Advancing Patient Care in Oncology: Integrating Multiscale Transcriptomics for Sarcoma Classification, and Beyond


Human Health


Cancer is responsible for more than 1 in 4 deaths in Canada, with more than 600 new cases diagnosed daily. Sarcomas – tumours of the bone and soft tissue – are the most challenging cancers to diagnose. The many sarcoma types all have intrinsically different molecular pathogenesis (the process by which a disease develops). Patients with sarcomas, which are proportionately more common in children, face delays of weeks to months until they can be referred to a specialist centre and there are few clinical trials. Current histomorphology and immunohistochemistry approaches to diagnosis are also extremely subjective, requiring clinicians to order 10-20 tests per patient. These challenges lead to ultimately higher health system costs and lower patient survival rates.

Pathologists need a comprehensive approach with better tools to diagnose sarcoma. Project researchers have recently developed a platform to accurately diagnose (with 85-95% accuracy) any sarcoma using its ribonucleic acid (RNA). The highly scalable RNA-Seq-based tumour classification system has been trained on >13,000 tumours and normal samples, and improves with every sample analyzed. This project will validate and implement the platform at two major Toronto hospitals, which together treat around 1,000 patients with tumours of soft tissue and bone each year. The team will implement the initial web platform and will work with two commercial partners: DNAstack to expand the platform to the cloud; and Illumina to expand access to this platform outside of Ontario. It will also compare the platform to World Health Organization classifications to support future global adoption of the platform. In 3-5 years, the platform will be expanded to other types of cancer and altogether better streamline the diagnosis of sarcomas of cancer patients.