Work

Value Analysis of MRI Habitat Analysis Combined Model in the Diagnosis of Ovarian Tumors

Year: 2026

Type: article

Abstract: Objective This study aimed to investigate the clinical diagnostic performance of a combined classification model incorporating magnetic resonance imaging (T1WI-CE) habitat and human epididymis protein 4 (HE4) for differentiating borderline ovarian tumors (BOTs) from malignant epithelial ovarian tumors (MEOTs). Methods A retrospective analysis was conducted on 127 patients with pathologically confirmed ovarian tumors, including 62 with BOTs and 65 with MEOTs, all of whom underwent preoperative magnetic resonance imaging examination. Twenty habitat features, including the original images, were extracted. T1WI-CE was used to extract 2395 radiomics features from two habitat subregions. Feature selection was performed using correlation analysis and least absolute shrinkage and selection operator regression. Results The combined classification model had the highest area under the curve, 0.941 in the training group and 0.880 in the test group, thus outperforming the habitat area and clinical data classification model. The DeLong test demonstrated statistically significant differences between the combined classification model and the clinical classification model, with P values of 0.041 in the training group and 0.023 in the test group. Additionally, a statistically significant difference was observed in the DeLong test results between the cystic habitat subregion (H2) and the overall habitat region. Conclusions The combined classification model of habitat analysis and clinical data effectively improved the diagnostic efficacy of differentiating borderline from malignant ovarian tumors. The diagnostic efficacy of the habitat subregion (H1) dominated by solid components and the habitat overall region was superior to that of the habitat subregion dominated by cystic components.

Source: Technology in Cancer Research & Treatment

DOI: 10.1177/15330338261439633

Authors: Junying Wang, Kang Liu, Yan Sui, Siyu Ma, Dewu Yang

Institutions: Beijing Health Vocational College, Capital Medical University

Language: English

Topic: Radiomics and Machine Learning in Medical Imaging

Subfield: Radiology, Nuclear Medicine and Imaging

Field: Medicine

Domain: Health Sciences

SDG: Good health and well-being

Open Access status: gold

Estimated APC paid: $2,650