Harnessing Artificial Intelligence for Risk Stratification and Outcome Prediction in Urologic Cancers: A Systematic Review
Bibliographic record
Abstract
BACKGROUND AND OBJECTIVE: Digital pathology-based artificial intelligence (DP-AI) biomarkers are emerging as transformative tools to guide clinical management of patients affected by various malignancies. We aimed to synthesise current evidence regarding their prognostic and predictive utility in urologic cancers. METHODS: In this prospectively registered systematic review (PROSPERO: CRD420251036536), we searched MEDLINE, Embase, and Web of Science in April 2025 for studies evaluating the prognostic and predictive values of DP-AI models in patients with prostate (PCa), bladder (BCa), renal cell (RCC), testicular (TCa), or penile (PeCa) cancer. The risk of bias was assessed using the Risk of Bias in Non-randomized Studies of Interventions (ROBINS-I) tool. Results were tabulated and summarised qualitatively. KEY FINDINGS AND LIMITATIONS: Of the 1537 screened individual records, we included 31 studies validating DP-AI models in 21 155 patients. Nineteen studies were conducted in PCa (n = 17 541), six in BCa (n = 2349), five in RCC (n = 1176), and one in TCa (n = 89) patients. Ten PCa studies (n = 8951) utilised the ArteraAI model, including two (n = 2786) showing that it allows identification of patients treated with radiotherapy for clinically localised PCa that can safely omit short-term (subdistribution hazard ratio [sHR] 0.34; 95% confidence interval [CI]: 0.19-0.63) or long-term (sHR 0.55; 95% CI: 0.41-0.73) androgen deprivation therapy. Two studies (n = 894) developed and validated a model allowing identification of patients with non-muscle-invasive BCa poorly responding to Bacillus Calmette-Guérin (HR 2.3; 95% CI: 1.9-2.8), including one study (n = 253) validating a predictive biomarker for patients who may benefit from upfront gemcitabine/docetaxel. Many DP-AI models showed a prognostic association in localised PCa (n = 16 863), metastatic PCa (n = 678), non-muscle-invasive BCa (n = 2069), muscle-invasive BCa (n = 280), localised RCC (n = 1176), and germline TCa (n = 89) settings. None of the included studies assessed DP-AI models prospectively. CONCLUSIONS AND CLINICAL IMPLICATIONS: DP-AI biomarkers hold promise to improve treatment personalisation through integration into clinical practice. Prospective validation is now required.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.005 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.002 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".