Background: Lung nodules caused by mycobacteria can resemble lung cancer on chest imaging. The advent of lung cancer screening with low-dose Computed Tomography is accompanied by high false-positive rates, making it necessary to establish criteria to differentiate malignant from benign nodules.
Methods: We conducted a retrospective case-control study of 52 patients with mycobacterial lung nodules and 139 patients with lung cancer, diagnosed between 2010 and 2012. We compared clinical and radiographic characteristics to identify predictors of disease by univariate and multivariate analysis. The discriminatory power of maximum Standardized Uptake Values from Positron-Emission-Tomography was also evaluated.
Results: Several variables were correlated with a diagnosis of mycobacterial infection or lung cancer on univariate analysis. Such variable include smoking status and history, lesion size and imaging evidence of tree-in-bud opacities, lymphadenopathy or emphysema on computed tomography. Upon author consensus, the most clinically-relevant variables were selected to undergo multivariate analysis. A history of current or former smoking [OR 4.4 (95 % CI 1.2-15.6) and 2.7 (95 % CI 1.1-6.8), respectively P = 0.04] was correlated with diagnoses of lung cancer. Contrarily, the presence of tree-in-bud opacities was less likely to be correlated with a diagnosis of malignancy [OR 0.04 (95 % CI 0.0-1.0), P = 0.05]. Additionally, higher maximum standardized uptake values from positron emission tomography were associated with malignancy on multivariate analysis [OR 1.1 (95 % CI 1.0-1.2), P = 0.04]; but the accuracy of the values in differentiating between diseases was only 0.67 as measured by the area under the curve. Lesion size was not independently associated with diagnosis [OR 0.5 (95 % CI 0.2-1.2), (P = 0.12)].
Conclusions: Establishing the likelihood of malignancy for lung nodules based on isolated clinical or radiographic criteria is difficult. Using the variables found in this study may allow clinicians to stratify patients into groups of high and low risk for malignancy, and therefore establish efficient diagnostic strategies.