Commercial nodule AI cut chest CT reporting time by about 15%

Design

Retrospective before-and-after cohort, single tertiary centre

Population

39,323 chest CT examinations in 19,433 patients

Primary outcome

Radiology reporting time

Effect

Median 21.3 to 18.2 minutes (−14.6%); aHR 1.17 (1.14 to 1.21)

A Dutch tertiary centre compared chest CT reporting times before and after deploying a commercial pulmonary nodule AI, across 39,323 examinations in 19,433 patients from 2021 to 2024, clustering by radiologist.

Adjusted median reporting time fell from 21.3 to 18.2 minutes (14.6%; HR for faster completion 1.17, 1.14 to 1.21). Thoracic radiologists gained most (25%) and ECG-gated thoracic CT fell 41%. Emergency department examinations took 7.1% longer. At their volumes the authors estimate about half a full-time radiologist saved.

The finding is the gain is uneven. Specialists benefited most; emergency work slowed, perhaps from reviewing AI flags in a setting where nodules are incidental.

• Measure reporting times before and after deploying any AI tool.
• Expect gains to vary by reader and setting.
• Agree how incidental nodules flagged in emergency CT should be handled.
• Report nodule management per Fleischner or local guidance, not AI output.

Why it matters

It gives real-world evidence that AI saves time, and shows where it does not.

Don't overread it

Before-and-after comparison — other workflow changes over three years could contribute.

The statistics, in plain English

A hazard ratio of 1.17 here means reports were completed about 17% faster at any point. Three minutes per scan is modest per case but adds up at 20,000 scans a year. The staffing estimate is modelled, not measured.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.