Skip to main content
. Author manuscript; available in PMC: 2023 Sep 27.
Published in final edited form as: Cancer Cytopathol. 2023 Jun 26;131(9):561–573. doi: 10.1002/cncy.22725

Figure 1: Study Overview:

Figure 1:

A) UC images are acquired for patients across the study period and are processed using B) AutoParis-X, which extracts imaging predictors, e.g., ABS; C) Imaging predictors are aggregated across collection periods to form fixed predictors which are then used to assess time-to-recurrence using Cox models; D) Imaging predictors were also studied dynamically considering results/extracted features from individual tests and their recurrence potential; risk of recurrence was also studied within specific time periods to demonstrate how the importance of these predictors varies with time; E) Scatterplots for two patients with time from the first positive primary versus the Atypia Burden Score as assessed using AutoParis-X; points were labeled by the UC categories assigned through manual examination of urine cytology