Imaging predictors of treatment outcomes in rectal cancer: An overview

Radiomics performs comparable to morphologic assessment by expert radiologists for prediction of response to neoadjuvant chemoradiotherapy on baseline staging MRI in rectal cancer

“…(1) Using minimum Redundancy Maximum Relevance (mRMR), implemented in Python package 'mifs' [28], a set of candidate features was selected from the training set, which (2) were fitted into a logistic regression model with l2 regularization and balanced class weights. To approximate the mutual information between the outcome and continuous features during mRMR selection, we employed the nearest neighbor method as described by Ross et al [29] Optimum number of features to select [5][6][7][8][9][10], as well as the k neighbors parameter [5][6][7][8] in mRMR and the C regularization parameter [10 −7 -10 2 ] in the logistic regression model were determined by 5-fold stratified cross-validation on the training set. Finally, the performance of the radiomics model to predict a 'complete' and 'good' response, respectively, was assessed using the Wilcoxon rank-sum test and by calculating the area under the ROC curve (AUC).…”

“…Several studies have shown that imaging may play a role in the pre-treatment prediction of response, with a particular focus on MRI being one of the main imaging modalities used to stage rectal cancer. "Semantic features" including the T-stage, N-stage, Circumferential Resection Margin (CRM), Extra-Mural Venous Invasion (EMVI) and baseline tumor volume have been shown to be associated with the chance of response to varying degrees [7][8][9][10]. Promising (though inconsistent) results have also been reported for the use of more novel functional MR imaging sequences such as diffusion-weighted imaging (DWI) and dynamic contrast enhanced (DCE) MRI, that can provide quantifiable information on biological tumor properties such as tumor cellularity and tumor perfusion [11][12][13].…”

Radiomics as a New Frontier of Imaging for Cancer Prognosis: A Narrative Review

“…Although magnetic resonance imaging (MRI) is the standard imaging technique for local staging and re-evaluation after nCRT in RC [162][163][164][165], its clinical utility in predicting pCR after nCRT is still uncertain. In this scenario, radiomics has emerged as a promising tool and several studies have shown promising results in the prediction and early assessment of response to chemotherapy using MRI [166][167][168]. Following the aim of the review, we will focus on studies investigating the prediction of response and the prognosis of LARC patients.…”

Value of combined multiparametric MRI and FDG-PET/CT to identify well-responding rectal cancer patients before the start of neoadjuvant chemoradiation

“…There is a growing interest in the value of imaging as a potential source for these biomarkers, with numerous reports exploring the potential of metabolic imaging (FDG-PET/CT) [11][12][13][14] and MRI with the addition of functional imaging sequences such as diffusion-weighted imaging (DWI) [15][16][17][18][19][20]. Most studies so far have focused on singlemodality imaging and included only one or a few imaging markers.…”