Tissue Around Breast Microcalcifications Emerges as a Key Source of Diagnostic Information

Breast microcalcifications are small calcium deposits that are often detected on mammography and digital breast tomosynthesis. Clinicians use their morphology and distribution to assess suspicious breast lesions and decide whether further diagnostic investigation is needed. However, this image-based assessment has focused mainly on the calcified deposits themselves, leaving the potential diagnostic contribution of nearby tissue unclear. A new study sheds light on this surrounding tissue, showing that the area immediately around breast microcalcifications may contain stronger diagnostic information than the calcified core alone.

Researchers from Vrije Universiteit Brussel (VUB)-imec and UZ Brussel used high-resolution three-dimensional micro-computed tomography to examine individual breast microcalcifications at the micrometer scale. They analyzed 94 paraffin-embedded breast biopsy blocks at a spatial resolution of approximately 8 μm, segmenting 3,504 individual microcalcifications. Quantitative radiomic features extracted from these images were then used to train machine-learning classifiers to predict whether the corresponding breast lesions were histopathologically benign or malignant.

To distinguish the diagnostic contribution of the calcified core from that of nearby tissue, the researchers varied the image preprocessing window while keeping the segmentation mask constant. They also expanded and contracted the microcalcification masks and extracted features from concentric shells inside and outside the calcified core. Together, these approaches allowed them to assess whether diagnostic information persisted when the calcium deposit itself was excluded from analysis.

Classification performance improved when surrounding tissue was incorporated into the analysis. Enlarging the preprocessing window around an unchanged calcification mask increased the area under the receiver operating characteristic curve (AUC) from 0.688 to 0.811. Similarly, moderate dilation of the segmentation masks to include adjacent tissue raised the AUC from approximately 0.689 to 0.813. Classifiers trained exclusively on concentric tissue shells maintained an AUC of approximately 0.81, comparable to models that included both the calcified core and its surrounding microenvironment.

The findings were published in Breast Cancer Research under the title “Micro-CT radiomics reveals the microcalcification-surrounding tissue interface as the diagnostic epicenter of breast microcalcifications.” The study involved VUB’s Department of Electronics and Informatics (ETRO)-imec, VUB’s Laboratory of Experimental Pathology, and the Departments of Pathology and Radiology at UZ Brussel. 

The researchers noted that the ex vivo micro-computed tomography approach does not mean the same structures can currently be resolved with conventional clinical breast imaging. Further validation and histological analysis will be needed to identify the tissue structures and cell types that contribute to the diagnostic signal.

Related Links
Vrije Universiteit Brussel
UZ Brussel