Objectives: In Japan, invasive ductal carcinomas, which account for 75% of breast cancer cases, are sub-classified as solid, tubule-forming, scirrhous, and other types based on the histopathological findings. Although time-intensity curve (TIC) analysis of magnetic resonance (MR) images has shown diagnostic ability in differentiating benign and malignant tumors, its ability to diagnose different tumor tissue types has not yet been achieved. In this study, we report a histological classification of invasive ductal carcinoma using the TIC analysis of dynamic MR images of the mammary gland.
Material and methods: A total of 312 invasive ductal carcinomas were analyzed, and each tissue type that indicated malignancy in the washout parts of the tumors was classified and characterized using the TIC.
Results: The tissue was classified, and the results were then compared to the pathohistological diagnosis. Using this method, the accuracy of tissue classification by quantitative analysis of TIC-MR images was 86.9% (271/312), which was higher than that obtained by ultrasonography 68.9% (215/312).
Conclusion: This method is effective for classifying tissue types in invasive ductal carcinoma.
Keywords: Breast cancer; Computer-aided diagnosis; Histological classification; Invasive ductal carcinoma; Time-intensity curve.
© 2020 Published by Scientific Scholar on behalf of Journal of Clinical Imaging Science.