Advertisement

Phase II randomised discontinuation trial of cabozantinib in patients with advanced solid tumours

Open AccessPublished:October 20, 2017DOI:https://doi.org/10.1016/j.ejca.2017.09.011

      Highlights

      • Cabozantinib was evaluated in a phase II randomised discontinuation trial (RDT).
      • Cabozantinib demonstrated antitumour activity in multiple solid-tumour types.
      • Results from this RDT supported initiation of phase III programs in castration-resistant prostate cancer and hepatocellular carcinoma.

      Abstract

      Background

      Cabozantinib is an inhibitor of tyrosine kinases, including MET, vascular endothelial growth factor receptor, AXL and RET. This multi-cohort phase II randomised discontinuation trial explored anticancer activity of cabozantinib in nine tumour types.

      Patients and methods

      Cabozantinib was administered (100 mg, once daily) to patients with advanced, recurrent or metastatic cancers. Those with stable disease at week 12 were randomised 1:1 to cabozantinib or placebo. Primary end-points were objective response rate (ORR) at week 12 and progression-free survival (PFS) in the randomised phase.

      Results

      A total of 526 patients were enrolled. The highest ORR was observed in ovarian cancer (OC) (21.7%); the largest PFS benefit was observed in castration-resistant prostate cancer (CRPC) (median 5.5 versus 1.4 months for placebo; hazard ratio 0.14, 95% confidence interval: 0.04, 0.52). Disease control rates were >40% for CRPC, OC, melanoma, metastatic breast cancer (MBC), hepatocellular carcinoma (HCC) and non–small cell lung cancer. Median duration of response ranged from 3.3 (MBC) to 11.2 months (OC). Encouraging efficacy results and symptomatic improvements prompted early suspension of the randomised stage and conversion to open-label non-randomised expansion cohorts. Dose reductions to manage adverse events (AEs) occurred in 48.7% of patients. The most frequent grade III–IV AEs were fatigue (12.4%), diarrhoea (10.5%), hypertension (10.5%) and palmar-plantar erythrodysesthesia syndrome (8.7%).

      Conclusions

      Clinical antitumour activity of cabozantinib was observed in a subset of tumour types: CRPC and OC were evaluated further in expansion cohorts. Phase III programs were initiated in CRPC and HCC. Interpretation of efficacy outcomes was limited by early termination of the randomised portion of the trial.

      Trial registration number

      NCT00940225.

      Keywords

      1. Introduction

      Cabozantinib, a small-molecule tyrosine kinase inhibitor (TKI), inhibits multiple receptor tyrosine kinases that are important in cancer pathobiology, including MET, the three vascular endothelial growth factor receptors (VEGFRs), AXL and RET. In preclinical studies, cabozantinib treatment was shown to inhibit tumour angiogenesis, invasiveness, metastasis and tumour progression [
      • Yakes F.M.
      • Chen J.
      • Tan J.
      • Yamaguchi K.
      • Shi Y.
      • Yu P.
      • et al.
      Cabozantinib (XL184), a novel MET and VEGFR2 inhibitor, simultaneously suppresses metastasis, angiogenesis, and tumor growth.
      ,
      • Nguyen H.M.
      • Ruppender N.
      • Zhang X.
      • Brown L.G.
      • Gross T.S.
      • Morrissey C.
      • et al.
      Cabozantinib inhibits growth of androgen-sensitive and castration-resistant prostate cancer and affects bone remodeling.
      ,
      • Dai J.
      • Zhang H.
      • Karatsinides A.
      • Keller J.M.
      • Kozloff K.M.
      • Aftab D.T.
      • et al.
      Cabozantinib inhibits prostate cancer growth and prevents tumor-induced bone lesions.
      ]. In xenograft models, the primary effect of cabozantinib was the inhibition of tumour growth rather than volumetric shrinkage, suggesting that the primary clinical benefit of cabozantinib may be disease stabilisation. Therefore, classical oncology paradigms for phase II clinical evaluation (e.g. single-arm, non-controlled studies using objective response rate [ORR] as the primary end-point) might not adequately demonstrate the antitumour activity of cabozantinib.
      Tumour stabilisation is best measured through the use of a controlled study, which ensures that differences detected between experimental and control arms are due to cytostatic activity and not the inherent indolent nature of a tumour [
      • Gray R.
      • Manola J.
      • Saxman S.
      • Wright J.
      • Dutcher J.
      • Atkins M.
      • et al.
      Phase II clinical trial design: methods in translational research from the Genitourinary Committee at the Eastern Cooperative Oncology Group.
      ,
      • Rosner G.L.
      • Stadler W.
      • Ratain M.J.
      Randomized discontinuation design: application to cytostatic antineoplastic agents.
      ]. In settings without an appropriate active control, a randomised discontinuation trial (RDT) attempts to assess cytostatic effect while limiting placebo exposure [
      • Ratain M.J.
      • Eisen T.
      • Stadler W.M.
      • Flaherty K.T.
      • Kaye S.B.
      • Rosner G.L.
      • et al.
      Phase II placebo-controlled randomized discontinuation trial of sorafenib in patients with metastatic renal cell carcinoma.
      ,
      • Stadler W.M.
      The randomized discontinuation trial: a phase II design to assess growth-inhibitory agents.
      ,
      • Stadler W.M.
      • Rosner G.
      • Small E.
      • Hollis D.
      • Rini B.
      • Zaentz S.D.
      • et al.
      Successful implementation of the randomized discontinuation trial design: an application to the study of the putative antiangiogenic agent carboxyaminoimidazole in renal cell carcinoma–CALGB 69901.
      ]. In RDTs, patients achieving tumour stabilisation while receiving experimental treatment during an initial lead-in period are identified as potentially experiencing cytostatic benefit. At a predefined time point, these patients are randomised to either continue experimental treatment or to receive placebo. Through this process, heterogeneity to drug response is decreased within the randomised population compared with studies carrying out randomisation at initiation.
      The objective of the present multi-cohort phase II RDT was to evaluate the clinical anticancer activity of cabozantinib for multiple solid-tumour types. The trial used an adaptive design, whereby a study oversight committee reviewed ongoing efficacy and safety data and made recommendations for cohort enrolment and study conduct.

      2. Patients and methods

      2.1 Patients

      This phase II RDT evaluated the antitumour activity of cabozantinib in nine solid-tumour types, which defined individual cohorts (ClinialTrials.gov, NCT00940225). The RDT design consisted of a 12-week lead-in stage with open-label treatment of cabozantinib followed by a placebo-controlled randomised stage for patients who achieved stable disease (SD) during the lead-in stage (Fig. 1). The study protocol was approved by all local/institutional review boards for human investigations as appropriate, and all patients provided written informed consent. Eligible patients had pathologically confirmed malignant solid tumours in an advanced, recurrent or metastatic stage with measurable disease by Response Evaluation Criteria in Solid Tumours (RECIST) version 1.0 [
      • Therasse P.
      • Arbuck S.G.
      • Eisenhauer E.A.
      • Wanders J.
      • Kaplan R.S.
      • Rubinstein L.
      • et al.
      New guidelines to evaluate the response to treatment in solid tumors. European Organization for Research and Treatment of Cancer, National Cancer Institute of the United States, National Cancer Institute of Canada.
      ]. Patients were enrolled in the following tumour cohorts: castration-resistant prostate cancer (CRPC), melanoma (MEL), ovarian cancer (OC; epithelial ovarian cancer, primary peritoneal or fallopian tube carcinoma), non–small cell lung cancer (NSCLC), metastatic breast cancer (MBC), hepatocellular carcinoma (HCC), gastric/gastroesophageal junction adenocarcinoma (GC), small-cell lung cancer (SCLC) and pancreatic adenocarcinoma (PAC). Additional eligibility criteria for the tumour cohorts are included in the Supplementary Information-Tumour-specific cohort eligibility.
      Fig. 1
      Fig. 1Multi-cohort randomised discontinuation trial design. CR, complete response; CRPC, castration-resistant prostate cancer; GC, gastric/gastroesophageal junction adenocarcinoma; HCC, hepatocellular carcinoma; MBC, metastatic breast cancer; MEL, melanoma; NSCLC, non–small cell lung cancer; OC, ovarian cancer; PAC, pancreatic adenocarcinoma; PD, progressive disease; PR, partial response; SCLC, small-cell lung cancer; SD, stable disease.

      2.2 Study drug administration

      During the 12-week lead-in stage, patients received 100-mg daily doses of cabozantinib capsules. After this lead-in stage, study sites randomised patients via an interactive web response system, which provided a blinded study drug package identifier. Cabozantinib and placebo were colour-, size- and shape-matched; patients and study site personnel were blinded to treatment assignment. The randomisation schedule was stratified by tumour type and within each tumour type permuted blocks were used to maintain the 1:1 ratio between cabozantinib and placebo arms.
      Protocol-defined dose modifications, including interruptions (up to 3 weeks) and dose reductions, were used to manage adverse events (AEs). Treatment could be reduced at the discretion of the investigator to 60 mg, 40 mg and then 20 mg. Patients could continue treatment until disease progression or unacceptable toxicity. Study treatment in the randomised stage was administered in a double-blind fashion.

      2.3 Study assessments

      The primary end-point of the lead-in phase was ORR at week 12, and the primary end-point of the randomised phase was PFS. Additional end-points included the disease control rate (DCR) at week 12, duration of response (DOR) and PFS from first dose of cabozantinib in all patients. Tumour assessments by computed tomography or magnetic resonance imaging were performed at screening and every 6 weeks after the first dose of study treatment. Tumour response and progression were determined by the investigator using RECIST 1.0. Imaging continued until permanent discontinuation of study treatment due to documented progressive disease (PD) or initiation of subsequent anticancer therapy. Changes in tumour markers were not considered for determination of progression. Although study treatment could be discontinued upon clinical deterioration, every effort was made to document radiographic progression.
      Safety was evaluated by assessment of AEs, vital signs, electrocardiograms and standard laboratory tests. Seriousness of AEs, severity grade per Common Terminology Criteria for Adverse Events, version 3.0, and relationship to study treatment were assessed by the investigator.

      2.4 Statistical considerations

      The study used an adaptive design with an assumed SD rate of 35% at week 12 to estimate the target enrolment from the number of patients planned for randomisation (Supplementary Fig. S1). The target enrolment of 200 patients per cohort was chosen to achieve the goal of 70 randomised patients and 52 events post randomisation. The design had an 80% power to detect a hazard ratio (HR) of 0.5 for progression-free survival (PFS) between cabozantinib and placebo with a two-sided alpha of 0.1. Under this design, the minimum observed effect that would demonstrate statistical significance for PFS was a 58.7% improvement (HR = 0.63). Given that this was a phase II trial intended to generate hypotheses for further study, there was no adjustment for multiple comparisons. Up to an additional 300 patients could be enrolled in cohorts deemed not suitable for full accrual to the randomised stage. The study oversight committee was responsible for prioritisation of tumour cohorts based on continuous review of emerging efficacy and safety data. Stopping rules were based on conditional probability to project futility (e.g. enrolment was suspended in tumour cohorts deemed unlikely to benefit from cabozantinib or if the ORR after the 12-week lead-in stage was higher than anticipated such that the randomised stage was unwarranted).
      ORR was defined as the proportion of patients with a best overall response of confirmed complete response (CR) or partial response (PR) during the lead-in stage. PFS in the randomised stage was defined as the time from randomisation until the earlier of radiographic progression or death. The comparison of PFS between the cabozantinib and placebo arms was performed using the log-rank test in the randomised population. The median duration of PFS and the associated 95% confidence interval (CI) for each treatment arm were estimated using the Kaplan–Meier method. The HR was estimated using a Cox regression model with treatment as the only covariate.
      DOR was summarised for patients treated in the lead-in and/or randomised stage and was defined as the time from the tumour assessment of the first documented PR or CR that was subsequently confirmed at least 28 d later until the date of disease progression. Median DOR and the associated 95% CI were estimated using the Kaplan–Meier method. DCR was defined as the proportion of patients treated in the lead-in stage with a best overall response of CR, PR or SD at Week 12.
      PFS from first dose of cabozantinib in all patients was estimated using a piecewise method. For this analysis, all patients contributed to the PFS estimate for the first 12 weeks of therapy. PFS after week 12 was estimated using a weighted average of group-specific PFS from the two groups continuously treated with cabozantinib: 1) continued open label and 2) randomised to cabozantinib. When combining the group-specific PFS estimates, the weights were the proportion of the sample size of patients in the group of continued open label versus patients in the randomised population [
      • Ratain M.J.
      • Eisen T.
      • Stadler W.M.
      • Flaherty K.T.
      • Kaye S.B.
      • Rosner G.L.
      • et al.
      Phase II placebo-controlled randomized discontinuation trial of sorafenib in patients with metastatic renal cell carcinoma.
      ].

      3. Results

      3.1 Patients and treatment

      Between September 2, 2009 and July 19, 2011, 795 patients were screened for eligibility, and 526 patients were enrolled in nine tumour-specific cohorts at 42 sites in the United States, Belgium, Israel and Taiwan. Baseline characteristics and cohort distribution are summarised in Supplementary Table S1, and a breakdown of metastatic areas can be found in Supplementary Table S2. Among 526 patients who received ≥1 dose of study treatment, 233 (44.3%) discontinued before completing the 12-week lead-in stage, 135 (25.7%) underwent randomisation at the end of week 12 and 158 (30.0%) continued cabozantinib in the open-label extension without being randomised, either because they experienced PR or CR during the lead-in stage or because randomisation was discontinued before they reached week 12. Among 135 randomised patients, 67 were randomly assigned to receive cabozantinib and 68 to placebo. Of patients randomised to placebo, 53 crossed over to receive open-label cabozantinib, either after progression on placebo or when the study was unblinded, and 15 discontinued without crossing over (Supplementary Fig. S2). Median duration of exposure to cabozantinib was 86 d (range: 4–1232 d), and median average daily dose was 69.8 mg (range: 15.4–100 mg). The most common reasons for treatment discontinuation in the study were radiographic progression (54.4%), AEs (20.3%) and clinical deterioration (10.1%).

      3.2 Tumour response during the lead-in stage per investigator

      An ORR ≥10% was observed in three tumour cohorts: OC, 21.7% (15/70); MBC, 13.6% (6/45) and NSCLC, 10% (6/60) (Table 1). One confirmed CR was observed in the OC cohort. A reduction in the sum of target lesion size was observed in 301 of 526 (57.2%) patients across all tumour types (Fig. 2). DCRs at week 12 ≥ 40% were observed for HCC, CRPC, OC, MBC, SCLC and MEL. Low rates (≤15%) of PD as best response were observed in HCC, CRPC and OC.
      Table 1Overview of efficacy results.
      CRPC
      For patients who achieved the best response of a confirmed CR/PR, the first response date and confirmation date may occur at different study stages. Only first CR/PR occurrence is being summarised.
      MEL
      For patients who achieved the best response of a confirmed CR/PR, the first response date and confirmation date may occur at different study stages. Only first CR/PR occurrence is being summarised.
      OC
      For patients who achieved the best response of a confirmed CR/PR, the first response date and confirmation date may occur at different study stages. Only first CR/PR occurrence is being summarised.
      ,
      For each cohort, one patient was excluded from the analysis during the lead-in stage due to lack of measurable disease at baseline.
      NSCLC
      For patients who achieved the best response of a confirmed CR/PR, the first response date and confirmation date may occur at different study stages. Only first CR/PR occurrence is being summarised.
      MBC
      For patients who achieved the best response of a confirmed CR/PR, the first response date and confirmation date may occur at different study stages. Only first CR/PR occurrence is being summarised.
      ,
      For each cohort, one patient was excluded from the analysis during the lead-in stage due to lack of measurable disease at baseline.
      HCC
      For patients who achieved the best response of a confirmed CR/PR, the first response date and confirmation date may occur at different study stages. Only first CR/PR occurrence is being summarised.
      GCSCLCPAC
      (N = 171)(N = 77)(N = 70)(N = 60)(N = 45)(N = 41)(N = 21)(N = 21)(N = 20)
      Lead-in stage
      Best overall response, n (%)
       Confirmed CR001 (1.4)000000
       Confirmed PR8 (4.7)4 (5.2)14 (20.3)6 (10.0)6 (13.6)2 (4.9)1 (4.8)1 (4.8)0
       SD129 (75.4)43 (55.8)35 (50.7)29 (48.3)25 (56.8)31 (75.6)8 (38.0)12 (57.1)11 (55.0)
       PD19 (11.1)19 (24.7)9 (13.0)12 (20.0)9 (20.5)3 (7.3)10 (47.6)7 (33.3)5 (25.0)
       Unable to evaluate1 (0.6)03 (4.3)1 (1.7)01 (2.4)000
       Missing14 (8.2)11 (14.3)7 (10.1)12 (20.0)4 (9.1)4 (9.8)2 (9.5)1 (4.8)4 (20.0)
      ORR (CR + PR), n (%)8 (4.7)4 (5.2)15 (21.7)6 (10.0)6 (13.6)2 (4.9)1 (4.8)1 (4.8)0
       95% CI2.0, 8.71.8, 12.413.1, 32.44.4, 20.36.0, 25.70.9, 16.10.2, 21.80.2, 21.80%, 15.4
      DCR (ORR + SD), n (%)
      DCR determined at week 12.
      112 (65.5)33 (42.9)35 (50.0)23 (38.3)21 (46.7)27 (65.9)7 (33.3)9 (42.9)7 (35.0)
       95% CI58.0, 72.332.0, 54.738.5, 61.526.1, 51.831.7, 61.650.0, 79.514.6, 55.121.8, 66.015.4, 58.9
      Randomised stage
      Patients (cabo, pbo), n31 (14, 17)26 (13, 13)13 (7, 6)16 (8, 8)10 (5, 5)22 (10, 12)5 (3, 2)6 (4, 2)6 (3, 3)
      PFS cabo event count975439132
       Median PFS (95% CI), mo5.5 (4.0, 8.5)4.5 (1.8, 15.4)4.9 (3.9, 13.6)2.4 (1.4, 2.9)6.9 (5.1, 13.5)2.5 (1.3, 6.8)UE (2.3, UE)5.5 (1.9, 6.9)2.8 (1.5, 2.8)
      PFS pbo event count11114659221
       Median PFS (95% CI), mo1.4 (1.3, 2.7)2.8 (1.5, 5.5)1.4 (1.3, UE)2.4 (1.4, 2.66)1.1 (0.2, 1.4)1.4 (1.3, 4.2)1.0 (0.7, 1.34)1.5 (0.9, 2.2)0.7 (UE, UE)
       HR (95% CI)0.14 (0.04, 0.52)0.47 (0.16, 1.35)UE0.94 (0.25, 3.54)UE0.82 (0.31, 2.12)UE0.24 (0.02, 2.67)UE
       Log-rank test P value0.0010.150.0040.930.0020.660.090.210.08
      Overall study
      DOR, cabo patients, n15615663NA
      Median DOR was not determined for this cohort because of the low number of responses observed.
      NA
      Median DOR was not determined for this cohort because of the low number of responses observed.
      NA
      Median DOR was not determined for this cohort because of the low number of responses observed.
       Median DOR (95% CI), mo5.6 (4.3, 8.3)5.6 (3.0, 13.9)11.2 (7.2, UE)6.9 (5.3, 10.7)3.3 (2.8, 4.2)4.2 (3.0, UE)NA
      Median DOR was not determined for this cohort because of the low number of responses observed.
      NA
      Median DOR was not determined for this cohort because of the low number of responses observed.
      NA
      Median DOR was not determined for this cohort because of the low number of responses observed.
      PFS from start of study, cabo patients, n1717770604541212120
       Median PFS
      Estimated using a piecewise method as previously described [6].
      , mo
      6.82.85.54.05.24.31.43.42.7
      CI, confidence interval; cabo, cabozantinib; CR, complete response; CRPC, castration-resistant prostate cancer; DCR, disease control rate; DOR, duration of objective response; GC, gastric/gastroesophageal junction adenocarcinoma; HCC, hepatocellular carcinoma; HR, hazard ratio; MBC, metastatic breast cancer; MEL, melanoma; NA, not applicable; NSCLC, non–small cell lung cancer; OC, ovarian cancer; ORR, objective response rate; PAC, pancreatic adenocarcinoma; pbo, placebo; PD, progressive disease; PFS, progression-free survival; PR, partial response; SCLC, small-cell lung cancer; SD, stable disease; UE, un-evaluable.
      Best overall response was the proportion of patients treated in the lead-in stage with CR or PR through the week 12 tumour assessment; DCR was the proportion of patients treated in the lead-in stage with an overall response of CR, PR and SD at the week 12 tumour assessment.
      a For patients who achieved the best response of a confirmed CR/PR, the first response date and confirmation date may occur at different study stages. Only first CR/PR occurrence is being summarised.
      b For each cohort, one patient was excluded from the analysis during the lead-in stage due to lack of measurable disease at baseline.
      c DCR determined at week 12.
      d Median DOR was not determined for this cohort because of the low number of responses observed.
      e Estimated using a piecewise method as previously described
      • Ratain M.J.
      • Eisen T.
      • Stadler W.M.
      • Flaherty K.T.
      • Kaye S.B.
      • Rosner G.L.
      • et al.
      Phase II placebo-controlled randomized discontinuation trial of sorafenib in patients with metastatic renal cell carcinoma.
      .
      Fig. 2
      Fig. 2Best change of the sum of target lesions from baseline in all evaluable patients (A) and in evaluable patients who achieved ≥30% reduction (B). Note: 3 of 459 patients who were evaluable for objective response at week 12 were not evaluable for this analysis. CRPC, castration-resistant prostate cancer; GC, gastric/gastroesophageal junction adenocarcinoma; HCC, hepatocellular carcinoma; MBC, metastatic breast cancer; MEL, melanoma; NSCLC, non–small cell lung cancer; OC, ovarian cancer; PAC, pancreatic adenocarcinoma; SCLC, small-cell lung cancer.

      3.3 Progression-free survival

      Assessment of PFS during the randomised stage indicated improvements in disease control with cabozantinib compared with placebo in CRPC (median PFS, 5.5 versus 1.4 months; HR = 0.14, 95% CI: 0.04, 0.52), MEL (4.5 versus 2.8 months; HR = 0.47, 95% CI: 0.16, 1.35) and HCC (2.5 versus 1.4 months; HR = 0.82, 95% CI: 0.31, 0.12) (Table 1, Supplementary Fig. S3). Characterisation of PFS during the randomisation stage for the other tumour cohorts was not possible, given the low number of events.
      In the evaluation of PFS from the first dose of cabozantinib in all patients, the highest median PFS results were observed in the CRPC (6.8 months), OC (5.5 months), HCC (5.2 months), MBC (4.3 months) and NSCLC (4.0 months) cohorts.

      3.4 Safety

      AEs, irrespective of causality, occurred in 99.6% of patients with 392 (74.5%) experiencing at least one grade ≥III AE (Table 2). The most common grade ≥III AEs were fatigue, 65 (12.4%); diarrhoea, 55 (10.5%); hypertension, 55 (10.5%) and palmar-plantar erythrodysesthesia syndrome, 46 (8.7%). Overall, 390 patients (74.1%) had AEs leading to dose reduction or interruption. Dose reductions across all tumour cohorts occurred in 256 (48.7%) patients, with 85 (16.2%) having two dose reductions and 20 (3.8%) having three dose reductions. Median time to first dose reduction was 47.5 d (range, 11–470). Deaths occurring through 30 d after the last dose of cabozantinib were mainly related to disease progression. Eleven patients (2.1%) continuously treated with cabozantinib expired because of causes other than disease progression (1 patient each died of acute peritonitis, bilateral pneumonia, enterocutaneous fistula, gastrointestinal bleeding, haemorrhage, perforated bowel, pneumonia, possible respiratory compromise, respiratory insufficiency, septic shock and unexplained death). One patient who crossed over from placebo to cabozantinib had an unknown cause of death, and one patient died of intracerebral haemorrhage while receiving treatment with placebo.
      Table 2Adverse events in ≥10% of patients (any grade) or ≥5% of patients (grade ≥III) irrespective of whether the event was considered by the investigator to be related to cabozantinib treatment (N = 526).
      Preferred termAll, n (%)Grade ≥III, n (%)
      Patients with at least one event524 (99.6)392 (74.5)
      Fatigue345 (65.6)65 (12.4)
      Diarrhoea317 (60.3)55 (10.5)
      Nausea273 (51.9)16 (3.0)
      Decreased appetite271 (51.5)26 (4.9)
      Vomiting196 (37.3)18 (3.4)
      Palmar-plantar erythrodysesthesia syndrome179 (34.0)46 (8.7)
      Constipation171 (32.5)11 (2.1)
      Weight decreased165 (31.4)11 (2.1)
      Dysgeusia156 (29.7)0
      Dysphonia142 (27.0)0
      Hypertension127 (24.1)55 (10.5)
      Abdominal pain126 (24.0)29 (5.5)
      Asthenia118 (22.4)32 (6.1)
      Dyspnoea109 (20.7)21 (4.0)
      Rash102 (19.4)5 (1.0)
      Stomatitis99 (18.8)4 (0.8)
      Aspartate aminotransferase increased94 (17.9)16 (3.0)
      Pain in extremity90 (17.1)11 (2.1)
      Cough88 (16.7)2 (0.4)
      Dry mouth87 (16.5)0
      Dizziness87 (16.5)4 (0.8)
      Mucosal inflammation84 (16.0)4 (0.8)
      Dehydration84 (16.0)29 (5.5)
      Headache80 (15.2)4 (0.8)
      Back pain79 (15.0)15 (2.9)
      Oedema peripheral78 (14.8)4 (0.8)
      Alanine aminotransferase increased71 (13.5)10 (1.9)
      Hypomagnesaemia70 (13.3)2 (0.4)
      Oral pain68 (12.9)2 (0.4)
      Anaemia66 (12.5)21 (4.0)
      Hypothyroidism66 (12.5)2 (0.4)
      Dyspepsia64 (12.2)2 (0.4)
      Hypokalemia64 (12.2)15 (2.9)
      Urinary tract infection64 (12.2)7 (1.3)
      Oropharyngeal pain60 (11.4)1 (0.2)
      Thrombocytopaenia58 (11.0)11 (2.1)
      Muscle spasms56 (10.6)0
      Dry skin54 (10.3)0
      Pulmonary embolism29 (5.5)29 (5.5)

      4. Discussion

      The goal of this RDT was to use the multi-tumour design as a screening tool to identify up to two tumour-specific cohorts that could be expanded to provide a sufficient number of patients with SD during the lead-in stage for generation of clinically meaningful PFS results during the randomised stage.
      The study accrued within the estimated timelines. During the lead-in stage, cabozantinib demonstrated preliminary clinical activity in multiple tumour types. ORR ≥10% was observed in three tumour cohorts (OC, MBC and NSCLC), and DCR at week 12 ≥ 40% occurred in six tumour cohorts (HCC, CRPC, OC, MBC, NSCLC and MEL). Responses in patients treated with cabozantinib appeared to be durable and ranged from 3.3 months in MBC to 11.2 months in OC. In the three tumour cohorts with the highest numbers of patients in the randomised stage (CRPC, MEL and HCC), improved median PFS for cabozantinib treatment compared with placebo was observed. The analysis of overall PFS from the start of the study in the CRPC and HCC cohorts supported the PFS results obtained in the randomised stage.
      The RDT design required periodic review of efficacy and safety across all tumour cohorts to adequately maintain enrolment within the scope of the study. Review of interim efficacy results led to the closure of the GC, SCLC and PAC cohorts after approximately 20 enrolled patients because these participants were deemed unlikely to benefit from cabozantinib. The other cohorts were allowed to continue with enrolment beyond 20 patients. The highest enrolment was achieved in the CRPC cohort (N = 171). Suspension of the randomised stage was based on better than anticipated efficacy results (e.g. ORR in the OC cohort and high DCRs) as well as reports of symptomatic improvements (e.g. clinically relevant pain palliation in patients with CRPC with bone metastases), bone scan resolution and rapid clinical deterioration among patients randomised to placebo after achieving SD with cabozantinib during the lead-in stage (Supplementary Fig. S1). For these reasons, the RDT was converted to an open-label, non-randomised study with two expansion cohorts in CRPC and OC. Detailed results for the CRPC, MBC, MEL, HCC and OC cohorts as well as the CRPC non-randomised expansion cohort are reported separately [
      • Smith D.C.
      • Smith M.R.
      • Sweeney C.
      • Elfiky A.A.
      • Logothetis C.
      • Corn P.G.
      • et al.
      Cabozantinib in patients with advanced prostate cancer: results of a phase II randomized discontinuation trial.
      ,
      • Smith M.R.
      • Sweeney C.J.
      • Corn P.G.
      • Rathkopf D.E.
      • Smith D.C.
      • Hussain M.
      • et al.
      Cabozantinib in chemotherapy-pretreated metastatic castration-resistant prostate cancer: results of a phase II nonrandomized expansion study.
      ,
      • Tolaney S.M.
      • Nechushtan H.
      • Ron I.G.
      • Schoffski P.
      • Awada A.
      • Yasenchak C.A.
      • et al.
      Cabozantinib for metastatic breast carcinoma: results of a phase II placebo-controlled randomized discontinuation study.
      ,
      • Kelley R.
      • Verslype C.
      • Cohn A.
      • Yang T.
      • Su W.
      • Burris H.
      • et al.
      Cabozantinib in hepatocellular carcinoma: results of a phase 2 placebo-controlled randomized discontinuation study.
      ,
      • Daud A.
      • Kluger H.
      • Kurzrock R.
      • Schimmoller F.
      • Weitzman A.
      • Samuel T.
      • et al.
      Phase II randomised discontinuation trial of the MET/VEGF receptor inhibitor cabozantinib in metastatic melanoma.
      ,
      • Vergote I.B.
      • Smith D.C.
      • Berger R.
      • Kurzrock R.
      • Vogelzang N.J.
      • Sella A.
      • et al.
      A phase 2 randomised discontinuation trial of cabozantinib in patients with ovarian carcinoma.
      ].
      The termination of the randomised stage of the study before achieving the maximum-allowed cohort enrolment limited the ability to draw definitive conclusions about PFS results. Nonetheless, the totality of the efficacy results throughout the study warranted further evaluation of cabozantinib activity in selected tumour types, including phase III studies in CRPC (ClinialTrials.gov, NCT01605227 [
      • Smith M.
      • De Bono J.
      • Sternberg C.
      • Le Moulec S.
      • Oudard S.
      • De Giorgi U.
      • et al.
      Phase III study of cabozantinib in previously treated metastatic castration-resistant prostate cancer: COMET-1.
      ] and NCT01522443) and an ongoing phase III study in HCC (NCT01908426).
      The safety results with cabozantinib were consistent with those for patients with advanced cancer treated with VEGFR TKIs [
      • Chen H.X.
      • Cleck J.N.
      Adverse effects of anticancer agents that target the VEGF pathway.
      ]. The observed AEs were similar to findings in other cabozantinib trials, and AEs were generally manageable with dose modifications and optimised supportive care.
      In addition to the tumour types evaluated in this study, cabozantinib has been evaluated and granted regulatory approval for other oncology indications. Based on results from randomised phase III trials, cabozantinib tablets (60 mg) are approved for patients with advanced renal-cell carcinoma who have received prior anti-angiogenic therapy [
      • Choueiri T.K.
      • Escudier B.
      • Powles T.
      • Mainwaring P.N.
      • Rini B.I.
      • Donskov F.
      • et al.
      Cabozantinib versus everolimus in Advanced Renal-Cell Carcinoma.
      ,
      • Choueiri T.K.
      • Escudier B.
      • Powles T.
      • Tannir N.M.
      • Mainwaring P.N.
      • Rini B.I.
      • et al.
      Cabozantinib versus everolimus in advanced renal cell carcinoma (METEOR): final results from a randomised, open-label, phase 3 trial.
      ], and cabozantinib capsules (140 mg) are approved for patients with progressive, metastatic medullary thyroid cancer [
      • Elisei R.
      • Schlumberger M.J.
      • Muller S.P.
      • Schoffski P.
      • Brose M.S.
      • Shah M.H.
      • et al.
      Cabozantinib in progressive medullary thyroid cancer.
      ].
      In summary, this phase II multi-cohort RDT allowed simultaneous evaluation of clinical activity and safety for cabozantinib in several tumour types. CRPC and OC were identified for further evaluation in open-label, non-randomised expansion cohorts, and the data collected for the CRPC and HCC cohorts supported the initiation of phase III programs in these indications.

      Conflict of interest statement

      P. Schöffski is the Lead Investigator on Exelixis, Inc., clinical trials. He consulted for Exelixis, Inc. and received financial support for advisory functions, educational activities and related travel. D. Smith gained research support from Exelixis, Inc. R. Kurzrock received research funding from Exelixis (payment to institution for the conduct of the study). A. Daud received grant support from Exelixis, Inc. N. Vogelzang is a speaker for Exelixis, Inc. Y. Lee was an employee of Exelixis, Inc. at the time this study was conducted. C. Scheffold is an employee of Exelixis, Inc. G. Shapiro received research funding from Exelixis (payment to institution for the conduct of the study). All remaining authors have declared no conflicts of interest.

      Funding

      This work was supported by Exelixis, Inc. 210 E. Grand Avenue, South San Francisco, CA 94080. Financial support for medical editorial assistance was provided by Exelixis, Inc. No grant number is applicable.

      Acknowledgements

      The authors thank the patients for their participation in this study. They are grateful to the SOC (P. Schöffski, University Hospitals Leuven, Belgium; M. Gordon, Pinnacle Oncology, Scottsdale, AZ, USA; D. Smith, University of Michigan, Ann Arbor, MI, USA) and all investigators and institutions involved in the study. A full list of investigators and institutions can be found in the Appendix. They would also like to thank Douglas Clary and Charles Blazey (Exelixis, Inc.) for their assistance in the preparation of this manuscript. Editorial assistance was provided by Michael Raffin (Fishawack Communications) and supported by Exelixis, Inc.

      Appendix A. Supplementary data

      Appendix.

      The authors are grateful to the following investigators and institutions for their participation in this study:

      Belgium:

      • A. Awada (Institut Jules Bordet, Brussels), J. De Grève (Universitaire Ziekenhuis Brussel), S. Holbrechts (Center Hospitalier Universitaire Ambroise), G. Jerusalem (Centre Hospitalier Sart Tilman Liège) and K. Nackaerts, P. Schöffski, E. Van Cutsem and I. Vergote (University Hospitals Leuven, Leuven Cancer Institute)

      Israel:

      • R. Berger (Sheba Medical Center, Tel Hashomer), H. Nechushtan (Hadassah Hebrew University Medical Center), I. Ron (Tel Aviv Sourasky Medical Center) and A. Sella (Assaf Harofeh Medical Centre)

      Taiwan:

      • C.C. Lin (National Taiwan University Hospital), W.C. Su (National Cheng Kung University Hospital), C.J. Tsao (Chi Mei Hospital) and T.S. Yang (Chang Gung Memorial Hospital LinKou)

      US:

      • A. Cohn (Rocky Mountain Cancer Centers), H.A. Burris III (Tennessee Oncology, PLLC), A. Daud (University of California, San Francisco), J. Davis (University of Kansas Cancer Center), G. Edelman (Mary Crowley Medical Research Center), C. Freter and M. Perry (University of Missouri, Ellis Fischel Cancer Center), S. Gadgeel and R.T. Morris (Karmanos Cancer Institute), M. Galsky and N.J. Vogelzang (US Oncology, LLC, Comprehensive Cancer Centers of Nevada), M.S. Gordon (Pinnacle Oncology Hematology), N. Haas (University of Pennsylvania), L. Hart (Florida Cancer Specialists), B. Hellerstedt (Texas Oncology Cancer Center), D.S. Hong (The University of Texas, MD Anderson Cancer Center), R. Kurzrock (The University of Texas, MD Anderson Cancer Center and University of California, San Diego, Moores Cancer Center), H. Kluger (Yale University School of Medicine), P. Lara Jr (University of California Davis Comprehensive Cancer Center), J. Lutzky (Mount Sinai Comprehensive Cancer Center), A. Moussa (Cancer Care Associates), D. Richards (Texas Oncology/Tyler Cancer Center), T. Samuel and A Nayak (Georgia Regents University), M. Sgroi (Indiana University Health, Central Indiana Cancer Center), G. Shapiro (Dana-Farber Cancer Institute), B.I. Sikic (Stanford Cancer Center), D.C. Smith (University of Michigan Health System), A.I. Spira (US Oncology Research, LLC and Virginia Cancer Specialists, PC), J.J. Stephenson Jr (Institute for Translational Oncology Research) and C. Yasenchak (Willamette Valley Cancer Institute)

      References

        • Yakes F.M.
        • Chen J.
        • Tan J.
        • Yamaguchi K.
        • Shi Y.
        • Yu P.
        • et al.
        Cabozantinib (XL184), a novel MET and VEGFR2 inhibitor, simultaneously suppresses metastasis, angiogenesis, and tumor growth.
        Mol Cancer Ther. 2011; 10: 2298-2308
        • Nguyen H.M.
        • Ruppender N.
        • Zhang X.
        • Brown L.G.
        • Gross T.S.
        • Morrissey C.
        • et al.
        Cabozantinib inhibits growth of androgen-sensitive and castration-resistant prostate cancer and affects bone remodeling.
        PLoS One. 2013; 8: e78881
        • Dai J.
        • Zhang H.
        • Karatsinides A.
        • Keller J.M.
        • Kozloff K.M.
        • Aftab D.T.
        • et al.
        Cabozantinib inhibits prostate cancer growth and prevents tumor-induced bone lesions.
        Clin Cancer Res. 2014; 20: 617-630
        • Gray R.
        • Manola J.
        • Saxman S.
        • Wright J.
        • Dutcher J.
        • Atkins M.
        • et al.
        Phase II clinical trial design: methods in translational research from the Genitourinary Committee at the Eastern Cooperative Oncology Group.
        Clin Cancer Res. 2006; 12: 1966-1969
        • Rosner G.L.
        • Stadler W.
        • Ratain M.J.
        Randomized discontinuation design: application to cytostatic antineoplastic agents.
        J Clin Oncol. 2002; 20: 4478-4484
        • Ratain M.J.
        • Eisen T.
        • Stadler W.M.
        • Flaherty K.T.
        • Kaye S.B.
        • Rosner G.L.
        • et al.
        Phase II placebo-controlled randomized discontinuation trial of sorafenib in patients with metastatic renal cell carcinoma.
        J Clin Oncol. 2006; 24: 2505-2512
        • Stadler W.M.
        The randomized discontinuation trial: a phase II design to assess growth-inhibitory agents.
        Mol Cancer Ther. 2007; 6: 1180-1185
        • Stadler W.M.
        • Rosner G.
        • Small E.
        • Hollis D.
        • Rini B.
        • Zaentz S.D.
        • et al.
        Successful implementation of the randomized discontinuation trial design: an application to the study of the putative antiangiogenic agent carboxyaminoimidazole in renal cell carcinoma–CALGB 69901.
        J Clin Oncol. 2005; 23: 3726-3732
        • Therasse P.
        • Arbuck S.G.
        • Eisenhauer E.A.
        • Wanders J.
        • Kaplan R.S.
        • Rubinstein L.
        • et al.
        New guidelines to evaluate the response to treatment in solid tumors. European Organization for Research and Treatment of Cancer, National Cancer Institute of the United States, National Cancer Institute of Canada.
        J Natl Cancer Inst. 2000; 92: 205-216
        • Smith D.C.
        • Smith M.R.
        • Sweeney C.
        • Elfiky A.A.
        • Logothetis C.
        • Corn P.G.
        • et al.
        Cabozantinib in patients with advanced prostate cancer: results of a phase II randomized discontinuation trial.
        J Clin Oncol. 2013; 31: 412-419
        • Smith M.R.
        • Sweeney C.J.
        • Corn P.G.
        • Rathkopf D.E.
        • Smith D.C.
        • Hussain M.
        • et al.
        Cabozantinib in chemotherapy-pretreated metastatic castration-resistant prostate cancer: results of a phase II nonrandomized expansion study.
        J Clin Oncol. 2014; 32: 3391-3399
        • Tolaney S.M.
        • Nechushtan H.
        • Ron I.G.
        • Schoffski P.
        • Awada A.
        • Yasenchak C.A.
        • et al.
        Cabozantinib for metastatic breast carcinoma: results of a phase II placebo-controlled randomized discontinuation study.
        Breast Cancer Res Treat. 2016; 160: 305-312
        • Kelley R.
        • Verslype C.
        • Cohn A.
        • Yang T.
        • Su W.
        • Burris H.
        • et al.
        Cabozantinib in hepatocellular carcinoma: results of a phase 2 placebo-controlled randomized discontinuation study.
        Ann Oncol. 2017; 28: 528-534
        • Daud A.
        • Kluger H.
        • Kurzrock R.
        • Schimmoller F.
        • Weitzman A.
        • Samuel T.
        • et al.
        Phase II randomised discontinuation trial of the MET/VEGF receptor inhibitor cabozantinib in metastatic melanoma.
        Br J Cancer. 2017; 116: 432-440
        • Vergote I.B.
        • Smith D.C.
        • Berger R.
        • Kurzrock R.
        • Vogelzang N.J.
        • Sella A.
        • et al.
        A phase 2 randomised discontinuation trial of cabozantinib in patients with ovarian carcinoma.
        Eur J Cancer. 2017; 83: 229-236
        • Smith M.
        • De Bono J.
        • Sternberg C.
        • Le Moulec S.
        • Oudard S.
        • De Giorgi U.
        • et al.
        Phase III study of cabozantinib in previously treated metastatic castration-resistant prostate cancer: COMET-1.
        J Clin Oncol. 2016; 34: 3005-3013
        • Chen H.X.
        • Cleck J.N.
        Adverse effects of anticancer agents that target the VEGF pathway.
        Nat Rev Clin Oncol. 2009; 6: 465-477
        • Choueiri T.K.
        • Escudier B.
        • Powles T.
        • Mainwaring P.N.
        • Rini B.I.
        • Donskov F.
        • et al.
        Cabozantinib versus everolimus in Advanced Renal-Cell Carcinoma.
        N Engl J Med. 2015; 373: 1814-1823
        • Choueiri T.K.
        • Escudier B.
        • Powles T.
        • Tannir N.M.
        • Mainwaring P.N.
        • Rini B.I.
        • et al.
        Cabozantinib versus everolimus in advanced renal cell carcinoma (METEOR): final results from a randomised, open-label, phase 3 trial.
        Lancet Oncol. 2016; 17: 917-927
        • Elisei R.
        • Schlumberger M.J.
        • Muller S.P.
        • Schoffski P.
        • Brose M.S.
        • Shah M.H.
        • et al.
        Cabozantinib in progressive medullary thyroid cancer.
        J Clin Oncol. 2013; 31: 3639-3646