ESTRO 2025 - Abstract Book

S3838

Physics - Radiomics, functional and biological imaging and outcome prediction

ESTRO 2025

Results:

The radiomic-cricoid and radiomic-larynx models showed moderate performance, with validation AUCs of 0.714 and 0.749, respectively. The radiological model demonstrated superior performance compared to both radiomic models, achieving a validation AUC of 0.875, which was significantly better compared to the radiomic-cricoid model (training: p = 0.004, validation: p = 0.001) and the radiomic-larynx model (training: p = 0.007, validation: p = 0.003). The integrated model demonstrated the best accuracy, with a validation AUC of 0.975, significantly outperforming all individual models (p < 0.001). Conclusion: The integration of radiomic and radiological features significantly enhances the accuracy of preoperative detection of thyroid and cricoid cartilage invasion in LSCC, providing a robust tool for personalized treatment planning in LSCC. Further external validation on larger, multicenter cohorts is necessary to confirm the model's clinical applicability and ensure robustness. References: (1) Forastiere AA, et al. Use of Larynx-Preservation Strategies in the Treatment of Laryngeal Cancer: American Society of Clinical Oncology Clinical Practice Guideline Update. J Clin Oncol. 2018 Apr 10;36(11):1143–69. (2) Patel et al. Comparing surgical and nonsurgical larynx-preserving treatments with total laryngectomy for locally advanced laryngeal cancer. Cancer. 2019 Oct 1;125(19):3367–77 (3) Timme DW, et al. Treatment Selection for T3/T4a Laryngeal Cancer. Annals of Otology, Rhinology & Laryngology. 2015 Nov 19;124(11):845–51 Keywords: Larynx, Radiomic