Researchers at UCLA and in Canada have developed a novel biopsy-driven genomic signature for prostate cancer prognosis.
BACKGROUND:
Prostate cancer is the second most common cause of male cancer deaths in North America. While most prostate cancers are indolent and are unlikely to cause, about 30% of men will experience an aggressive cancer that will recur and become lethal. Predicting which patients are likely to have a relapse is critical in determining whether patients will need additional (adjuvant) treatments after initial treatment. Despite several attempts to predict recurrence rate using simple clinical tests like pre-treatment prostate-specific antigen, the disease variability between patients and the spatial variability within each individual tumor has made these tests inaccurate. A pre-treatment, biopsy-based genomic signature of a patient’s tumor could help decide on treatment type and help predict the rate of recurrence among prostate cancer patients.
INNOVATION:
Researchers at UCLA developed a novel machine learning algorithm that can determine the risk of cancer recurrence in a patient based on information from a standard biopsy. This program works by comparing the genomic profile of a patient’s tumor to the genomic profiles of previous prostate cancer patients with recurring or non-recurring cancer. Through calculations comparing the similarity of the tumor’s profiles to those two groups, a prediction can be made on the likelihood that the patient will have recurring cancer.
POTENTIAL APPLICATIONS:
• Identifying patients with aggressive prostate cancer who need additional therapy
• Identifying patients with indolent prostate cancer who can safely watch their tumour without treatment
ADVANTAGES:
• Use of a tumor’s genomic profile to predict recurrence rates
• Highly accurate
• Uses samples routinely collected in clinical workflows.
DEVELOPMENT-TO-DATE:
A machine-learning program has been trained on a set of 100+ prostate cancer patients and validated on three datasets comprising over 500 patients.
RELATED PAPERS:
Böttcher, R. et al. Cribriform and intraductal prostate cancer are associated with increased genomic instability and distinct genomic alterations. BMC Cancer 18, 8 (2018).
Fraser, M. et al. Genomic hallmarks of localized, non-indolent prostate cancer. Nature 541, 359–364 (2017).
Lalonde E. et. al., Tumour genomic and microenvironmental heterogeneity for integrated prediction of 5-year biochemical recurrence of prostate cancer: a retrospective cohort study. Lancet Oncology 15, 13(2014).
Lalonde E. et. al., Translating a Prognostic DNA Genomic Classifier into the Clinic: Retrospective Validation in 563 Localized Prostate Tumors, European Urology 72, 1 (2017).