Researchers have developed a preoperative contrast-enhanced computed tomography classification system and diagnostic nomogram capable of distinguishing signet ring cell carcinoma from non-signet ring cell carcinoma in patients with gastric cancer ovarian metastases.
The retrospective cohort comprised 137 female patients with pathologically confirmed gastric cancer ovarian metastases who underwent preoperative contrast-enhanced computed tomography imaging. The team initially categorized each observed tumor into one of three general morphological patterns based on overall structure, grouping lesions as cystic-dominant, mixed cystic-solid, or solid-dominant presentations.
Beyond broad morphological groupings, the investigators introduced a dedicated classification scheme specifically focused on the enhancement behavior of solid tumor components after contrast administration. This novel system stratified solid areas into three distinct enhancement patterns: type I was defined by weak enhancement exhibiting a change in attenuation of less than 40 Hounsfield units; type II was defined by strong homogeneous enhancement with an attenuation increase of 40 Hounsfield units or greater; and type III was characterized by heterogeneous enhancement. In addition, the protocol established specific criteria for identifying what the researchers termed critical regions. These critical regions were formally defined as hypodense zones measuring between 20 and 30 Hounsfield units on unenhanced baseline scans that subsequently exhibited minimal contrast enhancement, documented as an attenuation increase of only 10 to 20 Hounsfield units post-contrast.
To construct a predictive tool while appropriately managing the statistical complexities of the dataset, the authors applied generalized estimating equations featuring an exchangeable correlation structure. Through both univariable and multivariable generalized estimating equation evaluations, the researchers identified three independent predictive variables for differentiating signet ring cell carcinoma: overall lesion size, solid component enhancement type, and the presence of identified critical regions. These three independent factors were then synthesized into a practical diagnostic nomogram.
The discriminatory performance of the resulting nomogram was evaluated via receiver operating characteristic curve analysis and validated through 1,000 cluster bootstrap resamples. In initial testing, the diagnostic nomogram achieved an area under the receiver operating characteristic curve of 0.802, accompanied by a 95 percent confidence interval spanning from 0.743 to 0.872. Following cluster bootstrap validation to account for model overfitting, the optimism-corrected area under the curve remained robust at 0.795, confirming stable discriminatory accuracy for distinguishing signet ring cell histology within ovarian metastases.
In addition to evaluating the primary ovarian lesions, the investigators conducted secondary exploratory analyses to determine whether the identified computed tomography features could predict the histological characteristics of the primary gastric tumor or the pathological concordance between primary and metastatic sites. These secondary assessments demonstrated comparatively limited discriminatory utility. The predictive performance for primary gastric pathology achieved area under the curve values ranging between 0.651 and 0.667, while the prediction of primary-metastatic concordance yielded area under the curve values between 0.674 and 0.702. The research team noted that these constrained secondary results likely reflect underlying tumor biologic heterogeneity between primary gastric malignancies and their corresponding metastatic deposits.
The authors concluded that their proposed nomogram offers an informative, non-invasive exploratory framework for estimating the preoperative probability of signet ring cell carcinoma in patients presenting with gastric cancer ovarian metastases. However, they explicitly underscored that the findings originate from a single retrospective cohort, noting that prospective multicenter validation studies will be essential before the classification model can be integrated into routine clinical decision-making.
Subject of Research: Computed tomography evaluation of gastric cancer ovarian metastases
Article Title: A novel contrast-enhanced CT classification system for differentiating signet ring cell carcinoma in gastric cancer ovarian metastases: a pathologically correlated study
Article References: Han, Z., Hao, Y., Wu, X., Zhu, X., Chen, H., & Yu, P. (2026). A novel contrast-enhanced CT classification system for differentiating signet ring cell carcinoma in gastric cancer ovarian metastases: a pathologically correlated study. Cancer Imaging. https://doi.org/10.1186/s40644-026-01155-0
Image Credits: AI Generated
DOI: 10.1186/s40644-026-01155-0
Keywords: computed tomography, gastric cancer, ovarian metastases, signet ring cell carcinoma, diagnostic nomogram, cancer imaging, generalized estimating equations, radiology, novel, contrast-enhanced, classification, system
News Source: Nathaniel Bowman. (November 6, 2026). CT Classification Model Differentiates Signet Ring Cell Carcinoma in Ovarian Metastases. Scienmag.
