ESTRO 2024 - Abstract Book

S1438

Clinical - Head & neck

ESTRO 2024

Pre-treatment TTMV-DNA fragment scores were significantly associated with tumor volume in this diverse patient population. However, TTMV-DNA was undetectable in 10.3% of patients with HPV-mediated oropharynx cancer. Surveillance NavDx testing had a high NPV and was able to detect recurrence prior to detection by imaging or biopsy, with only one false negative result in our cohort.

Keywords: HPV ctDNA, oropharynx

2974

Proffered Paper

Generalised Competing Risk models predict relative risk of cancer events in p16+ oropharynx cancer

James M Price 1,2 , David J Thomson 1,2 , Loren K Mell 3

1 The Christie NHS Foundation Trust, Clinical Oncology, Manchester, United Kingdom. 2 The University of Manchester, Division of Cancer Sciences, Manchester, United Kingdom. 3 University of California San Diego, Department of Radiation Medicine and Applied Sciences, San Diego, USA

Purpose/Objective:

Approaches to define groups of patients with good-prognosis p16-positive oropharynx squamous cell carcinoma (OPSCC) for treatment de-escalation have been unsuccessful to date 1 . Existing prognostic and predictive models do not differentiate between cancer events (i.e., residual, or recurrent OPSCC) versus competing events (i.e., death from other causes) 2 . In this work, we assessed the performance of the Head and Neck Cancer International Group (HNCIG) Omega Score Classifier, a novel model that uses Generalised Competing Events (GCE) regression to stratify patients according to their risk of cancer- or competing events 3 .

Material/Methods:

A retrospective review of prospectively-collected data. Inclusion criteria: patients with locally-advanced (T3-4 or N1+), non-metastatic, p16-positive OPSCC treated with curative intent radiotherapy +/- concurrent cisplatin chemotherapy at The Christie NHS Foundation Trust between 2012 and 2019. Patients who underwent primary surgery, or who had opted-out of their data being analysed for research purposes, were excluded. Intensified treatment was defined as the use of concurrent cisplatin chemotherapy 100mg/m2 (induction chemotherapy and concurrent cetuximab were considered non-intensified treatment). Two patient stratification methods were assessed, (i) the HNCIG classifier, producing Omega (Ω) scores (higher Ω scores: higher risk of cancer vs competing events), and (ii) eligibility criteria for the NRG HN005 Phase III RCT of treatment de- escalation (‘low - risk’ - eligible for trial). To assess the performance of the two methods in prognostication, multivariable Cox and GCE models were fitted; to assess the models’ ability to predict whether patients received intensified vs non-intensified treatment, logistic regression models were fitted.