恶性肿瘤
对比度(视觉)
对比度增强
医学
乳房磁振造影
磁共振成像
放射科
肿瘤科
乳腺癌
内科学
乳腺摄影术
计算机科学
癌症
人工智能
作者
Kamber Göksu,Ahmet Vural,Ahmet Kahraman
出处
期刊:Journal of Cancer
[Ivyspring International Publisher]
日期:2025-07-04
卷期号:16 (10): 3103-3111
摘要
Background: Magnetic resonance imaging (MRI) has a limited role in distinguishing non-mass enhancement (NME) lesions as benign or malignant and determining whether the lesions are invasive or not. In this study, we aimed to investigate the differences in MRI of benign and malignant NME lesions and to determine the relationship between the pattern of enhancement in NME lesions and histopathologic diagnosis. Materials and methods: Breast MRI examinations (n=5214) performed at the study institution between January 2018 and July 2024 were evaluated. We enrolled 460 patients in the study. NME lesions were classified according to the BI-RADS atlas. In addition, linear enhancements were divided into branching and non-branching. Factors showing significant associations in univariate analyses were evaluated with multivariate analyses using the logistic regression model. The assessments were performed by two radiologists who are experienced in breast imaging. Results: This study included 460 NME lesions (342 benign and 118 malignant). Focal and segmental distribution, dynamic enhancement features, Type I (persistent) and Type III (wash-out) dynamic curve modes, and clustered-ring internal enhancement pattern features showed statistically significant differences in terms of differentiating benign from malignant (P<0.05). Heterogeneous enhancement gave significant results in distinguishing invasive carcinoma from ductal carcinoma in situ (DCIS) (P<0.05). Wash-out type curve from dynamic enhancement curves was also seen at a higher rate in invasive carcinomas. Although the general results are similar to previous studies, in our study, unlike other studies, enhancements showing linear distribution were divided into two groups branching and non-branching, and lesion size was measured. It was observed that branching enhancements and lesion sizes greater than 15 mm significantly indicated malignancy (p<0.05). Conclusions: MRI is a valuable way to identify malignant NME lesions and may be useful in determining whether the lesions are invasive or not. Evaluating NME lesions with breast MRI can help decide on biopsy when branching types of lesions with linear distribution and lesions greater than 15 mm are detected.
科研通智能强力驱动
Strongly Powered by AbleSci AI