NET-02 trial protocol: a multicentre, randomised, parallel group, open-label, phase II, single-stage selection trial of liposomal irinotecan (nal-IRI) and 5-fluorouracil (5-FU)/folinic acid or docetaxel as second-line therapy in patients with progressive poorly differentiated extrapulmonary neuroendocrine carcinoma (NEC)

Introduction Poorly differentiated (PD), extrapulmonary (EP), neuroendocrine carcinomas (NECs) are rare but aggressive neuroendocrine neoplasms. First-line treatment for advanced disease is an etoposide and platinum-based chemotherapy combination. There is no established second-line treatment for patients with PD-EP-NEC, and this is an area of unmet need. Methods and analysis NET-02 is a UK, multicentre, randomised (1:1), parallel group, open-label, phase II, single-stage selection trial of liposomal irinotecan (nal-IRI)/5-fluorouracil (5-FU)/folinic acid or docetaxel as second-line therapy in patients with progressive PD-EP-NEC. One hundred and two eligible participants will be randomised to receive either nal-IRI/5-FU/folinic acid or docetaxel. The primary objective is to determine the 6-month progression-free survival (PFS) rate. The secondary objectives of this study are to determine PFS, overall survival, objective response rate, toxicity, quality of life and whether neuron-specific enolase is predictive of treatment response. If either treatment is found to have a 6-month PFS rate of at least 25%, that treatment will be considered for a phase III trial. If both treatments meet this target, prespecified selection criteria will be applied to establish which treatment to take forward. Ethics and dissemination This study has ethical approval from the Greater Manchester Central Research Ethics Committee (reference no. 18/NW/0031) and clinical trial authorisation from the Medicine and Healthcare Products Regulatory Agency. Results will be published in peer-reviewed journals and uploaded to the European Union Clinical Trials Register. Trial registration numbers ISRCTN10996604, NCT03837977, EudraCT Number: 2017-002453-11


GENERAL COMMENTS
The manuscript is an important randomized phase 2 selection design trial. Although appropriately described about study planning, treatment rationale and to take forward to future phase 3 trial, there are some concerns remain.
-There is concerned about how to calculate the primary endpoint. Calculating the proportion, if censor patients are excluded from the denominator, the estimated PFS may be overestimated. The authors should describe the handling of censor patients when calculating 6M PFS. Because it is also related to the interpretation of Primary endpoint.
-Why does the primary endpoint define as a binary variable (progression-free or not at 6month)? It is natural to define the 6M PFS rate by estimating the rate using the Kaplan-Meier method instead of the binary proportion. I wonder why the authors defined the PFS rate as a binary endpoint and thought it was appropriate. Please let me know if you have a reason.

GENERAL COMMENTS
The study proposal is well written and addresses an important topic.

NET-02 protocol paper response
We would like to thank the reviewers for their positive response and constructive comments. Please find responses to specific comments itemised below.
There is concerned about how to calculate the primary endpoint. Calculating the proportion, if censor patients are excluded from the denominator, the estimated PFS may be overestimated. The authors should describe the handling of censor patients when calculating 6M PFS. Because it is also related to the interpretation of Primary endpoint.
Details regarding censoring of patients for the primary endpoint have now been added to the manuscript in the statistical analysis section as follows: 'An individual is defined to have achieved the primary endpoint if they do not progress within the timeframe of treatment start date until 6 months post-randomisation. If an individual dies or is lost to follow-up, without confirmation of disease progression, within 6-months post-randomisation, they will be considered to have not achieved this endpoint and will be censored at the date of death or date last known to be alive and progression-free.' Why does the primary endpoint define as a binary variable (progression-free or not at 6month)? It is natural to define the 6M PFS rate by estimating the rate using the Kaplan-Meier method instead of the binary proportion. I wonder why the authors defined the PFS rate as a binary endpoint and thought it was appropriate. Please let me know if you have a reason.
The primary endpoint is defined as a binary variable as this is related to the statistical design using an A'Hern design. This means that the sample size was calculated using this binary endpoint approach. Therefore, the analysis will use the binary variable as the primary endpoint to reflect the initial design of the study. Additionally, this binary variable is being used for the primary endpoint because it is simple to understand and monitor. The data monitoring and ethics committee (DMEC), will review the primary endpoint in each DMEC report, therefore using a binary endpoint counting the number of patients achieving and failing to achieve progression-free or not at 6m will allow ease of monitoring because the DMEC will be able to clearly see if we have reached our 'target' PFS rate of 25% with a lower 95% confidence interval >15%. Finally, the progression-free rate at a specified time-point, defined as a binary variable, is a commonly used endpoint in phase II trials. To support this assertion we provide two examples where this endpoint has been utilised in other phase II cancer trials below: