Preoperative Prediction of Tumor Deposits in Rectal Cancer Using Cellular Microstructural Parameters Derived from Time-Dependent Diffusion MRI
Purpose To evaluate the efficacy of time-dependent diffusion MRI (td-dMRI)-derived quantitative parameters in preoperatively predicting tumor deposits (TD) in rectal cancer. Methods A total of 108 patients with rectal cancer who underwent preoperative td-dMRI, classified by pathology as TD-negative (n = 86) or TD-positive (n = 22). td-dMRI-derived parameters, including cell diameter (d), intracellular volume fraction (Vin), extracellular diffusion coefficient (Dex), cellularity, and ADC-related metrics, were measured. Firth-penalized logistic regression was performed to identify predictors and construct a predictive model. Diagnostic performance was evaluated using receiver operating characteristic (ROC) curves with bootstrap internal validation. Correlations between td-dMRI-derived parameters and pathological measurements were assessed by pearson correlation analysis in a subset of 19 patients. Results The Pos-TD group showed a significantly higher d value than the Neg-TD group (14.35 (IQR, 13.40-15.44) vs. 13.31 (IQR, 12.63–14.21), P = 0.041). Logistic regression analyses showed that d and ADC OGSE25Hz were associated with TD positivity (all P < 0.05). The combined model achieved an area under the curve (AUC) of 0.791 (0.693, 0.890) and a bootstrap-corrected AUC of 0.785 (0.675, 0.886), outperforming individual parameters and mrTD ( all P < 0.05). td-dMRI-derived d, Vin, and cellularity showed moderate-to-strong correlations with pathological findings ( r = 0.702, 0.646, and 0.694, respectively). Conclusion The combined model incorporating d and ADC OGSE25Hz demonstrated preliminary potential for preoperative prediction of TD in rectal cancer.
Authors
- Meining Chen
- Siqi Yi (ORCID: https://orcid.org/0000-0002-8706-4319)
- Thorsten Feiweier (ORCID: https://orcid.org/0009-0005-3537-9844)
- Peng Zhou (ORCID: https://orcid.org/0000-0002-1281-6088)
- Yanwan Li
- Ping Zheng (ORCID: https://orcid.org/0000-0001-9753-3062)
- Fan Yang (ORCID: https://orcid.org/0000-0002-8111-4880)
- Xiaolei Dong (ORCID: https://orcid.org/0009-0000-3959-6014)
- Yehan Zhou
- Xiaoli Chen
Institutions
- Sichuan Cancer Hospital (CN)
- Siemens Healthineers (Germany) (DE)
Publication Details
- Journal
- Research Square
- Published
- 2026-11-18
- DOI
- https://doi.org/10.21203/rs.3.rs-11593487/v1
- Primary Topic
- MRI in cancer diagnosis
- Type
- preprint