This study develops a pan-species cancer digital pathology atlas (pan-species.ai) and uses a supervised convolutional neural network trained on human samples for computational comparative pathology. The AI algorithm accurately classifies cells in two transmissible cancers and 18 other vertebrate species, with accuracy influenced by cell morphological similarity. A spatial immune score is associated with prognosis in canine melanoma and prostate tumors. A 'morphospace overlap' metric guides technology deployment. The study provides a foundation for transferring AI to veterinary pathology, accelerating veterinary medicine and comparative oncology.
Publisher
Nature Communications
Published On
Apr 26, 2023
Authors
Khalid AbdulJabbar, Simon P. Castillo, Katherine Hughes, Hannah Davidson, Amy M. Boddy, Lisa M. Abegglen, Lucia Minol, Selina Issuch, Elizabeth P. Murchison, Trevor A. Graham, Simon Spilo, Carlo C. Maley, Luca Aresu, Chiara Palmieri, Yinyin Yuan
Tags
digital pathology
cancer
comparative oncology
artificial intelligence
veterinary pathology
cell classification
prognosis
Related Publications
Explore these studies to deepen your understanding of the subject.