From the Journals

AI May Improve Radiologist Reports

Radiologists completed reports 15% faster without reducing diagnostic accuracy,

Share

A prospective cohort study involving 23,960 radiographs at a tertiary academic health system assessed the impact of generative AI on radiology reporting. The study found that radiologists completed reports 15% faster with AI assistance without compromising diagnostic accuracy. Mean documentation time improved to 160 seconds per report, with no significant difference in clinical outcomes or report integrity. The AI system achieved 72.7% sensitivity for identifying clinically significant pneumothorax, demonstrating rapid prioritization capabilities. This research highlights the potential of AI in enhancing radiological efficiency and accuracy.

Original Source(s)

AI May Improve Radiologist Reports

Radiologists completed reports 15% faster without reducing diagnostic accuracy,

  • by Kathryn Wighton

  • August 14, 2025

  • 2 min

Related Content

Leopardlike PET Pattern and Sarcoidlike Reaction

A rare but often favorable immune-related adverse event associated with immune checkpoint inhibitors can mimic cancer progression on imaging and typically resolves with conservative management.

Top 10 States to Practice Radiology

The following list highlights the states that offer the most favorable conditions for radiologists.

How Safe Are FDA-Cleared AI Tools?

“These findings indicate that standardized efficacy, safety, and risk assessment reporting remains inadequate for FDA-cleared AI/ML devices, underscoring the need for dedicated regulatory pathways and robust postmarket surveillance to ensure patient safety."