Deep-learning-based 3D super-resolution MRI radiomics model: superior predictive performance in preoperative T-staging of rectal cancer

Abstract

Objectives To investigate the feasibility and efficacy of a deep-learning (DL)-based three-dimensional (3D) super-resolution (SR) MRI radiomics model for preoperative T-staging prediction in rectal cancer (RC). Methods Seven hundred six eligible RC patients (T1/2 = 287, T3/4 = 419) were retrospectively enrolled in this study and chronologically allocated into a training cohort (n = 565) and a validation cohort (n = 141). We conducted a deep-transfer-learning network on high-resolution (HR) T2-weighted imaging (T2WI) to enhance the z-resolution of the images and acquired the preoperative SRT2WI. The radiomics models named modelHRT2 and modelSRT2 were respectively constructed with high-dimensional quantitative features extracted from manually segmented volume of interests of HRT2WI and SRT2WI through the Least Absolute Shrinkage and Selection Operator method. The performances of the models were evaluated by ROC, calibration, and decision curves. Results ModelSRT2 outperformed modelHRT2 (AUC 0.869, sensitivity 71.1%, specificity 93.1%, and accuracy 83.3% vs. AUC 0.810, sensitivity 89.5%, specificity 70.1%, and accuracy 77.3%) in distinguishing T1/2 and T3/4 RC with significant difference (p < 0.05). Both radiomics models achieved higher AUCs than the expert radiologists (0.685, 95% confidence interval 0.595–0.775, p < 0.05). The calibration curves confirmed high goodness of fit, and the decision curve analysis revealed the clinical value. Conclusions ModelSRT2 yielded superior predictive performance in preoperative RC T-staging by comparison with modelHRT2 and expert radiologists’ visual assessments. Key Points • For the first time, DL-based 3D SR images were applied in radiomics analysis for clinical utility. • Compared with the visual assessment of expert radiologists and the conventional radiomics model based on HRT2WI, the SR radiomics model showed a more favorable capability in helping clinicians assess the invasion depth of RC preoperatively. • This is the largest radiomics study for T-staging prediction in RC.

A bibliometric and visual analysis of publications on artificial intelligence in colorectal cancer (2002-2022)

“…( 71 ) also conducted a related study and showed that AI performed far better than experienced gastrointestinal pathologists. In addition, CT- or MRI-based imaging histology has many applications, including assessing pathological responses after radiotherapy or chemotherapy ( 72 ) and predicting colon cancer infiltration and metastasis ( 73 , 74 ). The duration of keyword bursts was long before 2016 and became shorter after 2016 ( Figure 6C ).…”

Deep Learning‐Based Multiparametric MRI Model for Preoperative T‐Stage in Rectal Cancer

“…Hou et al constructed a threedimensional super-resolution model based on deep learning to enhance the z-resolution of T2W images, with an AUC of 0.86. 23 However, this research was limited to manual segmentation. Our work employs a multi-parametric MRI image fusion approach, extracting features from the ROI of the tumor and surrounding tissue.…”

“…20 Although there have been a few studies that have utilized deep learning technology to predict the preoperative T-stage of rectal cancer, most of these studies have only focused on using a single imaging sequence, such as ADC or T2W. 23,24 For instance, Hou et al investigated a deep-learning model for generating the super-resolution image and extracted radiomics features extracted to predict preoperative T-staging. 23 Wang et al used a support vector machine (SVM) model to predict the T-stage of rectal cancer on the T2W image.…”

“…23,24 For instance, Hou et al investigated a deep-learning model for generating the super-resolution image and extracted radiomics features extracted to predict preoperative T-staging. 23 Wang et al used a support vector machine (SVM) model to predict the T-stage of rectal cancer on the T2W image. 24 However, the optimal combination of sequences for this type of prediction has been largely ignored.…”

Automatic treatment outcome prediction with DeepInteg based on multimodal radiological images in rectal cancer

“…In recent years, deep learning has shown promising value in image analysis to assist the clinical decision on cancer or other treatment [ [16] , [17] , [18] , [19] ]. Currently, deep learning-based tools have been developed for diagnosis of disease status in the serial management of patients by using multiple types of medical image [ [20] , [21] , [22] , [23] ].…”