Data Science Meets FTIR Imaging: a Promising Probe to Improve the Diagnosis of Human Uterine Muscle Lesions

Uterine smooth muscle tumors include a broad range of neoplasms, from benign leiomyomas (LMs) to malignant leiomyosarcomas (LMS), as well as intermediate forms classified as Smooth Muscle Tumors of Uncertain Malignant Potential (STUMP). An accurate diagnosis of these tumor types is essential for their appropriate clinical management; however, it remains challenging due to possible overlapping of histological features. In this study, a multidisciplinary approach combining Fourier Transform Infrared Imaging (FTIRI) spectroscopy, a label-free and non-destructive analytical technique, with histology and statistical analyses have been exploited for investigating the morpho-chemical characteristics of these uterine smooth muscle tumors. The analysis aimed to identify new reliable and diagnostic spectral markers, complementary to traditional histology, and thus useful for improving accuracy in cases with uncertain morphological features. Tissue samples including different leiomyoma histological subtypes, such as usual, cellular, apoplectic, and bizarre, were analyzed and compared with LMS and healthy myometrium. The analysis of IR data, submitted to univariate and multivariate statistical approaches, such as Principal Component Analysis (PCA) and Hierarchical Cluster Analysis (HCA), revealed distinctive spectral profiles associated with each tumor type and indicated changes in collagen content and organization as key features for a reliable discrimination not only between benign and malignant tissues but also among different LM histotypes.