Morphologic Severity of Atypia Is Predictive of Lung Cancer Diagnosis

SIMPLE SUMMARY: Lung cancer patients are treated by multidisciplinary care teams which rely upon the primary care provider to coordinate care and explain next steps to patients. In patients with nodule sampling that results in cytologic atypia, it is important that the primary care provider has a thorough understanding of the implications of this diagnosis so that they can advise the patient of appropriate follow up. This study investigated the correlation of atypia severity with diagnosis of lung cancer or benign respiratory process. We found that atypia severity, smoking pack years, and modified Herder score independently predicted cancer diagnosis. Patients with severe atypia may benefit from repeat sampling of lung nodules for cytologic confirmation within one month due to their high likelihood of malignancy, while those with less severe atypia may be followed clinically. ABSTRACT: In cytologic analysis of lung nodules, specimens classified as atypia cannot be definitively diagnosed as benign or malignant. Atypia patients are typically subject to additional procedures to obtain repeat samples, thus delaying diagnosis. We evaluate morphologic categories predictive of lung cancer in atypia patients. This retrospective study stratified patients evaluated for primary lung nodules based on cytologic diagnoses. Atypia patients were further stratified based on the most severe verbiage used to describe the atypical cytology. Logistic regressions and receiver operator characteristic curves were performed. Of 129 patients with cytologic atypia, 62.8% later had cytologically or histologically confirmed lung cancer and 37.2% had benign respiratory processes. Atypia severity significantly predicted final diagnosis even while controlling for pack years and modified Herder score (p = 0.012). Pack years, atypia severity, and modified Herder score predicted final diagnosis independently and while adjusting for covariates (all p < 0.001). This model generated a significantly improved area under the curve compared to pack years, atypia severity, and modified Herder score (all p < 0.001) alone. Patients with severe atypia may benefit from repeat sampling for cytologic confirmation within one month due to high likelihood of malignancy, while those with milder atypia may be followed clinically.