AUTHOR=Al Khader Ali , Seghetti Christian , Oumlil Fatine , Tollit Anna , Tirabosco Roberto , Amary Fernanda , O’Donnell Paul , Flanagan Adrienne M. TITLE=The value of morphology: osteoclast-like cells in soft tissue tumours JOURNAL=Pathology and Oncology Research VOLUME=Volume 31 - 2025 YEAR=2025 URL=https://www.por-journal.com/journals/pathology-and-oncology-research/articles/10.3389/pore.2025.1612175 DOI=10.3389/pore.2025.1612175 ISSN=1532-2807 ABSTRACT=Recognition of unusual histological features can augment and hasten a diagnosis but also stimulate ideas about physiological and pathological cellular interactions. Osteoclasts resorb mineralised tissue and therefore can be found at sites of heterotopic bone formation. However, multinucleated giant cells with morphological features of osteoclasts, so called ‘osteoclast-like cells’ can also be encountered in a variety of soft tissue tumours unrelated to ossification and calcification. Prompted by the presence of osteoclast-like cells in undifferentiated pleomorphic sarcoma while undertaking our Artificial Intelligence project for classifying sarcoma, we reviewed the English literature for these cells in soft tissue tumours and we found that this was poorly documented, and much was published before the release of the WHO essential diagnostic criteria in 2020. There were numerous single case reports and small series of a broad range of soft tissue tumours with osteoclast-like cells but only a limited number of diagnoses in which these cells were reported recurrently. We provide a comprehensive update of osteoclast-like cells and mineralisation in soft tissue tumours from the literature. We also present real-world incidence of osteoclast-like cells from selected tumour types in our Whole Slide Image (WSI) library of soft tissue tumours. Assessment of WSI from 1100 different patients showed that osteoclast-like cells were relatively common and under-recognised in nodular fasciitis (18.5 of 200), angiomatoid fibrous histiocytoma (17.5% of 40), undifferentiated pleomorphic sarcoma (15% of 261) and epithelioid sarcoma (9% of 68) while they were never encountered in myxofibrosarcoma (0/250) and clear cell sarcoma of soft tissue (0/80). Awareness of this phenomenon not only helps shape the differential diagnosis but also can be used to stimulate pathobiological questions and to enhance the performance of AI models for classifying disease.