MRI texture analysis in predicting treatment response to neoadjuvant chemoradiotherapy in rectal cancer

Abstract

To evaluate the importance of MRI texture analysis in prediction and early assessment of treatment response before and early neoadjuvant chemoradiotherapy (nCRT) in patients with locally advanced rectal cancer (LARC). This retrospective study comprised of 59 patients. The tumoral texture parameters were compared between pre- and early nCRT. Area Under receiver operating characteristic (ROC) Curves [AUCs] were used to compare the diagnostic performance of statistically significant difference parameters and logistic regression analysis predicted probabilities for discriminating responders and nonresponders. The Standard Deviation (SD), kurtosis and uniformity were statistically significantly difference between pre- and early nCRT (p = 0.0012, 0.0001, and < 0.0001, respectively). In pathological complete response (pCR) group, pre-uniformity and pre-Energy were significantly higher than that of nonresponders (p = 0.03 and p < 0.01, respectively), while the pre-entropy in nonresponder was reverse (p = 0.01). The diagnostic performance of pre-kurtosis and pre-Energy were higher in tumor regression grade (TRG) and pCR group (AUC = 0.67, 0.73, respectively). Logistic regression analysis showed that diagnostic performance for prediction responder and nonresponder did not significantly improve compared with to pre-uniformity, energy and entropy in pCR group (AUC = 0.76, p = 0.2794, 0.4222 and 0.3512, respectively). Texture parameters as imaging biomarkers have the potential to prediction and early assessment of tumoral treatment response to neoadjuvant chemoradiotherapy in patients with LARC.

Radiomic Features of Primary Rectal Cancers on Baseline T2‐Weighted MRI Are Associated With Pathologic Complete Response to Neoadjuvant Chemoradiation: A Multisite Study

“…For the specific problem of associating CRT response via pretreatment T 2 w MRI, a majority of these studies were limited to single‐site, single‐scanner cohorts 32 . While these studies used different texture feature sets, their performance in classifying patients who later achieved pCR using T 2 w MRI is comparable to our own (AUCs ranging from 0.71–0.80), 33–35 but these studies did not report on how their model generalized to data from other sites. Notably, some of these studies 37,38 reported that multiparametric radiomic features could offer enhanced response classification performance for rectal cancers.…”

Texture analysis of vertebral bone marrow using chemical shift encoding–based water-fat MRI: a feasibility study

“…Besides its recent multifold use in oncological imaging in terms of tissue entity discrimination, characterization, and treatment response monitoring, texture analysis also was described as a reproducible tool to quantitatively assess paraspinal fatty infiltration in MRI [ 33 – 36 ]. In these and other studies, texture heterogeneity was described to be associated with therapy response and clinical outcome [ 35 ]. Besides MRI, the use of texture analysis was investigated on using mammography and in CT in the past to analyze its capability to contribute to computer-aided cancer diagnosis or bone quality measurements [ 23 , 24 ].…”

“…“Entropy” accounts for a measure of randomness in pixel distribution and may depict clinically relevant changes in vertebral micro-architectural alterations. Other groups also investigated on the reliability of different texture parameters and proved that features like kurtosis, skewness, and uniformity showed good results in diagnostic and monitoring quality in cancer imaging [ 34 , 35 ]. The described arbitrarily detected texture features are inherently dependent on imaging properties like resolution, noise, and scan parameters (repetition time, echo time, and receiver bandwidth) [ 39 ].…”

Prediction of Response to Neoadjuvant Chemoradiotherapy by MRI-Based Machine Learning Texture Analysis in Rectal Cancer Patients

“…Nie et al using artificial neural network as classifier found that radiomic features extracted from T1/ T2W, diffusion-weighted (DW) and dynamic contrastenhanced (DCE) MR images could enhance the predictive power of pathologic response after preoperative nCRT for LARC [34]. In another study, Meng et al used MRI texture analysis for nCRT response prediction and found several textures such as standard deviation (SD), kurtosis, and energy; and uniformity were statistically different between responder and non-responder groups [35]. Horvat N et al used MR images to compare value of T2W radiomic textures compared with qualitative assessment at T2W and DW imaging for diagnosis of clinical complete response in patients with LARC after nCRT.…”