Time saving with artificial intelligence-assisted lumbar spine magnetic resonance imaging reporting: A preliminary study

Introduction: Lumbar spine magnetic resonance imaging (MRI) is a high-volume diagnostic examination, yet increasing caseloads and reporting complexity continue to strain radiology workflows. Emerging artificial intelligence (AI)-assisted reading tools may help streamline interpretation and reduce report turnaround times, but their real-world impact on efficiency remains insufficiently quantified. Objective: To evaluate the impact of an AI-based reading tool on lumbar spine MRI interpretation and reporting time. Methods: We randomly selected 236 lumbar spine MRI examinations performed between 2018 and 2023 in patients aged 18 and older. Cases with prior lumbar surgery or scoliosis were excluded. Digital imaging and communications in medicine (DICOM) data were processed using a commercial deep-learning software package, and outputs were reviewed in a standard DICOM viewer. Five radiologists participated. Studies 1 and 2 assessed the effect of AI on interpretation time using a within-reader design: radiologists interpreted each examination with AI support and then reinterpreted the same examinations 2 months later without AI, enabling direct comparison of interpretation times. Study 3 evaluated the effect of AI by comparing AI-assisted and unassisted interpretations in 146 randomly selected examinations. Results: AI assistance significantly accelerated report generation. Across the full dataset, AI-supported interpretation reduced time by approximately 52% compared with unassisted reading. AI-assisted generation of preliminary reports reduced radiologists’ overall time by nearly 30%. Linear mixed-effects modeling indicated that these reductions were statistically significant. The smaller reduction observed in Study 3 (9.21%) may reflect limited familiarity with the software’s reporting style and occasional instances in which the AI outputs did not fully support the radiologists’ findings. Conclusion: AI assistance improves the efficiency of lumbar spine MRI reporting and shortens reporting time.