Dr. Girish Kulkarni was awarded a $100,000 2023 Temerty Innovation Grant for his project to treat bladder cancer.
Current monitoring strategies for the recurrence and progression of bladder cancer are costly, patient unfriendly, and lack supporting evidence. Accurate and timely prediction of these patient outcomes remains a clinically important but unmet need.
This study aims to integrate the complete cancer timeline of each patient with artificial intelligence approaches to improve prognostication for tumour recurrence and progression. This project will markedly enhance personalization of bladder cancer treatment and follow-up surveillance, while alleviating the economic burden of this disease on our healthcare system.
Dr. Girish Kulkarni
Dr. Girish Kulkarni is a urologic surgeon in the Department of Surgical Oncology at the Princess Margaret Cancer Centre, University Health Network. He is also a surgeon-scientist who is affiliated with the Department of Surgery, Faculty of Medicine at the University of Toronto. At the University of Toronto, he is an assistant professor in the Department of Surgery, as well as in the Institute of Health Policy, Management and Education. Dr. Kulkarni's primary research interests revolve around the epidemiology of genitourinary malignancies, particularly prostate and bladder cancer. His investigations are dedicated towards the understanding of population-level quality of care from urologic malignancies, quality of life and the health economics associated with urologic malignancies, as well as determining the efficacy of clinical evaluation and treatment towards prostate and bladder cancer.
Proposal
Bladder cancer is the 5th most common cancer in Canada with over 13,000 newly diagnosed cases each year. It is categorized into non-muscle invasive (NMIBC) and muscle-invasive bladder cancer (MIBC), based on the presence of tumour invasion into the bladder muscularis propria. Most newly diagnosed cases are NMIBC, which typically confers a favourable prognosis. However, 70% of these patients will experience cancer recurrence and up to 20% will progress to MIBC.
Patients who progress to these higher stages of disease have worse cancer-specific survival and often require major surgery, which is associated with significant morbidity and mortality. Given the high recurrence and progression rates, patients require lifelong cystoscopic surveillance – making bladder cancer the most expensive cancer to treat per patient, estimated at almost $4 billion per year in the United States. However, current surveillance strategies are too frequent, patient unfriendly, costly, and lack supporting evidence. Therefore, tools to facilitate accurate and timely prediction of NMIBC recurrence and progression remain a clinically important but unmet need.
This study expands on our previous work by applying current best practices in AI to develop NIMBLE (Non-Muscle Invasive Bladder Cancer Longitudinal Evaluation). The objectives of this study include:
1. Develop NIMBLE: a time-series forecasting model uniquely modeling longitudinal data to predict NMIBC recurrence and progression.
2. Estimate the potential clinical benefit and cost-savings by using NIMBLE instead of the current standard of care.
3. Assess for potential biases within NIMBLE with respect to age, sex, ethnicity, and NMIBC risk classification.
4. Deploy NIMBLE as a web application and run a silent trial to prospectively evaluate NIMBLE.
NIMBLE can be integrated into existing clinical workflows and provide a novel data- and AI-driven approach to inform follow-up surveillance and management. Our open-source web-application will allow other institutions to easily and securely deploy NIMBLE within their infrastructure.
