Purpose: To determine whether MRI T2-weighted sequences-based texture analysis (TA) can predict histopathological tumor regression grade (TRG) in patients with locally advanced rectal cancer (LARC) undergoing neoadjuvant chemo-radiotherapy (nCRT).
Methods: Data on patients undergoing curative-intent surgery for LARC were collected. Patients with a complete pathological response, or TRG1 according to Mandard's system were classified as responders, while patients with TRG ≥ 2 were classified as non-responders. Tumor TA was performed on each patient's paraxial T2w MRI in both pre- and post-nCRT scans, in order to extract histograms, gray-level co-occurrence matrix (GLCM) and run-length matrix (RLM) texture parameters. For features that showed a significant difference between the two groups, a receiver operating characteristic (ROC) curve was drawn.
Results: Overall, 62 patients with LARC, treated with nCRT and resective surgery at our institution between 2013 and 2019 were identified. Only post-nCRT GLCM maximum probability showed a significant difference between the two groups (2909 ± 4479 in responders vs. 6515 ± 8990 in non- responders; p = 0.039); at the ROC curve, Youden index showed a sensitivity of 14% and a specificity of 100% for this parameter.
Conclusions: MRI T2-weighted sequences-based TA was not effective in predicting pathological complete response to nCRT in patients with LARC. Further studies are needed to thoroughly investigate the potential of MRI TA in this setting.
Keywords: Clinical response; Locally advanced rectal cancer; Magnetic resonance imaging; Neoadjuvant chemo-radiotherapy; Pathological response; Texture analysis.