A study was aimed to establish a prediction model that can predict malignancy in lung nodules in a real-world setting. It was imperative to develop an accurate method to differentiate between malignant and benign solitary pulmonary nodules.
The researchers retrospectively examined the computed tomography (CT) and clinical data of 121 patients with lung nodules. These patients were given percutaneous CT-guided transthoracic biopsies between 2014 and 2015. To evaluate the probability of malignancy, multiple logistic regression formula was utilized. It helped identify independent malignancy predictors and established a clinical prediction model.
A total of 121 patients, out of which 75 (62%) were men with a mean age of 64.7 years. The multiple logistic regression analysis discovered 6 independent malignancy predictors with the area under the curve (AUC) at 0.8573. The 6 predictors were gender, age, current extra-pulmonary cancer, air bronchogram, smoking status, and nodule size (p<0.05).
According to this study, the established prediction model can identify the malignancy probability and help diagnose lung nodules and follow-up interventions.
Reference: sciencedirect.com/science/article/pii/S2531043720301483
