AI-SCOPE: Artificial Intelligence for SarComa Outcome PrEdiction
Research type
Research Study
Full title
An Artificial Intelligence (AI) solution for diagnosing, prognosticating as well as predicting outcome of sarcomas and their mimics: a multi-centre study
IRAS ID
328987
Contact name
Adrienne M Flanagan
Contact email
Sponsor organisation
University College London
Clinicaltrials.gov Identifier
Z6364106/2023/08/89 , UCL Data Protection Registration Number
Duration of Study in the UK
9 years, 11 months, 31 days
Research summary
Cancer treatment is determined by how tumours are classified by pathologists. Recent discoveries have resulted in new treatments and improved survival for different cancer types, highlighting the importance of tumour classification, which encompasses diagnoses, prognoses and predictions of responses to therapies. However, classification is now more complex, requiring ever-more specialist knowledge when there is a growing shortage of pathologists. It's more time consuming and costly; it consumes more of the tumour sample preventing sophisticated tests e.g. whole genome sequencing being undertaken, that can guide personalised treatment. Artificial Intelligence (AI) is a solution to these challenges and can help support (not replace) pathologists as already shown for common cancers (e.g. prostate and bowel).
We aim to develop AI models to classify sarcomas (rare cancers of bone and soft tissue) that provide improved or similar classification performance compared to pathologists. Limited AI research has been conducted on sarcomas because of their rarity (<2% of all cancers; approximately 70 subtypes; involving all age groups).
AI requires large numbers of cases, and pathology slides must be digitised to generate whole-slide images (WSI). The availability of digital scanners across the UK has made this possible. We propose a multi-site project to gather large numbers of sarcoma cases using archived slides/data from the last 40 years and prospectively. Members of direct clinical care teams will gather pathology images/data from up to 50,000 patients. The project will run for 10 years (collecting images/data for 5 years and 5 years follow up). Consent will not be requested. Only members of clinical care teams will access patients’ identifiable data which they will de-identify and pseudonymise before it is shared with UCL researchers for developing tumour classification algorithms. The researchers will request updates on clinical outcomes/ treatment responses from the clinical care teams to ensure that the model is being trained correctly.
REC name
HSC REC B
REC reference
23/NI/0166
Date of REC Opinion
12 Dec 2023
REC opinion
Favourable Opinion