In a groundbreaking study published in Nature Communications, scientists from The Institute of Cancer Research in London have successfully utilized an artificial intelligence (AI) algorithm trained on human lung cancer samples to accurately categorize tumor and immune cells of various non-human animal species. This innovative research, which marks the first of its kind, establishes a vital link between human clinical care and wildlife studies, offering valuable insights into the development and progression of cancer in humans. By leveraging AI technology, this study paves the way for new possibilities in human cancer treatment by enhancing our understanding of evolutionary similarities and differences across species.
Cancer affects both humans and other animals, and the field of comparative oncology holds immense potential for advancing cancer research. By comprehending the reasons behind certain species evolving highly effective anti-cancer mechanisms, we can gain valuable knowledge in cancer biology. This understanding may reveal similar mechanisms that can be exploited for developing novel treatments for human cancer.
The researchers trained their algorithm using human lung tumor samples and employed it to identify and classify individual lymphocytes, stromal cells, and tumor cells in 120 tumor samples sourced from 20 different animal species. These samples were obtained from various institutions, including the Zoological Society of London (which oversees London Zoo and Whipsnade Zoo), the University of Cambridge, and two international collaborators from the University of Turin in Italy and the University of Queensland in Australia.
To validate the algorithm’s performance, the scientists compared its predictions with 14,570 annotations made by two expert veterinary pathologists. The algorithm demonstrated impressive accuracy and reliability, underscoring its potential as a valuable tool for characterizing tumor and immune cells across different animal species.
Across-species cancer study
The remarkable algorithm achieved impressive results in distinguishing cancer cells among the tested species, exhibiting a remarkable accuracy rate. Notably, it achieved a staggering 94% accuracy in identifying canine transmissible venereal tumor and an 88% accuracy in detecting Tasmanian devil facial tumor disease.
The AI algorithm leveraged the similarities shared by these cancer cells with human lung cancer, including factors such as nuclei shape, size, and coloration. These common features enabled the algorithm to accurately identify cancer cells in various species.
A deeper comprehension of the shared characteristics and disparities in cancer tumors across species holds the potential to shed light on the fundamental processes involved in cancer development and evolution.
Moreover, understanding cancer in animals and its resemblance to human cancer is pivotal for selecting appropriate animal models in the study of human diseases and for preclinical drug development.
The researchers intend to expand the application of their algorithm to a wider range of species. Their aim is to unravel why certain species possess superior mechanisms for suppressing cancer development compared to others. By gaining insights into these mechanisms, they hope to enhance our understanding of cancer treatment in humans.
Answering fundamental questions in cancer biology
Dr. Khalid Abdul Jabbar, one of the study’s co-leads and a former Postdoctoral Training Fellow in Computational Pathology and Integrative Genomics at The Institute of Cancer Research London, expressed that despite training the deep learning model on human samples, it successfully distinguished three major cell types with remarkable accuracy across numerous animal species. This achievement was made possible by the shared characteristics of these cell types throughout the animal kingdom.
Dr. Jabbar further expressed optimism that their model would be instrumental in identifying additional similarities and differences across species, thereby contributing to the exploration of fundamental questions in cancer biology. By improving our understanding of how cancer develops and evolves, this research has the potential to drive significant advancements in the field.
Another co-lead of the study, Dr. Simon Castillo, a Postdoctoral Training Fellow in Computational Pathology and Integrative Genomics at The Institute of Cancer Research in London, emphasized the occurrence of cancer in both human and non-human animal species. He emphasized the importance of considering animals in our comprehensive understanding and treatment of cancer. For instance, understanding cancer in animals like mice, which are commonly used in research, plays a critical role in selecting appropriate experimental models for studying human diseases and in the development of preclinical drugs.
Dr. Castillo also highlighted the significance of cross-species comparisons in addressing fundamental questions in cancer biology and evolution. These insights can be translated into innovative therapeutic approaches for human cancer treatment, opening new avenues for improved outcomes.
‘One Health’ cancer approach
Professor Yinyin Yuan, the Director of the Computational Pathology Research Program at The University of Texas MD Anderson Cancer Center, and former Team Leader in Computational Pathology and Integrative Genomics at the Institute of Cancer Research in London, emphasized the interconnectedness of human health, animal health, and the environment. She highlighted the relevance of this research in bridging the gap between human clinical care and wildlife care, advocating for a “One Health” approach to healthcare.
Professor Yuan highlighted that this study has broader implications beyond cancer research. It opens up avenues for understanding key evolutionary questions, such as why certain species have higher cancer rates than others. By embracing a holistic perspective, this research has significant potential to advance human healthcare, improve animal welfare, and contribute to wildlife conservation efforts. This interdisciplinary approach will ultimately benefit multiple species and ecosystems.
Source: Institute of Cancer Research