Lung Resection Without Tissue Diagnosis: A Pragmatic Perspective on the Indeterminate Pulmonary Nodule

Abstract

Background: The indeterminate pulmonary nodule is a common clinical problem. Preoperative tissue diagnosis is not always possible, despite all attempts. The objectives of this study were to determine the frequency of a malignant diagnosis in this scenario and whether attempted preoperative biopsy impacted estimation of the risk of malignancy.

Patients and methods: We reviewed 500 consecutive cases of pulmonary resection without a preoperative tissue diagnosis at a tertiary care center from 2009 to 2013. Age, sex, smoking status, prior malignancy, tumor size, and whether or not tissue diagnosis had been attempted were recorded. Logistic regression models were constructed to determine factors associated with a malignant diagnosis.

Results: There were 297 males (59.4%), the mean age was 64.9 years, and 412 had a smoking history (82.4%). Also, 203 patients (40.6%) had a malignancy history, and 36 patients (7.2%) had previous lung cancer. Biopsy was attempted for 102 patients (20.5%). The final diagnosis was lung cancer in 336 patients (67.2%), metastatic cancer in 93 patients (18.6%), and benign tumour in 71 patients (14.2%). Male sex, increasing age, smoking history, and prior lung cancer were positive predictors of lung cancer. Model discrimination was good (c-statistic, 0.83). Attempted biopsy did not alter model discrimination.

Conclusion: In this cohort, 86% of resected lesions were malignant. The decision to pursue preoperative tissue diagnosis did not change the predictive ability offered by clinical factors. These findings are reassuring in the scenario when a patient is operable but the diagnosis remains unknown.

Keywords: Lung biopsy; Lung cancer; Lung nodule; Predictive model; Thoracic surgery.