Improving Automating Quality Control in Radiology: Leveraging Large Language Models to Extract Correlative Findings in Radiology and Operative Reports.

Radiology Imaging plays a pivotal role in medical diagnostics, providing clinicians with insights into patient health and guiding the next steps in treatment. The true value of a radiological image lies in the accuracy of its accompanying report. To ensure the reliability of these reports, they are often cross-referenced with operative findings. The conventional method of manually comparing radiology and operative reports is labor-intensive and demands specialized knowledge. This study explores the potential of a Large Language Model (LLM) to simplify the radiology evaluation process by automatically extracting pertinent details from these reports, focusing especially on the shoulder's primary anatomical structures. A fine-tuned LLM identifies mentions of the supraspinatus tendon, infraspinatus tendon, subscapularis tendon, biceps tendon, and glenoid labrum in lengthy radiology and operative documents. Initial findings emphasize the model's capability to pinpoint relevant data, suggesting a transformative approach to the typical evaluation methods in radiology.

Utilization of Renal Mass Biopsy for T1 Renal Lesions across Michigan: Results from MUSIC-KIDNEY, A Statewide Quality Improvement Collaborative.

BACKGROUND: Renal mass biopsy (RMB) has had limited and varied utilization to guide management of renal masses (RM). OBJECTIVE: To evaluate utilization of RMB for newly diagnosed cT1 RMs across diverse practice types and assess associations of outcomes with RMB. DESIGN SETTING AND PARTICIPANTS: MUSIC-KIDNEY commenced data collection in September 2017 for all newly presenting patients with a cT1 RM at 14 diverse practices. Patients were assessed at ≥120 d after initial evaluation. OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS: Demographics and outcomes were compared for patients undergoing RMB versus no RMB. Clinical and demographic characteristics were summarized by RMB status using a χ2 test for categorical variables and Student t test for continuous variables. A mixed-effects logistic regression model was constructed to identify associations with RMB receipt. RESULTS AND LIMITATIONS: RMB was performed in 15.5% (n = 282) of 1808 patients with a cT1 RM. Practice level rates varied from 0% to 100% (p = 0.001), with only five of 14 practices using RMB in >20% of patients. On multivariate analysis, predictors of RMB included greater comorbidity (Charlson comorbidity index ≥2 vs 0: odds ratio [OR] 1.44; p = 0.025) and solid lesion type (cystic vs solid: OR 0.17; p = 0.001; indeterminate vs solid: OR 0.58; p = 0.01). RMB patients were less likely to have benign pathology at intervention (5.0% vs 13.5%; p = 0.01). No radical nephrectomies were performed for patients with benign histology at RMB. The limitations include short follow-up and inclusion of practices with low numbers of RMBs. CONCLUSIONS: Utilization of RMB varied widely across practices. Factors associated with RMB include comorbidities and lesion type. Patients undergoing RMB were less likely to have benign histology at intervention. PATIENT SUMMARY: Current use of biopsy for kidney tumors is low and varies across our collaborative. Biopsy was performed in patients with greater comorbidity (more additional medical conditions) and for solid kidney tumors. Pretreatment biopsy is associated with lower nonmalignant pathology detected at treatment.