Table 15 Describes the variables included in each mixed-effects model.

From: Critical evaluation of artificial intelligence as a digital twin of pathologists for prostate cancer pathology

 

Variables

Model (goal)

Fixed effects

Random effect (random intercept)

A

Tumor volume in percentage

Year of the pathology report

Proportion of positive slides

The total number of slides

Gleason grade made by the pathologists

Pathologists

B

Tumor volume in percentage

Year of the pathology report

Proportion of positive slides

The total number of slides

Pathologists

Gleason grade made by the pathologists

C

Tumor volume in percentage

Year of the pathology report

Proportion of positive slides

Gleason grade made by AI

Pathologists

D

Tumor stage (pT)

Locoregional lymph node metastases status (pN)

Cancer proximity to prostate capsule

Gleason grade made by the pathologists/by AI

Tumor volume in percentage

Proportion of positive slides

Pathologists