Systematic review of machine learning-based radiomics approach for predicting microsatellite instability status in colorectal cancer

Abstract

This study aimed to systematically summarize the performance of the machine learning-based radiomics models in the prediction of microsatellite instability (MSI) in patients with colorectal cancer (CRC). It was conducted according to the preferred reporting items for a systematic review and meta-analysis of diagnostic test accuracy studies (PRISMA-DTA) guideline and was registered at the PROSPERO website with an identifier CRD42022295787. Systematic literature searching was conducted in databases of PubMed, Embase, Web of Science, and Cochrane Library up to November 10, 2022. Research which applied radiomics analysis on preoperative CT/MRI/PET-CT images for predicting the MSI status in CRC patients with no history of anti-tumor therapies was eligible. The radiomics quality score (RQS) and Quality Assessment of Diagnostic Accuracy Studies 2 (QUADAS-2) were applied to evaluate the research quality (full score 100%). Twelve studies with 4,320 patients were included. All studies were retrospective, and only four had an external validation cohort. The median incidence of MSI was 19% (range 8–34%). The area under the receiver operator curve of the models ranged from 0.78 to 0.96 (median 0.83) in the external validation cohort. The median sensitivity was 0.76 (range 0.32–1.00), and the median specificity was 0.87 (range 0.69–1.00). The median RQS score was 38% (range 14–50%), and half of the studies showed high risk in patient selection as evaluated by QUADAS-2. In conclusion, while radiomics based on pretreatment imaging modalities had a high performance in the prediction of MSI status in CRC, so far it does not appear to be ready for clinical use due to insufficient methodological quality.

Artificial Intelligence in Predicting Microsatellite Instability and KRAS, BRAF Mutations from Whole-Slide Images in Colorectal Cancer: A Systematic Review

“…AI applied to radiologic images, called radiomics, was also evaluated to predict the MSI status in colorectal cancer with no exploration of BRAF and KRAS mutation status [ 51 ]. Contrary to pathomics, whole images in radiomics are smaller and consequently less informative but can be entirely used without preprocessing to train algorithms.…”

“…Contrary to pathomics, whole images in radiomics are smaller and consequently less informative but can be entirely used without preprocessing to train algorithms. Recent retrospective studies on the subject showed an AUC with a range from 0.78 to 0.96 AUC [ 51 ]. Some studies combined imaging with clinical and/or histological data (Ki-67, gender, age, tumor localization, differentiation degree of tumor, smoking history, hypertension, diabetes, and family history of cancer), allowing a better prediction of the MSI status [ 52 ].…”

A radiomics signature derived from CT imaging to predict MSI status and immunotherapy outcomes in gastric cancer: a multi-cohort study

“…However, information on MSI expression status obtained postoperatively has limited impact on treatment planning prior to surgery. Moreover, the limited samples obtained through biopsy may not comprehensively reflect tumor heterogeneity, leading to false-negative results (2.1–5.9%) [ 16 , 17 ]. Additionally, biopsies and surgeries are invasive, time-consuming, expensive, and carry risks of complications, making repeated monitoring inconvenient [ 17 ].…”

“…Moreover, the limited samples obtained through biopsy may not comprehensively reflect tumor heterogeneity, leading to false-negative results (2.1–5.9%) [ 16 , 17 ]. Additionally, biopsies and surgeries are invasive, time-consuming, expensive, and carry risks of complications, making repeated monitoring inconvenient [ 17 ]. Hence, there is an urgent need to develop a non-invasive, reliable, and cost-effective method to identify MSI status.…”

Development and validation of a contrast-enhanced CT-based radiomics nomogram for preoperative diagnosis in neuroendocrine carcinoma of digestive system

“…Medical images contain a large amount of invisible data, and it is the value of radiomics to reveal these invisible disease features. Radiomics has been defined as the use of mathematical algorithms to transform the underlying pathophysiological information contained in medical images into quantitative, high-dimensional image features and to explore the correlation of these image features with clinical outcomes or biological properties ( 6 , 7 ). When radiomics is applied to cancer research, it is possible to characterize the imaging of tumor patients non-invasively, quantify the heterogeneity between tissues, describe the microenvironment of the tumor, assess the effectiveness of treatment, and predict survival after obtaining radiological images by CT, MRI, and other examination methods ( 8 , 9 ).…”