Giovanni M. Bariani

Non-inferiority cancer clinical trials: scope and purposes underlying their design

BACKGROUND Non-inferiority clinical trials (NIFCTs) aim to demonstrate that the experimental therapy has advantages over the standard of care, with acceptable loss of efficacy. We evaluated the purposes underlying the selection of a non-inferiority design in oncology and the size of their non-inferiority margins (NIFms). PATIENTS AND METHODS All NIFCTs of cancer-directed therapies and supportive care agents published in a 10-year period were eligible. Two investigators extracted the data and independently classified the trials by their purpose to choose a non-inferiority design. RESULTS Seventy-five were included: 43% received funds from industry, overall survival was the most common primary end point and 73% reported positive results. The most frequent purposes underlying the selection of a non-inferiority design were to test more conveniently administered schedules and/or less toxic treatments. In 13 (17%) trials, a clear purpose was not identified. Among the trials that reported a pre-specified NIFm, the median value was 12.5% (range 4%-25%) for trials with binary primary end points and Hazard Ratio of 1.25 (range 1.10-1.50) for trials that used time-to-event primary outcomes. CONCLUSION Cancer NIFCT harbor serious methodological and ethical issues. Many use large NIFm and nearly one-fifth did not state a clear purpose for selecting a non-inferiority design.BACKGROUND Non-inferiority clinical trials (NIFCTs) aim to demonstrate that the experimental therapy has advantages over the standard of care, with acceptable loss of efficacy. We evaluated the purposes underlying the selection of a non-inferiority design in oncology and the size of their non-inferiority margins (NIFms). PATIENTS AND METHODS All NIFCTs of cancer-directed therapies and supportive care agents published in a 10-year period were eligible. Two investigators extracted the data and independently classified the trials by their purpose to choose a non-inferiority design. RESULTS Seventy-five were included: 43% received funds from industry, overall survival was the most common primary end point and 73% reported positive results. The most frequent purposes underlying the selection of a non-inferiority design were to test more conveniently administered schedules and/or less toxic treatments. In 13 (17%) trials, a clear purpose was not identified. Among the trials that reported a pre-specified NIFm, the median value was 12.5% (range 4%-25%) for trials with binary primary end points and Hazard Ratio of 1.25 (range 1.10-1.50) for trials that used time-to-event primary outcomes. CONCLUSION Cancer NIFCT harbor serious methodological and ethical issues. Many use large NIFm and nearly one-fifth did not state a clear purpose for selecting a non-inferiority design.

Phase 2 Trial of Metformin Combined With 5-Fluorouracil in Patients With Refractory Metastatic Colorectal Cancer.

BACKGROUND Observational and preclinical studies have suggested that metformin has antitumor effects in solid tumors, including colorectal cancer (CRC). However, the effects of metformin in CRC have not been tested in clinical trials. PATIENTS AND METHODS This was a single-center, single-arm phase 2 clinical trial where histologically confirmed CRC patients with measurable and progressing metastatic disease previously treated with 5-fluorouracil (5-FU), irinotecan, oxaliplatin, and an anti-epidermal growth factor receptor (if the tumor was RAS wild type) were enrolled to receive metformin 850 mg orally continuously 2 times a day plus 5-FU 425 mg/m2 and leucovorin 50 mg intravenously weekly until disease progression, unacceptable toxicity, or withdrawal of consent. The primary end point was disease control rate at 8 weeks. RESULTS Among 50 patients included, 11 (22%) met the primary end point. The median progression-free survival was 1.8 months and the median overall survival 7.9 months. Analyzing only the 11 patients who experienced disease control at 8 weeks, their median progression-free survival was 5.6 months and their median overall survival was 16.2 months. There was a trend for prolonged median survival for obese patients (12.4 vs. 5.8 months) and those longer off 5-FU. The treatment was well tolerated; the main adverse effects were diarrhea, nausea, vomiting, and myelotoxicity. CONCLUSION Metformin and 5-FU showed an overall modest but intriguing activity in patients with refractory CRC in this phase 2 study. Some patients experienced long-term disease control. Further trials are needed to confirm these results, particularly in obese patients with CRC.

Poor Evidence to Standardize Adjuvant Treatment for Patients With Biliary Tract Cancer

or both) compared with surgery alone for the entire group of biliary tract cancers (odds ratio [OR], 0.74; P .06). A sensitivity analysis demonstrated significant benefit in survival in patients with nodepositive and R1 disease (OR, 0.49; P .004; and OR, 0.36; P .002, respectively). With these results, the authors concluded that adjuvant therapy is the standard of care for high-risk patients and recommend againstano-treatmentarminfuturerandomizedtrialsinthissetting. We respectfully disagree with their conclusions. Selection bias, stagemigration,andheterogeneoustreatmentsarelikelytohaveconfoundedtheauthors’findings.Westronglybelievethatadjuvanttherapy for biliary tract cancer is not yet a standard and that the authors were hasty when they condemned a no-treatment arm as the control group in forthcoming trials. Basically, the quality of studies included in their meta-analysis was poor. Of the 20 studies included, there was only one randomized trialthatfoundanimproved5-yearsurvivalratewithadjuvantchemotherapyforpatientswithgallbladdercancer,butnotinthesubgroup of patients with cholangiocarcinomas. 2 The great majority of data were pooled from uncontrolled retrospective studies, in which mixed groups of patients with various tumor types (some studies included pancreatic and ampullary cancers) were treated with different adjuvant therapies. These factors combine to result in several intrinsic biases.First,becauseoftheretrospectivenatureofthestudiesincluded in this meta-analysis, selection bias could explain the better outcomes reported for adjuvant therapy. The authors discuss that patients with high-risk features could have been more likely to receive adjuvant therapy, which makes their results even more striking. We argue the opposite: similar to the case with other cancer types, patients with better performance status and/or younger age could have been more likely to receive adjuvant treatment in comparison with elderly patients, those with many comorbid illnesses, those who experienced serious postoperative complications, and/or patients who had poor clinical conditions. 3,4 Such confounders may have led to an overestimation of the adjuvant treatment and misinterpretation of the data. Second, the variability of diagnostic/staging tests used across the time spancoveredbythemeta-analysismayhaveledtostagemigrationasa result of instrumental bias. It is quite possible that patients who were stagedashavinglocalizeddiseaseinoldstudiesactuallyhadmetastatic disease; imbalances in the number of patients with metastatic disease between the adjuvant and control groups could have compromised the results. Third, the meta-analysis is based on literature-extracted data instead of individual patient information, which may produce different efficacy estimations and levels of evidence. 5 Fourth, many studies included do not report on resection margins. When margins are compromised, postoperative treatment is acceptable with the aim ofimprovinglocalcontrol,butthiscannotbeconsideredanadjuvant intervention per the definition of adjuvant. It simply involves early treatment of advanced disease. Other drawbacks such as unpublished negativetrials,shortfollow-up,andvariable,unscheduledtimepoints for recurrence evaluation across studies may have also generated biased interpretation with respect to the adjuvant treatment. The authors highlight some of these limitations in their discussion session. Nevertheless, they still endorse the use of adjuvant therapy for patients with node-positive disease and R1 resection. Although we agree that this could be an acceptable strategy, we advise against a definitive recommendation—particularly in the subset of node-positive tumors. High-risk status does not mean that a patient will benefit from adjuvant treatment. In fact, adjuvant treatments have been reportedly ineffective in certain tumor types that harbor high risk for recurrence such as rectal cancer and ampullary adenocarcinoma. 6,7

Combination of irinotecan and a platinum agent for poorly differentiated neuroendocrine carcinomas.

Extrapulmonary poorly differentiated neuroendocrine carcinoma (PDNEC) is a rare and highly aggressive neoplasm for which the optimal chemotherapy remains unclear. The objective of this study was to evaluate the outcomes of patients with PDNEC treated with cisplatin and irinotecan (IP) and perform a review of the literature. From 2008 to 2012, patients with advanced PDNEC (Ki67≥20%) who received the IP combination were selected for analysis. Radiologic responses were determined through Response Evaluation Criteria In Solid Tumors criteria. Twenty-eight patients were included. The median age at diagnosis was 57 years and the most common presentation was pancreatic PDNEC. Twenty-five patients (89%) received chemotherapy with cisplatin and irinotecan and three received carboplatin and irinotecan. Forty-six percent of the patients achieved objective response and the median time to tumor progression was 3.7 months. The median overall survival was 11.7 months. Thirteen patients (46%) had treatment interruptions or dose reductions due to grade 3/4 toxicity. This retrospective cohort of advanced extrapulmonary PDNEC patients suggests that the IP combination is feasible and resulted in similar response rate and median survival to other treatments previously reported.

Sample Size Calculation in Oncology Trials: Quality of Reporting and Implications for Clinical Cancer Research.

Objectives:Sample size calculation (SSC) is a pivotal step in clinical trial conception and design. Herein, we describe the frequency with which oncology phase III trials report the parameters required for SSC. Materials and Methods:We systematically searched for phase III trials published in 6 leading journals, which were accompanied by editorials from January 2008 to October 2011. Two blinded investigators extracted required and optional parameters for SSC according to the primary endpoint. Results:We retrieved 140 eligible phase III trials. The median target sample size was 596 subjects (50 to 40,000); in 66.4% of cases, the number of enrolled subjects was at least 90% of the target. The primary endpoint was a continuous variable in 5.7%, categorical in 30.0%, and a time-to-event variable in 64.3% of phase III trials. Although nearly 80% reported a target sample size, only 27.9% of the trials provided all the required parameters for proper SSC. The most commonly reported parameters for sample size computation were &agr; (93.6%) and &bgr; (90.7%) errors. The parameters least reported were the expected outcomes in the control or experimental groups, each provided in only 57.9% of trials. Conclusions:The quality of SSC reporting in phase III cancer trials is poor. Such incomplete reporting may compromise future study designs, pooling of data, and interpretation of results. Lack of transparency in SSC reporting may also have ethical implications.

Antitumor Effect of Everolimus in a Patient with Type 3 Gastric Neuroendocrine Tumor

Background: Gastric neuroendocrine tumors (NET) are rare and are classified into 3 types: type 1 and 2 (characterized by hypergastrinemia), and type 3 (characterized by normal gastrin). Surgery is the standard procedure, and systemic treatment is reserved for unresectable disease. Currently, targeted therapies are being evaluated in NET. The activity of everolimus, an mTOR inhibitor, has been shown in pancreatic NET but not reported in type 3 gastric carcinoid tumors. Case Report: Here we report a case of a patient who, after multiple lines of systemic therapy, had a prolonged disease control of nearly 1 year, significant clinical benefit, and minor tumor shrinkage with oral everolimus 10 mg continuously. Conclusion: There is no effective treatment for type 3 gastric carcinoid tumors. The frequency of mTOR expression in these tumors is not known, but the case reported here suggests that inhibition of this pathway may play an important role.

Reply to R. Brierley et al

ACKNOWLEDGMENT R.B. and D.C. are Deputy Editor and Editor-in-Chief, respectively, of The Lancet Oncology. AUTHORS’ DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST Employment or Leadership Position: Rob Brierley, Elsevier (C); David Collingridge, Elsevier (C) Consultant or Advisory Role: None Stock Ownership: Rob Brierley, Reed Elsevier; David Collingridge, Reed Elsevier Honoraria: None Research Funding: None Expert Testimony: None Patents: None Other Remuneration: None