Eric Wolsztynski, Insight at UCC

Prostate cancer is among the most common cancers for men globally. One in seven men in Ireland is affected. The primary treatment method is Radical Prostatectomy (RP). RP fails in 20 to 40 per cent of patients, who subsequently develop biochemical recurrence (BCR). Accurate time-to-BCR prediction is critical for patient treatment decision making. The goal of this research is to identify and use mRNA markers to increase predictive performance in determining BCR-free survival before prostatectomy. The researchers at Insight at UCC and UL examined different modelling approaches, some employing mRNA, some employing clinical variables, some a combination of both.

No statistically significant difference was found between any of the clinical only models. Machine learning models outperformed traditional models when mRNA variables were incorporated. The conclusion? mRNA information should be considered before performing a Radical Prostatectomy to adjudicate on the likely success of the procedure in individual patients, or flag the potential need to consider alternative or complementary treatment options earlier. This personalised approach to treatment is just one of the supporting functions that AI can bring to cancer therapy.

Researchers: Eric Wolsztynski, Autumn O’Donnell, Micheal Cronin, Shirin Moghaddam