Reporting of endpoints and tracking of failures in randomized trials of radiotherapy or concurrent chemoradiotherapy for locally advanced head and neck squamous cell cancer (LA-HNSCC)

6072 Background: Due to their anatomical complexity, the interactive effects of multiple treatment modalities, the difficulties in differentiating scar tissues versus residual disease and second cancers versus tumor recurrence, LA-HNSCC represent a challenging disease for the reporting of endpoints and the tracking of failures. Methods: We retrieved all randomized trials that began accrual on or after 1978, and enrolled previously untreated nonmetastatic HNSCC patients receiving primary (chemo)radiotherapy. The reporting of endpoints and the tracking of failures in these trials were analyzed. Failures were defined as events meeting a pre-specified endpoint definition. Results: Forty trials involving 13,892 patients fulfilled our inclusion criteria. A total of 125 endpoints were identified: primary versus secondary = 34:91, survival-based (e.g., overall survival [OS]) versus surrogate (e.g. locoregional control [LRC]) = 47:78. In 6 trials, no primary endpoint was identified. LRC and OS accounted for 70% of primary endpoints. All but one trial reported at least one secondary endpoint, with a median of 2 per trial (range: 0–5), and as many as 17 different types of secondary endpoints were reported. Among 72 endpoints tracking locoregional failures, 21/72 (29%) did not define locoregional failure, while 46/72 (64%) specified the absence of complete response as a failure. Whether salvage surgery or elective node dissection was performed or not was reported in less than half of the trials. Furthermore, it was usually not specified if residual disease found during these procedures would account for failure or not. The means (i.e. clinical and/or radiological examinations) to ascertain failures and the protocol-specified timing to track failures were reported in 41% and 67% of surrogate endpoints, respectively. The tracking of other types of failure beyond the first failure is not reported by any of the trials. The reporting of second cancers was found in 15/40 (38%) trials, whereas the duration of follow-up was quantified in 31/40 (78%) trials. Conclusions: These results demonstrate the vast heterogeneity in endpoint reporting and tracking of failures in clinical trials of LA-HNSCC. No significant financial relationships to disclose.