PulseExploreJournal ClubResearchersJournals
Instagram
HomeJournal ClubExplore
Synapse
⌘+K
Synapse
February 28, 2026PLOS Digital HealthOpen Access

A novel statistical framework for quantifying risks and benefits of AI automation in screening mammography

View Full Paper
Ask AI
Bookmark
Share

Authors

SGSonia GiebelMCMaggie ChungAYAdam Yala

Discussion

Loading...

Member takes

Overview

This retrospective study quantifies risks and benefits of AI thresholds in screening mammography, suggesting optimal usage in practice.

Key Points

  • The aim is to establish a statistical framework to determine the optimal AI rule-out threshold for screening mammography.
  • Conducted a retrospective analysis of 114,229 mammograms from 2006-2023.
  • Applied an AI scoring system using the Mirai deep-learning model to assign risk scores.
  • Examined metrics using two different thresholds for ruling out cases: 0.2 and 0.05.
  • At the 0.20 threshold, the Caseload Reduction Rate (CRR) was 75%, with G-FOR of 0.26% and 121 missed cancer cases.
  • At the 0.05 threshold, CRR was 36%, with G-FOR of 0.12% and no additional missed cancer cases.

Cite This Study

Giebel et al. (2026) studied this question.

synapsesocial.com/papers/69a286b80a974eb0d3c01da6https://doi.org/10.1371/journal.pdig.0001231
View Full Paper
Ask AI
Bookmark
Share