Summary
Predicting how lesions seen on CT scans change over time can be hard when only a few scan results are available. The authors studied a dataset of CT scans from patients and tested several methods to predict lesion growth. They found that having two prior scans helps, but more scans do not improve predictions much. Also, different prediction methods varied in how well they estimated uncertainty, and some predictions struggled depending on lesion location. This work suggests that adding more historical scans alone isn’t enough to improve forecast accuracy or reliability in sparse data situations.
What this means in practice
- •For hospital data teams: Improve CT lesion tracking by choosing better forecasting methods for sparse follow-up data based on uncertainty reliability and accuracy.
- •For medical imaging software developers: Incorporate calibrated prediction intervals in lesion size forecasts to better inform clinicians with reliable uncertainty estimates in sparse longitudinal CT data.
Tested on one dataset.
Abstract
Sparse longitudinal CT follow-up limits lesion-size forecasting when only a few prior observations are available. We constructed a five-visit DLT-derived same-lesion trajectory benchmark from DeepLesion and Deep Lesion Tracker (DLT), yielding 205 trajectories from 129 patients. We compared an exploratory conventional sparse-to-final analysis with a primary fixed visit-index horizon design predicting the common log change from T3 to T4 while progressively adding earlier observations, evaluating predictive accuracy, uncertainty reliability, post-hoc conformal interval calibration, subgroup performance, and Gompertz-inspired trajectory regularization. The evaluated methods showed partially overlapping point-prediction accuracy but distinct uncertainty behavior. Mean held-out RMSE across ten training seeds was 0.4726, 0.4305, 0.4499, and 0.4513 for m = 1, 2, 3, 4, indicating the lowest mean RMSE at m = 2; additional history did not improve RMSE. At m = 4, raw Cohort-Level Feature GP coverage was near the 95% nominal level, whereas MC Dropout, Deep Ensemble, and residual-scale intervals were conservative. Patient-level conformal calibration generally produced near-nominal or conservative coverage at the cost of wider intervals. Patient-grouped development cross-validation selected lambda* = 0 for the Gompertz-inspired term. A global population reference frequently opposed lesion-level change directions, and prediction difficulty varied across anatomical subgroups. Overall, additional historical observations provided limited predictive benefit once the prediction horizon was controlled, while predictive accuracy, uncertainty reliability, and trajectory consistency did not necessarily improve together, and should be evaluated jointly in sparse longitudinal imaging.