Raman imaging as a tool for the chemical and spatial characterization of breast microcalcifications to improve lesion assessment

Background: microcalcifications (MC) are common findings on screening mammography and are among the earliest signs of breast cancer. At the same time, from the use of well-known radiographic risk score systems, which include MC assessment, such as Breast Imaging-Reporting and Data System (BI-RADS), only 20% of screened patients are further associated with malignancy. This leads to repeated biopsies and to unnecessary surgeries with discomfort for patients and increased costs for the healthcare system. Furthermore, the definitive diagnosis by histological and immunohistological evaluations is still laborious and time-consuming. Raman Spectroscopy (RS) is a photonic approach capable to provide detailed chemical information of analysed samples without complex tissue preparation or staining. RS has a proven ability to distinguish different crystal structures, including those commonly present in MC. In this context some studies based on RS suggested a correlation between MC chemical features and pathology. On the other hand, previous Raman-based studies mainly investigated the overall MC chemical composition (by single-point scans) while an extensive MC characterization by Raman imaging approaches (mapping) for diagnostic purposes is still lacking. The aim of this study is to assess the usefulness of Raman imaging as a quick and accurate tool for a complete spatial characterization of MC detected on screening mammography and sampled by breast biopsy in order to better distinguish malignant vs. benign lesions. Method: 30 patients with breast calcifications detected on mammography with radiological classification BI-RAD 3-5 where selected and evaluated by core biopsy. 10 μm formalin-fixed paraffin-embedded (FFPE) histological sections obtained from biopsies were dewaxed with a specifically developed protocol that allows the removal of paraffin in less than 15 minutes. All MC present in each tissue section (usually from 2 to >10 per section) were then characterized by a Raman microscope thus obtaining Raman maps with lateral resolution between 5 and 10 μm. After pre-processing steps the Raman maps were analysed by both clustering and multivariate analysis approaches used to produce false-colour images and to perform automated features identification. Results: Our results confirm that hydroxyapatite is the prevalent form of calcium phosphate in MC and that MC composition correlates with lesion malignancy. In addition, thanks to the Raman imaging approach used here, we report for the first time that hydroxyapatite is more homogeneously distributed in malignant lesions and that, on the contrary, benign lesions show a heterogeneous distribution of hydroxyapatite, whitlockite and calcium-carbonate, inside the lesion and in the surrounding tissue. Conclusion: These evidences suggest that the characterization of MC by Raman imaging is a potential tool for the definition of new diagnostic signatures of breast cancer, especially if we consider that these evaluations can be performed by the simple and relative fast scanning of dewaxed slices, without altering the clinical workflow and without the need of staining or antibodies. Further studies with a larger cohort will be done to validate these results.