Objectives: To investigate the feasibility of radiomics models for predicting the source of hepatic metastases from gastrointestinal (GI) vs. non-gastrointestinal (non-GI) primary tumors on contrast enhanced CT(CECT).
Methods: 347 patients with liver metastases (180 from GI and 167 from non-GI) and abdominal CECT including arterial, portal venous, and delayed phases were divided into training (221) and validation (96) sets at a ratio of 7:3 and an independent testing set (30). Radiomics features were extracted from volumes of interest (VOIs) including tumoral (Vtc) and peritumoral (Vpt) regions on CECT. Optimal radiomics features were used in logistic regression models using receiver operating curve (ROC) analysis to evaluate the diagnostic efficiency.
Results: The best single-phase model was a venous phase peritumoral VOI with 11 features. Area under the curve (AUC), sensitivity and specificity were 0.817, 0.740 and 0.761, respectively in the validation set. While the best arterial phase tumoral VOI gave an AUC of 0.677 in the validation set. For the combined models, peritumoral VOI in arterial and venous phases (15 features) achieved the best prediction performance with an AUC of 0.926 in the validation set and 0.884 in the testing set.
Conclusion: Liver peritumoral radiomics features extracted from CECT were able to identify the source of hepatic metastases as GI versus non-GI.
Advances in knowledge: Peritumoral radiomics features showed a correlation with source of liver metastases. The radiomics features from liver peritumoral arterial and venous phases CT were promising in differentiating the source of hepatic metastases from GI vs. non-GI primary tumors.
Keywords: Liver Neoplasms; Neoplasm Metastasis; Tomography; Unknown Primary; X-Ray Computed.
© The Author(s) 2024. Published by Oxford University Press on behalf of the British Institute of Radiology. All rights reserved. For permissions, please email: journals.permissions@oup.com.