Pulmonary Nodule Characteristics and Workup Rates in a Population of 5.5 Million Patients Across 7 Institutions in New York

Purpose: Annually, 1.6 million patients in the U.S. have a pulmonary nodule (PN) identified, but only about one-third of patients with a PN received recommended workup. Workup depends on radiological risk factors though these features are not standardized in radiology reports. The objective of this study is to describe the differences in radiology reporting and PN workup rates across institutions.

Methods/Materials: We used a retrospective cross-sectional observation design, where structured demographic data and 58 million unstructured radiology reports and clinical notes were used to characterize PNs across seven New York healthcare institutions between 2018-2022. We abstracted malignancy risk factors (e.g. PN size, attenuation, and margins) from radiology reports using a clinical natural language processing application, CLiX unlock. Variations in workup rates across index imaging, PN characteristics, and symptoms were evaluated to assess gaps in PN documentation and management. This study was funded by AstraZeneca.

Results: We analyzed 5.5 million unique patients from 7 institutions in New York state and identified 151,436 patients with PNs. Of these, 40% (60,749) had evidence of a PN in the upper lung and 35% (52,789) had a PN ≥6 mm. Workup rates were 55% (28,958) in patients with PNs ≥6 mm and 43% (42,113) in patients with PNs < 6mm. The overall workup rate in patients with PNs was 47% (71,071), but merely 36% (6,697) in patients with only radiological risk factors. Across institutions, there were notable variations of PN morphology, attenuation and spiculation, though the evidence of PN sizing was more consistent (Figure 1). For the index imaging in PNs ≥6 mm, workup rates for non-chest CT were 49% compared to 58% for chest CT.

Conclusions: We found variations in radiological characteristics between the populations across institutions and workup rates by index imaging type. Findings of this descriptive analysis suggest areas to improve standardization in PN radiology reports and workup. Further research is needed to explore ways to better standardize PN data across institutions and determine how variations in PN management affect patient outcomes.