Contrast-enhanced CT useful for classifying histologic subtypes in epithelial ovarian carcinoma: JAMA

Histologic subtypes and International Federation of Gynecology and Obstetrics (FIGO) stages are crucial characteristics for treatment stratification as well as disease prognostication. Epithelial ovarian carcinoma can be classified as high-grade serous carcinoma (HGSC) and non-HGSC according to the different pathways of carcinogenesis.

High-grade serous carcinoma is the most frequent and lethal subtype, accounting for 70% of EOC. Non-HGSC consists of low-grade serous carcinoma, mucinous carcinoma, endometrioid carcinoma, clear cell carcinoma, and malignant Brenner tumors. Non-HGSC presents as an indolent behavior and progresses through a stepwise mutation process, whereas HGSC tends to be more aggressive with greater genetic instability and thereby metastasizes rapidly. With better understanding of the molecular events that support ovarian carcinogenesis, newer tailored therapies and subtype-specific research could be investigated.

Accurate identification of histologic subtypes is vital and will be beneficial for personalized management. In clinical practice, histologic diagnosis is made through surgery or tissue biopsy. Intraoperative frozen section could assist in histologic classification but could still result in misdiagnosis, not to mention the invasive nature of these procedures and associated increase in intraoperative time.

Epithelial ovarian carcinoma has intratumor heterogeneity, especially in large ovarian masses. Radiomics analysis is emerging as a noninvasive and useful tool to assess highly heterogenous malignant tumors, such as HGSCs. Radiomics is a mathematical quantitative analysis that converts medical images into minable, high-dimensional data by extracting large amounts of mathematical features. Previous studies reported the ability of computed tomography (CT)–based radiomics in the evaluation of EOC, including differentiating between EOC and non-EOC, as well as predicting clinical outcome or survival.

To validate these findings based on a large cohort across multiple centers in East Asia, with different machines and varied imaging parameters, the aim of this multicenter study by Mandi Wang and team was to assess the value of CT-based radiomic features in histologic subtyping of EOC.

In this diagnostic study, 665 patients with histologically confirmed epithelial ovarian carcinoma were retrospectively recruited from 4 centers (Hong Kong, Guangdong Province of China, and Seoul, South Korea) between January 1, 2012, and February 28, 2022. The patients were randomly divided into a training cohort (n = 532) and a testing cohort (n = 133) with a ratio of 8:2. This process was repeated 100 times. Tumor segmentation was manually delineated on each section of contrast-enhanced CT images to encompass the entire tumor. Selected features were used to build the logistic regression model for differentiating high-grade serous carcinoma and non–high-grade serous carcinoma.

In this study, 665 female patients (mean [SD] age, 53.6 [10.9] years) with epithelial ovarian carcinoma were enrolled and analyzed. The Dice similarity coefficients of intraobserver and interobserver were all greater than 0.80. Twenty radiomic features were selected for modeling. The areas under the curve of the logistic regression model in differentiating high-grade serous carcinoma and non–high-grade serous carcinoma were 0.837 (95% CI, 0.835-0.838) for the training cohort and 0.836 (95% CI, 0.833-0.840) for the testing cohort.

This multicenter diagnostic study investigated the clinical utility of radiomic features extracted from ceCT based on an LR model in differentiating histologic subtypes of EOC. The proposed CT radiomics model exhibited excellent performance in both the training and testing cohorts, with AUCs of 0.837 and 0.836, respectively. Histologic subtypes of EOC were classified as HGSC and non-HGSC in our study, according to the different pathways of ovarian tumorigenesis. High-grade serous carcinoma tends to have a poorer prognosis compared with non-HGSC, with an increased risk of death. Histologic subtyping also helps in treatment stratification and subtype-specific research in EOC. Therefore, the preoperative prediction of the histologic subtypes in EOC could notably benefit the clinical management and prognosis evaluation.

This multicenter diagnostic study assessed the utility of CT radiomics in discriminating histologic subtypes of EOC. The proposed LR model of CT radiomic features selected by the voted LASSO method demonstrated excellent diagnostic performance in differentiating HGSC and non-HGSC.

Source: Mandi Wang, PhD; Jose A. U. Perucho, PhD; Yangling Hu; JAMA Network Open. 2022;5(12):e2245141. doi:10.1001/jamanetworkopen.2022.45141 (Re