Histogram analysis of apparent diffusion coefficient with multi-b-value DWI at 3.0T MRI: correlation with prognostic factors and subtypes of breast cancer
10.3969/j.issn.1000-8179.2019.01.164
- VernacularTitle:3.0T MRI多b值扩散加权成像ADC直方图与乳腺癌分子分型及预后因素的相关性研究
- Author:
Shumeng SUN
1
;
Zhenzhen SHAO
;
Peifang LIU
Author Information
1. 天津医科大学肿瘤医院乳腺影像诊断科
- Keywords:
breast cancer;
diffusion-weighted magnetic resonance imaging;
molecular subtyping;
prognosis
- From:
Chinese Journal of Clinical Oncology
2019;46(1):39-43
- CountryChina
- Language:Chinese
-
Abstract:
Objective: To investigate the correlations between parameters of histograms of the apparent diffusion coefficient (ADC) with multi-b-value diffusion-weighted imaging (DWI) at 3.0T MRI and prognostic factors and molecular subtypes of breast cancer, to evaluate the diagnostic performance of ADC histograms at different b values. Methods: A total of 114 patients (116 lesions) with inva-sive ductal carcinomas confirmed by surgical pathology who underwent breast magnetic resonance imaging from March 2015 to Janu-ary 2016 in Tianjin Medical University Cancer Institute and Hospital were analyzed retrospectively. The histograms of ADC with b val-ues of 0, 500, 800, 1000, and 1,500 s/mm2 were generated using Image J software. Various parameters were calculated: for example, the minimum, mean, mode, skewness, and kurtosis. Different groups were based on the molecular subtypes, tumor size (T1 vs . T2-3), histologic grade (high vs. low), and lymph node status (positive vs. negative) that were recorded. Mann-Whitney U tests were used to compare the differences in ADC histogram parameters between two different groups. Receiver operating characteristic curves (ROC) were constructed. Results: The skewness was lower in Luminal tumors than that in non-Luminal tumors with b values of 500, 800, 1, 000, and 1,500s/mm2 (P<0.05). The ADCmin was higher in human epidermal growth factor receptor-2 (HER-2) over-expression than in non-HER-2 over-expression (P<0.05). The kurtosis was lower in stage T1 tumors than stage T2-3 tumors (P<0.05), and kurtosis was cor-related with tumor size (P<0.05). ADCmode and ADCmean were different between different histological subtypes with a b value of 500 s/mm2 (P<0.05). Under different b values, there were no significant differences in terms of areas under the curve for each histogram pa-rameter, which had statistically significant differences (P>0.05). Conclusions: Multi-b-value DWI ADC histogram analysis, as a quantita- tive method to characterize tumor heterogeneity, can reflect the biological behavior and prognosis of breast cancer to some extent, and the diagnostic performance of ADC histograms showed no significant differences in differentiating molecular types and prognostic factors of breast cancer at different b values.
