Magnetic Resonance-Based Texture Analysis Differentiating KRAS Mutation Status in Rectal Cancer

Area of Science:

• Oncology
• Radiology
• Genetics

Background:

• Kirsten Ras (KRAS) oncogene mutations are prevalent in colorectal cancers (30%-40%) and carry prognostic implications in rectal cancer.
• Accurate determination of KRAS mutation status is crucial for guiding treatment strategies and predicting patient outcomes.

Purpose of the Study:

• To investigate the utility of radiomics features from T2-weighted magnetic resonance (MR) images in distinguishing between rectal tumors with mutant KRAS and wild-type KRAS.
• To develop a predictive model for KRAS mutation status using imaging biomarkers.

Main Methods:

• Retrospective analysis of T2-weighted MR images from 60 rectal cancer patients (25 mutant KRAS, 35 wild-type KRAS).
• Manual segmentation of regions of interest by two radiologists, followed by texture analysis to extract radiomic features.
• Identification of significant features using t-tests and construction of a decision tree model for KRAS mutation prediction.

Main Results:

• Three radiomic features, including skewness with gradient filter and standard deviations of medium texture (SSF3, SSF4), were significantly associated with KRAS mutational status (p < 0.05).
• The decision tree model achieved a sensitivity of 84%, specificity of 80%, and accuracy of 81.7% in predicting KRAS mutation status.
• Mutant KRAS tumors exhibited significantly higher mean skewness and standard deviations for medium texture compared to wild-type tumors.

Conclusions:

• T2-weighted MR imaging-based texture analysis can effectively differentiate between mutant and wild-type KRAS in rectal cancer.
• Radiomics offers a non-invasive method for preoperative prediction of KRAS mutation status, potentially influencing clinical management.