A new study suggests that an artificial intelligence (AI) model may help doctors better predict survival outcomes for patients with bladder cancer after surgical removal of the bladder.
Researchers published their findings in JMIR Perioperative Medicine after reviewing health records from 370 patients. The study population reflected real-world complexity, as nearly 80% of participants were smokers and 21% had diabetes. Using an algorithm called CatBoost, the team analyzed a wide range of data, from the physical characteristics of the tumor to markers found in routine blood work, to identify which factors had the biggest impact on a patient’s recovery path.
The study found that the cancer stage and tumor type, along with levels of internal inflammation, were the most significant clues to how a patient might recover. By looking at a blood marker called the Systemic Immune-Inflammation Index (SII), researchers noticed that when inflammation levels passed 1,000, the risk of complications tended to rise more noticeably.
Read more about bladder cancer treatment and care
Interestingly, the analysis also revealed an “age paradox,” where younger patients sometimes faced more aggressive forms of the disease. Body weight was another key factor, as patients who maintained a moderate weight fared better than those who were either significantly underweight or overweight.
Identifying high-risk factors is crucial because approximately 50% of patients develop metastases within two years of undergoing a radical cystectomy — the surgical removal of the bladder. “Integrating readily available biomarkers, such as [the Systemic Immune-Inflammation Index], represents a cost-effective approach to improving prognostic tools,” the researchers stated.
Although the model shows promising accuracy, the researchers cautioned that it is not yet a “crystal ball” for individual predictions. Because the average margin of error in survival predictions remains relatively high, the algorithm is currently best suited for organizing patients in clinical trials rather than providing precise day-to-day forecasts. Further studies with larger groups of people are needed before this tool becomes a standard part of clinical decision-making.
Sign up here to get the latest news, perspectives, and information about bladder cancer sent directly to your inbox. Registration is free and only takes a minute.