The majority of patients with squamous cell carcinoma of the head and neck present with locally advanced disease. Despite treatment with curative intent, approximately 40% of patients eventually relapse. It is currently difficult to predict those who will relapse at treatment outset. Nomograms can improve prognostic discussions for individual patients and potentially guide future research to individualise treatment. We aim to develop OS and PFS nomograms using retrospective data from patients with LA OPSCC treated in our institution.
We performed a retrospective analysis of baseline characteristics and outcomes of LA OPSCC patients who underwent curative-intent treatment in our institution from January 2008 to December 2017. Nomograms were constructed to estimate OS and PFS incorporating age, gender, performance status, smoking history, stage, grade, and p16 status. Multivariable Cox models were selected by backward elimination using the Akaike information criterion, and validated internally using bootstrap with 1000 resamples. Nomogram points were assigned proportional to variable effect size.
A total of 417 patients have been identified so far. An initial cohort of 74 patients with complete clinical annotations were analysed. Median follow-up was 24 months, with 9 deaths and 25 events during follow-up. ECOG 0, N stage 0-1, histological grade 1-2 and p16 positive status were favourable predictors in the OS nomogram. Age ≤ 50, male gender, smoking history ≤ 10 pack-years and p16 positive status were favourable predictors in the PFS nomogram. The concordance index for OS was 0.86 (95% CI 0.76 to 0.99) and for PFS was 0.72 (95% CI 0.65 to 0.83). Bias-corrected indices were 0.82 and 0.69, respectively.
We present the first OS and PFS nomograms for Australian patients with LA OPSCC. The nomograms demonstrated clinically useful prediction of OS and PFS in patients with LA OPSCC, relying on four routinely collected variables, with superior concordance to previously reported models. Expansion of the retrospective cohort is ongoing and validation in an external patient cohort is planned.
Clinical trial identification
Legal entity responsible for the study
St Vincent's Hospital, Sydney.
Has not received any funding.
All authors have declared no conflicts of interest.