Oncological outcomes, quality of life outcomes and complications of partial cystectomy for selected cases of muscle-invasive bladder cancer

To evaluate the oncological results, associated complications, and postoperative health-related quality of life (HR-QoL) in patients treated with partial cystectomy (PC) for muscle-invasive bladder cancer (MIBC). 27 patients who underwent open PC for cT2 MIBC were included. A simple Cox’s proportional hazards regression model was used to assess the association of several potential prognostic factors with survival. Postoperative HR-QoL was assessed with the EORTC (European Organisation for the Research and Treatment of Cancer) QLQ-C30 questionnaire version 3.0. Final pathological tumour stages in PC specimen were: pT0: 18.5%, non-MIBC: 3.7%, MIBC: 74.1%, pCIS: 14.8%. Estimated 5-year overall- and progression-free survival rates were 53.7% and 62.1%. Five (18.5%) patients experienced local recurrence with MIBC. Overall, the salvage cystectomy rate was 18.5%. The 90-day mortality rate was 0%. Significant risk factors for progression-free survival were vascular invasion (HR 5.33) and tumour multilocularity (HR 4.5) in the PC specimen, and a ureteric reimplantation during PC (HR 4.53). The rates of intraoperative complications, 30- and 90-day major complications were 7.4%, respectively and 14.8% for overall long-term complications. Postoperatively, median (IQR) global health status and QoL in our PC cohort was 79.2 (52.1–97.9). Open PC can provide adequate cancer control of MIBC with good HR-QoL in highly selected cases. Open PC can lead to long-term bladder preservation and shows an acceptable rate of severe perioperative complications, even in highly comorbid patients.

Outcome measures. Surgery related complications: Postoperative complications were classified according to  . A score ≥3 was considered a major complication. Early and late postoperative complications were defined as complications before or after 90 days following PC.
Oncological outcomes: Patients were followed according to the EAU guidelines for patients with MIBC following RC, with some risk-adapted modifications in selected individuals. Additionally, patients underwent cystoscopy and urinary cytology at 3 months post PC and subsequently every 3 months for a period of 2 years, and every 6 months thereafter until 5 years, and then yearly following the EAU guidelines for non MIBCs with high risk of recurrence and progression 12 .
Survival rates: Overall survival (OS; time between PC and last follow-up or death), cancer-specific survival (CSS; time between PC and BC related death), progression-free survival (PFS; time between PC and first tumour progression, defined as local MIBC recurrence and/or metachronous metastasis and/or BC related death), local recurrence-free survival (RFS; time between PC and first local recurrence of BC, irrespective of whether superficial urothelial cancer [non-MIBC ≈ pTa, pT1, pCis] or MIBC), local MIBC-recurrence-free survival (MRFS; time between PC and first local recurrence of MIBC), and metastasis-free survival (MFS; time between PC and first detection of a metachronous metastasis) rates were assessed.
Health-related quality of life (HR-QoL) outcomes: Postoperative HR-QoL was assessed with the cancer-specific EORTC (European Organisation for the Research and Treatment of Cancer) QLQ-C30 questionnaire version 3.0 13 in 12 patients who completed the questionnaire independently at a median of 4.6 (2.8-7.7) years postoperatively. The QLQ-C30 questionnaires were analysed according to the instructions in the EORTC scoring manual 14 . HR-QoL data of the 12 PC patients were contrasted with EORTC-QLQ-C30 questionnaire data of 58 patients treated with RC and urinary diversion (ileal conduit: n = 24, orthotopic ileal neobladder: n = 34) at the Charité between 1993 and 2007, as published previously 15 .
Additionally, following data was assessed for each patient: patient demographics, prior history of BC management, tumour characteristics of the PC specimen, surgical management, adjuvant therapies, and the management of tumour recurrence and/or progression. All patients were staged according to the 2009  data is presented as mean with standard deviation (SD). An independent samples t-test analysis (in case of normally distributed data) and a Mann-Whitney U-test of differences (in case non-normally distributed data) were used to compare differences between HR-QoL outcomes of the 12 PC-patients in this study and the 58 RC-patients in a previously published study 15 . The 5-year survival rates for OS, CSS, PFS, RFS, MRFS, and MFS were estimated using Kaplan-Meier analysis. Simple Cox's proportional hazards regression model was used to test the significance of several potential prognostic factors associated with OS, CSS, PFS, and RFS. Due to the limited number of events, a multivariate regression model was not meaningful. All tests were performed at a significance level of α = 0.05. Surgical technique. Surgery was performed with an exptraperitoneal technique, as previsously described 16 .
Pelvic lymphadenectomy (PLND) was carried out including the external iliac nodes, the obturatoric nodes, the internal iliac nodes, and the presacral nodes 17 . The bladder tumours were removed with a 1-2 cm safety margin, which resulted in ureteric reimplantation in some cases. Negative surgical margins were confirmed by separate intraoperative frozen sections of the resection margin or by intraoperative assessment of the full PC specimen.
Data availability. The datasets generated and analysed during the current study are not publicly available due to relevant data protection formalities but are available from the corresponding author J.E. on reasonable request. Table 1 lists patient characteristics and comorbidities (Table 1A), prior histories of urothelial carcinoma (Table 1B), tumour characteristics of the PC specimen (Table 1C), surgical characteristics (Table 1D), and hospitalization details (Table 1E). Significant comorbidities were identified throughout the cohort. Overall, 63.0% of the patients had a Charlson comorbidity grade ≥3, and 48.1% were considered severe comorbidity in the Adult Comorbidity Evaluation-27 (ACE-27) score (Table 1A). No patient received NAC due to contraindications or patients' lack of willingness. The tumour stage of the PC specimen showed pathological up-staging to pT3 disease in 12 (44.1%) patients and down-staging in 6 (22.2%) patients, with 5 (18.5%) patients being staged as pT0. Four (14.8%) patients had CIS on histology of the PC specimen (Table 1C). A median number of 8 (3-10) lymph nodes were identified with unilateral PLND, performed on the tumour side in 8 (29.6%) cases, and 12 (9-17) lymph nodes when bilateral PLND was performed in 17 (63.0%) cases (Table 1D). Overall, 6 (22.2%) patients were found to have regional lymph node metastases. One (3.7%) patient suffered from an incidental subcutaneous metastasis in the abdominal wall, which was treated with local resection during PC. A transperitoneal approach to PC was only necessary in 1 (3.7%) patient. The most common tumour locations were the lateral walls (48.1%), the bladder dome (37.0%), and the posterior wall (22.2%). Overall, 9 (33.3%) tumours were located close to the ureteric orifices, requiring ureter reimplantation in all 9 cases to remove tumour entirely. The median maximum tumour diameter was 30 (20-46) mm, with 59.3% of the tumours being solitary and 40.7% multilocular (Table 1C). Surgical and hospitalization details are summarized in Table 1D and E. Intra-and postoperative complications. The intra-and postoperative complications are shown in Table 2. The intraoperative complication rate was 7.4% due to severe bleeding necessitating blood transfusion in 2 patients (Table 2A). The 30-and 90-day postoperative complication rate was 55.6% and 59.3%, respectively, with Clavien grade I and grade II being the main complication grades. Major complications were observed in 7.4% of the patients due to wound healing issues and/or wound infection, necessitating reintervention under general anaesthesia. We identified 12 grade I complications in 10 patients, 11 grade II complications in 8 patients, and 2 grade IIIb complications in 2 patients within the first 30 days postoperatively. One additional grade II complication occurred between day 30 and 90. An overview of the 30-and 90-day complications is given in Table 2B. The late complication rate was 14.8%. Hydronephrosis was the most common late complication, appearing in 3 (11.1%) patients. In all 3 patients the reason for hydronephrosis was surgery-related but, in only 1 case related to surgical reimplantation of the ureter (Table 2C).

Results
Oncological follow-up, adjuvant and palliative treatment, and outcomes. The oncological follow-up, adjuvant and palliative treatments, and the outcome data are summarized in Table 3. Follow-up data was available for all patients. Overall, median follow-up time was 36.5 (23.3-78.8) months and 51.7 (33.9-90.9) months for survivors. Overall, 12 (44.4%) patients continued to be cancer-free, including all pT0 patients (18.5%), with a median disease-free time of 56.2 (36.1-90.9) months. Nine (36.0%) patients died from cancer, and the overall mortality rate was 55.5% (Table 3A). During the follow-up period, 22 (81.5%) patients preserved their bladder. Figure 1 shows a chart of patients who developed local tumour recurrence. Table 3B gives additional information on tumour stages in first local recurrence. There was no evidence of local tumour seeding with pelvic recurrences outside the bladder post PC. Altogether 10 (37.0%) patients developed metastasis after a median of 8.9 (4.8-27.7) months (Table 3C). Sixteen (59.3%) patients remained free from any progression. For the 11 (40.7%) patients who had tumour progression, the median progression free time was 36.1 (11.8-73.8) months (Table 3A). Patients with metastatic disease were treated with palliative systemic chemotherapy in 4 (14.8%) cases, with local radiation therapy in 2 (7.4%) cases, and with combined radiochemotherapy in 1 (3.7%) case. Three (11.1%) patients with metastases underwent additional local resection (Table 3D). The imaging performed during the course of the oncological follow-up is shown in Table 3E.
Survival rates. Estimated    subgroup staged pT0 -pT1 in the PC specimen, 56.3% (45.0%) for the subgroup staged pT2 in the PC specimen and 58.3% (40.0%) for the subgroup staged >pT2 in the PC specimen, log rank p = 0.21 (log rank p = 0.09) (Fig. 2b,c). Simple Cox's proportional hazards regression model identified MIBC tumour stage and vascular invasion of the PC specimen as well as the development of metachronous metastases to be significantly related to OS and CSS (Table 4A and B). Local recurrence of MIBC was found to constitute an additional risk factor for CSS (Table 4B) and borderline significance for OS (Table 4A). In addition, ureter reimplantation during PC was marginally significantly related to CSS. Variables significantly associated with PFS were MIBC tumour stage and vascular invasion of the PC specimen, multilocularity of the tumour, and reimplantation of the ureter during PC (Table 4C). Local RFS was significantly related to MIBC tumour stage of the PC specimen as well as to the time from primary diagnosis of BC to PC (Table 4 D ).

Discussion
Historical PC series published in the 1970's reported high rates of local recurrence and poor oncological outcome [18][19][20][21] . However, PC has experienced some resurgence as a less morbid and oncologically effective treatment 6 . With more stringent selection criteria that limits PC to patients with a solitary primary MIBC-tumour, without concomitant CIS or a history of non-MIBC, and feasible for full-wall resection with an adequate safety margin, 5-year OS rates of 53-70% and RFS rates of 40-64% have been reported in more recent series 7-10,22-24 . However, it is estimated that only 3-10% of patients with MIBC meet these criteria 9,25-27 . During the last ten years, the utilization rate of PC has remained stable in the United States of America, accounting for 7%-10% of all cystectomies performed 28 . European data on PC rates is currently limited. Taking into account that only 52.5% of patients with MIBC received aggressive treatment therapy in the series (2004-2008) by Gray et al. 28 , only 2.8% of patients with MIBC underwent a PC. In this series, elderly patients aged 81-90 years were the group that had the lowest rate of aggressive treatment therapies, with only 35% receiving aggressive therapy defined as RC or PC or definitive radiotherapy (RT)/chemoradiotherapy. It is this group that might benefit the most from a less morbid approach. Thus, these numbers suggest that in general, all aggressive therapies for BC including PC are underutilized, although older data (1988)(1989)(1990)(1991)(1992)(1993)(1994)(1995)(1996)(1997)(1998)(1999)(2000) from the Surveillance, Epidemiology and End Results (SEER) registry and the Nationwide Inpatient Sample (NIS) showed a PC rate of 13-17% disproportionately used in certain medical centres (nonteaching, rural, low volume) and patient populations (elderly, black, females, stage I disease) reflecting selective referral or overuse 27 .
However, in our series PC was mainly performed in patients with a high Charlson Comorbidity Grade or ACE-27 Score, taking into consideration that complications related to RC may be directly related to pre-existing comorbidity as well as to bowel anastomosis or urinary diversion [3][4][5] , and that comorbidity was identified to be an independent prognostic factor for perioperative mortality, overall mortality, and cancer-specific mortality in RC patients [3][4][5] . Thus, the selection of patients in our series differed from the stringent selection criteria for PC since patients with multifocal disease, the necessity to reimplant the ureter, and a prior history of non-MIBC were   (Table 6). These differences may reflect differences in patient selection, patient numbers, age, extend of pelvic lymph node dissection, and impact of neo-adjuvant or adjuvant treatment. In addition, definitions of RFS are inconsistent and include also distant metastases in some publications. Furthermore, Koga's series 30 is the only one based on a very stringent bladder-sparing protocol consisting of neo-adjuvant low-dose radiochemotherapy (NARC) followed by PC and PLND with well-defined inclusion criteria for PC. Knoedler's series 7 is the only one also reporting comorbidity data, showing that severe comorbidities are associated with an impaired OS and CSS. Differences in selected patient cohorts between the series make comparisons in oncological outcomes challenging, however some conclusions might be drawn. Despite a higher rate of node positive disease in our cohort  rates of the discussed series, but positive pelvic lymph node involvement was much lower (12% 24 and 0% 30 vs. 22.2% in our cohort). Of note, patients in Koga's series 30 treated with a completed protocol including PC after NARC developed only superficial local recurrence and 5-year MFS was 100% while 23.1% of patients treated with the same protocol but declined PC developed local MIBC recurrence, and 5-year MFS was significantly reduced to 83%.  Ma et al. 23 and Zhang et al. 29 concluded that lymphovascular invasion was a risk factor for poor prognosis after PC. Holzbeierlein et al. 8 , Ma et al. 23 , Zhang et al. 29 , and Koga et al. 30 identified multifocality to be associated with local recurrence. Kassouf et al. 9 and Ma et al. 23 reported a history of prior cancer to also be a risk factor for tumour recurrence and impaired survival. In addition, Ma et al. 23 identified ureter reimplatation to be associated with poor prognosis. Despite our less stringent patient selection, we also identified ureter reimplantation, vascular invasion, and tumour multilocularity to be risk factors for tumour progression. In the 3 series 8,9,24 that showed better survival outcomes, ureter reimplantation was absent in all and tumour multifocaltiy in two 9,24 . In Koga's cohort 30 , ureter reimplantation (26%) was also performed but not included in the regression analysis; however, this cohort showed no local MIBC recurrence. Notably, in Zhang's cohort 29 , ureter reimplantation (22%) was associated with a higher CSS rate, but the authors recommend to consider this "unreasonable" result cautiously. Eventually, the need for ureter reimplantation and a multifocality of the tumour should be regarded as comparative contraindications.
However, survival data after PC has been shown to compare favourably with RC series 7,10 or series after a trimodal therapy (TMT) approach, including maximal TURBT followed by concurrent radiosensitising chemotherapy and RT 31 . A TMT approach is the most studied bladder-sparing strategy, and it is considered a valuable alternative treatment for selected patients with T2 MIBC 2 . However, in our institution, we estimate that only 3-5 patients undergo TMT per year, compared to about 60 RCs per year. Overall, the 5-year CSS ranges from 50% to 82% and the 5-year OS is approximately 50%, ranging from 36% to 74%. The local recurrence rate after TMT ranges from 24% to 43%. Of these recurrent tumours, the rate of MIBC recurrences ranges from 11% to 18.5%, resulting in an average overall salvage cystectomy rate of approximately 25-30%. Overall, the rate of RC for late effects of RT ranges from 0% to 2% 31 .  The 5-year CSS and OS rates after RC are approximately 62%-66% 7,10,32-34 . RFS is roughly 58-68% after RC, but varies between 34% and 74% depending on local tumour stage and nodal status 33,34 . Two matched analyses from Capitanio et al. 10 and Knoedler et al. 7 comparing PC with RC demonstrated no statistically significant differences in cancer control, neither in CSS nor in OS rates. Nevertheless, patients treated with PC remain at risk for intravesical local recurrence. In our cohort, 55.5% of patients remained free of local recurrence, 18.5% developed local MIBC recurrence, the rest superficial recurrence, resulting in a salvage cystectomy rate of 18.5%. Notably, 2 of the 12 patients who experienced local recurrence in our study developed late local MIBC recurrence 32 and 45 months after PC, emphasizing the importance of a life-long surveillance after PC. The aforementioned publications showed similar results with local MIBC recurrences ranging between 8 and 31%, superficial recurrences between 8 and 24% 8,9,23,24 , and a rate of late cystectomies of 7-19% 7-9,23,24 .   In our series, major 30 to 90-day complications occurred in 7.4%, and major late complications in 11.1% of patients, with wound healing disorder or wound infections constituting the most common short-term and hydronephrosis the most common long-term complications requiring intervention. Little data currently exists regarding complications of PC. To our knowledge, we are the first group that uses the Clavien-Dindo classification to describe complications in an open PC series. Smaldone et al. 24 and Holzbeierlein et al. 8 rudimentarily describe postoperative complications in their open PC cohorts. In Smaldone's study 24 , a long-term rate of distal ureter stenosis requiring ureteroneocystostomy was observed in 4% of cases. Kates et al. 35 utilized the Nationwide Inpatient Sample (NIS) to examine 10780 patients who underwent PC for BC. They found that 15.8% of patients experienced an in-hospital complication and 3.9% a "never event" -in-hospital complication with a mortality rate of 1.8%. RC series which used the Clavien-Dindo scoring describe a rate of 13-36% major (Clavien-Dindo ≥3) 90-day complications depending on age, surgical technique, and type of urinary diversion investigated [36][37][38][39][40][41] .
The only other PC study group (Golombos et al. 42 ) using the Clavien-Dindo score in a robot-assisted PC approach (also including non-MIBC) describes an overall 90-day complication rate of 24.1% with no occurrence of major complications.
In the past years, there has been an increasing interest on QoL outcomes in urological malignancies, developing new specific instruments with the aim of evaluating the HR-QoL of the patients and the impact of the respective health condition on their lives. Nevertheless, QoL outcomes are difficult to compare. Many different questionnaires are used to measure the QoL in oncological urology (e.g. SF-36 (Short Form-36), FACT-G (Functional Assessment of Cancer Therapy -General), EORTC QLQ-C30); the questionnaires differ considerably in some of the topics covered 43,44 . Another difficulty is that QoL is a multi-dimensional concept, incorporating many domains that are weighted by their importance to the individual and may change over time 45,46 . In addition, continent and incontinent forms of urinary diversion after radical surgery for bladder cancer were shown to influence postoperative QoL with different impact, eventually favouring orthotopic neobladder, but consistent evidence from systematic reviews and meta-analyses is lacking 43,44 . An interaction could even be demonstrated between patients and the investigating institution which has hypothetically been attributed to glorification and idealization of the attending urologists 47 . Thus, we used EORTC QLC-C30 QoL data from patients treated with RC at our institution and faced the data with this PC series to minimize confounding factors and to achieve high comparability. Compared to patients treated with RC and orthotopic ileal neobladder, in our institution postoperative global health status and QoL outcomes showed equally high scores in this series of patients treated with PC but higher scores compared to patients treated with RC and ileal conduit. This result might be due to the surgical procedures, resulting in continent or incontinent urinary conditions. On the other hand, patients following PC suffered more often from emotional problems compared to patients after RC and orthotopic ileal neobladder. However, patients following PC experienced financial difficulties caused by their physical condition or medical treatment less frequently compared to patients following RC. This difference was most obvious after urinary diversion with a conduit. However, this observation has to be interpreted with caution since the number of cases is small in each group. In addition, the three groups differ in terms of age, gender distribution, and comorbidity grade. Patients provided with orthotopic ileal neobladder were a median four years younger (62 [56-66]) and patients with ileal conduit four years older (70 [64-75]) compared to patients treated with PC. Patients following PC or RC with ileal conduit had higher Charlson comorbidity grades at time of surgery (3 [2][3][4]; 3 [2][3]) than patients after RC with orthotopic ileal neobladder (2 [2][3]), and 96% of the patients treated with RC and orthotopic ileal neobladder were males compared to 70% in the PC series. Hence, a statistically robust comparison is ineligible, and firm conclusions cannot be drawn. Furthermore, since no baseline QoL data were available, a comparison of pre-and postoperative scores was not possible.
Considering external data, the systematic review and meta-analysis by Yang et al. 44 showed a median global health status/QoL scale in the EORTC QLQ-C30 questionnaire of 72 for RC with continent urinary diversion, and of 65 for RC with incontinent urinary diversion, which is in line with our presented data.
Although aiming to help to improve QoL if the bladder function remains preserved after treatment, QoL data after TMT are very limited and are mainly focused on late toxicity. Data comparing QoL after PC and TMT in patients with MIBC are lacking. However, better QoL was shown after TMT compared to RC with incontinent urinary diversion when using a self-administered questionnaire 48 .

Reference
Year of publication  Table 6. Fife-year oncological outcomes in different PC series for the treatment of muscle-invasive bladdercancer. (OS) overall survival, (CSS) cancer-specific survival, (RFS) recurrence-free survival, (IQR) interquartile range, (SD) standard deviation. 1 Recurrence defined as local recurrence within the bladder (non-muscle invasive and/or muscle-invasive). 2 Recurrence defined as local recurrence within the bladder (non-muscle invasive and/or muscle-invasive) and/or development of pelvic or distant metastasis. Our study is limited in its retrospective nature and small sample size. However, we confirmed that PC can provide adequate cancer control of MIBC in selected cases and showed that it offers patients a good chance of a long-term bladder preservation, providing a good postoperative global health status and QoL and that it is associated with only a moderate rate of severe perioperative complications, even in comorbid patients. Though, survival data of our presented study and of the cited PC series refer to selected patient cohorts, making a direct comparison to RC or TMT series difficult.
However, the outcome of MIBC is determined more by the extent of disease (stage) than by the extent (type) of surgery. Hence, to cure MIBC, removal of the tumour with negative surgical margins and no residual tumour is required, regardless of the procedure used 49 . That might also include an en-bloc tumour resection in the previous transurethral staging operation to better define the margin during PC, but this approach is still considered experimental 50,51 .
In addition, there is no doubt that neoadjuvant chemotherapy (NAC) improves survival of MIBC over RC alone [52][53][54] . Thus, although proven for RC, NAC should also be considered in MIBC patients who are candidates for PC and suitable for NAC. Since patients who decide to undergo PC are likely also more willing to take the risk of under-treatment (rather than the risk of side effects or restrictions in QoL), it is not surprising that none of our PC patients was treated with neoadjuvant chemotherapy, even though we offered NAC to suitable candidates. This stresses the fact that PC cohorts, including our series, represent highly selected cases.
In conclusion, patient selection with clearly defined inclusion and exclusion criteria are essential to further improve oncological outcomes of PC and MIBC patients in general.
We need prospective randomized controlled trials (RCT) in MIBC comparing RC with PC and TMT in the subgroup of patients potentially suitable for a bladder-sparing therapy protocol with strictly defined inclusion criteria (e.g. primary, small, solitary tumours with response to NAC, including pre-therapeutic genetic tests of the tumour heterogeneity to predict the response to NAC und to increase the acceptance of NAC). This could clarify the essential oncological and quality of life endpoints between the different therapies and could answer the question who is likely to benefit from a less invasive approach and for which patient a RC might be an unnecessary overtreatment. Unfortunately, previous attempts of RCTs, like the SPARE (Selective Bladder Preservation Against Radical Excision) trial, which compares outcomes in patients who receive NAC followed by RC or RT, failed to recruit patients. Strong clinician and patient preferences for treatments impacted willingness to undergo randomization and acceptance of treatment allocation 55 , which underlines the fact that high patients' preferences for specific treatment options is not only a bias of retrospective studies, but also impacts the design and success of RCTs.