European Journal of Cancer (2014) 50, 902– 911

Available at www.sciencedirect.com

ScienceDirect journal homepage: www.ejcancer.com

Capecitabine and streptozocin ± cisplatin in advanced gastroenteropancreatic neuroendocrine tumours Tim Meyer a,b,1, Wendi Qian c,d, Martyn E. Caplin a, Graham Armstrong c, Si-Houy Lao-Sirieix c, Richard Hardy c, Juan W. Valle e, Denis C. Talbot f, David Cunningham g, Nick Reed h, Ashley Shaw i, Shaunak Navalkissoor a, Tu-Vinh Luong a, Pippa G. Corrie c,i,⇑,1 a

Neuroendocrine Tumour Unit, The Royal Free Hospital, Pond Street, London, UK UCL Cancer Institute, London, UK c Cambridge Cancer Trials Centre, Cambridge Clinical Trials Unit – Cancer Theme, Addenbrooke’s Hospital, Cambridge, UK d Medical Research Council Biostatistics Unit Hub for Trials Methodology, Cambridge, UK e Department of Medical Oncology, The Christie, Manchester, UK f Oxford Neuroendocrine Tumour Centre, Churchill Hospital, Oxford, UK g Gastrointestinal Unit, The Royal Marsden, London, UK h Beatson Oncology Centre, Glasgow, UK i Oncology Centre, Addenbrooke’s Hospital, Cambridge, UK b

Available online 17 January 2014

KEYWORDS Neuroendocrine tumour Randomised clinical trial Chemotherapy Capecitabine Streptozocin Cisplatin Phase II

Abstract Background: Cytotoxic chemotherapy is widely used for advanced, unresectable pancreatic and other gastrointestinal foregut neuroendocrine tumours (NETs) and the most commonly used regimen combines 5-fluorouracil with streptozocin. The NET01 trial was designed to investigate whether capecitabine combined with streptozocin was an acceptable regimen with or without adding cisplatin. Methods: Patients with advanced, unresectable NETs of pancreatic, gastrointestinal foregut or unknown primary site were randomised to receive three-weekly capecitabine (Cap) 625 mg/m2 twice daily orally, streptozocin (Strep) 1.0 g/m2 intravenously on day 1, with or without cisplatin (Cis) 70 mg/m2 intravenously on day 1. The primary outcome measure was objective response. Secondary outcome measures included progression-free and overall survival, quality of life, toxicity and biochemical response. Results: 86 (44 CapStrep, 42 CapStrepCis) patients were randomised. Best objective response rate was 12% (95% confidence interval (CI) = 2–22%) with CapStrep and 16%

⇑ Corresponding author at: Oncology Centre, Addenbrooke’s Hospital, Hills Road, Cambridge CB2 0QQ, UK. Tel.: +44 1223 274401; fax: +44 1223 217094. E-mail address: [email protected] (P.G. Corrie). 1 Both authors contributed equally.

0959-8049/$ - see front matter Ó 2014 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ejca.2013.12.011

T. Meyer et al. / European Journal of Cancer 50 (2014) 902–911

903

(95% CI = 4–27%) with CapStrepCis. Disease-control rate was 80% with CapStrep and 74% with CapStrepCis. The estimated median progression-free and overall survival were 10.2 and 26.7 months for CapStrep and 9.7 and 27.5 months for CapStrepCis. 44% of CapStrep and 68% of CapStrepCis patients experienced grade P3 adverse events. Interpretation: The efficacies of the novel CapStrep ± Cis regimens were very similar. CapStrep was better tolerated than CapStrepCis. The trial was registered as EudraCT: 2004-005202-71 and ISRCTN: 35124268. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction Cytotoxic chemotherapy is widely used for advanced, unresectable pancreatic and other gastrointestinal foregut neuroendocine tumours (NETs), although few randomised trials have been undertaken in this rare cancer [1–4]. The first of these trials [1] established streptozocin (Strep) and 5-fluorouracil (5-FU) as the standard of care for advanced pancreatic NETs. The remarkable 63% response rate [1] was based on combined clinical and biochemical response criteria in addition to radiological response. Subsequent studies using conventional WHO or RECIST response criteria reported lower response rates for Strep-based regimens ranging between 6 and 55% [5–12]. Cisplatin (Cis) was combined with Strep and 5-FU in a retrospective study of 49 patients with advanced pancreatic NETs and an objective partial response rate of 38% was reported [11]. However, no prospective trials of cisplatin-containing chemotherapy for NETs have previously been undertaken. Over the last 10 years, capecitabine (Cap) has been effectively substituted for 5-FU in many regimens used to treat gastrointestinal tract cancers [13–15]. Cap has reported activity as a single agent or in combination regimens for non-pancreatic NETs [16,17] as well as in combination with temozolomide for pancreatic NETs [18]. The NET01 trial was designed to investigate whether Cap combined with Strep was an acceptable regimen with or without adding cisplatin to treat advanced gastroenteropancreatic (foregut) NETs. 2. Patients and methods 2.1. Main eligibility Patients with chemonaive, histologically confirmed, unresectable, advanced and/or metastatic NETs of the pancreas, other gastrointestinal foregut, or unknown primary site suggestive of abdominal foregut origin were eligible. Patients were required to have measurable disease by RECIST (version 1.0), Eastern Cooperative Oncology Group performance status (ECOG PS) 62, adequate bone marrow, hepatic and renal function with creatinine clearance >60 ml/min. Radiological confirmation of progressing disease between consecutive imaging was not mandated. Ethics committee approval of the

protocol and written informed consent from all patients were obtained. Paraffin-embedded (blocks/slides) or fresh-frozen biopsy material taken from the patient’s tumour (primary/secondary) prior to treatment was collected for central pathology review and immunohistochemical assessment of Ki67 expression and mitotic index as previously described [19]. 2.2. Treatment Patients were randomised (1:1), using the stratified random block method, to receive six cycles of threeweekly CapStrep or CapStrepCis regimens. The stratification factors were functional tumour (yes versus no), previous treatment received (somatostatin analogues/ interferon versus none) and primary tumour site (known versus unknown). The CapStrep regimen comprised capecitabine 625 mg/m2 administered orally, twice daily on days 1–21, and streptozocin 1.0 g/m2 (2-h infusion intravenously in normal saline) on day 1. The CapStrepCis regimen comprised CapStrep plus cisplatin 70 mg/m2 (2-h infusion intravenously in normal saline with hydration) on day 1, directly after the Strep infusion. Patients could continue to receive the same treatment beyond six cycles if there was evidence of benefit. 2.3. Assessments Patients were assessed for adverse events (AEs) every cycle, using the NCI CTC (version 3), and followed for disease progression and survival every 12 weeks. Tumour assessments with CT scans were performed at the baseline, every three cycles while on treatment and every 12 weeks until progression. Retrospective central radiology review was undertaken for objective tumour response assessments in 10% of randomly selected patients who completed at least three treatment cycles. Biochemical measurements of 24 h urinary 5-hyroxyindoleacetic acid (5HIAA) and serum chromogranin A (CgA) were undertaken by radioimmunoassay prior to treatment and if above the normal range were repeated every three cycles while on treatment and at 12 weeks from the end of treatment. Patient quality of life (QoL) using the EORTC QLQ-C30 questionnaires was assessed before randomisation, after three

904

T. Meyer et al. / European Journal of Cancer 50 (2014) 902–911

and six cycles of treatment or at the time of stopping treatment and at 12 weeks post-treatment. 2.4. Statistical considerations The primary outcome measure was objective response rate by RECIST; secondary outcome measures were biochemical response, safety, progression-free survival (PFS), overall survival (OS) and QoL. The trial was powered to test whether either CapStrep or CapStrepCis (or both regimens) would be sufficiently active (response rate >60%, equivalent to the overall response rate reported by Moertel et al. [1]) to take forward for further evaluation; a response rate below 40% would be considered less interesting in clinical terms, based on published data at the time of initiating this trial. Forty-two patients were required to test the response rate 60% with a significance level of 5% and 80% power in each arm, a total of 84 patients. Although the trial was not powered for comparisons between two arms, the differences could be estimated in balanced arms achieved by randomisation. Best objective response during treatment was the greatest reduction in measurable disease achieved from the start of the treatment until disease progression (if treatment stopped early) or the end of treatment. In addition to objective response measurement, the biochemical markers, 5HIAA and CgA, were included for overall response evaluation. Patients were eligible for biochemical response assessment if the baseline level of either 5HIAA or CgA was at least twice the upper limit of the normal range (2ULN). Biochemical response was defined as >50% reduction compared with the baseline level in either biochemical marker during treatment. 86 patients randomised

CapStrep (N=44)

CapStrepCis (N=42)

Chemotherapy received 26 (59%) 6 cycles 1 received no chemotherapy due to worsening of symptoms 17 stopped early 7 AE/intercurrent illness 1 death 8 progression 1 holiday

Chemotherapy received 14 (33%) 6 cycles 2 received no chemotherapy due to worsening of symptoms 26 stopped early 12 AE/intercurrent illness 11 progression 3 refused treatment

Follow up Primary outcome measure 4 not included

Follow up Primary outcome measure 3 not included 1 no treatment 1 stopped (1 cycle) not due to disease 1 no adequate baseline assessment

2 no treatment 2 stopped (1 cycle) not due to disease

5 partial response 28 stable disease 8 progressive disease

6 partial response 22 stable disease 10 progressive disease

Progression-free survival 34 progressed/died 10 alive without progression

Progression-free survival 35 progressed/died 7 alive without progression

Overall survival 25 died 19 alive

Overall survival 22 died 20 alive

Fig. 1. Trial profile.

Overall response was defined as evidence of objective CR or PR, or >50% reduction in 5HIAA or CgA in patients with baseline levels P2ULN. Objective and overall response were determined for all patients who received at least one cycle of chemotherapy. PFS was calculated from the date of randomisation to the date of the first progression or date of death from any cause, whichever occurred first; survivors without disease progression were censored at the date of last known alive. OS was calculated from the date of randomisation to the date of death from any cause; survivors were censored at the date of last known alive. Kaplan–Meier estimates were applied to both PFS and OS analyses together with the estimated hazard ratio (HR) and 95% confidence intervals (CI) for all patients randomised. The changes from the baseline to the middle cycle and at 6 months on the global QoL score from the QLQ-C30 (range 0–100, with 100 being best health-related QoL) were summarised in each arm.

3. Results 3.1. Patients Between November 2006 and October 2010, 86 (44 CapStrep, 42 CapStrepCis) patients were randomised from 13 United Kingdom (UK) centres (Fig. 1). Tissue for central pathology review was available for 69 (80%) patients. The two groups were well matched at the baseline (Table 1). Primary disease site was pancreas (48%), foregut (20%) or unknown (33%). Overall, 31 (36%) patients had functional tumours, of which 25 had received prior somatostatin analogues for symptom control. Two patients in the CapStrep arm had received prior systemic therapy with interferon. For the subgroup of patients in whom central pathology review was performed, 48% had Ki67 6 9%, 71% had