Abstract
ABSTRACTAimsThis study assessed an artificial intelligence (AI) model’s performance in predicting elevated brain natriuretic peptide (BNP) levels from chest radiograms and its effect on human diagnostic performance.Methods and resultsPatients who underwent chest radiography and BNP testing on the same day were included. Data were sourced from two hospitals: one for model development, and the other for external testing. Two final ensemble models were developed to predict elevated BNP levels of >= 200 pg/mL and >= 100 pg/mL, respectively. Humans were evaluated to predict elevated BNP levels, followed by the same test, referring to the AI model’s predictions. The 8390 images from 1334 patients were collected for model creation, and 1713 images from 273 patients for tests. The AI model achieved an accuracy of 0.855, precision of 0.873, sensitivity of 0.827, specificity of 0.882, f1 score of 0.850, and receiver-operating-characteristics area-under-curve of 0.929. The accuracy of the testing with the 100 images by 35 participants significantly improved from 0.708±0.049 to 0.829±0.069 (P < 0.001) with the AI assistance (an accuracy of 0.920). Without the AI assistance, the accuracy of the experts was higher than that of non-experts (0.728±0.051 vs. 0.692±0.042, P = 0.030); however, with the AI assistance, the accuracy of the non-experts was rather higher than that of the experts (0.851±0.074 vs. 0.803±0.054, P = 0.033).ConclusionThe AI model can predict elevated BNP levels from chest radiograms and has the potential to improve human performance. The gap in utilizing new tools represents one of the emerging issues.Graphical AbstractWe developed AI models using an ensemble method to predict elevated BNP levels. The AI model achieved a higher accuracy rate than any individual participant. While the accuracy of experts was higher in the non-assisted test, with the AI assistance, the accuracy of non-experts surpassed that of the experts. AI, artificial intelligence; AUC, area-under-curve; BNP, brain natriuretic peptide; GPU, graphic processing unit; PR, precision-recall; ROC, receiver-operating-characteristics.
Publisher
Cold Spring Harbor Laboratory