clients the opportunity to talk, provided time for themselves and made them feel supported and understood. Results: Barriers to implementation included getting the course up-and-running regularly, incorporating the programme into the care pathway, buy in from professionals, time, money, venue, supporting people returning to work, recruiting men, younger people and BME groups. Successes of implementing the course occurred when Macmillan staff, healthcare professionals, volunteers and management were involved in the plans from the beginning, actively referring and encouraging course retention. Cancer Information Centres were useful hubs to recruit onto the course. Patients liked the friendly neutral space and professionals found it useful to signpost patients. The most successful time to refer patients to the course was after being discharged from treatment when patients often felt isolated like "falling off of a cliff". Professionals liked having the HOPE Programme to refer patients to meet their unmet psychosocial needs. Conclusion: These findings help understand barriers further and provide examples of how some sites have successfully overcome them. Some felt that more of a nationwide recruitment strategy and advertisement scheme would be beneficial, encouraging regions to incorporate the programme into the patient's care pathway. Sharing these lessons of implementation will hopefully improve implementation consistency across the nation, increase the impact of the course and inform the implementation science field.
Disclosure and raw HNA data for 26 patients (50%). 45 TYA (88%) reported concerns: 35 (69%) reported 1-5 concerns; 2 (4%) reported 6-10 concerns; 5 (10%) reporting 11-15 concerns; 1 (2%) reported 1 concern and 2 (4%) reported >20 concerns: general appearance, physical fitness and emotional issues, were the most numerous. One third of TYA were referred for psychological support; twothirds benefitted from third sector support programmes. Conclusion: Almost 90% of TYA reported at least one concern after cancer treatment, with more than two-thirds of patients reporting five to 10 concerns, and 10% reporting up to 15 concerns. While support services, largely provided by third sector, are in place for many patients, further evaluation of the Health Needs Assessments, exploring relationships to diagnoses, and identification of gaps in services, are required to inform future developments. Background: There is a growing cohort of people who, although they cannot be cured of their cancer, are on treatments that can reduce cancer burden, alleviate symptoms and prolong life. This is a heterogeneous group with different prognoses and treatments, which could be described as living with treatable but not curable cancer (TNCC). This project aims to build a search criterion to identify the TNCC population using the England cancer registry. Method: A set of possible search criteria were developed for evaluation by 20 oncologists, haematologists and specialist nurses. Through their expertise and analysis, we will refine this into a single, comprehensive search criterion.
Results: Our first criterion was based on distant metastatic cancer at diagnosis. There were 76,000 people in England in 2015 diagnosed either at stage IV cancer, with a secondary malignancy or with certain haematological cancers. Our next criterion focused on metastatic disease developed post-diagnosis. It included 135,000 people identified in inpatient HES and 10,000 from the Cancer Waiting Times dataset.
Our next set of criteria focused on treatment. This selected 46,000 people who, in 2015, received one of 236 chemotherapy regimens thought to target TNCC; or had a second chemotherapy treatment over a year from their first treatment. It found 7,000 people who had a second round of radiotherapy after a six-month gap. Another criterion pinpointed 99,000 people who received palliative chemotherapy or radiotherapy.
In the final searches we identified 57,000 people diagnosed with an intermediate survival cancer and 78,000 with a shorterterm survival cancer. McConnell et al. (2017) hypothesised that intermediate survival cancers are often long-term conditions. New treatments will make the shorter-term survival group an important part of TNCC. Conclusion: The criteria will not cover all circumstances but, ultimately, this project will allow better recognition of people with TNCC and help appropriate services to be provided for them. These symptoms can lead to severe difficulties including limiting work, travel or socialising. To improve the care for patients with bowel problems, the EAGLE study implemented an innovative gastroenterological service in three NHS centres (two led by a specialist nurse). The purpose of the qualitative analysis was to assess the acceptability of the new service and monitor the experiences of patients and professionals. Method: Semi-structured interviews were held with healthcare professionals and patients at baseline, 6 months and 12 months. Interviews were audio-recorded and transcribed verbatim; data were analysed using framework analysis.
Results: Thirty-five healthcare professionals (Baseline n = 17; 6 months n = 13; 12 months n = 2) and 16 patients (including their companions: Baseline n=9; 6 months n = 5; 12 months n = 2) were interviewed. Four key themes emerged from the qualitative analyses: 1) Making a difference: patients reported the information they were given to manage their symptoms to be useful and effective, making a positive difference and providing support they did not have previously; 2) capacity: both patients and professionals were supportive of the service, hoping that this service could continue after the research study; 3) Expertise: patients and professionals were supportive of the nurse-led model -the professionals found this freed-up consultation time; 4) Barriers and facilitators: patients were happy to both join the service and the research study to evaluate the service, although some patients did not differentiate a difference between the two. Conclusion: The new service was well received by both patients and professionals. Patients felt the new service has provided them with support, helping them to relieve and improve long-standing side effects of bowel radiotherapy. We will triangulate additional health economics and statistical data to determine the overall value of the service and explore its expansion beyond research. University of Bristol, 2 University of bristol Background: Dietary factors and physical activity may alter prostate cancer progression. We explored the feasibility of lifestyle interventions following radical prostatectomy for localised prostate cancer. Method: We recruited patients into a pre-surgical observational cohort; following their radical prostatectomy, we offered the men randomisation into a 2x3 factorial randomised controlled trial (RCT). This involved randomisation into both a modified nutrition group (either increased vegetable and fruit and reduced dairy milk; or lycopene supplementation; or control) and a physical activity group (brisk walking or control). Outcomes were collected at trial baseline, three and six months, with daily adherence reported throughout. Primary outcomes were measures of feasibility: randomisation rates and intervention adherence at six months follow-up. Results: 108 men entered the pre-surgical cohort and 81 the postsurgical RCT (randomisation rate: 93.1% problems are difficult to diagnose and treat in clinical oncology settings, often because of a lack of routine assessment practices. The aim of this study was to gain insight from people living with and beyond cancer about their experiences of chemotherapy sideeffects and strategies used to manage their eating behaviours and diet. Method: Data was collected through in-depth face to face semistructured interviews with people living with and beyond cancer in the East Midlands of England. Data were analysed using interpretative phenomenological analysis (IPA).
Results: Data analysis led to a narrative organized in two parts. The first part (retrospective understanding of chemotherapy sideeffects experiences) reports on the experience of temporary and permanent reduced taste and smell, and their impact on eating behaviours and quality of life, as well as feelings of desolation contrasted to exceptional received support. The second part explains strategies used by participants to manage their health, barriers to and facilitators of access to information and professional advice.
Conclusion: This retrospective investigation of survivors' experienced chemotherapy side-effects gave access to aspects of their experience that often remain undiscussed with healthcare professionals. Further research is needed to develop a taxonomy of taste and smell alterations and food hedonics, which may give clinicians better diagnostic clues to the precise nature of these challenges and inform the design of interventions to ameliorate specific treatment-related side-effects. . Median OS was shorter (log-rank pvalue <0.001) in the "relapse" group (17.5 months (95%CI 15.9-20.9) vs. "non-relapse" (not reached).
Disclosure University of Southampton, 2 Institute of Cancer Research, 3 University of Southampton, University Hospital Southampton NHS Foundation Trust, 4 Swansea University, 5 University of Nottingham Background: The ColoREctal Well-being (CREW) study is the first study to prospectively recruit a representative sample of colorectal cancer (CRC) patients, carry out the first assessment pre-treatment and then follow up longitudinally over five years in order to explore the impact of treatment on health and wellbeing. Method: CRC patients from UK cancer centres received questionnaires at baseline (pre-surgery), 3, 9, 15, 24, 36, 48 and 60 months. Quality of life (QOL), self-efficacy (confidence to manage illness-related problems), mental health, social support, affect, socio-demographics, clinical and treatment characteristics were assessed. Data were analysed using multivariate statistical techniques. Results: A representative cohort of 872 non metastatic CRC patients were recruited from 29 UK cancer centres. Most participants recovered well after curative treatment but around 30% had poor psychosocial outcomes and this persisted up to the 5 year follow-up 1 . Baseline psychosocial factors (particularly selfefficacy and depression) were more important than disease stage and location of tumour in determining who is most likely to have problems over the first 5 years 1 . Self-efficacy remained stable over time whereas social support declined for a significant number of participants. Both low self-efficacy and perceived decline in social support were associated with poorer outcomes 1-3 . Comorbidities that limit an individual's typical daily activities were associated with poorer health and wellbeing outcomes. Conclusion: This unique study provides robust evidence that psychosocial factors, such as self-efficacy, are important predictors for the longer term outcomes of CRC patients. We call for early assessment and intervention, including assessment of depression and confidence to manage illness related problems and limiting co-morbidities, from diagnosis onwards. Early assessment would identify those most likely to need support in their recovery. Early intervention has the potential to reduce need and improve outcomes throughout treatment and beyond. Disclosure

University of Leeds
Background: There is concern that standard chemotherapy pathways of six cycles scheduled every two or four weeks reflect administrative and operational needs rather than patients' personal and biological needs. Process mining of routine data can help identify and explore common pathway variants. Method: We extracted anonymised records from routine data at Leeds Cancer Centre for breast cancer patients with a first diagnosis between 2004 and 2013 with an adjuvant chemotherapy pathway (n = 738). This produced an event log data file (containing events, dates and times) which was analysed using the ProM and DISCO process mining tools. We used the Inductive Miner plug-in and constructed statistical and visual models of the clinical pathways. The data covered a ten-year period and we created multiple splits of the event log to examined statistically significant variations over time.
Results: Most patients varied from the expected pathway (712 variants for 738 patients). We produced a pathway model which included these variants and checked conformance. Overall fitness of data to model was high (97.1%) but we noted significant changes to the fit in 2006 (a 5.1% change) and 2011 (8.9% change) which require further investigation. In total 51% (n = 376) of patients did complete all six cycles, less than half (21% of total, n = 158) completed the cycles without an adverse event while many (30%, n = 218) experienced at least one adverse event including missed appointments, neutropenic sepsis and emergency admissions. Of the 49% (n=362) who did not complete six cycles, 28% (n=207) experienced adverse events with the remainder (21%, n=155) not completing for other reasons.
Conclusion: Process mining of routine data showed extensive variation from standard chemotherapy pathways including incomplete treatment and adverse events. Future work is needed to explore potential causal links and understand changes in the pathway over time. Background: Insufficient time, competing priorities and the limitations of drawing inferences from underpowered, singlecentre populations are commonly cited barriers to the effective study of real-world cancer outcomes. Doctors not in a dedicated academic post may also lack sufficient training to undertake research, despite recognition from medical bodies of the importance of research education. In a number of surgical disciplines, trainee-led research collaboratives spanning multiple institutions are now well established as a means to meeting these challenges. To our knowledge, there have been no previous attempts at establishing similar approaches in non-surgical oncology (NSO). The Northern Oncology Trainees Collaborative for Healthcare Research (NOTCH) has recently been established to promote trainee-led NSO outcomes research. We provide here an overview of this group's aims, scope and early activities. Method: Two clinical and two medical oncology trainees were invited to lead the initiative from each of six large cancer centres within the north of England and Northern Ireland. Trainees from each centre are supported by a wider advisory group consisting of senior clinicians and academics. A call for project proposals was launched between March-May 2018, with proposals submitted by trainees in partnership with senior clinicians. Results: Twelve proposals were received, ten of which focussed on a specific malignancy and two of which focussed on outcomes across a number of cancers. All but one of the proposals involved retrospective data analysis. Proposed outcomes included analysing compliance with existing guidelines, the development of prognostic indices and analyses of toxicity, recurrence and survival patterns. One or more of the proposals will be selected to be undertaken by trainees from across the six centres.
Conclusion: NOTCH provides a novel trainee-led approach for collaborative multi-centre real-world NSO outcomes research for which there is evidence of engagement from trainees and senior clinicians alike. Mummy's Star, 4 University College London Hospitals NHS Foundation Trust and UCL Background: Cancer during pregnancy affects approximately 1:1000 pregnancies, but is becoming more common. More women are becoming pregnant having survived childhood cancer. The lack of UK data on cancer and pregnancy outcomes has meant management is often idiosyncratic and not evidence-based. We aim to establish a UK Registry for pregnant women affected by cancer to improve cancer and pregnancy outcomes for affected women and children.
Method: We report data collected by 'Mummy's Star', a UK charity dedicated to the support of women and their families affected by cancer during pregnancy and up to one-year post-partum. We also collected data on pregnancies affected by cancer at University College London Hospital (UCLH).
Results: Mummy's Star reported a total 419 women of whom 257 (61%) were diagnosed with cancer during pregnancy and 162 (39%) were diagnosed postnatally. Breast cancer was the commonest malignancy affecting 180 (43%) women. Thirty two (8%) women had died and approximately one third of women had metastatic cancer. In a preliminary data collection at UCLH we identified 30 women; 23 were cancer survivors and 7 were diagnosed with cancer during pregnancy. The majority, 17 (57%) were aged between 32 and 40 years old. Hodgkin lymphoma was the commonest malignancy in 7 (23%) patients. Chemotherapy was delayed until childbirth for 3 (10%) patients. Three (10%) patients suffered from anxiety and depression during pregnancy. All women were alive 42 days post-delivery. Conclusion: Cancer during pregnancy is a heterogeneous and relatively rare condition. No single centre can gain enough experience to provide evidence-based care for optimal pregnancy outcome. A national Registry in association with Mummy's Star will build on an established data-set to accelerate our understanding of best practise for these vulnerable women and their families.
Acknowledgement Univariable and multivariable regression analyses were performed to assess association between both polygenic risk score and the 352 individual variants with radiotherapy toxicity at 2 years including telangiectasia, breast oedema, photographic shrinkage, induration, pigmentation, breast pain, breast sensitivity and overall toxicity.
Results: After correction for multiple testing, no association was found between polygenic risk score and development of late radiotherapy toxicity. On multivariable analysis of individual variants, rs138944387 was significantly associated with breast pain (beta = 1.12; 95% CI 0.62-1.61; p = 1.09 x 10 −5 ) and rs17513613 was associated with risk of breast oedema (beta=-0.21; 95% CI -0.31 to -0.12; p = 2.01 x 10 −5 ). Conclusion: Cancer patients with a high polygenic predisposition to breast cancer calculated using an updated polygenic risk score do not have increased risk of radiotherapy toxicity up to two years following radiotherapy. University of Surrey, 2 Prostate Cancer UK, 3 Northumbria University Background: Evidence that physical activity improves prostate cancer treatment outcomes is now recognised. This research explores the feasibility and acceptability of a community pharmacy lifestyle intervention for improving physical activity and health related outcomes. Pharmacy-led health assessment provide opportunities for lifestyle prescription in cancer survivors, we used a computerised algorithm to support personalisation for physical activity and diet. This was repeated at three months and supported by two phone calls and instructional materials in nine community pharmacies. Method: Particpants included men with non-metastatic prostate cancer, completed cancer treatment and at least one of the three risk factors: underweight, overweight or obese; active androgen deprivation therapy; hypertension. Physical and functional fitness were measured with Siconolfi step test and physical activity was assessed objectively using accelerometry. Upper body strength (grip strength), lower body strength (sit-to-stand test) and cardiovascular health (QRISK2 algorithm including weight, body mass index, blood pressure and cholesterol) were also assessed. Feasibility and acceptability were evaluated. Results: 403 men were eligible, 172 (43%) responded and 116 (29%) were subsequently recruited. We report 15% attrition as 99 (85%) men completed the pharmacy intervention. The intervention was feasible and acceptable for men. At three months compared to baseline scores, men's moderate to vigorous physical activity (MVPA) increased by an average of 34 minutes (95% CI: 6 to 62, P = 0.018). The level of MVPA was not sustained at six months (average increase was 14 minutes, 95% CI: -27 to 54, P = 0.509). QRISK2 score was reduced at 3 months (P = 0.001).
Selected Abstracts from the 2018 NCRI Cancer Conference of National. . .

Conclusion:
This study demonstrates that a community pharmacy-led intervention is feasible and acceptable to nonmetastatic prostate cancer patients. The increase in physical activity and improved health, support the development of this delivery approach after cancer treatment. An adequately powered randomised controlled trial is needed to assess the relative costefficacy of this intervention. University of Leicester, UK Background: BRCA1 associated protein 1 (BAP1) is a tumour suppressor that is commonly inactivated in the majority of mesotheliomas. We have previously reported that loss of BRCA1 expression in mesothelioma is a common event, and is associated with resistance to spindle checkpoint activator vinorelbine, a drug with relevance to treatment of mesothelioma. However, 1. The mechanism of BRCA1 loss is unknown, and 2) the potential functional interaction between BRCA1 and BAP1 linked to spindle checkpoint is unknown. The aim of this study was to assess the functional relationship between BAP1 and BRCA1 and examine their role in genome stability in mesothelioma cells. Method: We conducted functional genetic analysis of BAP1 and BRCA1 in two MPM cell lines, MSTO and H2452, the latter carrying an inactivating A95D mutation in the UCH domain of BAP1. BAP1 knockdown was achieved by siRNA transfection, while BRCA1 knockdown was achieved by doxycycline induction of an integrated shRNA.
Results: Loss of BAP1 expression led to reduced expression of BRCA1. Treatment with the proteasome inhibitor, MG132, restored BRCA1 expression in the absence of BAP1 indicating that BAP1contributes to post-translational stabilization of BRCA1 protein. Consistent with previous data, knockdown of BAP1 induced SAC deficiency and vinorelbine resistance concurrent with reduced expression of BRCA1. Loss of BAP1 and BRCA1 also led to an increased frequency of amplified centrosomes. Unexpectedly though, additional defects were observed in mitotic spindle architecture in response to BAP1 loss that were not seen upon loss of BRCA1.
Conclusion: Our data demonstrate that BAP1 controls BRCA1 expression through regulating its protein stability. They also demonstrate both BRCA1-dependent and independent roles for BAP1 in mitotic progression. These findings suggest that BAP1 loss may disrupt spindle checkpoint function and predict resistance to agents such as vinorelbine, a hypothesis that we will test in our randomised trial VIM. Disclosure: Funded by University of Leicester Corresponding author: Anita Singh

Glycolysis supports EGFR-mutant Lung Adenocarcinoma
Cell Survival by blocking Autophagy-mediated EGFR Degradation Wonjun Ji 1 , Jin Kyung Rho 1 , Jaekyoung Son 1 , Jae Cheol Lee 1 , Chang-Min Choi 1   1 Asan Medical Center Background: Oncogenic epidermal growth factor receptor (EGFR) is essential for the development and growth of non-small cell lung cancer (NSCLC), but the precise roles of EGFR in lung cancer metabolism remain unclear. Method: We studied the effect of EGFR on metabolism via targeted liquid chromatography-tandem mass spectrometry (LC-MS/MS) metabolomic analysis, glucose consumption and lactate production assay, and extracellular acidification and oxygen consumption assay in NSCLCs. Mechanisms regulating EGFR stability were investigated using RNA interference/pharmacological inhibitors followed by immunoblotting, cell death assay, and functional assays. Therapeutic potential of JNK activator was determined by monitoring cell growth and death and in mouse xenograft models. Immunohistochemistry was used to analyze JNK phosphorylayion and EGFR expression in NSCLC tissues (n = 244). All statistical tests were two-sided.
Results: EGFR knockdown in EGFR-mutant NSCLC significantly decreased the levels of glycolytic pathway intermediates (p < 0.05) via transcriptional regulation of glycolytic genes. EGFR-mutant NSCLCs exhibited significantly elevated glucose uptake and lactate production compared with EGFR-WT NSCLCs. Glucose deprivation markedly triggered EGFR-mutant NSCLC cell death through robust reduction of EGFR levels, but had no significant effects on EGFR-WT NSCLCs. EGFR-mediated enhanced glycolysis was a major source of carbon for TCA cycle in EGFR-mutant NSCLC, which is essential for maintaining EGFR levels. Glucose Metastasis is driven by inflammation, which potentiates the invasive and migratory behaviour of pancreatic cancer cells.
Our lab is interested in a rare subset of immune cells known as gamma delta (gd) T cells that have pro-metastatic properties in breast cancer mouse models. Pro-metastatic gdT cells are defined by their ability to produce IL-17 and by the lack of expression of the costimulatory molecule CD27. IL-17producing (CD27-) gdT cells promote breast cancer metastasis by expanding and polarising the neutrophils to an immunosuppressive phenotype, which inhibits anti-tumour CD8+ T cells. Additionally, recent reports have shown that gdT cells are crucial for pancreatic cancer progression in transplantable mouse models, and they constitute a large proportion of tumour-infiltrating lymphocytes within human PDAC. However, the mechanisms by which IL-17-producing (CD27-) gdT cells function are largely unknown. Method: To investigate the mechanisms of gdT cell function, we use the Kras G12D ;Trp53 R172H ;Pdx1-Cre (KPC) mouse model, which develops PDAC and liver metastasis at around 150 days of age. Results: We have found that gdT cells are absent from normal pancreas, but relatively abundant in tumours from KPC mice. gdT cells are also reduced in livers from KPC mice when compared with livers from wild-type control mice. Conclusion: Current efforts are underway to determine whether PDAC progression and metastasis are affected in gdT cell-deficient mice and to examine the potential crosstalk between gdT cells and neutrophils. These studies may uncover specific immunotherapeutic strategies to counteract metastatic pancreatic cancer. Method: We used an in vitro model of quiescence in mouse neural stem cells, incorporating a conditional human FOXG1 overexpression cassette, to identify, through high content pharmacological screening, a synergistic relationship between high FOXG1 expression and inhibition of glycogen synthase kinase 3 (GSK3) in driving cells into an active, proliferative state. We quantified this effect using EdU incorporation and colony forming assays. Wnt inhibitors were used to abrogate the effect and a genetic approach, using a constitutively active beta-catenin cassette, was used to elucidate the nature of the synergy. Patient-derived human glioblastoma stem cells (GSCs), with or without CRISPR-Cas9 excision of FOXG1, were used to confirm the relevance of the effect. Results: EdU incorporation, following a 2 hour pulse, was increased from 4% in growth factors alone, 5.9% with FOXG1 overexpression, 5.5% with GSK3 inhibition, to 31.3% with FOXG1 overexpression and GSK3 inhibition combined. Colony forming assays confirm high efficiency cell cycle re-entry in cells with FOXG1 overexpression treated with a GSK3 inhibitor. We subsequently show that this effect is present in patientderived human GSCs and is abolished by excision of FOXG1.
The effect of GSK3 inhibition can be phenocopied both by a ligand of the canonical Wnt signalling pathway and by inducible constitutively active beta-catenin, suggesting that the synergy is effected through beta-catenin, the key downstream effector of canonical Wnt signalling.
Furthermore, the combined effect of FOXG1 overexpression and GSK3 inhibition on exit from quiescence can be abrogated by Wnt inhibitors. Conclusion: Targeting the synergistic relationship between FOXG1 and beta-catenin may provide an exciting therapeutic opportunity in preventing relapse and improving the prognosis of glioblastoma.
Disclosure: Funded by Cancer Research UK Corresponding author: Faye Robertson 52. Investigating Aspirin and Ticagrelor for the prevention of tumour cell-induced platelet aggregation Meera Chauhan, David Adlam, Anne Thomas, Alison Goodall, Joy Wright University of Leicester Background: Tumour cell induced platelet aggregation (TCIPA) may affect the metastatic potential of cancer. Mechanisms include protection of circulating tumour cells from immune destruction, interaction of platelet receptors with tumour ligands to facilitate adhesion, and enrichment of the tumour microenvironment, promoting extravasation and proliferation. Reducing these interactions using anti-platelet agents could alter metastatic progression. This study investigated the effects of Ticagrelor and Aspirin as monotherapy and dual therapy on TCIPA in metastatic breast cancer patients compared to healthy controls. Method: Participants recruited to this randomised, crossover study received Aspirin or Ticagrelor for 2 weeks, followed by a 2week washout and crossover to the other monotherapy, before completing 2 weeks of dual therapy. Platelet rich plasma was prepared from blood samples taken at baseline and the end of each treatment. Flow cytometry measured platelet activation markers (P-selectin expression, fibrinogen and Annexin-V binding) at rest and after agonist stimulation of the platelets in vitro. Method: To investigate which molecules govern the activation of pro-tumorigenic γδ T cells, we measured the expression of various activating receptors on CD27 + and CD27γδ T cells from tumour naïve or tumour-bearing mice, as CD27 stratifies IFNγ-producing (CD27 + ) from IL-17-producing (CD27 -) γδ T cells.

Results
Results: Surprisingly, we found that expression of NKG2D, a receptor involved in recognition of stressed or malignantly transformed cells is higher on CD27γδ T cells compared with CD27 + cells. This increased expression of NKG2D on CD27γδ T cells was specific, as other cytotoxic receptors were elevated on CD27 + γδ T cells. Furthermore, in the K14-Cre;Brca1 F/F ;Trp53 F/F (KB1P) model of breast cancer, we found that NKG2D ligands are upregulated in the myeloid compartment at metastatic sites, such as the lung of tumour-bearing mice. Conclusion: We are currently investigating the interplay between these NKG2D ligand-expressing myeloid cells and IL-17-producing γδ T cells to determine whether the NKG2D axis plays a role in breast cancer metastasis.

University of Nottingham
Background: DNA methylation (5-methylcytosine, 5mC) is the major epigenetic modification involved in transcriptional regulation during the early stages of development in eukaryotes. The patterns of 5mC are frequently altered in cancer. TET proteins can enzymatically oxidize 5mC producing 5-hydroxymethylcytosine (5hmC), 5-formylcytosine (5fC) and 5-carboxylcytosine (5caC). Although the exact biological roles of these oxidized forms of 5mC in cancer pathogenesis are still unknown, there are indications that they might contribute to malignant transformation. According to several reports, 5hmC levels are reduced in human tumours; however, the distribution of 5fC and 5caC in cancers is poorly studied. Although our previous studies showed that 5caC is surprisingly enriched in a proportion of breast cancers and pediatric brain tumors, the distribution and the biological role of this mark in glioblastoma multiforme (GBM) has not been systematically assessed. Method: Here, using mass spectrometry and immunofluorescence, we examine the global levels of 5hmC and 5caC in four human GBM cell lines (LN18, LN228, U251 and U87MG).
Results: We show that while the GBM cell lines exhibit low levels of 5hmC, they are, rather unexpectedly, characterized by relatively high immunochemistry and mass-spec detectable 5caC levels paralleled by the absence of 5fC. Remarkably, 5caC content in GBM does not correlate with 5hmC levels but corresponds to elevated levels of TET2 transcript in these cancers where its transient siRNA mediated knockdown leads to a dramatic decrease in 5caC levels.  (5mC) is an epigenetic modification usually associated with transcriptional repression. The Ten-Eleven Translocase proteins (Tet1/2/3) can oxidise 5mC (oximC) to 5-hydroxymethylcytosine (5hmC), 5-formylcytosie (5fC) and 5-carboxylcytosine (5caC) in vertebrate DNA. These oxidised forms of 5-methylcytosine exhibit distinct genomic distributions in different biological systems. Whilst embryonic stem cells exhibit elevated levels of oxi-mC modifications, adult somatic tissue genomes display depletion for 5fC and 5caC modifications. Curiously, neoplasms of the central nervous system, particularly ependymoma, medulloblastoma and glioblastoma possess significant enrichment of these oxidised forms, however the precise functional roles of these modifications are not fully understood. Transcription factor and tumour suppressor/oncogenic protein Wilm's Tumour 1 (WT1) possesses preferential binding affinity for 5mC and 5caC over unmodified cytosines, 5hmC and 5fC in vitro and exhibits elevated expression in multiple cancers. Expression of WT1 in embryonic neural epithelial progenitors during lineage commitment but absence in terminally differentiated neurons and astrocytes may be indicative of aberrant developmental signalling resulting in malignant neoplasm formation. Method: By employing mass spectrometry-validated immunohistochemistry and confocal microscopic quantification, we demonstrate that high levels of both 5caC and WT1 are characteristics of brain tumour cell lines. Moreover, 5caC is present on promoters of core brain tumour signalling pathway genes. Furthermore, we endeavour to perform CRISPR cas9 facilitated WT1 targeted knockout and subsequently implement next generation sequencing technologies to globally map 5caC genomic distribution and influence on brain tumour transcriptomes. Results: Ultimately this will facilitate the elucidation of a mechanistic relationship between WT1 and 5caC which may aetiologically induce or functionally perpetuate clinical aspects of brain tumour pathogenicity. Barrow Neurological Institute Background: Glioblastoma (GBM) is characterized by rapidly proliferating and invasive cells that infiltrate normal brain regions. Following exposure to aggressive treatment regimens, GBMs frequently shift their biological features upon recurrence, acquiring a more resistant phenotype. However, the dynamics and molecular mechanisms that facilitate GBM recurrence are still poorly understood. Considering the unchanged dismal prognosis for GBM patients, there is a need to understand, at a systems level, how plastic processes (molecular switches) in glioma stem-like cells (GSCs) may drive tumor maintenance and cancer cell adaptability in GBM. The objective our study was to determine how GSCs temporally adjust their expression profile and phenotype in response to ionizing radiation in vitro and in vivo using patient-derived xenograft (PDX) models of GBM. Method: We established PDX GBM models by intracranially implanting two patient-derived GSC lines belonging to different GBM molecular subgroups into immunocompromised mice. The tumor-bearing mice were treated with single doses of ionizing radiation to assess acute responses to treatment. Mice from each cohort were be sacrificed at multiple distinct time points following treatment. Using immunohistochemical methods, we assessed changes in the expression of GBM subclass markers, stemness and differentiation markers, and DNA damage/repair proteins across the entire tumor population over time. To understand how GSCs respond to radiation at a molecular level, we employed mass cytometry (CyTOF) and RNA-seq to determine how important cellular signaling pathways and transcriptional programs necessary for GSC self-renewal, invasion and growth are altered at various time points post-treatment. Results: We demonstrate that GSCs, both in vitro and in vivo, undergo an immediate response following exposure to radiation that results in a global modulation of the expression of key stemness and proliferation genes under adverse conditions. Conclusion: Our results suggest that this acute response allows GSCs to enter a transient semi-differentiated state that favors GSC adaptability and resistance to therapy Disclosure: Funded by National Institutes of Health Method: C1orf106 gene and protein expression was assayed by qRT-PCR and western blotting following exogenous stimulation or inhibition of TGF-β signalling. The role of the canonical TGF-β pathway in regulating C1orf106 transcriptional induction was explored by siRNA-mediated gene silencing. C1orf106 mRNA expression levels and influence on clinical outcomes were assessed in the Oncomine™ and KM Plotter platforms. The phenotypic effects of stable overexpression or shRNA mediated knockdown of C1orf106 expression was investigated in breast cancer cell lines. Anchorage-independent growth and colonyforming assays, and a limiting dilution xenograft were carried out to assess tumourigenicity in vitro and in vivo, respectively. Results: C1orf106 mRNA and protein induction in response to TGF-β stimulation was observed in a plethora of cell types and determined to be SMAD3-dependent. Analysis of publically available datasets indicated that elevated C1orf106 mRNA expression is associated with poor clinical outcomes in breast cancer. C1orf106 expression correlated with metastatic progression in the 4T1 murine mammary carcinoma model. High C1orf106 was associated with enhanced anchorage-independent growth ability, colony forming capacity and clonogenicity in vitro in this model system and trended towards increased tumour initiation frequency in vivo. Background: Progressive accumulation of mutations in oncogenic and tumour suppressor pathways (e.g., KRAS and p53) and also microRNA (miR) deregulation are associated with colorectal cancer (CRC) development. Very little has been done to dissect specific CRC pathways involved in miR regulation in response to stress and metabolic changes. Therefore, we investigated the interaction between miR-mediated regulation of bone marrow stromal cell antigen 1 (BST1) or CD157, a metabolic enzyme involved in the conversion of nicotinamide adenine dinuclease (NAD) to paracrine factor cyclic ADP ribose (cADPR), following the acquisition of KRAS mutation in metastatic CRC.
Method: A combination of array analysis data, in-silico prediction tools and nuclear magnetic resonance (NMR) system were used to analyse gene (mRNA) and miR expression, as well as metabolic changes associated with different rounds of knock in/out mutations in Apc, Kras, and p53 CRC mouse models and their tumour-derived organoids (TDOs). Subsequently, we genetically modified human CRC cell lines and patients KRAS mutant (mut) TDOs from CRC metastases to modulate the candidate gene and miR. Organoids formation, growth rate and viability were measured with Live-Cell imaging systems. Results: We identified mRNA/miR networks involved in cancer metabolic pathways and assessed the most significant candidates. Further target validation and analysis of human tissues microarrays showed an association between KRAS mutations, miR-203 down-regulation and over-expression of BST1, which was identified as a direct target of miR-203 regulation in our studies. Repressing BST1 and over-expressing miR-203 had a significant effect on the proliferation and migration abilities of organoids in 3D culture in normal and calorie-deprived conditions. Conclusion: This project aimed to find a promising candidate as a therapeutic target in KRAS mut CRC. Although we showed KRAS mut CRC cells lost a growth advantage with miR-203 and BST1 University of Glasgow, 2 Cardif University Background: Colorectal cancer (CRC) is a heterogeneous group of malignancies that arise in the same organ. The 5-year survival rate for CRC is 60% and this is significantly reduced in stage four metastatic disease. A common site of metastasis is the liver. In 2017, phenotypic subtypes for CRC were developed in an effort to move towards precision medicine. The phenotypic subtypes (immune, canonical, latent and stromal) are derived from three features; inflammation, stromal invasion and proliferation. This study aimed to investigate the relationship between primary tumour and liver metastases phenotypic subtypes and association with clinicopathological outcomes. Method: Matched patient-derived colonic primary tumours and liver metastases were stained for Ki67 proliferation index and an H&E analysed for Klintrup-Makinen grade and tumour-stromapercentage to determine phenotypic subtype. The relationship between the primary tumour subtype (PS-primary) and liver metastases (PS-met) was investigated using bivariate correlations and paired sample T-tests. The relationship between PS-primary/ PS-met and clinicopathological features was analysed using Chisquared tests.

Conclusion:
The results suggest that PS-primary is predictive of PS-met, and that each phenotypic subtype associates with different specific clinicopathological features differing between primary or metastatic lesions. This may provide a step towards the development of precision medicine for CRC. We first constructed a backbone network of transcription factor (TF) -target gene pairs inferred from chromatin landscape data and TF binding motifs, and applied nonlinear regression using expression profiles of each subtype to derive subtype-specific regulatory parameters for each TF-gene pair. Next, we mined for co-regulatory TF pairs using correlation analysis. Mechanisms responsible for subtype-specific behaviour of TFs were then inferred from expression, regulatory and co-regulatory signatures. Finally, we simulated the effects of perturbing druggable signature TFs and their partners by propagating changes in expression to the corresponding target genes and then to the protein signaling layer using a random walk-based algorithm.
Results: We show that subtype-specific repurposing of TFs explains a significant proportion of subtype-specific transcription landscapes. At least two mechanisms for TF repurposing are Selected Abstracts from the 2018 NCRI Cancer Conference of National. . .
implied -differential expression of the TF itself and differential partnering with co-regulatory TFs. Using effectors of the apoptosis pathway as readout in our in silico perturbation analysis, we show that targeting a subset of druggable signature TFs and/or their partners may be specifically beneficial for treating the corresponding subtypes of GBM. British Columbia Cancer Research Centre, 2 Dalhousie University Background: The shape and organization of tumour cell nuclei has been shown to be associated with aggressiveness in several cancers, including prostate and cervical cancer. Fortunately, this data can be obtained using a single quantitative nuclear stain on FFPE tissues. We sought to determine whether nuclear features could be utilized to distinguish cell types within the lung tumour microenvironment and to classify outcome groupings.
Method: A tissue microarray of non-small cell lung cancer samples was stained with the quantitative nuclear dye, Feulgen/thionin. The stained slides were scanned using a hyperspectral imaging platform, images were spectrally unmixed, and cell nuclei were identified using a segmentation algorithm. Nuclei were then analyzed to identify 245 nuclear morphometry features. Nuclear features were assessed in order to dichotomize good outcome (alive at 5 years) and bad outcome (survival of 3 years or less) patients, as well as several cell types within the tumour microenvironment. Finally, we assessed the spatial organization of these different classes of nuclei in good and poor outcome cases.
Results: Nuclear features derived from thionin stain classified epithelial, stromal, and immune cells. Of the 245 nuclear features assessed, 87 differed significantly between good and poor outcome groups. While many distinguishing features were identified in tumour cell nuclei, we were also able to identify distinguishing features of non-tumour cell nuclei. Using pairs of features, good and poor outcome cases were stratified with up to 69% accuracy. However, the addition of spatial information to our classifier resulted in the improvement of sample stratification to 86% accuracy. Conclusion: Using a single nuclear stain on FFPE lung cancer histology specimens, we were not only able to sub-classify cell types, but also to distinguish between good and poor outcome cases. The spatial organization of these cell types greatly improved our classifier, indicating the potential of this approach for prognostic purposes. University of Wolverhampton, 2 Royal Preston Hospital, 3 The Walton centre Background: Over recent years breast cancer survival rates have improved. However, even after many years of apparent diseasefree health, cancer can recur. Many of these tumours occur specifically in the brain and metastatic brain tumours have very poor prognoses. Identifying genomic alterations that occur in breast primary tumours that eventually metastesise to the brain will provide new opportunities for treatment and prognosis. Method: Whole-exome sequencing (WES) was carried out in 18 brain metastasis samples that originated from breast tumours. Each sample was sequenced to a depth >100X. Bioinformatic analysis was carried out to identify common recurrent mutations. We are in the process of validating the candidate mutations by Sanger sequencing and screening a larger cohort of Breast to Brain Metastasis (BBMs) and non-metastatic primary breast tumours to confirm metastasis-associated alterations.
Results: Each of the 18 BBMs analysed by WES contain >7000 nonsynonymous variants. All variants were screened for their consequence on the protein product (via Polyphen and the Exome Aggregation Consortium(ExAc)). Those variants with high scores relating to pathogenicity were retained. Following this filtering, potential germline polymorphisms were excluded by removing those variants with a minor allele frequency (MAF) of >0.1%. This screening has generated a long list of 300 variants found across all 18 tumours analysed. A final screen identified genes that contained pathogenic variant in more than 2/18 tumours and had been described in any other cancer type (via the catalogue of somatic mutations (COSMIC)). Via this stringent screen, we have identified 22 candidate metastasis-associated genes. we are currently screening nonmetastatic primary breast tumours and BBM tumours to determine how frequently these occur. The genes identified have varied Selected Abstracts from the 2018 NCRI Cancer Conference of National. . . cellular roles, including cell surface proteins, migration and gene regulation. Conclusion: We expect that this analysis will identify genes that are frequently mutated in BBMs, but infrequently in nonmetastatic breast tumours. The Beatson Institute for Cancer Research Background: The elongation phase of mRNA translation is the most energy and amino acid consuming part of protein synthesis. This process is controlled by the Eukaryotic elongation factor 2 kinase (eEF2K), an atypical calmodulin-dependent protein kinase that slows down translation elongation by phosphorylating eEF2. eEF2K functions as a negative regulator of protein synthesis and cell growth. Cancer cells may possess mechanisms to inhibit eEF2K. We have previously shown that in colorectal cancer Rapamycin-induced reduction of elongation suppresses intestinal regeneration and tumourigenesis following APC loss. This process was shown to be controlled by eEF2K. Method: We generated in vivo mouse models with acute deletion of either a single or both copies of Apc with either conditional expression of eEF2K kinase dead allele (D273A) or conditional deletion of eEF2K in the mouse intestine. We followed the rates of intestinal proliferation and tumourigenesis and the response to various drug therapies.
Results: In both of these models we observed loss of the inhibitory phosphorylation of eEF2. Intestinal crypt organoids from these mice showed increased ribosome run-off rate associated with inhibition of elongation control due to the loss of eEF2K function. We have also demonstrated that loss of eEF2K function or total eEF2K levels, following Apc loss, led to significant increase in survival compared to mice with Apc deletion alone. Deletion of eEF2K in addition to Apc and Kras, one of the most frequently altered genes following Apc, did not have an effect on survival. Background: MCL-1 is a pro-survival member of the BCL-2 protein family that has been shown to be up-regulated in a range of cancers. We have shown that high levels of MCL-1 predict poor prognosis in breast cancer. This is particularly relevant in triple negative breast cancer where treatment resistance and disease recurrence remain a major challenge. A new class of drugs specifically targeting MCL-1 have been developed and are currently in clinical trials for haematopoietic malignancies. We are investigating the therapeutic potential of targeting MCL-1 in breast cancer and the role of MCL-1 in breast cancer stem cells. Method: We have used a combination of approaches to systematically evaluate the role of MCL-1 in breast cancer. This includes the use of small molecule MCL-1 inhibitors and knock-down/ knock-out of MCL-1 in vitro and in vivo xenograft/genetically engineered mouse models to investigate the therapeutic potential of targeting MCL-1 in breast cancer. Results: We find that in addition to the poor prognosis indicated by high levels of MCL-1, breast cancer cells are highly dependent on MCL-1. Therapeutic targeting or genetic deletion of MCL-1 is sufficient to delay tumour development and regress established tumours in vivo. We find that MCL-1 is important in breast cancer stem cells and targeting MCL-1 could be particularly important in the context of treatment resistance and recurrence.

The role of ELTD1/ADGRL4 in tumour angiogenesis
Koon Hwee Ang University of Oxford Background: ELTD1/ADGRL4, is an orphan adhesion GPCR. ELTD1 is expressed in endothelial and vascular smooth muscle cells but its expression in the tumour vasculature is significantly increased. Our aims were to analyse ELTD1's function in endothelial cells and breast cancer to explore its potential as an anti-cancer therapeutic target.
Method: Primary human breast cancers (n = 245) and matched primary & nodal secondary breast cancers (n = 79) were stained for ELTD1. Staining intensity was scored and compared with survival. ELTD1 was expressed in breast cancer cell lines to assess the effect on tumour growth in xenograft experiments and knockout mice were generated with 1) tamoxifen-inducible knockout of Eltd1 in all tissues 2) constitutive Eltd1 knockout in vessels or 3) inducible Eltd1 knockout in vessels. Extracellular vesicles (EV's) were harvested from ELTD1 expressing cells and their effects assessed in vitro and in vivo.
Results: Human breast cancer staining revealed a higher intensity vascular ELTD1 staining within the tumour stroma compared to normal stroma and~15% of the tumours had ELTD1 expression within tumour cells. Higher ELTD1 expression in both the tumour stroma vasculature (n = 241; HR = 0.68; p = 0.04) and within the subset of tumour positive cases (n = 24; HR = 0.3; p = 0.02) correlated with improved relapse free survival (RFS). ELTD1 expression in human breast cancer cell lines did not affect proliferation or spheroid growth, but reduction in tumour growth was seen in xenograft models. Tumours grown in the constitutive and inducible Eltd1 knockout mice showed no difference in growth but marked changes in vasculature and necrosis. ELTD1 is incorporated into EV's. Vesicles isolated from ELTD1 overexpressing breast cancer cell lines and HUVECs were pro-angiogenic in vitro and reduced tumour growth when injected into xenografts.

Imperial College London
Background: Neutrophil counts have been linked to the progression of cancer due to their tumourigenic role in the cancer microenvironment. Numerous meta-analyses and individual studies explore the association between neutrophil counts and cancer prognosis, contributing to a large body of evidence with variable strength and validity. Uncertainty exists around the association between neutrophils and cancer outcomes depending on the site, outcome and treatments considered. Method: For this umbrella review we searched Medline, EMBASE, and the Cochrane Database of Systematic Reviews for metaanalyses of observational studies evaluating the association between neutrophil to lymphocyte ratio (NLR) or tumour associated neutrophils (TAN) and specific cancer outcomes related to prognosis. The available evidence was graded as strong, highly suggestive, suggestive, or weak through the application of pre-set grading criteria. For each included meta-analysis, the grading criteria considered the significance of the random effects estimate, the significance of the largest included study, the number of studies and individuals included, the heterogeneity between included studies, the 95% prediction intervals, presence of small study effects, excess significance and credibility ceilings. Results: Ultimately, 81 meta-analyses from 36 studies met the criteria for inclusion. All meta-analyses suggested a hazard ratio in the same direction of effect (HR>1). When assessed for significance and bias related to heterogeneity and small study effects, only three (4%) associations between NLR and overall survival and progression-free survival in gastrointestinal and nasopharyngeal cancers were supported by strong evidence. Conclusion: Despite many publications exploring the association between NLR and cancer prognosis, the evidence is limited by significant heterogeneity and small study effects. There is a lack of evidence on the association between TAN and cancer prognosis, with all nine meta-analyses identified arising from the same study. Results: The biggest cause of concern to patients was the process of waiting for and obtaining test results. Most had experienced swift referral/testing, and it was difficult for participants to understand how the new standard could impact upon time spent progressing through the system. Responsibility for meeting the standard was also a concern: patients did not see their own behaviours such as accepting a cancellation appointment as a form of involvement in the standard being met. The GP's role was conceptualised as communicating with the patient about their referral, establishing their preferences for information and continued involvement at each stage of the referral process. The standard legitimised chasing for test results, but 28 days was considered too long. Conclusion: Patients should be asked what they would like to know about their referral. Where appropriate, GPs should be more transparent about the referral process and the potential for lack of clarity around next steps, timescales and outcomes. Patients should know that it's ok to make use of opportunities perceived as 'manipulating the system' but this needs to be balanced against adding to existing patient burden. University of Lincoln, 2 University of Oxford Background: The BIRADS score for mammographic density does not currently indicate which breast has the higher density, and may be at higher risk of developing (or having) cancer. A focal density ('developing') is a region that is apparent on a recent mammogram, is present in both CC and MLO views, but can only be seen in one breast. Using focal density (FD) quantification, we assign a density score to each breast separately in order to assess breast asymmetry. Method: Our method, which substantially improves our previous method (RICE) [1], suppresses normal breast parenchyma in order to highlight ROIs (see Figure 1). It embodies the reasonable assumption that the neighborhood typically has a similar tissue density to that of the tissue encompassing the candidate masked tumor. In this preliminary study, the method was applied to 11 patient cases of very dense mammograms, each of which contained at least one cancer. Our method produces a density score for each individual breast and determines asymmetry as a percentage score. Results: Table 1  Method: Patients with suspected intrinsic glioma discussed at neuro-oncology Multidisciplinary Team meetings and suitable for fluorescence guided cytoreductive surgery were eligible. 5aminolevulinic acid (5-ALA) was used to generate visible fluorescence. Tissue samples were sent for peri-operative histopathological analysis to establish an intra-operative diagnosis of LGG or HGG. Presence of visible fluorescence was collected. These data were compared with the final central pathological diagnosis. Results: From Feb 2015 to March 2017 in the UK, 106 patients were recruited: median age 59 (range 23-77); 59% male; 25% WHO radiological grade II transforming to a higher grade and 55% grade IV. 5-ALA were given for 103 patients with a median dose of 1500mg (range 960-2200mg). 67% of patients classified as HGG at local per-operative diagnosis were confirmed by the central review (weighted Kappa 0.37 (95%CI = 0.21-0.54)). 88 patients were evaluable for the primary endpoint: 81 had visible fluorescence of the tumour with central histopathology diagnosis as 1 LGG, 78 HGG (a 99% concordance in HGG classification with the 99%CI = 91%-99.9%) and 2 not assessed; 7 patients had no visible fluorescence and were diagnosed as 6 LGG and 1 HGG. Conclusion: There is an urgent need to improve the local perioperative diagnosis. The presence of visible fluorescence can be used as an additional pragmatic intra-operative diagnostic surgical biomarker of high-grade disease within a tumour mass. Use for assessment of low-grade disease needs further investigation. Disclosure

Results
: Cytoplasmic, granular TOMM20 and nuclear Ki67 were highly expressed at the advancing front of OPSCCs, independent of HPV status. Co-expression of membranous MCT1 and MCT4 at the advancing front was observed in 84% and 30% of the HPV(+) and HPV(-) tumours, respectively (p=0.001). While all HPV(+) tumours showed MCT4(+) stromal spindled cells, regarded as myofibroblasts or cancer-associated fibroblasts, adjacent to the advancing front, only 70% of the HPV(-) tumours did so (p=0.002). Conclusion: Tumour cells at the advancing front of both HPV(+) and HPV(-) OPSCC are rich in mitochondria, show similar proliferative activity, and consume mitochondrial fuel (lactate), imported via MCT1 expression. The different frequencies of coexpression of MCT1 and MCT4 in HPV(+) and HPV(-) OPSCC indicate that the three compartment model is more compatible with data from HPV(-) OPSCC. The metabolism of the stromal component differs between HPV(+) and HPV(-) tumours, but these differences are not reflected in the proliferative activity of the tumour cells in the advancing front. We conclude that tumour metabolism in OPSCC may be related to HPV status, but is complex and requires further examination. Here we describe a novel blood-based biomarker for OAC based on the PIG-A mutation assay, which has the further potential of predicting early treatment response. We hypothesise that cancer patients will have a higher level of this circulating mutation compared to controls. Furthermore, will patients undergoing chemo/radio therapy who display an increase in mutation frequency during their treatment respond better to such therapies? Method: Blood samples were obtained from consenting patients attending endoscopy including patients with reflux disease, Barrett's metaplasia (BM) and newly diagnosed OAC. Multiple blood samples were obtained from patients with OAC throughout their treatment course. Ten microliters of whole blood was stained Selected Abstracts from the 2018 NCRI Cancer Conference of National. . . with antibodies against the erythrocyte marker CD235a and two GPI-linked proteins (CD55 and CD59) and analysed using flow cytometry. Results: With over 300 participants currently recruited to the study including healthy volunteers, data shows that treatment naïve OAC patients had a 3-fold increase in PIG-A mutant frequency compared to reflux patients (p<0.001). Reflux (n=75), Barrett's (n=63) and treatment naïve cancer patients (n=38) had a median PIG-A mutant frequency of 3.2 (95% CI: 1.51-5.43), 4.52 (95% CI: 2.53-6.09) and 9.75 (95% CI: 3.76-17.52) respectively.
OAC patients undergoing treatment (n=13) had intra-individual variation of 1 to 8 fold increase in mutation frequency compared to their pre-treatment levels, with mutant frequencies in the range of 1.0-29.8 mutant RBC's/million. Correlation with therapeutic response is on-going. Conclusion: The PIG-A mutation assay has potential as a lessinvasive biomarker for OAC as well as a treatment-response tool to identify patients with limited response to current therapies. University of Cambridge, 5 Anglia Ruskin University Background: Breast cancer, a heterogeneous disease, may be subtyped using immunohistochemistry (e.g. estrogen receptor (ER) and human epidermal growth factor receptor 2 expression). Differences in survival occur between apparently similar cases. Thus, better resolution of the disease into clinically meaningful subtypes is needed. Breast cancer lacks a biomarker for invasive disease that will metastasise and the discovery of such a biomarker would represent a significant step forward. Transglutaminase 2 (TG2), a multifunctional protein implicated in cell-cell adhesion, apoptosis and metastatsis, warrants investigation. Levels have been reported to vary with tumour progression. Method: We used tissue-micro-arrays (TMAs) to evaluate TG2 expression in the tumours of 1,942 patients for whom we have associated data for survival time (72% alive; 28% dead), and tumour grade and stage. 1,917 samples also have hormone status data (79%, positive; 21%, negative). Hormone status was derived from ER and progesterone receptor (PR) expression data (designated 'positive' if either was positive, 'negative' if both were negative and where only the status of ER or PR was available this was assigned individually as the hormone status). TMAs were scored manually, and tumours expressing TG2 in more than 10 per cent of cells were categorised as 'TG2 positive'. Results: AlthoughTG2 was expressed in 37% of tumours it was not associated with outcome in hormone receptor positive disease. This was not the case in hormone receptor negative disease where its expression was associated with decreased breast cancer specific mortality (HR, 0.57; 95% CI, 0.37 − 0.89; p value, 0.012). Conclusion: These results need confirming in a larger study. Nevertheless, the molecular basis of this finding if true would be intriguing, given the established links between TG2 and tumourigenesis. TG2 might be a novel drug target. Furthermore, given the prognostic effect-size, TG2 could prove a useful marker for identifying patients most likely to benefit from adjuvant chemotherapy.
Disclosure Flinders University, 2 University of South Australia, 3 University of Waterloo, 4 University of Glasgow Background: Breast cancer (BC) incidence in 'middle-aged' women (45-64 years) is increasing in industrialised countries, despite ongoing warnings about modifiable risk factors such as alcohol, the consumption levels of which have also increased in middle-aged women. Alcohol use, entrenched in Australian society, is directly linked to BC yet is seldom proposed as a causal mechanism by women at risk. We investigated the place of alcohol in the lives of South Australian women at various levels of perceived risk for BCexploring the ways in which women think about BC in the context of alcohol consumption as a socially-entrenched practice. Method: In 2017, 122 women completed a brief online purposive sampling survey. After cross-characterising responses according to perceived cancer risk, level of alcohol consumption and education level, 35 women were interviewed. The audio-recorded interviews were transcribed, coded and thematically analysed. Results: Socialising, relaxing and coping were the main purposes of alcohol consumption in this group. Women with higher socioeconomic status (SES) tended to talk about alcohol as a pleasurable social lubricant, while lower SES women tended to talk about alcohol as a mechanism to cope and manage stress. The principal marker of problematic alcohol use overall was 'functionality' in work, care-giving and social roles.
BC candidacy was determined by family history with BC development generally seen as unpreventable and framed as 'bad luck. Where a role for alcohol in BC was accepted, the risk was weighed against the important social function of alcohol. Conclusion: In this group, most aspects of alcohol use were framed as social phenomena firmly outside of which context was the concept of alcohol-related BC risk. Prevention activities will need to respond to the social functions of alcohol, and appropriately target their SES differences, if they are to enact meaningful behaviour change to reduce BC risk.

Cardiff University
Background: Huntington's disease (HD) is a neurodegenerative genetic disorder caused by the trinucleotide expansions of CAG that results in the formation of elongated polyQtracts. It is hypothesised that these polyQtracts accumulate intracellularly causing cellular apoptosis prior to tumour-induced cell division. 1 Thus, expanded polyQtracts may be a potential protective mechanism against cancer. This theory is based on evidence from studies in Denmark and Sweden who both report a reduction of cancer incidence in HD patients. However, to date, no study has explored the relationship between CAG length and cancer incidence. 2,3 Method: 6540 HD patients were identified using data from the European Huntington's Disease Registry. The age-standardised incidence ratio (SIR) for specific types of cancer was calculated by comparing risks to general population data from the WHO. Conclusion: This cross-national study provides further evidence to suggest a possible link between HD and a reduction of cancer incidence. 2,3 Individuals with early-onset HD have a lower cancer incidence in comparison to late-onset HD patients as they experience its protective effects much sooner. Non-cancer patients have longer polyQtracts that protect against cancer by accumulating intracellularly, restricting cellular proliferation and increasing apoptosis. Furthermore, other polyQ diseases such as hereditary ataxia showed a similar reduction in cancer incidence, albeit to a lesser extent. 5 Further research is warranted to investigate these exact mechanisms in the hope to establish the next break-through in cancer research.
Disclosure: The European Huntington's Disease Network is funded by the CHDI Foundation, Inc.: this project was EHDN project 478, and this work was supported by the MRC (MR/ L010305/1). Background: Studies investigating the association of Mediterranean diet (MD) adherence with pancreatic cancer risk are limited and had inconsistent results. We examined the association between MD adherence and pancreatic cancer incidence by pooling individual subject data from two Dutch cohorts. Method: The Netherlands Cohort Study (NLCS, 120852 subjects) and the Dutch cohort of the European Prospective Investigation into Cancer and Nutrition (EPIC-NL, 40011 subjects) were included in this pooled analysis. MD adherence was assessed using alternate and modified Mediterranean diet scores (aMED and mMED, respectively), including and excluding alcohol. After median follow-ups of 20.3 (NLCS) and 19.2 (EPIC-NL) years, 449 cases of microscopically confirmed pancreatic cancer (MCPC) were included in study-specific multivariable Cox proportional hazards models. Study-specific estimates were pooled using a random-effects model. Results: MD adherence was not significantly associated with MCPC risk in pooled and study-specific analyses, regardless of sex and MD score. Pooled hazard ratios (95% confidence interval) for high (6-8) compared to low (0-3) values of mMED excluding alcohol were 0.66 (0.40 -1.10) in men and 0.94 (0.63 -1.40) in women. A two-point increment in mMED excluding alcohol was Selected Abstracts from the 2018 NCRI Cancer Conference of National. . . borderline significantly associated with a reduced MCPC risk in never smokers (p = 0.07), but not in ever smokers (pheterogene-ity=0.03). Hazard ratios were consistent across strata of other potential effect modifiers. Considering MD scores excluding alcohol, mMED-containing models generally fitted better than aMED-containing models, particularly in men. Although associations somewhat differed when all pancreatic cancers were considered instead of MCPC, the overall conclusion was similar. Conclusion: MD adherence was not associated with pancreatic cancer risk according to a pooled analysis of two Dutch cohorts.
Disclosure University College London, 2 Imperial College London Background: In England, uptake of the human papillomavirus (HPV) vaccine for the prevention of HPV-related cancers is suboptimal among girls from ethnic minority backgrounds and in some regions, particularly London. As part of the school-based HPV immunisation programme, a consent form with parental signature has to be returned regardless of whether consent for vaccination is given or withheld. Lack of a signed consent form is the main reason for adolescent girls not receiving the vaccine in the UK. We aimed to determine the feasibility of undertaking a cluster randomised controlled trial (RCT) of incentivising consent form return to improve HPV vaccine uptake.
Method: An equal-allocation, two-arm cluster RCT design was used. We invited 60 London schools to participate. Those agreeing were randomised to either a standard invitation or incentive intervention arm, in which Year 8 girls had the chance to win a £50 shopping voucher if they returned a vaccination consent form, regardless of whether consent was provided. We collected data on school and parent participation rates, questionnaire response rates, consent form return and vaccine uptake. Analyses were descriptive. Results: Six schools completed the trial and only 3% of parents opted out. The response rate was 70% for the girls' questionnaire and 17% for the parents' questionnaire. In the intervention arm, 87% of girls returned a consent form compared to 67% in the standard invitation arm. The proportion of girls whose parents gave consent for vaccination was higher in the intervention arm (76%) than the standard invitation arm (61%). Conclusion: A RCT of an incentive intervention is feasible. This incentive intervention has the potential to substantially improve HPV vaccination uptake, which should reduce HPV-related cancer incidence, with minimal work from immunisation providers, but a fully-powered RCT is needed.

Northwestern University
Background: The effectiveness of preventive therapy for breast cancer depends on adequate uptake, but initiation rates remain low. Little is known about factors influencing the decision to use chemoprevention. We examined whether women at increased risk of breast cancer can be categorised into groups with similar medication beliefs and evaluated whether belief group membership was associated with uptake. Method: Women (n = 732) attending an appointment at one of 20 centres in England to discuss breast cancer risk were approached; 55.7% (408/732) completed a survey containing the Beliefs about Medicines Questionnaire (BMQ) and the Perceived Sensitivity to Medicines (PSM) scale. Self-reported uptake of tamoxifen at 3-month follow-up was reported in 258 (63.2%). The optimal number of medication belief groups was identified using Latent Profile Analysis (LPA).
Results: Among baseline respondents (mean age = 45.3 years, SD = 7.8), 59.6% were at moderately high risk of breast cancer (17-30% lifetime risk) and 39.0% were at high risk of breast cancer (≥30% lifetime risk). Uptake of tamoxifen was 14.7% (38/ 258). The LPA model supported a 2-group model. Both groups held weak beliefs about their perceived need for tamoxifen. Group 2 (38% of the sample) reported stronger concerns about tamoxifen and medicines in general, and stronger perceived sensitivity to the negative effects of medicines compared with Group 1 (62%). Women with low necessity and lower concerns (Group 1) were more likely to initiate tamoxifen (18.3%; 33) than those with low necessity and higher concerns (Group 2) (6.4%; 5). After adjusting for demographic and clinical factors, the OR was 3.37 (95% confidence interval: 1.08 -10.51, p = .036).
Conclusion: Uptake of breast cancer preventive therapy was low. An important sub-group of women reported low need for preventive therapy and strong medication concerns. These women were less likely to initiate tamoxifen. Medication beliefs are modifiable targets to support informed treatment decisionmaking. Disclosure: SGS was supported by a Cancer Research UK postdoctoral fellowship (C42785/A17965) during the collection of these data. He also acknowledges funding support from a Yorkshire Cancer Research University Academic Fellowship.
Corresponding author: Samuel Smith

University of Bristol
Background: Feasibility trials are preliminary trials that assess the viability and acceptability of intervention studies and the effects of the intervention on intermediate endpoints. Due to their short duration, they are unable to establish the effects of the intervention on long-term clinical outcomes. We propose a novel method that could transform the interpretation of feasibility trials using modified two-stage randomization analyses. Method: In this two-stage process, we explored the effects of a 6month feasibility factorial randomised controlled trial (RCT) of lycopene and green tea dietary interventions (ProDiet) on 159 serum metabolic traits in 133 men with raised PSA levels but prostate cancer (PCA) free. In the first stage, we conducted an intention-to-treat analysis, using linear regression to examine the effects of the interventions on metabolic traits, compared to the placebo group. In the second stage, we used a two-sample Mendelian Randomization (MR) approach to assess the causal effect of metabolic traits altered by the interventions, on PCA risk, using summary statistics data from an international PCA consortium of 44,825 cancer cases and 27,904 controls.
Selected Abstracts from the 2018 NCRI Cancer Conference of National. . .
Conclusion: Using a two-stage randomization analysis in a feasibility RCT, we found that lycopene lowered levels of pyruvate, which our Mendelian randomization analysis suggests may be causally related to reduced PCA risk. The carcinogenic effect of sleep patterns has also been investigated among night shift workers, with inconsistent findings.
Here we conducted a Mendelian randomization (MR) analysis using genetic variants robustly associated with chronotype (morning/evening preference), insomnia and sleep duration to investigate causal links with breast cancer. Method: Observational associations between sleep characteristics and breast cancer in UK Biobank (9,599 cases, 170,616 controls) were first assessed using logistic regression with adjustment for age, assessment centre, ancestry and several socio-economic, lifestyle and reproductive factors. For the MR analysis, breast cancer status was regressed against predicted values of sleep characteristics based on genotype in a logistic regression model, with adjustment for age, assessment centre, ancestry and genotyping chip. Two-sample MR was also conducted using summary data from a large genome-wide association study of breast cancer ( Background: Cancer is associated with socio-economic disadvantage. Yet many interventions designed to reduce risk and improve health fail to reach those with the greatest needs and the most vulnerable. Disadvantaged women, including those who have suffered domestic abuse or who are within the judicial system, represent a group that is particularly poorly accessed in prevention strategies and in research. Our study focused on such disadvantaged women, at two women's centres that provide support and training. Method: This qualitative study involved thirty participants (23 women and seven staff) in individual interviews and two focus groups. It sought to understand perceptions of, and vulnerability to, cancer; decision making (including screening); cancer symptom awareness and views on health promotion within the context of the women's daily lives. Verbatim transcripts were analysed thematically.
Results: Mental distress dominated our findings. Risk factors of alcohol use, smoking, physical inactivity and unhealthy eating were common but reported within the context of distressing experiences of mental ill-heath, poverty, addition and abuse. Walking, for example, was reported due to lost driving licences or a symptom of anxiety; smoking was reported as part of other additive behaviours such as alcohol abuse. Women's views of themselves such as self-worth were often negative, shaped by experiences of neglect and abuse. Health-seeking behaviours such as accessing screening services or being aware and presenting with symptoms needed to be understood in the context of highly complex and difficult to navigate, and sometimes even obstructive, health services. Conclusion: Women in this study were at high risk of chronic diseases, including cancer. Their experiences of social disadvantage and lack of control profoundly shaped their practices, aspirations and attitudes towards risk, health and healthcare. Our findings will inform the design of a feasibility study to test a cancer prevention strategy co-designed by and tailored to vulnerable women.

University of Bristol
Background: Lycopene, plant-based diets (PBD) and physical activity (PA) have been previously associated with reduced risk and slower progression of prostate cancer (PCA), however, the potential mechanisms are not completely understood. Method: We explored the effects of the PrEvENT randomised controlled trial (RCT) with a 6-month dietary (lycopene supplementation and PBD advice) and brisk walking(BW) intervention on 155 serum metabolites in 74 men with PCA who had undergone prostatectomy, using linear regression and instrumental variable (IV) analysis. One-stage-individual-participant meta-analysis was Selected Abstracts from the 2018 NCRI Cancer Conference of National. . . performed using a subset of data from ProDiet, an RCT of men with raised PSA levels but PCA free who were randomised to lycopene supplements (n=85). The causal effect of the metabolic traits on PCA was assessed by Mendelian Randomization (MR) on 44,825 cancer cases and 27,904 controls in the PRACTICAL consortium.
Results: The effects of lycopene supplementation and PBD advice on the serum metabolic profile were comparable (R 2 =.64). There were no strong differences in metabolite levels in either the BW or the dietary intervention. After adjustment for baseline metabolites, there was evidence for decreases of triglycerides in intermediate-density lipoproteins, large, medium and small lowdensity lipoproteins and saturated fatty acids (p<0.00385) in the BW arm. When accounting for the effect of the dietary intervention on serum lycopene (IV analysis), pyruvate decreased, and acetate increased (p-value<0.05). After pooled meta-analysis with ProDiet, there was strong evidence (p-value<0.004) of decreased pyruvate and alanine, and increased acetate levels, in the lycopene arm. Using genetic instruments, the MR analysis showed evidence for a causal effect of pyruvate on PCA (OR 1.

University of Leicester
Background: Cancer chemoprevention is the use of natural or chemical agents to prevent or delay the development of cancer. However, there is variation in response to chemopreventive agents amongst individuals because it is a complex trait controlled by multiple genes and environmental factors.
Here, we aimed to decipher the genetic factors controlling the variation in response to aspirin, metformin, curcumin and eicosapentaenoic acid using a yeast-based genetic screen before translating findings to humans. This approach was feasible due to the conservation of genes between these two organisms. Method: A multiparent quantitative trait loci (QTL) mapping approach was applied. A panel of 111 F12 meiotic segregants generated from a cross of four S. cerevisiae wildtype isolates were genotyped by whole genome sequencing. Segregants were phenotyped using an automated pipeline which measured yeast growth in different chemopreventive agent treatments. Subsequently, linkage-based fine QTL mapping strategies were performed to locate regions of the genome correlating to the observed phenotype and the identification of causative genes. Results: Linkage analysis has mapped hundreds of genetic loci in the yeast genome responsible for the variation in response to the agents tested. Conserved homologues to human genes with DNA damage repair, histone modification and kinase functions were identified. Some hits have been previously supported in the literature such as the effect of aspirin and metformin on MTOR thus validating this screening approach. Novel genes identified revealed different pathways by which these agents may exhibit their anti-cancer properties. Conclusion: Detection of genetic variants influencing the differences in drug response could help identify individuals at risk or benefit of using chemoprevention agents. Current work includes validating the alleles of causative genes in human cells to assess their biological relevance. This study could aid the Method: Using in vivo mouse models we mutate APC specifically in the Lgr5+ve stem cell population of the intestine. We subsequently challenge these models with short term therapeutic strategies to influence early lesions, and assess the ability of chemotherapy to influence mutant clone establishment, as well as tumour formation and overall survival. Results: We demonstrate that APC deficient intestinal stem cells are more sensitive to chemotherapy when compared with wild type stem cells. In addition, we establish that short term therapeutic interventions, including those which cause DNA damage and those which specifically target apoptotic machinery, are able to influence tumour initiation, extend survival and reduce tumour numbers in mouse models. Furthermore, we identify that timing is crucial for therapeutic efficacy, as treatment of established tumours has no significant impact on tumour progression. Conclusion: Overall we provide evidence that chemoprevention is a tenable approach to combat CRC, and could be of significant benefit to high risk patients, such as those with FAP. Similar to our mouse models, chemoprevention could offer more tangible patient benefit than treatment of established tumours. This work highlights that early lesions are exquisitely sensitive to a variety of therapies and that mutant clones can be eliminated and readily replaced with healthy stem cells prior to tumour formation. Method: STAMPEDE is a multi-arm multi-stage platform protocol which included a randomised phase III comparison to test the above hypothesis. Standard-of-care (SOC) was lifelong ADT, with 6*3-weekly docetaxel permitted from 2016. Stratified randomisation allocated pts 1:1 to SOC or SOC+RT. Pts allocated to RT received daily (55Gy/20f over 4 weeks) or weekly (36Gy/6f over 6 weeks) RT according to investigators' pre-specified (before randomisation) choice. Randomisation was ≤12 weeks on ADT; RT started ≤8 weeks after randomisation or docetaxel. The primary outcome measure (OM) was death from any cause; secondary efficacy OMs included failure-free survival (FFS) & symptomatic local event-free survival (SLEFS). Comparison to control for survival had 90% power at 2.5% 1-sided alpha for hazard ratio (HR) of 0.75, requiring~267 control arm deaths, accounting for 3 intermediate lack-of-benefit analyses on FFS. Analyses used Cox proportional hazards & flexible parametric models, adjusted for stratification factors. Subgroup analyses will show effects by disease volume at entry and planned RT schedule. Results: 2061 pts from UK & Switzerland with newly-diagnosed M1 PCa were contemporaneously randomised to the 2 arms (Jan2013 -Sep2016). Randomised groups were well balanced: median age 68yrs; median PSA 97ng/ml; 18% early docetaxel; prespecified RT schedule: 52% daily, 48% weekly; volume 39% low, 54% high, 8% pending/NK. Conclusion: The pending results from this large randomised comparison will define the role of RT to the primary tumour in men with newly-diagnosed M1 PCa.
(CTU follows a controlled access approach for data sharing. The data are available after independent review to bona fide researchers following a formal application. They are not freely available because we have sensitive personal information. A data release request can be initiated by submitting an email to mrcctu. stampede@ucl.ac.uk or the CTU Enquiries address.) Disclosure: Funded by Cancer Research UK --dedicated grant; Medical Research Council --core trials unit funding; Contribution to the platform protocol from Astellas, Clovis, Janssen, Pfizer and Sanofi-Aventis.

13
Oncology Centre, Addenbrooke's Hospital, Cambridge, UK Background: The anti-retroviral agent, nelfinavir, demonstrates radiosensitising effects in pre-clinical models of pancreatic cancer. The primary objective of Stage 1 was to establish the maximum tolerated dose (MTD) of nelfinavir combined with capecitabinechemoradiation (CRT) after gemcitabine+nab-paclitaxel (GEMABX) induction chemotherapy. Other outcomes included overall survival and progression-free survival. Method: Patients with inoperable, histologically/cytologically proven locally advanced pancreatic cancer (LAPC) and WHO performance status 0-1 were eligible for this rolling-six doseescalation stage. After 3 cycles of induction GEMABX (28-day cycle of nab-paclitaxel 125mg/m 2 and gemcitabine 1000mg/m 2 on days 1, 8, and 15), patients with non-progressive disease had 1 further cycle followed by CRT (50.4Gy/28 fractions, capecitabine 830mg/ m 2 bd on radiotherapy days) and 1000mg or 1250mg nelfinavir bd continuously during CRT. Results: 27 patients were recruited from 8 UK centres (March 2016-June 2017). Median age was 62 years, 30% were male, 78% had head tumours, and 30% had biliary stents. Baseline median tumour diameter was 36mm. 67% commenced CRT. 11 patients received 1000mg and there was one dose-limiting toxicity (DLT) in this group: grade 3 acute coronary syndrome. The nelfinavir dose was escalated as per the rolling-six design. 7 patients received 1250mg nelfinavir and no DLTs were observed.
Survival analysis will be presented. Independant Researcher, 2 Newcaslte University, 3 Durham University Background: Platinum chemotherapy has an evolving role in the treatment of BRCA mutant and triple negative breast cancers (TNBC). Platinums (Pt) exert their cytotoxic effect by forming DNA crosslinks (Pt-DNA adducts), which impair DNA function if left unrepaired. AZD6738 is an oral inhibitor of ataxia telangiectasia and rad-3-related (ATR) protein kinase, a key component of the DNA damage response (DDR). ATR signals to prevent the onward transmission of DNA damage at cell division. Whilst the rationale for combining AZD6738 with Pt is one of potential enhanced efficacy, a greater understanding of the mechanism of AZD6738-Pt interaction, particularly with regards to the formation and repair of Pt-DNA adducts, may help guide the optimum dose and scheduling in the clinic. Method: The cytotoxicity of AZD6738 alone, and in combination with cisplatin was assessed by clonogenic survival assay in three human breast cell lines: MCF7 (BRCA1 heterozygous), HCC1806 (TNBC, BRCA wildtype) and T47D (BRCA wildtype). Pt-DNA adduct formation in MCF7 cells treated with cisplatin and AZD6738 was measured by inductively-coupled plasma mass spectrometry (ICP-MS). Results: Sensitivity to AZD6738 monotherapy varied: LC50s ranged from 0.25µM +/-0.07 in MCF7 to 0.64µM +/-0.04 in HCC1806. AZD6738 in combination with cisplatin was synergistic in MCF7 and HCC1806 cells. Neither pre-exposure, co-exposure nor post-exposure of AZD6738 in relation to cisplatin affected total numbers or longevity of Pt-DNA adducts formation in MCF7 cells. Conclusion: The synergistic action of AZD6738 with cisplatin in non-pathogenic BRCA mutant cells supports their combination in the clinic. However, more work is required to understand the mechanism of interaction. To this end, this synergistic effect is under investigation in additional cell lines. Investigation of baseline expression of key DDR proteins as a determinant of AZD6738 sensitivity is ongoing and will be presented. University of Dundee, 2 Concept Life Sciences Background: Use of molecularly-targeted agents in cancer treatment has significantly improved patient survival, but is severely constrained by toxicity and rapid onset of drug resistance. This can be attributed in part to the high potency of such drugs, which are used at doses (close to MTD), which can be several orders of magnitude > IC 50 . A key question is whether treatment regimens can be changed to retain efficacy and reduce the onset of drug resistance, greatly facilitating the development of effective combination treatments. Such studies require new animal models which more closely reflect human responses to drugs. Method: To address the limitations of mouse models in predicting drug efficacy in man, we developed transgenic mice nulled or humanised for the key P450 enzymes involved in drug disposition in man. We have characterised these models and carried out pharmacokinetic and drug/drug interaction studies with targeted anti-cancer drugs.
Results: Whereas murine Cyp2d enzymes metabolise the EGFR inhibitor osimertinib, we have shown that CYP3A4 is the major human enzyme involved, also uncovering a novel pathway of disposition involving human CYP1A1. Preliminary data using our 8HUM model -where 32 mouse P450s are replaced with six human enzymes accounting for >90% of P450-catalysed drug metabolism in man -found in vivo PK of the BRAF inhibitor dabrafenib to closely match that in patients, and that induction or inhibition of CYP3A4, the major human P450 involved in dabrafenib disposition, significantly alters dabrafenib PK. We also demonstrated that dabrafenib is a potent CYP3A4 inducer which will markedly affect its own disposition and that of comedications. Institute of Cancer Research Background: Mitomycin C (MMC) chemotherapy has a welldefined safety profile and is used in the treatment of intermediate and high risk NMIBC. CALIBER aimed to demonstrate that intravesical MMC (chemoablation) had sufficient activity to warrant further investigation as an alternative to surgery for low risk NMIBC. Method: CALIBER has a Simon two-stage design, incorporating a surgical control group to test feasibility of randomisation. Patients with recurrent low risk NMIBC were randomised 2:1 to chemoablation (4x 40mg weekly MMC) vs. surgery (standard of care). The primary endpoint was complete response (CR) to chemoablation by visual assessment and histological biopsy at 3 months posttreatment, aiming at excluding CR rate <45% (Stage 1). Quality of life (QoL) at 3, 6 and 12 months post-treatment was measured using EORTC QLQ-C30 and QLQ-NMIBC24. Results: 82 participants (54 chemoablation, 28 surgery) were recruited from 37 UK centres (28/01/2015 to 04/09/2017). Feasibility of randomisation was demonstrated with acceptance rates of 55%. Stage 1 CR criterion was not met and the trial closed to recruitment in September 2017. Estimated 3 month CR rate is 37.0% (95% CI: 24.3, 51.3) in the chemoablation group and 80.8% (95% CI: 60.6, 93.4) in the surgery group. No grade 3-4 toxicities were reported in either group. 70/78 (89.7%) patients participated in the optional QoL sub-study and completed baseline questionnaires (49 chemoablation, 21 surgery). 6months mean change from baseline in QLQ-C30 global health status was +1.5 surgery vs -4.8 chemoablation (no statistically significant difference). Conclusion: Chemoablation with MMC is safe, however it did not meet pre-specified activity levels to pursue further investigation. Whilst surgery appears more effective in this setting, the proportion of patients with residual disease at 3 months suggests surgery alone may be suboptimal. Further research is required to determine the role of chemoablation with other agents in patients with low risk NMIBC. Cambridge University Hospitals NHS Foundation Trust, 19 Tata Memorial Centre Background: Pre-clinical, observational and randomised evidence suggests aspirin may prevent or delay the development of cancer and metastases, and is strongest for colorectal and gastrooesophageal cancer. However, concerns around feasibility, adherence and tolerability (particularly serious bleeding) have limited aspirin use for cancer chemoprevention. Method: Add-Aspirin is a double-blind, randomised-controlled trial encompassing 4 individually powered phase III trials in earlystage breast, prostate, colorectal and gastro-oesophageal cancer, Selected Abstracts from the 2018 NCRI Cancer Conference of National. . . evaluating the effect of aspirin after radical therapy. All participants initially take open-label aspirin (100mg daily) for 8 weeks (run-in), to assess adherence and toxicity prior to randomisation (1:1:1, aspirin 300mg, aspirin 100mg or matched placebo for ≥5 years). A pre-planned feasibility analysis was performed to assess tolerability and adherence when >2000 participants had completed the run-in period. Results: Between October 2015 and October 2017, 3494 of a targeted 11000 participants were registered from 165 sites in the UK; recruitment rates differed across tumour sites compared to predictions. Run-in data (n=2253) showed good adherence: 95% took 6-7 tablets/week and 85% proceeded to randomisation, with rates consistent across tumour cohorts. Main reasons for not proceeding to randomisation were toxicity (mostly minor, grade 1/ 2) and/or patient choice, with only 0.7% (16/2253) of participants experiencing toxicity requiring discontinuation during the run-in. Fewer than 1% (13/2253) experienced a grade > 3 toxicity (including one lower gastrointestinal bleed in a prostate cancer participant; no upper gastrointestinal bleeds). Conclusion: The data demonstrate that aspirin is well-tolerated over an 8-week run-in, and acceptable to patients after radical cancer therapy, with low toxicity rates in all tumour cohorts, including gastro-oesophageal participants. A run-in approach may be useful in adjuvant (or prevention) studies for reducing the risk of non-adherence and participant attrition at a later date.  ) is a member of the matrix metalloproteinase (MMP) family involved in tissue remodelling through proteolysis of extracellular matrix components. Overexpression of MT1-MMP is seen in multiple tumour types including non-small cell lung cancer (NSCLC), triple negative breast cancer (TNBC) and sarcoma. BT1718 is a novel first in class bicyclic targeting peptide that binds MT1-MMP and is linked to the maytansinoid tubulin inhibitor DM1 by a cleavable disulfide linker. Bicyclic peptides have a low molecular weight in comparison to other conjugated toxin approaches, enabling rapid penetration and a short systemic half-life, potentially reducing toxicity. Method: This is an open label first in human phase I/IIa study. The primary objective is to propose a recommended phase 2 dose (RP2D) and schedule of BT1718. Secondary objectives include pharmacokinetic (PK) parameters and preliminary clinical responses in biomarker pre-defined cohorts of patients. Tertiary objectives include correlative studies related to predictive biomarkers of response.
Dose escalation (phase I) Stage 1: Exploring an initial twice-a-week schedule. There will be single patient cohorts until either Grade 2 drug related toxicity or dose exceeding 6 mg/m 2 twice a week. A 3+3 design will then be followed until the RP2D.
Stage 2: Exploring a once-weekly schedule, using a 3+3 design until the RP2D for this schedule has also been established.
Dose expansion (phase IIa) Part A: 14 patients with MT1-MMP expressing NSCLC or TNBC, treated with BT1718 at the twice-a-week RP2D. At least 6 will have pre-and post-treatment biopsies.
Part B: 14 patients with MT1-MMP expressing NSCLC or TNBC, treated with BT1718 at the once-weekly RP2D. At least 6 will have pre-and post-treatment biopsies.
Part C/D: following parts A & B, a decision will be made to explore the selected schedule in tumour-specific cohort(s) of around 15 patients, with refined MT1-MMP biomarker selection. Background: Osteoporosis in patients with cancer is common and considered multi-factorial in aetiology. A high frequency of symptomatic fractures was observed in patients treated with chemotherapy and cetuximab at our institution. We therefore performed a retrospective analysis of patients with metastatic colorectal cancer treated with cetuximab in combination with chemotherapy to determine the incidence and potential risk factors for the development of fractures. Method: Consecutive patients treated with fluorouracil, folinic acid and irinotecan in combination with cetuximab (FOLFIRIcetuximab) at the Edinburgh Cancer Centre were retrospectively analysed. Baseline characteristics, number treatment cycles and cumulative steroid dose were collected. FRAX scores were used to calculate 10-year probability of major fracture. Fractures were assessed by reviewing serial CT scan reports during treatment. Patients treated with capecitabine and oxaliplatin (CAPOX) were chosen as a comparison group. Results: 31 patients treated with FOLFIRI-cetuximab and 30 patients treated with CAPOX were reviewed. Patients received a median of 9 two weekly cycles of FOLFIRI-cetuximab (range 1-32) and 6 three weekly cycles of CAPOX (range 1-18). One patient in the FOLFIRI-cetuximab group and no patients in the CAPOX group had a fracture documented prior to treatment. Median FRAX score in the FOLFIRI-cetuximab group was 8%. Ten patients developed new non-metastatic fractures in the FOLFIRI-cetuximab group (28%), 90% were symptomatic. One patient in the CAPOX group (3%) developed a new vertebral fracture. Median cumulative steroid dose for the patients in the FOLFIRI-cetuximab group was dexamethasone 306mg and for the CAPOX group 204mg. Conclusion: Patients with metastatic colorectal cancer treated with FOLFIRI-cetuximab demonstrate a high incidence of symptomatic fractures with significant morbidity. The high incidence may be related to higher steroid exposure although an effect of the chemotherapy regime cannot be excluded. The introduction of bone protection and changes to anti-emetic protocols to reduce cumulative steroid dose should be considered. Cancer Research UK, 2 University of Birmingham Background: 36% of cancer diagnoses in 2015 were in people 75 and over. By 2035, this will rise to 46%. Cancer services in the UK must meet the needs of the patients they serve and adapt to a changing population.
However, cancer survival is generally lower for older patients, older patients are also more likely to be diagnosed in an emergency and are less likely to receive many different treatments.
The Cancer Strategy highlighted that current methods of assessing older patients are not fit for purpose, meaning older people's needs are not considered.
CRUK commissioned the University of Birmingham to identify solutions for improving the quality of decision-making. Method: A literature review to understand the current evidence base : (1) 15 qualitative interviews with national policymakers, 80 interviews with health professionals. (2) Three UK surveys of primary and secondary care health professionals and older patients.
(3)Clinical observations of MDT meetings and multidisciplinary clinics. Results: Older patients have more complex medical and social needs, however the most appropriate support is not always available; including social care, or Clinical Nurse Specialists. Health professionals are likely to include a wide range of clinical factors in treatment decision-making for older patients, but Comprehensive Geriatric Assessments (CGA) are not used consistently in primary or secondary care. Systemic issues also prevent the right information being included in decisions, including with links between primary and secondary care, or links into cancer MDTs. This group will also be hardest hit by wider pressures -a lack of time for consultations, and workforce shortages. Older patients are also typically under-represented in clinical trials, meaning there is limited evidence to support treatment choices. Conclusion: Cancer services are not meeting the needs of older patients and without mitigation, this will worsen. Addressing this is vital if we are to achieve our ambition of world-class cancer outcomes in the UK. Royal Marsden Hospital, 14 Southmead Hospital, 15 Princess Alexandra Hospital NHS Strust, 16 Imperial health and Imperial College London (ICL) Background: Focal HIFU is an emerging minimally invasive treatment option for localised prostate cancer. Phase II clinical trials and prospective databases have shown promising oncological outcomes. However, there is a paucity of data comparing HIFU to standard of care therapies such as radical prostatectomy. Method: All consecutive men undergoing either focal HIFU (n = 625) or laparoscopic radical prostatectomy (LRP) (n = 571) between 2007 -2017 had their pre-and post-operative data collected in prospective databases.
A propensity score was constructed with groups matched 1:1 using nearest neighbour matching based on the propensity score.
Primary outcome was Failure Free Survival (FFS) defined as transition to salvage or systemic therapy. Up to two focal-HIFUs was part of this intervention as defined in NCRN PART. Results: After inclusion/exclusion criteria, 425 HIFU and 194 LRP patients remained with matching leading to 86 in each cohort. At all-time points, there was no significant difference in FFS (p = 0.29). The actuarial FFS for those undergoing HIFU was 91% at 4-years and 86% for those undergoing LRP Figure 1 and Patients with favourable clinical and social circumstances and a low-risk MASCC score were discharged with oral antibiotics. An electronic register was used to study all emergency presentations of possible NS between April 2016 and April 2018 to Acute Oncology (AOS), Medical Admissions (MAU) and A&E.
Results: In total, 242 patients presented with potential NS. 170 attended AOS and 72 attended MAU/A&E. The most common primary tumours were breast (64%), lung (13%) and prostate (9.6%). The most common regimens were carboplatin and docetaxel (72%) and FEC (17%). Among AOS attendees 83 patients (49%) were discharged, all of whom had a low-risk MASCC score. Within these 83 patients, median age was 60, and 64 patients (77%) were female. 11 patients (13%) post-discharge were re-admitted within a week; none of these patients died or were admitted for more than one week. Significantly, those patients who were re-admitted had a lower MASCC score at the time of assessment (22 vs 24; p = 0.0028). Among non-AOS attendees, only 2 patients (2.7%) were not admitted suggesting that early discharge using MASCC score had not yet been adopted outside of AOS. Conclusion: MASCC score accurately identifies patients at risk of NS who can be safely managed at home. Re-admission of these patients is more likely if the absolute MASCC score nears 21 and this should inform counseling and follow-up in this patient group. Disclosure: None declared Corresponding author: Ahmed Eltinay 128. Structure-based virtual screening identifies pranlukast as a CD49f antagonist that reduces stemness in MDA-MB-231 breast cancer cells Marco Velasco-Velazquez 1 , Inés Velázquez-Quesada 1 , Angel Ruiz-Moreno 1 , Diana Casique-Aguirre 1 , Charmina Aguirre-Alvarado 1 , Fabiola Cortés-Mendoza 1 , Marisol de la Fuente-Granada 1 , Mayra Pérez-Tapia 2 , Aliesha Gonzalez-Arenas 1 , Aldo Segura-Cabrera 3 1 National Autonomous University of Mexico, 2 National Polytechnic Institute, 3 The European Bioinformatics Institute (EMBL-EBI) Background: Breast cancer is the neoplastic disease with higher mobility and mortality in women worldwide. The Cancer Stem Cell (CSC) model proposes that a subpopulation at the top of the tumor cell hierarchy drives the initiation, maintenance, and therapy response in breast cancer. CD49f is an integrin subunit that is expressed in breast CSCs and promotes maintenance of stemness. Thus, blockade of CD49f is a potential therapeutic approach for targeting breast CSCs. Method: To repurpose drugs as CD49f antagonists, we performed consensus molecular docking using a subdomain critical for heterodimerization and a collection of 13,000+ drugs. Five drugs were selected for in vitro biological validation using MDA-MB-231 cells. We performed limiting-dilution xenotransplantation using the drug with better CSC-selectivity. Results: Pranlukast, a cysteinyl leukotriene receptor 1 (CysLTR) antagonist that is used to treat asthma, inhibited cell adhesion to laminin and decreased the mammosphere-forming efficiency but had no impact on the viability of bulk tumor cells. Short exposure of breast cancer cells to pranlukast reduced CD49f-downstream signaling, including Focal Adhesion Kinase (FAK) and phosphatidylinositol 3-kinase (PI3K) activation. Pranlukast-treated cells showed reduced transactivation of Sox2 promoter and decreased tumorigenicity in vivo, indicating that this drug decreases the number of CSCs. Conclusion: Pranlukast antagonizes CD49f, impairing CSCassociated functions. Since the pharmacokinetics and toxicology of pranlukast are known, it is a potential adjuvant therapy for breast cancer patients.
Disclosure: This work was supported by CONACYT 221103 and PAPIIT IN228616 Corresponding author: Marco Velasco-Velazquez 129. A systematic review of eligibility criteria for phase II clinical trials for patients with lung cancer Gary Doherty 1 , Innocent Ogunmwonyi 2 , Rohan Shotton 1   1 Cambridge University Hospitals NHS Foundation Trust, 2 University of Cambridge Background: Significant barriers exist for the successful enrolment of Oncology patients in early phase clinical trials. These include restrictive eligibility criteria that often exclude patients with intracranial disease, abnormal organ function or laboratory tests, and particular co-morbidities. To investigate these further, we performed a systematic review of the eligibility criteria in recent phase II clinical trials for patients with lung cancer. Method: We searched the clinicaltrials.gov database for trials registered in 2016 using the criteria "lung cancer" and "phase 2." 223 trials were identified. Duplicate/incomplete entries, and those not investigating systemic anti-cancer agents, were excluded. Eligibility criteria were then determined for the remaining 173 trials. Results were matched against known pharmacokinetic/ toxicity data for the relevant agents, trial outcomes, sponsorship type, and country. Further analyses were performed to determine temporal trends. Results: Common eligibility criteria were highly variable. Highly significant variability existed for the exclusion of patients with liver function derangement (used in 79.2%; 12.7% even without known hepatic drug metabolism/hepatotoxicity), renal function derangement (used in 79.2%; 14.5% even without known renal drug clearance/nephrotoxicity), low blood cell/haemoglobin counts (used in 78.6%), or coagulation abnormalities (used in 16.8%). Wide ranges of cut-offs and methodologies were used to determine eligibility. While 89.6% of trials permitted patients with brain metastases, 41.2% of these excluded patients taking concomitant corticosteroids. Successfully suppressed HIV/HepB/ HepC was permitted in 41.6, 82.7% and 81.5%, respectively. Comprehensive eligibility criteria analyses, pairwise correlations with drug metabolism/toxicities, temporal trends, and correlation between eligibility restrictiveness and trial outcomes, will be presented.

Conclusion:
The safety of patients participating in clinical trials is paramount. However, our results suggest that the significant variability in eligibility criteria cannot be wholly explained by a priori knowledge of the investigational agents' metabolism and known toxicities, which may lead to irrational and unjustifiable exclusion of patients in a high-need population from access to experimental therapies. measurements. These are increasingly determined using isotope dilution mass spectrometry (IDMS) assays. We present validation of the CamGFR model that we developed originally with non-IDMS data (JCO, 2017) and a validated extension using IDMS creatinine data.
Method: The study was approved by the relevant ethics committees. Data on age, sex, height, weight, serum creatinine, and results for GFR from 51 Cr-EDTA excretion measurements were obtained from adult patients with cancer from one Swedish and two UK centres. Data were split 4:1 into development and validation datasets. For IDMS data, we refitted the CamGFR model using interaction terms between all creatinine terms and the creatinine measurement type. We assessed bias, accuracy, and precision for GFR using median residuals, root-mean-squared-error (RMSE), and residual interquartile range (IQR Overexpression of Glo1 induces anti-cancer multidrug resistance (MDR) in human tumour cells lines and human tumour cells in primary culture. We hypothesize that the cytotoxicity of anticancer drugs is mediated, in part, by inducing increase of MG to cytotoxic levels. This may be achieved by druginduced increased MG formation and/or decreased MG metabolism; the latter achieved by drug-induced direct or indirect Glo1 inhibition.
Method: Using a model human tumour cell line, HEK293 cells were stably vector-transfected to overexpress Glo1 and with empty vector as control. The effect of anticancer drugs on growth and toxicity of HEK293 cells in three conditions (wild-type, Glo1ioverexpression and empty vector) was studied in vitro and median growth inhibitory concentration (GC50) values determined. The effect of cell permeable Glo1 inhibitor, S-pbromobenzylglutathione cyclopentyl diester (BBGCp2) on the potency of anticancer drugs was also studied. The glyoxalase system and dicarbonyl metabolism were characterised by measuring cellular activities of Glo1, Glo2, MG reductase and MG dehydrogenase. The flux of formation of D-lactatea surrogate indicator of flux of MG formation, glucose consumption and net L-Lactate formation were measured in HEK293 cells cultures by endpoint enzymatic assays. Results: Doxorubicin, mitomycin C, paclitaxel, mechlorethamine and methotrexate had the highest resistance conferred by Glo1 overexpression: MDR was 16-fold, 15-fold, 8-fold, 7-fold and 7-fold, respectively. BBGCp2 potentiated the cytotoxicity of anti-cancer drugs. There was an increase in flux of formation of D-lactate and L-lactate and consumption of D-glucose in HEK293 cells treated with mechlorethamine, doxorubicin, paclitaxel and methotrexate, compared to untreated cells. However, cells treated with mitomycin C had a decrease and increase in D-lactate and Llactate formation respectively.  Leeds,20 The Royal Marsden NHS Foundation Trust, London, 21 Gustave Roussy Institute, Villejuif Background: Overall survival (OS) is considered the gold standard endpoint for controlled clinical trials but it requires extended follow-up (median OS> 40 months for first line therapy) and large sample sizes. The UK contributed 3 trials to this Gynaecological Cancer Intergroup (GCIG) meta-analysis. The objective was to evaluate whether progression free survival (PFS) based on CA125 measurements confirmed by radiological exam or combined GCIG criteria is a surrogate endpoint for OS in advanced ovarian cancer (AOC). Method: Using the meta-analytic approach on trials published after 2000, correlations between PFS and OS at the individual level, and between treatment effects on PFS and on OS at the trial level, were estimated using Kendall' tau and copula bivariate (R 2 Copula ) models respectively. Criteria for PFS surrogacy required R 2 Copula ≥ 0.80. Results: We analyzed individual patient data (IPD) from 10,502 patients in 16 randomized first line trials of standard (n = 7), intensification (n = 5) and maintenance (n = 4) . No heterogeneity in the treatment effects across trials was detected. High correlations were found at the individual level (tau = 0.77) but low correlation at the trial level (R 2 Copula = 0.2). Sub-group analyses led to similar results (see  (PC) succumb to the disease despite curative resection, many of whom recur within 6 months of surgery. This suggests current staging is inadequate and there is a need to better define the biology and clinical behaviour of PC prior to surgery. This study aimed to develop and validate gene signature sets and biomarkers that accurately predicts disease recurrence patterns in patients undergoing resection for PC.
Method: Disease patterns were defined as early recurrence after surgery (< 12 months), liver metastases, lung metastases (no evidence of liver recurrence) and local recurrence only. The molecular features of clinical disease patterns were investigated using transcriptomic analysis and immunohistochemistry (IHC). These were correlated with recurrence patterns, disease presentation (localised, locally advanced and metastatic) and molecular subtypes of PC. Results: Early recurrence, liver metastases and metastatic presentation were strongly associated with gene expression sets that define the squamous subtype of PC (P < 0.001). Lung recurrence, localised disease and long-term survival were associated with the classical pancreatic subtype and an anti-tumour immune response (P < 0.001). High S100A2 and S100A4 IHC expression was associated with the squamous subtype of PC and were independent poor prognostic factors in 3 independent patient cohorts of PC (totalling n = 1184 patients). Using these biomarkers, a molecular prognostic nomogram was generated to identify poor prognostic PC and validated using 3 independent cohorts of PC. Conclusion: Gene signature and biomarker expression sets generated from biological relevant processes and known poor prognostic features of PC can allow accurate prognostication of patients with operable PC. Defining poor prognosis can allow stratification towards systemic neoadjuvant therapy in operable disease. Conversely, if a patient is predicted to have a favourable prognosis, more aggressive and extensive surgery in the setting of borderline resectable or locally advanced disease could be justified.
Disclosure: Funded by Cancer Research UK Corresponding author: Stephan Dreyer 135. Improving patient experience of immunotherapy treatment for melanoma: the Leeds nurse-led immunotherapy telephone clinic Maria Marples, Helen Jackson, Helen Nicholson, Beverley Ryder, Karen Ingham, Jane Hook Leeds Cancer Centre Background: Immunotherapy has transformed the outcomes of metastatic melanoma, but the burden on patients of frequent hospital visits for assessments is high. We devised, implemented, developed and evaluated a nurse-led telephone clinic to assess patients being treated with immunotherapy. Method: We piloted a questionnaire, enquiring about a range of potential immunotherapy toxicities, as well as symptoms of progressive disease and medication changes. This formed the basis of a telephone clinic starting in November 2016, where patients who were established on pembrolizumab immunotherapy were telephoned by a clinical nurse specialist up to 1 week prior to treatment. Outcomes were emailed to a consultant for a prescription, blood tests were done locally, and annotations saved in the electronic patient record. In 2017, patients being treated with nivolumab were added to the clinics. A separate code was created for patients being monitored for immunotherapy toxicity (mainly hepatitis). We evaluated the clinic with a patient satisfaction questionnaire and a detailed review of 6 months' clinical activity. Results: In the first 6 months of the clinic, monthly assessments increased from 14 to 49. Patient satisfaction was high, with 98% patients reporting that the clinic was convenient and saved them time, and 100% that concerns were addressed and they and their GP had the necessary information. Review of the work done in the subsequent 6 months showed that 50% of calls resulted in no additional action, around 40% required telephone advice, and only 4% required medical review that week. Conclusion: Patients on immunotherapy can be safely and effectively assessed on the telephone, which is well-received by patients, and saves transport and clinic costs. This approach is applicable to other tumour types which are treated with Addenbrooke's Hospital, Cambridge Background: Adjuvant trastuzumab has significantly improved outcomes for HER2 positive early breast cancer patients, using the 12 month duration empirically adopted from the pivotal registration trials. Given an annual per patient cost of trastuzumab treatment of over £30,000 (Euro35,000), a shorter duration has the potential to improve cost-effectiveness if efficacy is maintained. Method: The within trial cost effectiveness analyses uses data collected as part of the PERSEPHONE trial, a randomised phase 3 non-inferiority trial comparing 6 to 12 month trastuzumab, the largest reduced-duration non-inferiority trial internationally. Comprehensive health service activity and costs were collected. Quality of life was measured using the EQ-5D. The analyses adopted the perspective of the health and social care sector over 2 years follow-up. Incremental cost effectiveness ratios (ICER) present the cost per quality adjusted life year (QALY). Additional analysis presents the ICER using the primary outcome of the trial, disease free survival at 4-years. Uncertainty around the ICER is estimated using the non-parametric bootstrap method. Results: 4088 patients were randomised between 4th October 2007 and 31st July 2015. The results of the cost effectiveness analyses will be presented.

Conclusions:
The cost effectiveness analyses will provide evidence to demonstrate the value for money of 6 versus 12m trastuzumab. The results will inform decisions around reduction of standard trastuzumab duration to 6 month. Disclosure: Funded by NIHR HTA Corresponding author: Peter Hall 137. Cancer Medicines Outcomes Programme (CMOP): Better use of existing data to understand outcomes in a local population Jennifer Laskey 1 , Julie Clarke 1 , Kelly Baillie 1 , Yvonne Semple 1 , Olivia Wu 2 , Christine Crearie 1 , Rob Jones 1 , Ashita Waterston 1 , Tanja Mueller 3 , Jiafeng Pan 3 , Marion Bennie 3 1 NHS Greater Glasgow and Clyde, 2 University of Glasgow, 3 University of Strathclyde Background: Cancer medicines comprise the highest proportion of new medicines introduced within NHS Scotland. There is increasing interest in 'real world' data to understand local population outcomes. The Cancer Medicines Outcomes Programme (CMOP) aims to maximise use of routinely collected healthcare data to determine outcomes in the local population.
Method: Melanoma and prostate cancer were year 1 exemplar projects. Quality and quantity of data available from electronic record linkage in a safe haven were compared to individual case note review. Outcomes were compared with trial data. Results: Record linkage data appeared broadly comparable to individual case note review (