Automated detection of lung nodules and coronary artery calcium using artificial intelligence on low-dose CT scans for lung cancer screening: accuracy and prognostic value

Abstract

Background Artificial intelligence (AI) in diagnostic radiology is undergoing rapid development. Its potential utility to improve diagnostic performance for cardiopulmonary events is widely recognized, but the accuracy and precision have yet to be demonstrated in the context of current screening modalities. Here, we present findings on the performance of an AI convolutional neural network (CNN) prototype (AI-RAD Companion, Siemens Healthineers) that automatically detects pulmonary nodules and quantifies coronary artery calcium volume (CACV) on low-dose chest CT (LDCT), and compare results to expert radiologists. We also correlate AI findings with adverse cardiopulmonary outcomes in a retrospective cohort of 117 patients who underwent LDCT. Methods A total of 117 patients were enrolled in this study. Two CNNs were used to identify lung nodules and CACV on LDCT scans. All subjects were used for lung nodule analysis, and 96 subjects met the criteria for coronary artery calcium volume analysis. Interobserver concordance was measured using ICC and Cohen’s kappa. Multivariate logistic regression and partial least squares regression were used for outcomes analysis. Results Agreement of the AI findings with experts was excellent (CACV ICC = 0.904, lung nodules Cohen’s kappa = 0.846) with high sensitivity and specificity (CACV: sensitivity = .929, specificity = .960; lung nodules: sensitivity = 1, specificity = 0.708). The AI findings improved the prediction of major cardiopulmonary outcomes at 1-year follow-up including major adverse cardiac events and lung cancer (AUCMACE = 0.911, AUCLung Cancer = 0.942). Conclusion We conclude the AI prototype rapidly and accurately identifies significant risk factors for cardiopulmonary disease on standard screening low-dose chest CT. This information can be used to improve diagnostic ability, facilitate intervention, improve morbidity and mortality, and decrease healthcare costs. There is also potential application in countries with limited numbers of cardiothoracic radiologists.

Radiation-Induced Lung Injury—Current Perspectives and Management

“…Artificial intelligence or machine-based learning algorithms have been evaluated recently for lung nodule and coronary artery calcium detection [ 81 , 82 ]. Giordano et al studied the performance of a machine-based learning algorithm in distinguishing radiation pneumonitis from COVID-19 pneumonia.…”

Computed tomography lung-cancer screening: does iterative reconstruction matter?

“…A severe limitation to traditional CAD approaches is an inability to acquire knowledge from new information. Thus, various research groups investigate deep learning ML approaches (23)(24)(25). In a proof-of-concept study reported in (25), the potential of deep learning ML software system has been investigated with the conclusion that results strongly agree with expert radiologist determination of lung nodule detection.…”

“…Thus, various research groups investigate deep learning ML approaches (23)(24)(25). In a proof-of-concept study reported in (25), the potential of deep learning ML software system has been investigated with the conclusion that results strongly agree with expert radiologist determination of lung nodule detection. Diagnosis of lung nodules on a per-nodule basis is highly sensitive, but poorly specific, with false-positive rates like those of radiologists.…”

“…The effective dose and convolution kernels effects on the detection of pulmonary nodules in IR were investigated in studies with anthropomorphic lung phantoms by quantifying the perception of expert radiologists' at a Likert scale, contrast-to-noise and signal-to-noise ratio (SNR) (19)(20)(21). Other studies evaluated machine learning (ML), computer-aided detection (CAD) software systems or commercially available deep learning ML based CAD systems by investigating receiver-operating-curve measures in artificial or ex-vivo lung phantoms (22)(23)(24)(25). In the future ML algorithms might enable risk assessment to determine the indication for lung scans in clinical routine and even define CT imaging parameters.…”

Application of AI and IoT in Clinical Medicine: Summary and Challenges

“…In the field of medical imaging, deep learning algorithms have been widely used in different modalities, such as CT, MRI, PET, and ultrasound, as well as the application of tumor detection, segmentation, disease prediction, etc [15] . As a result, excellent results have been achieved in the fields of lung nodule detection [16] , diabetic retinopathy (DR) screening [17] , and the accurate diagnosis and treatment of cardiovascular diseases [18] . This paper summarizes the relevant technical basis of AI and IoT in terms of their applications in clinical medicine, analyzes the main challenges thereof, and discussed various ideas and opinions for future research.…”