Lung cancer prediction by Deep Learning to identify benign lung nodules

TLDR

Deep learning is emerging as a promising tool for classifying malignant lung nodules. The study aimed to retrospectively validate the LCP‑CNN on an independent European multicentre dataset to rule out benign nodules while preserving high lung‑cancer sensitivity. The LCP‑CNN assigns a malignancy score to each CT nodule, and a rule‑out threshold was set to achieve at least 99 % sensitivity on NLST data, then applied to 2106 nodules from three European centres. The model achieved an AUC of 94.5 % and, at 99 % sensitivity, ruled out malignancy in 22.1 % of nodules, enabling 18.5 % of patients to avoid follow‑up scans, with only two false negatives (small typical carcinoids).

Abstract

IntroductionDeep Learning has been proposed as promising tool to classify malignant nodules. Our aim was to retrospectively validate our Lung Cancer Prediction Convolutional Neural Network (LCP-CNN), which was trained on US screening data, on an independent dataset of indeterminate nodules in an European multicentre trial, to rule out benign nodules maintaining a high lung cancer sensitivity.MethodsThe LCP-CNN has been trained to generate a malignancy score for each nodule using CT data from the U.S. National Lung Screening Trial (NLST), and validated on CT scans containing 2106 nodules (205 lung cancers) detected in patients from from the Early Lung Cancer Diagnosis Using Artificial Intelligence and Big Data (LUCINDA) study, recruited from three tertiary referral centers in the UK, Germany and Netherlands. We pre-defined a benign nodule rule-out test, to identify benign nodules whilst maintaining a high sensitivity, by calculating thresholds on the malignancy score that achieve at least 99 % sensitivity on the NLST data. Overall performance per validation site was evaluated using Area-Under-the-ROC-Curve analysis (AUC).ResultsThe overall AUC across the European centers was 94.5 % (95 %CI 92.6–96.1). With a high sensitivity of 99.0 %, malignancy could be ruled out in 22.1 % of the nodules, enabling 18.5 % of the patients to avoid follow-up scans. The two false-negative results both represented small typical carcinoids.ConclusionThe LCP-CNN, trained on participants with lung nodules from the US NLST dataset, showed excellent performance on identification of benign lung nodules in a multi-center external dataset, ruling out malignancy with high accuracy in about one fifth of the patients with 5−15 mm nodules.