Identification of Nonaggressive Pulmonary Nodules Using an Optimized Scoring System

Area of Science:

• Pulmonary medicine
• Radiology
• Medical imaging analysis

Background:

• Pulmonary nodules are common findings on computed tomography (CT) scans.
• Distinguishing benign intrapulmonary lymph nodes from malignant nodules is crucial for appropriate patient management.
• Current diagnostic criteria may lead to overutilization of CT scans.

Purpose of the Study:

• To define an optimal scoring method for accurately identifying benign intrapulmonary lymph nodes.
• To compare the performance of a simple scoring strategy versus an optimized strategy.

Main Methods:

• Retrospective analysis of 765 pulmonary nodules from 377 patients in the COPDGene study.
• Comparison of nodule characteristics between 636 benign nodules and 51 nodules adjacent to malignancies.
• Two radiologists scored nodules using a simple and an optimized scoring strategy.

Main Results:

• Nodules exhibiting the majority of lymph node characteristics were predominantly benign.
• The optimized scoring strategy significantly outperformed the simple strategy (AUC 0.80 vs. 0.55).
• A simulation suggested potential reductions in CT utilization for Fleischner criteria (36%) and LUNG-RADS (5%).

Conclusions:

• Pulmonary nodules with most lymph node characteristics are likely benign, irrespective of location.
• Weighting specific lymph node characteristics improves diagnostic performance.
• Incorporating lymph node characteristics into follow-up recommendations can reduce unnecessary CT scans.