A Proof-of-Concept Study of Digital Spatial Distance Analysis in Ovarian Mature and Immature Teratomas.

Abstract

Introduction: Mature (MT) and immature teratomas (ITs) contain diverse histological components; however, their tissue-level spatial organization remains poorly understood. This proof-of-concept study aimed to develop a reproducible digital pathology workflow for quantifying the spatial distances among histological components of ovarian teratomas.

Methods: Four representative mature teratoma sections and three ITs with neural elements were analyzed. Whole-slide images of hematoxylin and eosin-stained sections were manually annotated, and each morphologically recognizable component was represented by a centroid using a Python-based image analysis method. Pairwise centroid-based distances were calculated using brute-force and nearest-distance approaches, followed by an exploratory statistical analysis.

Results: Anatomically related tissue pairs, including bronchial epithelium-bronchial gland and epidermis-hair follicle, showed relatively short centroid-based distances, suggesting that the workflow can capture the expected tissue-level architectural relationships. Additional associations, such as bronchial gland-bone proximity, were also observed but should be interpreted as hypothesis-generating findings. No consistent spatial proximity was observed between the immature and mature components.

Conclusion: These findings support the technical feasibility of the proposed workflow for the exploratory tissue-level spatial analysis of teratomas, while emphasizing that spatial associations do not directly demonstrate biological interactions, developmental induction, or lineage relationships.