AI and Cancer Research ── Bladder ── 2026-11-17

Bladder cancer: treatment, trials and AI
- Abstract
BACKGROUND: Predicting BCG non-response in high-risk non-muscle-invasive bladder cancer (NMIBC) remains a genuine challenge in everyday urological practice. BCG engages the host immune system, and variation in individual immune responses may be reflected in measurable pre-treatment markers. OBJECTIVES: To evaluate exploratory machine learning (ML) models predicting BCG non-response at 12 months from routine pre-treatment data. DESIGN: Single-centre retrospective observational cohort study. METHODS: We reviewed 114 patients with NMIBC given intravesical BCG at one centre in Turkey, 2019-2024. By the 2021 European Association of Urology classification, all tumours were high grade and all patients high risk. BCG non-response was defined as histologically confirmed recurrence within a fixed 12-month window after at least five induction instillations, with first surveillance cystoscopy at three months. Fourteen candidate predictors, all available before the first instillation, were used; variables determined during or after treatment, including the number of BCG doses, were excluded. Penalized logistic regression, random forest and support vector machine (SVM) models were compared. All preprocessing was fitted within training folds. Performance was assessed by nested cross-validation and reported as pooled out-of-fold area under the curve (AUC) with 95% bootstrap confidence intervals. Reporting follows STROBE and TRIPOD+AI. RESULTS: Thirty-three patients (28.9%) did not respond; 18 recurrences (54.5%) were high grade. Tumour number (adjusted OR 2.701, 95% CI 1.424-5.122) and tumour size (adjusted OR 1.309 per 10 mm, 95% CI 1.012-1.693) retained an association with non-response. Random forest gave the highest pooled out-of-fold AUC (0.915, 95% CI 0.841-0.974), then SVM (0.874) and penalized logistic regression (0.703). Restricting the outcome to high-grade recurrence reduced random forest discrimination to 0.683-0.754. CONCLUSION: ML models discriminated between BCG responders and non-responders at 12 months, but discrimination was substantially lower for high-grade recurrence. Given the modest sample size and the absence of external validation, these findings should be regarded as hypothesis-generating and require prospective multicentre external validation before routine clinical implementation. Bladder cancer that has not grown into the muscle wall is usually treated by removing the tumour and then washing the bladder out with a medicine called BCG, which works by prompting the immune system to attack any remaining cancer cells. BCG helps many people, but in some patients the cancer comes back despite treatment, and at present doctors cannot tell in advance who those patients will be. We looked back at the records of 114 patients treated at a single hospital in Turkey between 2019 and 2024. All of them had aggressive tumours and all received BCG. We checked whether the cancer came back during the year after treatment finished, using the bladder examinations they had every three months. It came back in 33 of the 114 patients. We then used computer programs that learn patterns from data to see whether information already collected before treatment started — the patient's age, features of the tumour such as its size and how many there were, and results from routine blood tests — could indicate in advance who was likely to have the cancer return. We tested the programs carefully, always checking them on patients whose results they had not been shown, so that the performance we report is not simply the programs repeating what they had memorised. One of the programs separated the two groups reasonably well. However, when we looked only at the more serious returns of cancer, which are the ones that change treatment decisions, the programs performed considerably less well. Almost half of the cancers that came back were of the least aggressive type, which is normally managed by removing the tumour again rather than by more major treatment. Our study is small, comes from a single hospital and has not yet been tested anywhere else, so these programs are not ready to be used in patient care. They do suggest that routinely collected information may hold useful signals, and that this is worth testing in larger studies across several hospitals.
Journal IF-equivalent: 4.1 (OpenAlex 2-year mean citedness, value as of 2026-11-16, retrieved 2026-11-17; not the official Clarivate IF)Reference: Şığva H, Sevim M, Körpe K, Atış V, Arslan V, Görür S, et al. Machine learning–based prediction of BCG NON-RESPONSE in high-risk NON-muscle-invasive bladder cancer: An exploratory single-centre cohort study. Ther Adv Urol. 2026 Sep 28;18:17562872261493691. doi:10.1177/17562872261493691. PMID: 43036513.Checked: Abstract only