Molecular subtyping and genomic profiling expand precision medicine in refractory metastatic triple-negative breast cancer: the FUTURE trial

Triple-negative breast cancer (TNBC) is a highly heterogeneous disease, and molecular subtyping may result in improved diagnostic precision and targeted therapies. Our previous study classified TNBCs into four subtypes with putative therapeutic targets. Here, we conducted the FUTURE trial (ClinicalTrials.gov identifier: NCT03805399), a phase Ib/II subtyping-based and genomic biomarker-guided umbrella trial, to evaluate the efficacy of these targets. Patients with refractory metastatic TNBC were enrolled and stratified by TNBC subtypes and genomic biomarkers, and assigned to one of these seven arms: (A) pyrotinib with capecitabine, (B) androgen receptor inhibitor with CDK4/6 inhibitor, (C) anti PD-1 with nab-paclitaxel, (D) PARP inhibitor included, (E) and (F) anti-VEGFR included, or (G) mTOR inhibitor with nab-paclitaxel. The primary end point was the objective response rate (ORR). We enrolled 69 refractory metastatic TNBC patients with a median of three previous lines of therapy (range, 1–8). Objective response was achieved in 20 (29.0%, 95% confidence interval (CI): 18.7%–41.2%) of the 69 intention-to-treat (ITT) patients. Our results showed that immunotherapy (arm C), in particular, achieved the highest ORR (52.6%, 95% CI: 28.9%–75.6%) in the ITT population. Arm E demonstrated favorable ORR (26.1%, 95% CI: 10.2%–48.4% in the ITT population) but with more high grade (≥ 3) adverse events. Somatic mutations of TOP2A and CD8 immunohistochemical score may have the potential to predict immunotherapy response in the immunomodulatory subtype of TNBC. In conclusion, the phase Ib/II FUTURE trial suggested a new concept for TNBC treatment, demonstrating the clinical benefit of subtyping-based targeted therapy for refractory metastatic TNBC.


S1. Biopsy, TNBC subtyping and biomarker testing
The procedures of biopsy, TNBC subtyping, and genomic sequencing were as follows: a fresh core needle biopsy of a refractory metastatic tumor lesion was obtained from each patient by an interventional radiologist under computed tomography or ultrasound guidance. Written informed consent was obtained from each patient before the biopsy. Two fresh tumor specimens (approximately 2 cm long each) were collected. One sample was immediately formalin-fixed, then embedded in paraffin, for review by two pathologists at the Department of Pathology of FUSCC for TNBC IHC subtyping. We generally used AR, CD8, and FOXC1 markers to classify TNBC into four subtypes (Supplementary methods and Appendix Figure A1).
The other tumor specimen was frozen, stored and allocated for DNA-targeted sequencing.
Genomic alterations, such as HER2 somatic mutation, and BRCA1/2 germline mutation, were sequenced based on the FUSCC NGS panel (Supplementary methods and Appendix Table A1).
The genomic sequencing was conducted at Chinese National Human Genome Center at Shanghai, which passed the quality control assessment of Chinese government.

S2.1 Selection of markers for IHC-based classification
We have previously classified TNBC into four subtypes (BLIS, IM, LAR, MES) in our published paper 1 . Here, we conducted several steps to select optimal TNBC subtype specific markers based on the multi-omics data of our FUSCC TNBC cohort and the TCGA cohort 1 . 1) we performed differential expression analysis on RNA sequencing data using DESeq2 to identify the highly expressed genes (log2(fold change) > 1.5, adjusted P value < 0.05) in each subtype. 2) using RNA sequencing and mass spectrometry protein expression data from TCGA, we assessed the correlation between the mRNA and protein expression of these genes and retained those that satisfied the following criteria: a correlation coefficient of ≥ 0.5 and a P value of < 0.05. 3) we conducted receiver operating characteristic (ROC) analysis to test the accuracy of using the retained genes to identify the corresponding subtypes and focused on the top 10 genes ordered by the area under the curve (AUC) in each subtype. CD8A and FOXC1 were the top-ranked genes and were therefore selected as markers for the IM and BLIS subtypes respectively. AR was the gene rank fifth but has been demonstrated as a feasible detecting marker and therapeutic target in AR-positive TNBCs [2][3][4] . Thus, we selected AR as the marker 3/7 for the LAR subtype.

S2.2 Calculation of the optimal cutoff value of the markers
We performed IHC staining for both markers on the paraffin-embedded sections.

S2.3 IHC staining and IHC score definition in the FUTURE trial
We performed IHC staining on paraffin-embedded sections (4 μm thick) of tumor specimens from the patients in the FUTURE trial to evaluate the expression of AR, CD8 and FOXC1. As previously described, IHC staining was performed using an Ventana Benchmark

S3.2 Targeted sequencing using the FUSCC 484 gene panel
A custom-designed genetic panel, which constituted a hybridization-capture-based assay of 484 genes that are targets of approved and experimental therapies as well as frequently mutated genes in breast cancer, was used in this study. The panel was designed for the detection of mutations and small insertions and deletions. In-house generated RNA baits were utilized to capture all protein-coding exons of target genes. The RNA baits were generated from an oligo pool synthesized by Synbio Technologies (Suzhou, China). The detailed laboratory protocol was described previously 5 .
Both tumor and matched blood samples were sequenced. At least 10 ng of each DNA sample obtained after SYBR green quantification was fragmented using a Covaris M220 and then subjected to end-repair, A-tailing and adapter ligation using a KAPA HyperPlus kit (Kapa Biosystems) according to the manufacturer's recommended protocol. Subsequently, 750 ng of prepared DNA in a volume of 3.4 µl was captured by RNA baits, and the captured library was then sequenced using indexing primers. After quantification with a Multi-Mode Reader (BioTek), the libraries were pooled and sequenced using an Illumina HiSeq X TEN platform (Illumina Inc.,  probability of visceral metastasis and brain metastasis, rapid progress, limited treatment [1][2][3][4] . The reason for the dilemma of TNBC is the heterogeneity [5][6][7] . The naming of TNBC is an exclusive diagnosis, but it is in fact a group of heterogeneous diseases. The research direction is to distinguish the subtypes of TNBCs, to figure out potential therapeutic targets of each subgroup, to improve the treatment strategy and prognosis of these patients.

Definition of refractory TNBC in this study
Because of the lack of traditional effective endocrine and anti-HER2 targeted  [8][9][10] . In this study, refractory TNBC was defined as the failure of all the above chemotherapeutic drugs.

Further classification of TNBC based on multi-omics data
Because triple-negative breast cancer has high invasiveness, few treatments and poor prognosis, there have been a series of studies trying to evaluate the efficacy of targeted drugs for triple-negative breast cancer, but they have failed to achieve the desired results. One of the possible reasons for the analysis is that the patients in these studies were generally triple-negative breast cancer, without further accurate classification of triple-negative breast cancer, and lack of more accurate classification optimization treatment. These failed studies suggest that smarter and more precise treatment strategies should be targeted therapies for possible molecular drive events in each subtype. were uniquely sensitive to bicalutamide (an AR antagonist). These data may be useful in biomarker selection, drug discovery, and clinical trial design that will enable alignment of TNBC patients to appropriate targeted therapies. The team published a research paper at PloS One in 2016 [12] , using histopathological quantification and lasercapture microdissection to determine that transcripts in the previously described immunomodulatory (IM) and mesenchymal stem-like (MSL) subtypes, and they found that these transcripts were contributed from infiltrating lymphocytes and tumorassociated stromal cells, respectively. Therefore, they refined TNBC molecular subtypes from six into four tumor-specific subtypes (BL1, BL2, M and LAR) and  [9,28], [17,41], respectively).
2)The study of Bursteinet et al., published in Clin Cancer Res,2015 [13] , suggested that TNBC is a heterogeneous disease. In their study, RNA and DNA profiling analyses  [14][15] comprehensively analyzed clinical, genomic, and transcriptomic data of a cohort of 465 primary TNBC. PIK3CA mutations and copy-number gains of chromosome 22q11 were more frequent in our Chinese cohort than in The Cancer Genome Atlas (TCGA). We classified TNBCs into four transcriptome-based subtypes: (i) luminal androgen receptor (LAR), (ii) immunomodulatory (IM), (iii) basal-like immune-suppressed (BLIS), and (iv) mesenchymal-like (MES). Putative therapeutic targets or biomarkers were identified among each subtype. Importantly, the LAR subtype showed more ERBB2 somatic mutations, infrequent mutational signature 3 and frequent CDKN2A loss. The comprehensive profile of TNBCs provided here will serve as a reference to further advance the understanding and precision treatment of TNBC.
According to our previous data, the characteristics of each type and the potential therapeutic targets of the four subtypes are described below: infiltrating lymphocytes (TILs) and intra-tumor TILs [16][17] . Although mutation load was not significantly higher in IM subtype, Gene Set Enrichment Analysis (GSEA) between the IM TNBCs and other subtypes demonstrated the activation in antigen processing and presentation related pathways. In addition, Combing CIBERSORT and differential expression profiling, we demonstrated that immune activating cell and immuno-stimulators were enriched in IM subtype.
As both clinical and omics features had proved that immune recognition has been activated in IM subtype, the way in which these tumors achieved immune escape were likely to be the recruitment of immune suppressive cells or the activation of immune score was used to further classify the BLIS subgroup-HRD high score group and HRD low score group, in order to accurately screen those patients who were sensitive to DNA damage such as platinum drugs. The high HRD score was significantly correlated with the mutation of BRCA gene, who might benefit from Olaparib [18] . And the prognosis of these patients was significantly better than that of the patients with low HRD score.  This study is an exploratory Ib/II phase clinical trial, the main purpose of which is to screen valuable therapeutic arm, in order to develop phase III clinical studies with larger samples.

Sample size estimation
In this study, 3 or more than 3 of 20 patients in each arm group reached CR or PR, will be defined to reach the study end point.

Primary purpose
To evaluate the efficacy and safety of precision treatment of refractory TNBC based on molecular subtyping.

Exploratory purpose
To explore the pharmacokinetics of CDK4/6 inhibitor (SHR6390) in combination with AR inhibitor (SHR3680). A total of 18 blood collection points, each blood collection point to take 2 tubes blood (3 ml each), see 8.3 specific description.

Specific regimen
The specific treatment options and dosages of the 7 treatment arms in this study are shown in table 2. Participants will receive treatment until disease progression or intolerable toxicity. The evaluation will be based on MRI, CT and physical examination according to novel international standard set by RECIST committee. Evaluation will be done every two cycles (A, E and F arms 6 weeks ±3 days and other arms 8 weeks ±3 days). Since the primary endpoint of this study was total effective, patients who obtained CR or PR required imaging efficacy confirmation after 4 weeks.
In the clinical study process, based on efficacy, safety and existing treatment methods, according to the target and targeted drug availability, new treatment arm may be added during the treatment process, and then it will be submitted to the Ethics Committee for discussion and approval before implementation.

Pyrotinib
Pyrotinib (SHR1258) is an orally administered dual irreversible tyrosine kinase inhibitor of the epidermal growth factor receptor (HER1) and the human epiderminal growth factor receptor2 (HER2), which is developed by Jiangsu Hengrui Phamaceutical Co LTD. The recommaned dosage of Pyrotinib is 400mg by oral administration everyday.
Pyrotinib has completed phase I and randomized controlled phase II clinical trials in advanced breast cancer in China, and its results have been published in well-known international journals [19][20] . Based on the good tolerance of phase I study, a phase II domestic multicenter clinical study assigned patients with pyrotinib in combination with capecitabine or lapatinib in combination with capecitabine, which has

SHR6390
Cyclin-dependent kinase (  week (D22-28), and one dose is given every 28 days medicine cycle. Fasting is recommended (fasting should be guaranteed at least 1 hour before and 2 hours after taking the drug during the administration period).
In the above-mentioned Phase I study of advanced solid tumors, among all 20 subjects, most of the drug-related AEs were grades 1-2, grade 3 AEs included neutropenia (7 cases) and leukopenia (5 cases), No grade 4 AE was observed. No withdrew from the study due to AE, and no serious adverse events (SAEs) occurred in this study, which was well tolerated.

SHR3680 is a new oral androgen receptor (AR) inhibitor developed by Jiangsu
Hengrui Pharmaceutical Co Ltd. SHR3680 is recommended for oral administration by 240mg once daily.
Preclinical studies have verified that SHR3680 can effectively inhibit AR nuclear translocation, AR transcriptional activity, AR overexpression of prostate cancer cell proliferation and prostate-specific antigen secretion and has no AR partial agonistic effect. SHR3680 is intensively comparable and therefore significantly better than the first generation of AR antagonists bicalutamide.
A phase I/II clinical study designed to explore the tolerability, pharmacokinetics,

SHR1210
PD-1 is an immune checkpoint protein involved in the negative regulation of pain, increased bilirubin is mild to moderate, and clinically controllable.

SHR3162
The development of PARP inhibitors could be traced back as early as the 1990s.
At first, attempts were made to enhance the efficacy of chemotherapeutic drugs, but the combination of chemotherapeutic drugs and PARP inhibitors proved to be very toxic and research was interrupted. In 2005, two Nature articles made breakthrough progress.  After a Phase I clinical study, the recommended dose of SHR3162 is 150 mg orally twice daily.

Apatinib
Apatinib (Apatinib, trade name: Aitan) is a small molecule VEGFR tyrosine kinase inhibitor. It blocks the signal transmission after VEGF binding to its receptor by inhibiting the activity of VEGFR tyrosine kinase and inhibit tumor angiogenesis. proteinuria (13.6%) and increased ALT (11.9%).

Everolimus
Everolimus (trade name: Afinitor) derivatives of rapamycin selectively inhibit mammalian rapamycin target protein (mTOR), especially targeting the mTOR-raptor signaling complex. mTOR is a key serine-threonine kinase in the phosphatidylinositol- For TNBC, everolimus cannot be combined with endocrine or anti-HER2 drug, so it is feasible to choose everolimus combined with paclitaxel as illustrated on BOLERO-1.

Capecitabine
Capecitabine is a commonly used chemotherapeutic drug for breast cancer and is covered by medical insurance.

Albumin paclitaxel
Albumin paclitaxel is a chemotherapeutic drug approved for breast cancer indications in China and abroad, but its price is relatively high and it has not been The main organ functions are basically normal and meet the following conditions: The Have not received radiotherapy, endocrine therapy, molecular targeted therapy, and surgery within 3 weeks before the start of the study, and have recovered from the acute toxicity of previous treatment (if surgery, the wound has completely healed), no peripheral neuropathy or 1 degree peripheral neurotoxicity ECOG score ≤2, and life expectancy ≥3 months.
Fertile female subjects need to use a medically approved contraceptive during study treatment and at least 3 months after the last use of the study drug.

33/74
Subjects volunteered to join the study, signed informed consent, have good compliance, and cooperated with the follow-up.

Common exclusion criteria
Patients with any of the following were excluded from the study: Radiotherapy (except for palliative reasons), chemotherapy, and immunotherapy 3 weeks before treatment, except bisphosphonates (can be used for bone metastasis).
Uncontrolled central nervous system metastasis (referring to symptoms or the use of glucocorticoids or mannitol to control symptoms).
A history of clinically important or uncontrolled heart disease, including congestive heart failure, angina pectoris, myocardial infarction or ventricular arrhythmia in the past 6 months.
Adverse reactions with grade ≥1 that are ongoing due to previous treatment.
Exceptions to this are hair loss or the investigator's opinion should not be ruled out.
Such cases should be clearly documented in the investigator's notes.
Major surgery (except minor outpatient surgery, such as placement of vascular access) within 3 weeks of the first course of trial treatment.
Pregnant or lactating patients.
Malignant tumors in the past five years (except for cured skin basal cell carcinoma and cervical carcinoma in situ).

Special exclusion criteria for each arm
Special exclusion criteria for arm A LVEF≥50% (Echocardiography).
Have suffered any of the following heart diseases: (1) Angina pectoris, (2) Arrhythmias requiring medical treatment or clinical significance, Urine routine shows urinary protein ≥ ++ or confirmed 24-hour urine protein quantification> 1.0 g.
Those who have hypertension and cannot be reduced to normal range with antihypertensive medication (systolic blood pressure> 140 mmHg, diastolic blood pressure> 90 mmHg).

Suspension criteria
Patients need to stop treatment and continue to be followed up unnder one of the following conditions: Patients who cannot tolerate after two dose adjustments.
Patients with a delay of dosing for more than 3 weeks.
Intolerable adverse reactions occurred during the test. Pregnancy.
Patients who are not eligible for further treatment due to violation of the research protocol (at the discretion of the investigator).
Any other conditions that investigator consider it necessary to discontinue treatment.

Exit criteria
Patient should withdraw from this study under one of the following conditions: Patient withdraws informed consent.
Complicated diseases that seriously affect clinical evaluation.
Any case where the researcher considers it necessary to withdraw from the study.
Receive other systemic treatments or use of drugs forbidden in this study.

36/74
Tissue specimens of recurrent metastatic TNBC (mainly including breast, axillary lymph nodes, lungs, or liver) were re-obtained within 4 weeks before treatment, and the

37/74
sum of the baseline diameters. If the lymph node diameter is included, as mentioned above, only the short diameter is counted. The sum of the baseline diameters will be used as a reference value for the baseline level of the disease.
All other lesions, including pathological lymph nodes, can be considered as nontarget lesions and do not need to be measured, but should be recorded at baseline assessment, such as "existing", "missing" or in rare cases "clear progress". Extensive target lesions can be recorded with target organs (eg. extensive liver metastases).
Collection of basic data within one week before treatment, including medical history, physical examination, general condition score ECOG score, blood routine, urine routine, liver and kidney function (total bilirubin, ALT, AST, AKP, LDH, total protein, albumin, Urea nitrogen, creatinine, blood glucose), electrocardiogram,

8.2.2
The blood pressure was measured 3 times a day for 2 weeks before apatinib treatment. If the blood pressure is abnormal, the blood pressure will be monitored daily.
If the blood pressure is normal, the blood pressure will be measured twice a week

At the end of treatment
At the end of treatment: comprehensive tumor assessment, vital signs, physical examination, ECOG score, blood routine, urine routine, electrocardiogram, liver and kidney function, tumor markers (CEA, CA-153, CA-125), and quality of life score.

Follow-up
The patient entered the post-treatment follow-up period after the last use of the study drug. For patients who were excluded due to non-disease progression, all subjects were followed up from 21 to 35 days after the last dose. Thereafter, they were followed up every 8 weeks, and imaging tests were performed to observe whether the tumor progressed.
During the follow-up period, the following parameters should be recorded: the time of disease progression or death in patients who did not develop disease when taking the study drug, other tumor treatments, SAE related to the study drug, follow-up survival status every 3 months (phone follow-up available).

General dosage
When hematological toxicity reaches III or above or non-hematological toxicity reaches II or above, the investigator decides whether to suspend or reduce the dose. In non-hematological toxicity, nausea, vomiting and fever with a certain cause can be controlled (Such as infections, tumors, etc.), active symptomatic treatment and treatment can be carried out without dose suspension and dose reduction.

Provisions for treatment suspension
In the treatment, the drug administration should be suspended because the drug toxicity has not recovered. The pause time should not exceed 2 times per cycle to ensure the strength of the drug received by the subjects in the trial.

Dose down regulation
The drug is poorly tolerated and dose reductions can be performed on the first day of subsequent cycles. Subjects could adjust the dose twice. Once a dose level was lowered, the dose was not allowed to be increased for any reason, but the dose suspension was still allowed. See Table 3 for dose adjustments for each treatment arm.

Other regulations
To ensure the consistency of dose adjustment throughout the study, a dose pause was first performed during each dose cycle. After taking the dose suspension measure, if the subject is still poorly tolerated, the dose can be adjusted down on the first day of the next dosing cycle. Suspension of treatment).

Concomitant Therapy
In patients who do not have a high risk of infection or a risk of bleeding, it is recommended that colony cell stimulating factor or interleukin-11 or thrombopoietin be administered only in the presence of bone marrow suppression of grade III or higher.
Researchers have the right to decide whether they need to deal with it accordingly. This plan prohibits the use of SFDA-approved modern Chinese medicine preparations and immunomodulators (such as thymosin, interferon, interleukin-2, and lentinan) for the treatment of breast cancer.

Main endpoints and observation methods
Overall In this study, 20 patients are prepared for each arm. If 3 patients or more achieved CR or PR, this treatment arm would be considered valuable and reached the main endpoint of this study. It is then worth recommending to design a larger randomized phase III clinical study in the following study project.

Disease Control Rate (DCR):
Refers to the percentage of patients who can be evaluated for complete response, partial response, and stable disease for more than 4 weeks. Overal Survival (OS):

Progression
Refers to the time from enrollment to death for any reason.

1 Treatment of diarrhea
Diarrhea: Investigators should inform subjects the possibility of diarrhea and its management before starting treatment. Follow-up and observation (≤14 days) is considered first after onset of diarrhea. Oral montmorillonite powder TID is advised when subjects experiencing diarrhea. Those who experiencing severe diarrhea may be treated with electrolytes solution. Study Medication is withheld until diarrhea improves to Grade≤1 or described as Table 4.

Treatment of hand-foot skin reaction
Hand-foot skin reaction (HFSR) is a kind of dermatological adverse event, which may occur as a side effect of certain chemotherapy or targeted therapy. Symptoms of HFSR include numbness, tingling, burning, or itching sensation, swelling, redness, tenderness and rash. Grade: Grade 1: Numbness, dysesthesia/paresthesia, tingling, painless swelling or erythema of the hands and / or feet and / or discomfort, which does not disrupt normal activities.
Grade 2: Painful erythema and swelling of the hands and / or feet and / or discomfort affecting the patient's activities.
Grade 3: Moist desquamation, ulceration, blistering or severe pain of the hands and / or feet and / or severe discomfort that causes the patient to be unable to work or perform activities of daily living.

Management:
Supportive care is considered, including maintaining skin cleanness, avoiding pressure and rubbing, using topical steroid and urea cream, using topical antibiotics when necessary.
Subjects will be withdrawn from this study if they experience HFSR Grade≥2 for 3 times and exacerbate. Subjects with Gilbert syndrome must monitor TBIL and DBIL and monitoring closely when DBIL abnormality occurs.

Treatment of liver injury
If discontinuation is required because of elevated AST, ALT and / or bilirubin, close observation is recommended, including: i. Review liver enzymes and serum bilirubin 2 to 3 times a week. If abnormal results are stabilized or recover to normal range, reduce the review frequency to 1 or less times per week.
ii. Obtain more detailed medical history about current symptoms.
iii. Obtain more detailed medical history and / or history of concomitant diseases, including any pre-existing hepatic disease history or risk factors.

Treatment of proteinuria
Subjects with urine protein ++ or more twice should receive 24-hour urine protein test.
Subjects will be withdrawn from the study if they experience nephrotic syndrome.

Prevention and treatment of stomatitis/oral mucositis/mouth ulcers
For prevention the stomatitis, all subjects will be instructed to perform routine "good oral care" each day during the trial. Good oral care will consist of: brushing teeth at least twice daily with soft bristled toothbrush, continue current daily flossing routine (if patients were not already flossing daily, they should not be instructed to start flossing as this could cause oral trauma), and continue routine dental care/maintenance with their dentist, if they have one. It is recommended that patients should use 10mL of an alcohol-free, 0.5mg/5mL dexamethasone steroid mouthwash swishing and spitting QID, especially during the first 8 weeks of treatment (majority of stomatitis events occur within the first 8 weeks of treatment). The mouthwash is to be held in mouth and swished around mouth to cover the entire buccal mucosa surface for a minimum of two minutes, and then spat out.
Subjects with a clinical history of stomatitis/mucositis/mouth ulcers and those with gastrointestinal morbidity associated with mouth/dental infections, irritation of esophageal mucosa e.g. gastroesophageal reflux disease (GERD) and pre-existing stomatitis/mucositis must be monitored even more closely. Subjects should be instructed to report the first onset of buccal mucosa irritation/reddening to their investigators immediately.
General guidance and management include patient awareness and early intervention. Stomatitis/oral mucositis/mouth ulcers due to everolimus should be treated using local supportive care. Evaluation for herpes virus or fungal infection should be considered. Subjects should be informed about the possibility of developing mouth ulcers/ oral mucositis and instructed to report promptly any signs or symptoms to their investigators. Subjects should be educated about good oral hygiene, instructed to avoid spicy/acidic/salty foods, and should follow the following guidelines: (i) For mild toxicity (grade 1), use conservative measures such as non-alcoholic mouth wash or normal saline mouth wash several times a day until resolution.
(ii) For more severe toxicity (grade 2 in which case subjects have pain but are able to maintain adequate oral alimentation, or grade 3 in which case subjects cannot maintain adequate oral alimentation), the suggested treatments are topical analgesic mouth treatments (i.e., local anesthetics such as benzocaine, butyl aminobenzoate, tetracaine hydrochloride, menthol, or phenol) with or without topical corticosteroids, such as triamcinolone oral paste 0.1% (Kenalog in Orabase®).
(iii) Agents containing alcohol, hydrogen peroxide, iodine, and thyme derivatives may tend to worsen mouth ulcers. It is preferable to avoid these agents.
(iv)Antifungal agents must be avoided unless a fungal infection is diagnosed. In particular, systemic imidazole antifungal agents (ketoconazole, fluconazole, FUTURE Study Protocol

47/74
itraconazole, etc.) should be avoided in all patients due to their strong inhibition of Everolimus metabolism, therefore leading to higher Everolimus exposures. Therefore, topical antifungal agents are preferred if an infection is diagnosed. Similarly, antiviral agents such as Acyclovir should be avoided unless a viral infection is diagnosed.

Treatment of non-infectious pneumontis
Non-infectious pneumonitis is a known side effect of rapamycin analogues.
Clinically significant pneumonitis is typically accompanied by non-specific symptoms including dyspnea, nonproductive cough, fatigue, and fever. Diagnosis is generally suspected in individuals receiving mTOR inhibitors who develop these symptoms or in asymptomatic individuals in whom a routine chest CT scan reveals a new ground glass or alveolar infiltrate.
The frequency of symptomatic pulmonary toxicity (all grades) was approximately 13% in a phase III study of Everolimus in patients with metastatic renal cell carcinoma.
Severe (CTC grade 3) pneumonitis occurred in 4% of patients, and an occasional fatality was reported. The lung toxicity was partly or completely reversible in the majority of cases with interventions including drug interruption, discontinuation and the use of corticosteroids.
Subjects will be routinely questioned as to the presence of new or changed pulmonary symptoms consistent with lung toxicity. CT scans and pulmonary function test should be done, as clinically indicated, if there are symptoms that indicate that the patient has developed non-infectious pneumonitis. If non-infectious pneumonitis develops, the guidelines in Table 5 should be followed. Dose modification instructions are also provided in Table 5. Consultation with a pulmonologist is recommended for any case of pneumonitis that develops during the study. Discontinuation.
Note: Combined treatment of HMG-CoA reductase and Fenofibrate may lead to rare but severe musculoskeletal toxicity, whose symptoms and signs include rhabdomyolysis, significant elevation of CPK, myoglobinuria, acute kidney injury or even death. Thus, risk/benefit ratio for each subject should be calculated before treatment of hyperlipidemia.
There have been reports of hyperglycemia from patients receiving Everolimus. It is suggested to monitor fasting plasma glucose before and after subjects receiving FUTURE Study Protocol

50/74
Everolimus and monitor more frequently if subjects receiving Everolimus with other medication that could lead to hyperglycemia. Optimal glycemic control is suggested before subjects receiving Everolimus.

SHR-1316
Subjects will receive appropriate supportive care investigators consider necessary.
Supportive care dealing with potential Immune-Related Adverse Effects (irAE) will be listed below, including oral or intravenous corticosteroids, and other anti-inflammatory agents when symptoms are resistant to corticosteroids. Tapering corticosteroids may take several periods because of possible recurring of symptoms. Rule out other possible reasons that may need supportive care, e.g., metastatic diseases or bacterial or viral infection. Supportive treatment will be conducted when investigators ascertain AE is associated with SHR-1210 or SHR-1316 and will not be conducted when AE is not associated with SHR-1210 or SHR-1316.
If capillary endothelial proliferation occurs, conduct biopsy and pathological examination if possible. Subjects experiencing severe and enduring capillary endothelial proliferation are suggested endoscopy and MRI scan to ascertain if there is visceral or mucous involvement.

51/74
Symptoms and signs of enterocolitis (e.g., diarrhea, abdominal pain, hematochezia or mucous stool, with or without fever) and intestinal perforation (e.g. peritoneal irritation signs and intestinal obstruction) should be monitored closely.

Management of infusion related reaction
Infusion reaction may be triggered by infusion of immune checkpoint inhibitor SHR1210 or SHR1316, and it should be graded and managed according to Table 7. obtain medication number and corresponding study medication will be distributed.
Designated personnel of the hospital is put in charge of keeping compounds, fill in records of receiving and using compounds and retrieve the remaining compounds and empty bottles promptly during the test. The usage and records of compounds should be checked at regular intervals and deal with retrieved medication at any time.

Storage and Management of Compounds
According to Good Clinical Practice (GCP), study medication is uniformly stored, distributed and recycled by the study site. Compounds are stored sealed, protected from light at room temperature, with a tentative expiration date of 2 years.

Disposal of Remaining Compounds
Investigators should record the date and dosage of each subject's medication. The total amount of study medication is 120% of the pre-designed dosage. The remaining compounds should be returned to Hengrui Medicine at the end of the trial.
14 Adverse Events

Definition
An adverse event (AE) is any undesirable medical event experienced by clinical trial subjects after administration of certain compounds, however, AE does not always reflect a causal relationship with intervention.

Adverse event reporting period
The reporting period starts from enrollment and lasts till the final follow up. Any adverse event happens during this period should be filled in the case report form.

Serious adverse event (SAE)
An adverse event is defined as a SAE if it agrees with one or more following

Record and report
The investigator should explain to the patient in detail and ask the patient to truthfully reflect the change in the condition after the administration. Physicians should avoid induced question. While observing the curative effect, the physicians pay close attention to observing adverse events, analyzing the causes, making judgments, and following up observations and records to count the incidence of adverse reactions.
For adverse event occurred during the trial, its time of onset, symptoms, severity, duration, treatment and prognosis should be recorded in the case report, so as to evaluate its significance with test compound. The detailed record should be provided, signed and dated by investigators. The adverse events will be graded based on NCI-CTC 3.0. For each symptom, the highest grade experienced since last follow-up should be reported.
Determination of relationship between adverse events and clinical trial: Attribute to one of the five categories: definitely related, probably related, probably unrelated, unrelated and not appreciable. The first two categories are considered as adverse event and the proportions of adverse events will be calculated.
i. Definite attribution: The chronological order of adverse event after trial onset is reasonable and the reaction is consistent with the known type of reaction. The situation improves after dis-administration and the reaction reappears after readministration ii.
Probably related: The chronological order of adverse event after trial onset is reasonable and the reaction is consistent with the known type of reaction. the clinical condition or alternative intervention may also cause such reaction. iii.
Probably unrelated: The chronological order of adverse event after trial onset is less reasonable and the reaction is less consistent with the known type of reaction, the clinical condition or alternative intervention may also cause such reaction.
iv. Unrelated: The chronological order of adverse event after trial onset is unreasonable and the reaction is inconsistent with the known type of reaction, the clinical condition or alternative intervention may also cause such reaction.

Requirements of data filled by investigators
For all the subjects who have filled in the informed consent and selected to enter the trial, every item in the case report form should be recorded carefully and in detail, and no blank or missing item is allowed (the blank space without record should be crossed).
i. All data in the case report form should be checked with the subject's medical record data to ensure accuracy. ii.
As the original data, the case report form is only allowed to be crossed with any correction made, and the corrected data should be annotated with the signature of the investigator and the date. iii.
The copy of the laboratory test reports should be placed after the case report form.
iv. Data that are significantly higher or beyond the clinical acceptance range should be verified and necessary explanation should be made to subjects by the investigator.
v. Please refer to the case report form for instructions.

Data traceability and completion of CRF
The original record is the study medical record for proper preservation. The case report form from the research medical record is filled in by the researcher. Each included case must complete the case report form.

Data Set for Statistical Analysis
After the completion of the trial scheme and case report form, the analysis plan shall be formulated, and necessary modifications shall be made during the trial process.
The plan shall be completed before data locking, and the statistical analysis report will be provided after data analysis.

Analysis data set
i. Full Analysis Set (FAS): Full analysis set refers to the collection of qualified cases and shedding cases, with the exception of the cases excluded.
ii. Per-Protocol population (PP) Data Set: The PP data set is defined restricted to all the cases that meet the inclusion criteria and complete the treatment plan, with good compliance, no banned compounds and fulfilled required contents in case report form, as well as the observation record documents from subjects whose compliance satisfies the study protocol requirement.
iii. Safety Data Set: Safety Data Set refers to the data with safety records after receiving at least one treatment, with the exception of the cases excluded.

Method of statistics
i. Whether subject to normal distribution: if not, modify statistical methods or perform data transformation. ii.
Whether there is outlier: make statistical and professional analysis, and decide whether to include or not. iii.
Whether there is missing value: when a primary therapeutic index of individual subject fails to be measured, the last observation data should be transferred.
iv. The percentage of dropouts should not exceed 20%, otherwise it requires analysis and explanation.

Method of analysis
i. Measurement data: Use t test, paired t test, rank sum test, paired rank sum test, etc.
ii. Enumeration data: Use Fisher's exact test, etc., rank sum test is adopted to ranked data.
iii. Analysis of efficacy indicators: CMH test, chi square test or logistic regression will be used for enumeration data. Analysis of variance or rank sum test will be used for measurement data according to the feature of the data. Kaplan Meier method or Cox regression will be used for survival data.
iv. FAS analysis and PP analysis: PP analysis and FAS analysis will be conducted simultaneously for the main efficacy indicators.

Statistic expressions
i. The report is mainly represented by tables with title, annotation and number of cases, which are self-evident.
ii. Two-sided P values will be calculated for all statistical tests. A value of P <0.05 is considered significant.

Analysis software
All statistical analyses will be performed using R version 3.6.1 (Foundation for Statistical Computing, Vienna, Austria).

Interim Analysis
An interim analysis will be conducted when 20 subjects are enrolled in at least one arm, and at least one subject is enrolled in each arm, to preliminarily evaluate the efficacy and safety of the drug combination in each arm. With the estimated enrollment speed, around 50% of subjects would have been enrolled by the interim analysis time point.

Quality Control and Quality Assurance
Regular supervision and inspection will be performed during the trial to ensure the implementation of study protocol. The raw data will be reviewed to ensure the consistency with data in the case report form.

Ethical Principle
The study procedure must strictly conform to the requirement of Good Clinical Practice of SFDA and Declaration of Helsinki.

Institutional Ethics Committee (IEC)
This protocol and written informed consent as well as material directly related with subjects should be submitted to ethics committee. The trial can only initiate after achievement of written approval of the ethics committee.

Informed consent form (ICF)
Prior to enrollment the investigators are responsible for oral and written consent about information including objective, procedure and potential risks of the study to every subject. The subject should be informed about the right to decide whether to participate in the trial and that subject is free to withdraw from trial any time willingly.
Subjects or their legal representatives will read and understand the informed consent form and sign it, and keep the copy of signature page.

Trials progress
Duration of inclusion:30 months (From October 2018 to February 2021).
Duration of follow-up after treatment:12 months after inclusion of the last subject.
Date of primary analysis: February 2021.

Data retention
The case report forms will be confirmed with signature by investigators.

Introduction
The FUTURE trial is a phase Ib/II biopsy-mandated, subtyping-based and genomic biomarker-guided umbrella trial to evaluate if targeting the subtypes therapeutically might improve patient outcomes.

Study Design
This is a Phase Ib/II, open-label, umbrella study.

Sample Size
Based on the different multi-gene expression profiles and the potential molecular characteristics of different pathways, seven treatment arms and six treatment groups were initially set up to eroll 20 patients per treatment arm. Therefore, a total of 140 patients were enrolled in this study. 3 or more than 3 of 20 patients in each arm group reached CR or PR, will be defined to reach the study end point.

Objective response rate, ORR
The primary endpoint will be the objective response rate (ORR) [PR+CR], with responders requiring a confirmatory response assessment no sooner than 4 weeks after the first responseassessment.Assessment of tumor response is based upon on-site readings by local radiologists,using RECIST1.1.

Disease Control Rate, DCR
One of the secondary endpoints will be disease control rate(DCR) [CR+PR+SD].

Progression Free Survival, PFS
One of the secondary endpoints will be progression free survival (PFS).
Progression-free survival is defined as the interval from the first dose start date to the date of disease progression defined as documented PD or death from any cause, whichever occurs first.

Overall Survival, OS
Overall survival is defined as the time fromthe date of the first dose start date to the date of death due to any cause. Patients without documentation of death at the time of the data cut off for analysis will be censored at the date the patient was last known to be alive or the data cut off date,whichever is earlier. The last known alive date is the last record in the study database. This date may be the maximum of the last visit date or last contact date, including telephone follow-up where the patientis known to be alive.

Other Analyses Related to Efficacy
Duration of response (DOR) will be calculated as the date of the first evaluation showing documented PR or CR to the date of the first PD or death, whichever is earlier.
Swimmer plots of treatment duration showing the date of progression or death, whichever is earlier, will be presented. Treatment ongoing status will be marked at the end of the plot. Waterfall plots of the percent change from baseline in target lesion measurement will be presented.
Example of swimmer plot: Example of waterfall plot:

Safety Endpoints
Safety will be assessed for overall safety population. Data will be presented in terms of AEs, laboratory data, and vital signs.

 Adverse Events
 Clinical Laboratory Evaluations  Vital Signs

Statistical Analysis
All analyses were based on descriptive statistics, without formal statistical hypothesis testing. As mentioned above, 3 or more than 3 of 20 patients in each arm group reached CR or PR, will be defined to reach the study end point.

Analysis Sets
Per-Protocl (PP) Set: The PP data set is defined restricted to all the cases that meet the inclusion criteria and complete the treatment plan, with good compliance, no banned compounds and fulfilled required contents in case report form, as well as the FUTURE Statistical Analysis Plan

68/74
observation record documents from subjects whose compliance satisfies the study protocol requirement.
Intention to Treat (ITT) Set: The ITT population includes all patients who have signed informed consent, regardless of their adherence with the entry criteria, regardless of the treatment they actually received, and regardless of subsequent withdrawal from treatment or deviation from the protocol.
Safety Set: This will include all patients who received at least 1 dose of medication, irrespective of dose.

Methods for Handling Missing Data
Missing or partial dates will not be imputed except for AE and concomitant medication data. In this case the listings will show these dates as missing, but the following approach will be used to define whether an AE is treatment-emergent or a therapy is considered a prior medication.

Dates
If the stop date is non-missing and the imputed start date is after the stop date, the stop date will be used as the start date.
(1) Missing day only • If the month and year of the AE/the concomitant medication are the same as the month and year of the first dose date, the first dose date will be used.
• If the month and year are before the month and year of the first dose date, the last day of the month will be assigned to the missing day.
• If the month and year are after the month and year of the first dose date, the first day of the month will be assigned to the missing day.
(2) Missing day and month • If the year is the same as the year of the first dose date, the first dose date will be used.

69/74
• If the year is prior to the year of the first dose date, December 31 st will be assigned to the missing fields.
• If the year is after the year of the first dose date, January 1 st will be assigned to the missing fields.
(3) Missing day, month, and year • The first dose date will be used.

Medication Stop Date
If the start date is non-missing and the imputed stop date is before the start date, the start date will be used. If the death date is available and the imputed stop date is after the death date, the death date will be used.
(1) Missing day only • The last day of the month will be assigned as the missing day.
(2) Missing day and month • December 31 st will be assigned to the missing fields.

(3) Missing day, month and year
• The event will be regarded as ongoing.

Missing/Partial Dates during Screening Visit
The following rules apply to dates recorded during the screening visits (eg, prior therapies/medications, medical history): (1) Missing day only • The first day of the month will be used if the year and the month are the same as those for the first dose of study drug. Otherwise, the 15 th will be used.
(2) Missing day and month FUTURE Statistical Analysis Plan

70/74
• If the year is the same as the year of the first dose of study drug, the 15 th of January will be used unless it is later than the first dose, in which case the date of the first of January will be used.
• If the year is not the same as the year of the first dose of study drug, the 15 th of June will be used, unless other data indicates that the date is earlier.
(3) Missing day, month, and year • No imputation will be applied.

Missing Last Dosing Date
Missing/incomplete last dose date from the treatment discontinuation page will be imputed as follows: (1) Missing day only • If the treatment discontinuation reason is death, the death date will be used.
• Else if the last available dosing date from dosing data matches the partial last dose date from the treatment discontinuation page, the last available dosing date will be used.
• Else the first day of the month will be used.
(2) Missing day and month • If the treatment discontinuation reason is death, the death date will be used.
• Else if the last available dosing date from dosing data matches the partial last dose date from the treatment discontinuation page, the last available dosing date will be used.
• Else January 1 will be used.

(3) Missing day, month and year
• If the treatment discontinuation reason is death, the death date will be used.
• Else the last available dosing date will be used.

71/74
The imputed last dose date will be compared to the available study discontinuation date and the data cutoff date. Then the earliest date will be used.

Subjects Disposition
Disposition in terms of number of patients screened/entered into the study, treated, permanently discontinued treatment, and reasons for treatment discontinuation will be summarized for the ITT Population and the Safety Population. The number of patients included in each population will be summarized by way of a flow chart. A by-patient listing for disposition will be provided, including whether the patient is included in each of the analysis sets, treatment status, date of stopping treatment, date of stopping study participation, reason for treatment discontinuation, study completion status, and survival follow-up status.
A separate listing will be provided for patients who were registered into the study but did not receive study drug and the reason for not receiving study drug. Screen failures and entered-butnot-dosed patients will be excluded from all analyses.

Demographics
Baseline characteristics will include age, Eastern Cooperative Oncology Group (ECOG) performance status, numbers of metastatic organs and metastaic sites at initial diagnosis and at screening.

Medical History
Prior systemic anticancer therapy will be summarized. Number of prior anticancer therapies for metastatic disease (＜3 prior lines vs 3-6 prior lines vs ＞6 prior lines) and types of prior chemotherapies for metastatic disease will also be summarized categorically.

Major Protocol Deviations
A by-patient listing with major study protocol violations and deviations will be provided for patients.

Protocol Violations to be Programmed
• Inclusion/Exclusion Criteria FUTURE Statistical Analysis Plan

72/74
-<2 weeks between previous treatment, immune therapy, chemotherapy, or investigational therapy for metastatic disease and start of treatment.
• Informed Consent -Patient's written informed consent not available.
-Patient's written informed consent too late (after start of study-specific procedures) • Prohibited Medication -Anticancer therapy during treatment.
-Radiation during treatment.
-Prophylactic medication of hematopoietic growth factors or blood transfusions before Cycle 1.

Efficacy Analysis
Formal statistical hypothesis-testing will not be performed. Descriptive statistics, data summaries and graphical methods will be used to assess the efficacy. The primary endpoint will be the objective response rate (ORR) [PR+CR], with responders requiring a confirmatory response assessment no sooner than 4 weeks after the first response assessment. Assessment of tumor response is based upon on-site readings by local radiologists, using RECIST 1.1. Earliest imaging date of associated imaging methods was used as the response assessment date. Secondary efficacy endpoints will include disease control rate (DCR, CR+PR+SD), progression-free survival (PFS), and overall survival (OS). PFS, and OS data will be analyzed via Kaplan-Meier method and 95% CI from Clopper-Pearson method with log-log transformation.

Safety Analysis
Safety will be assessed for Safety Population. Data will be presented in terms of AEs, laboratory data, and vital signs.

Extent of Exposure
The extent of exposure mainly summarizes the drug exposure time, the cumulative dosage and the drug intensity.

Adverse Events
Treatment-emergent adverse events (TEAEs) are defined as any AEs that begin or worsen on or after the start of study drug through 30 days after the last dose of study drug. All AEs will be coded using Medical Dictionary for Regulatory Activities (MedDRA) Version 20.0 unless otherwise specified. The severity will be graded based on the National Cancer Institute's (NCI) Common Terminology Criteria for Adverse Events (CTCAE) Version 4.03. All AEs will be listed. Only TEAEs will be summarized and will be referred to as AEs hereafter. Timing of AEs and concomitant medications will take account of the date and the time of the AE or concomitant medication. The frequency and severity of AEs will be tabulated by MedDRA SOC and PT. For this purpose, an AE that occurs more than once within each patient will be counted only once (at the worst CTCAE grade and relationship category).
Additional by-patient listings will be provided for AEs leading to on-treatment death, serious AEs (SAEs), and AEs leading to discontinuation of treatment (excluding AEs leading to death).

Laboratory Evaluations
Clinical laboratory data results will be reported in standard international units and Chinese conventional units. Baseline is defined as the last observation occurring prior to the first treatment administration of medication. Observations occurring on the same day as first treatment administration may be the baseline assessment only if the time of assessment occurs prior to the time of treatment. If this cannot be determined, the observation will be assumed to have occurred after dosing. If a lab value is reported using a non-numeric qualifier (eg, less than [<] a certain value, or greater than [>] a certain value), the given numeric value will be used in the summary statistics, ignoring the non-numeric qualifier. Hematology and serum chemistry data will be listed by patient and summarized by study visit. Actual values by visit and change-from-baseline will be summarized by mean, median, standard deviation, minimum, maximum and number of patients. Shift tables from baseline to worst CTCAE grade on treatment and from worst to last CTCAE grade on treatment will be presented where CTCAE grade is available. Shift tables will be presented based on CTCAE v4.03 criteria, using the grades 1 through 4 as well as a grade 0 indicating no abnormality. These shift tables