Role of Radiomics to Predict Disease Recurrence in Sinonasal Squamous Cell Carcinoma: A Systematic Review and Meta-Analysis.

Abstract

Introduction: Radiomics offers the potential to predict oncological outcomes from pre-operative imaging, aiding in the identification of 'high risk' patients with sinonasal cancer who are at an increased risk of recurrence. This study aims to comprehensively review the current literature on the role of radiomics as a predictor of disease recurrence in sinonasal squamous cell carcinoma.

Methods: A systematic search was conducted in Medline, EMBASE and Web of Science databases. Retrospective and prospective studies examining the use of radiomics to predict post-operative recurrence in sinonasal cancer that met the inclusion criteria were included. Study quality was assessed using the QUADAS-2 and Radiomics Quality Score (RQS) tools.

Results: Five studies met the inclusion criteria, encompassing 638 participants. All studies were single-centre and utilised MRI-based radiomics in the construction of their models. Radiomic models demonstrated excellent predictive performance. The median AUC, sensitivity and specificity were 0.947, 0.86 and 0.923 in the training set, and 0.914, 0.833 and 0.878 in the validation set. A pooled meta-analysis estimated the combined AUC across training sets as 0.931 (95% CI, 0.898-0.963) and 0.922 (95% CI, 0.880-0.964) for validation sets.

Conclusion: Our systematic review provides evidence supporting the role of radiomics in predicting post-operative disease recurrence in sinonasal cancer. Radiomics shows promise in enhancing personalised treatment strategies by improving prognostic accuracy. However, further research is needed to standardise methodologies and validate these findings in larger, multicentre cohorts.