Related Experiment Video
Updated: May 11, 2025
Competing-Risk Nomogram for Predicting Cancer-Specific Survival in Multiple Primary Colorectal Cancer Patients after Surgery
Published on: September 27, 2024
MRI-based radiomics nomogram to predict complete response to definitive chemoradiation in patients with anal cancer
Hugo C Temperley1,2, Fariba Tohidinezhad3, Niall J O'Sullivan1
1Department of Radiology, St James's Hospital, Dublin D08 NHY1, Ireland.
Objectives:
Treatment response to definitive chemoradiation (dCRT) in patients with anal cancer varies significantly, with a subset experiencing persistent or progressive disease despite therapy. Radiomics extracts quantitative features from radiological images, with the potential to develop predictive tools to assess treatment response. We aim to develop and validate an MRI-based radiomics nomogram to predict response to dCRT in patients with anal cancer.
Methods:
A single-institutional retrospective analysis of 45 patients with anal cancer treated with dCRT was performed. Radiomic features were extracted from pre-treatment T2-weighted MRI scans, and predictive models were constructed. Clinical and radiomic features were analysed to develop the nomogram. Internal validation with 1000 bootstrap samples was performed to calculate optimism-corrected performance measures.
Results:
Overall, 30/45(66.7%) achieved a complete treatment response. Male gender was found to be an independent predictor of incomplete response to dCRT (OR 4.763,95% CI: 1.170-19.384,*P = .029). Two radiomic signatures emerged as strong predictors of treatment response to dCRT. The combined model outperformed the clinical and radiomic models. The combined model showed the highest predictive accuracy, achieving an apparent area under the receiver operating characteristic curve (AUC): 0.87 (0.75-0.99) and an optimism-corrected AUC: 0.85, mean absolute error: 0.029, positive predictive value (0.68) and negative predictive value (0.92), indicating excellent discriminative performance. It demonstrated a positive net benefit in decision analysis. The optimism-corrected calibration curves demonstrate that the radiomic and combined model provide well-calibrated predictions.
Conclusion:
This MRI-based radiomics nomogram offers a promising approach to predict response to dCRT in patients with anal cancer.
Advances In Knowledge:
This study is the first to integrate radiomics and clinical features into a validated predictive model for anal cancer.
