Hepatoid adenocarcinoma of the stomach: discrimination from conventional gastric adenocarcinoma with a computed tomography-based radiomics nomogram

Background: Previous studies found it difficult to differentiate hepatoid adenocarcinoma of the stomach (HAS) from conventional gastric adenocarcinoma (CGA). We aimed to assess the efficacy of a computed tomography (CT)-based radiomics nomogram in identifying HAS.

Methods: Portal phase CT images were collected from 59 patients with HAS and 122 patients with CGA. HAS and CGA were differentiated through univariate analysis of clinical characteristics, serum biochemical biomarkers, and CT features. The construction of the radiomics signature involved the application of the least absolute shrinkage and selection operator (LASSO) regression model. Multivariable logistic regression analysis was employed to establish the CT-based radiomics nomogram.

Results: The separation of HAS patients from CGA patients relied on the serum alpha-fetoprotein (AFP) level and radiomics signature. The area under the curve (AUC) of AFP was 0.726 [95% confidence interval (CI): 0.639-0.801] in the training cohort and 0.681 (95% CI: 0.541-0.800) in the test cohort, whereas the radiomic signature demonstrated a significantly higher AUC of 0.949 (95% CI: 0.895-0.980) in the training cohort and 0.868 (95% CI: 0.749-0.944) in the test cohort. The nomogram model yielded excellent accuracy for identifying HAS, achieving an AUC of 0.970 (95% CI: 0.923-0.992) in the training cohort and 0.905 (95% CI: 0.796-0.968) in the test cohort.

Conclusions: Radiomics analysis offers promise for differentiating HAS from CGA, and the CT-based radiomics nomogram is likely to have significant clinical implications on HAS distinction.