Skip to main content
. Author manuscript; available in PMC: 2023 Apr 21.
Published in final edited form as: Pattern Recognit. 2022 Jul 22;132:108919. doi: 10.1016/j.patcog.2022.108919

Table 2.

AUC values of using the four priors in LRP-NET and of using a single prior in two settings: Setting 1: LRP-NET with Single-Prior Input and Setting 2: Single-Prior Model. Here prior-1, prior-2, prior-3, and prior-4 denote from the most-recent to the most-distant examination relative to the current examination.

Model CC view AUC MLO view AUC MLO+CC view AUC
LRP-NET (prior-4 → prior-3 → prior-2 → prior-1) 0.57 0.60 0.67 (referent)
Setting 1:
LRP-NET with Single-Prior Input (prior-1) 0.56 0.55 0.61 (p = 0.05)
LRP-NET with Single-Prior Input (prior-2) 0.53 0.55 0.60 (p = 0.05)
LRP-NET with Single-Prior Input (prior-3) 0.55 0.57 0.59 (p = 0.03)
LRP-NET with Single-Prior Input (prior-4) 0.54 0.54 0.59 (p = 0.03)
Setting 2 [4]:
Single-Prior Model (prior-1) 0.61 0.52 0.60 (p = 0.04)
Single-Prior Model (prior-2) 0.53 0.57 0.59 (p = 0.02)
Single-Prior Model (prior-3) 0.53 0.59 0.53 (p = 0.01)
Single-Prior Model (prior-4) 0.54 0.60 0.54 (p = 0.04)