Utility of Brock model for risk stratification in patients with high-risk pulmonary nodules: A single center study

e18824 Background: Owing to the widespread implementation of low-dose CT screening (LDCT), an increasing number of pulmonary nodules are being identified. Per LDCT lung cancer screening protocol, radiologists assign Lung-RADS (LR) scores ranging from 1-4 with L1-2 being at low risk and L3-4 being at intermediate to high risk of malignancy. The Brock model derived from the Pan-Canadian Early Detection of Lung Cancer screening study is a mathematical model that incorporates both clinical (age, sex, family history of lung cancer) and radiological (emphysema, nodule size, location of nodule in the upper lobe, nodule type, nodule count and spiculation) factors to predict the risk of lung cancer, but is not commonly used in the United States. The goal of our study was to estimate whether a higher Brock score can be used to further stratify malignancy potential in patients with LR≥3. Methods: We performed a retrospective analysis at Mercy Catholic Medical Center, where we reviewed LDCT findings of 1090 patients, performed between 1/1/2018 - 6/30/2021 and identified 82 patients with LR≥3. Brock Model was used to calculate the malignancy probability of these pulmonary nodules. All patients were followed until biopsy, resolution of nodule or stability on follow-up imaging. In patients with multiple pulmonary nodules, probability was estimated for the largest nodule. An analysis of area under the receiver operating characteristic curve (AUC) was performed to evaluate efficiency of Brock model in our study population (using GraphPad Prism). We used the British Thoracic Society’s suggested threshold of 10% malignancy risk. Results: 82 patients (43 females and 39 males) were found to have pulmonary nodules with LR≥3. Among these, 10 patients were lost to follow-up and excluded from this study. Out of the 38 patients with LR 4, 16 patients had biopsy-proven lung cancer, 22 patients were found to have benign nodules. Among 34 patients with LR 3, only 3 were found to have lung cancer on follow-up. The mean calculated malignancy risk percentage was 6.69% for benign nodules (n = 51), but 27.14% for malignant nodules (n = 18). The calculated AUC in our study population was 0.85 (95%, CI 0.75-0.95, p < 0.0001). At 10% threshold, sensitivity and specificity were 78.95% and 83.02%, respectively. Conclusions: Based on our data, this model can be used to further assess risk of lung cancer in patients with intermediate to high-risk lung nodules (LR≥3). Similar to the use of Tyrer-Cuzick and Gail models in assessing risk of breast cancer, we may be able to use the Brock Model in predicting malignancy risk of lung nodules. However, a larger study is needed to estimate specific cut-offs. This can potentially lower the need for biopsies in lower risk groups and warrant more aggressive work-up and early diagnosis in higher risk populations.