Study limitations and strengths
Selection bias inherent to a retrospective design notwithstanding, we recognise that the principal limitation of our study relates to a relatively small sample size and, by extension, limited number of NFL events to analyse and to analyse subgroups (e.g. T stage). Not only does this carry a propensity for type I or II statistical error, but also restricted the number of predictive factors that could be examined. This was compounded by use of relatively crude outcome measures, with no availability of patient-reported functional outcome measures, leading potentially to a lack of data granularity. Nevertheless, our study included a homogenous group of patients with locally advanced LSCC for whom robust follow-up data was collated. Furthermore, as outlined previously we believe this study attempts to enrich the evidence-base around an important clinical question, currently informed by a paucity of data.