Longitudinal and Multimodal Radiomics Models for Head and Neck Cancer Outcome Prediction

SIMPLE SUMMARY: Machine learning based radiomics models for prediction of loco-regional recurrence today mostly rely on features extracted from pre-treatment imaging data. In this work, we investigate the predictive ability of such models when imaging data obtained during the course of treatment are used. This is achieved by extracting features from pre-treatment CT and FDG-PET images as well as from images obtained two (only CT) and three weeks after start of radiochemotherapy. Models comprised of combined features from both modalities and multiple timepoints are evaluated. We confirm that predictive model performance is improved when features from in-treatment imaging are used, finding that CT-based features allow for more accurate risk prediction, while FDG-PET based models are able to stratify patients into low- and high-risk groups more reliably. ABSTRACT: Radiomics analysis provides a promising avenue towards the enabling of personalized radiotherapy. Most frequently, prognostic radiomics models are based on features extracted from medical images that are acquired before treatment. Here, we investigate whether combining data from multiple timepoints during treatment and from multiple imaging modalities can improve the predictive ability of radiomics models. We extracted radiomics features from computed tomography (CT) images acquired before treatment as well as two and three weeks after the start of radiochemotherapy for 55 patients with locally advanced head and neck squamous cell carcinoma (HNSCC). Additionally, we obtained features from FDG-PET images taken before treatment and three weeks after the start of therapy. Cox proportional hazards models were then built based on features of the different image modalities, treatment timepoints, and combinations thereof using two different feature selection methods in a five-fold cross-validation approach. Based on the cross-validation results, feature signatures were derived and their performance was independently validated. Discrimination regarding loco-regional control was assessed by the concordance index (C-index) and log-rank tests were performed to assess risk stratification. The best prognostic performance was obtained for timepoints during treatment for all modalities. Overall, CT was the best discriminating modality with an independent validation C-index of 0.78 for week two and weeks two and three combined. However, none of these models achieved statistically significant patient stratification. Models based on FDG-PET features from week three provided both satisfactory discrimination (C-index = 0.61 and 0.64) and statistically significant stratification ([Formula: see text] and [Formula: see text]), but produced highly imbalanced risk groups. After independent validation on larger datasets, the value of (multimodal) radiomics models combining several imaging timepoints should be prospectively assessed for personalized treatment strategies.