Radiomics Models as Tools for Predicting Genetic Mutations in Colorectal Cancer: A Systematic Review and Meta-Analysis

Clinical Snapshot

Key Findings

• P Value: Not reported in abstract

• Effect Size: Pooled sensitivity 0.75; pooled specificity 0.78; overall AUC 0.79. Radio-clinical models (integrating clinical and radiomics features) demonstrated superior sensitivity compared to radiomics-only models

• Primary Outcome: Diagnostic accuracy of radiomics models for predicting genetic mutations in colorectal cancer across 16 included studies

• Nnt Or Sensitivity: Sensitivity 0.75 (75%), Specificity 0.78 (78%), AUC 0.79 — indicating that approximately 1 in 4 mutation-positive patients would be missed (false negative rate 25%) and approximately 1 in 5 mutation-negative patients would be incorrectly flagged (false positive rate 22%) using current radiomics models

• Confidence Interval: Sensitivity: 95% CI 0.67–0.81; Specificity: 95% CI 0.70–0.85

Clinical Application

Radiomics workflows require dedicated image segmentation software, feature extraction pipelines, and validated predictive models — infrastructure not routinely available in most clinical radiology departments. Standardisation of imaging protocols (slice thickness, reconstruction kernels, contrast timing) is a prerequisite for model deployment. Current evidence does not support standalone clinical use; radiomics outputs should be considered adjunctive to, not replacements for, molecular pathology. In Australia, molecular testing for CRC mutations (KRAS, NRAS, BRAF, MMR/MSI) is integral to treatment planning and is supported by PBS-listed targeted therapies including anti-EGFR agents (cetuximab, panitumumab — PBS-listed for RAS wild-type metastatic CRC) and pembrolizumab (PBS-listed for MSI-H/dMMR metastatic CRC). The RACGP and Cancer Australia guidelines emphasise tissue-based molecular profiling as the standard of care. Radiomics-based mutation prediction is not currently endorsed by the TGA, RACGP, or any Australian oncology guideline body. However, the technology may have future utility in the Australian context for patients with inaccessible or heterogeneous tumours, or as a triage tool to prioritise molecular testing. Australian centres participating in the AGITG or similar cooperative groups would be well-positioned to conduct prospective validation studies. Adult patients with confirmed colorectal cancer who are candidates for pre-operative CT, MRI, or PET/CT imaging and in whom genetic mutation status (e.g., KRAS, BRAF, MMR/MSI) is clinically relevant for treatment planning, particularly those in whom tissue biopsy is technically challenging, high-risk, or subject to sampling heterogeneity

Abstract

BACKGROUND: The evaluation of genetic mutations is crucial for personalized therapy in colorectal cancer (CRC), but the invasive tissue biopsy is subject to sampling bias and other complications. Radiomics has emerged as a non-invasive tool to predict these mutations from standard medical images. In this systematic review and meta-analysis, we aimed to evaluate the diagnostic accuracy and methodological quality of radiomics models for predicting key genetic mutations in CRC. METHODS: A comprehensive search of PubMed, Scopus, Web of Science, and Embase was conducted in accordance with PRISMA guidelines. Studies evaluating radiomics models for predicting genetic mutations in CRC patients using pre-operative CT, MRI, or PET/CT were included. A meta-analysis of diagnostic accuracy was performed to calculate the pooled sensitivity, specificity. Methodological quality was assessed using the Radiomics Quality Score (RQS) and QUADAS-2 tools. RESULTS: Sixteen studies were included in the quantitative analysis. The pooled sensitivity and specificity were 0.75 (95% CI, 0.67-0.81) and 0.78 (95% CI, 0.70-0.85), respectively, with an overall AUC of 0.79. Subgroup analyses revealed that radio-clinical models integrating both clinical and radiomics features achieved superior sensitivity compared to models with only radiological input. However, the overall methodological quality of the included studies was low, with a mean RQS of 45%. CONCLUSION: Conventional radiomics models demonstrate promising results for the non-invasive prediction of genetic mutations in CRC, with sensitivity enhanced by the integration of clinical data. Despite this potential, significant methodological shortcomings and heterogeneity across studies highlight the need for standardized protocols and large-scale, prospective validation before these models can be translated into routine clinical practice.

References

1. 1.Rahnama, Y., Shahbazi, A., Dadashi, A., Fathabadi, F., Salahshour, F., Delazar, S., Sedaghat, M., Keshvari, A., Kazemeini, A., Keramati, M. R., Fazeli, M. S., Behboudi, B., & Ahmadi-Tafti, S. M. (2026). Radiomics models as tools for predicting genetic mutations in colorectal cancer: A systematic review and meta-analysis. Journal of Gastrointestinal Cancer. https://doi.org/10.1186/s12880-020-00457-4