Click here to

Session: Imaging General ePoster Viewing [Return to Session]

MRI-Based Radiomics Model Improves Prognosis Prediction in Advanced Rectal Cancer Undergoing Neoadjuvant Chemoradiotherapy and Surgical Resection

Q Zhao1*, Radiation Therapy center, Xiangyang Central Hospital, Affiliated Hospital of Hubei University of Arts and Science, Xiangyang, Hubei, China


PO-GePV-I-34 (Sunday, 7/10/2022)   [Eastern Time (GMT-4)]

ePoster Forums

Purpose: The aim was to develop an MRI-based radiomic model to predict the prognostic value of imaging features in patients with locally advanced rectal cancer (LARC) patients receiving neoadjuvant chemoradiotherapy and surgical resection.

Methods: This retrospective observational cohort study included 220 patients (102 men and 118 women, age 63.2±8.4 years) with LARC undergoing neoadjuvant chemoradiotherapy and surgical resection. Patients were randomized to the training cohort (n=180) and the validation cohort (n=40). Radiomic features were extracted from prior treatment ADC-MR maps and T1w-MR images, wavelet filtering and Gaussian Laplacian (Log) filtered images by Pyramidics. Predictive models were constructed based on radiomic features, histological and clinical nomograms. The accuracy of the model was assessed by calibration curves and consistency index (C-index). In addition, we explored the correlation between radiomic expression patterns, quantitative histological features and clinical data.

Results: The model combining radiomic features with all clinical data outperformed the model based on clinical data only (C-index 0.732 vs. C-index 0.825, respectively), and the calibration curves showed good agreement. Heat maps showed that radiomic expression patterns and selected histological features correlated with clinical stage, T-stage and N-stage (p<0.05).

Conclusion: MRI-based radiomics can significantly improve the efficacy of traditional TNM staging and clinical data in predicting progression-free survival (PFS) in LARC patients, which may provide an opportunity for precision medicine.


Not Applicable / None Entered.


Not Applicable / None Entered.

Contact Email