Artificial Intelligence-Based Prognostic Model for Urologic Cancers: A SEER-Based Study

Abstract

Background: Prognostication is essential to determine the risk profile of patients with urologic cancers.

Methods: We utilized the SEER national cancer registry database with approximately 2 million patients diagnosed with urologic cancers (penile, testicular, prostate, bladder, ureter, and kidney). The cohort was randomly divided into the development set (90%) and the out-held test set (10%). Modeling algorithms and clinically relevant parameters were utilized for cancer-specific mortality prognosis. The model fitness for the survival estimation was assessed using the differences between the predicted and observed Kaplan-Meier estimates on the out-held test set. The overall concordance index (c-index) score estimated the discriminative accuracy of the survival model on the test set. A simulation study assessed the estimated minimum follow-up duration and time points with the risk stability.

Results: We achieved a well-calibrated prognostic model with an overall c-index score of 0.800 (95% CI: 0.795-0.805) on the representative out-held test set. The simulation study revealed that the suggestions for the follow-up duration covered the minimum duration and differed by the tumor dissemination stages and affected organs. Time points with a high likelihood for risk stability were identifiable.

Conclusions: A personalized temporal survival estimation is feasible using artificial intelligence and has potential application in clinical settings, including surveillance management.

Keywords: artificial intelligence; data-driven solution; machine learning; surveillance management; survival modeling; urologic cancers.

Figure 1

Shows how well the predicted Kaplan–Meier (KM) estimates fit to the observed KM estimates even after stratifying by cancers of different organs on the disjointed test set after the model calibration on the training set. A good overlap means a well-fitted model that facilitates the reconstruction of the unseen population’s survival history with cancer diseases. 95% Confidence intervals are visualized as bands around KM curves. For most cancers, we observed a max deviation of 5.5% between predicted and observed survival estimates for times lapse of 26 years after diagnosis. The model predictions become less reliable after 28 years from diagnosis for testicular cancers.

Figure 2

(A) the Kaplan–Meier (KM) Curves for the SEER staging groups and (B) KM curve after stratifying the prognosticated survival probabilities using quantile thresholds (10%, 50%, 90%) defined on the training set. The KM curves form A and B were significantly discriminative (Log-rank p < 0.005). For (A), we calculated Log-rank P using localized, distant, regional, and regional/localized stage groups. Localized and regional prostate cancer cases were combined due to their high 5-year survival rates (nearly 100%) compared to other entities. The KM curves were generated on the test set. 95% Confidence intervals are visualized as bands around KM curves.

Figure 3

Describes the results of the simulation for the follow-up duration assessment and the risk-profile stability of unique clinical conditions in prostate, kidney, and testicular cancers for all simulation cases (A,B) and after stratifying according to the tumor dissemination status (C,D). The lower quartile corresponds with the minimum recommended follow-up duration. The green bars highlight the years more likely have stable cancer-specific mortality risk compared to other years.