Diagnostic performance of artificial intelligence approved for adults for the interpretation of pediatric chest radiographs

Abstract

Artificial intelligence (AI) applied to pediatric chest radiographs are yet scarce. This study evaluated whether AI-based software developed for adult chest radiographs can be used for pediatric chest radiographs. Pediatric patients (≤ 18 years old) who underwent chest radiographs from March to May 2021 were included retrospectively. An AI-based lesion detection software assessed the presence of nodules, consolidation, fibrosis, atelectasis, cardiomegaly, pleural effusion, pneumothorax, and pneumoperitoneum. Using the pediatric radiologist's results as standard reference, we assessed the diagnostic performance of the software. For the total 2273 chest radiographs, the AI-based software showed a sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy of 67.2%, 91.1%, 57.7%, 93.9%, and 87.5%, respectively. Age was a significant factor for incorrect results (odds radio 0.821, 95% confidence interval 0.791-0.851). When we excluded cardiomegaly and children 2 years old or younger, sensitivity, specificity, PPV, NPV and accuracy significantly increased (86.4%, 97.9%, 79.7%, 98.7% and 96.9%, respectively, all p < 0.001). In conclusion, AI-based software developed with adult chest radiographs showed diagnostic accuracies up to 96.9% for pediatric chest radiographs when we excluded cardiomegaly and children 2 years old or younger. AI-based lesion detection software needs to be validated in younger children.

Figure 1

Flowcharts of diagnosis including (a) all lesions, (b) excluding cardiomegaly from all lesions, and (c) excluding cardiomegaly and patients ≤ 2 years old.

Figure 2

Age distribution for correct and incorrect diagnoses after excluding cardiomegaly. (a) Box-whisker plot comparing the median, interquartile ranges and entire range of age according to diagnosis. (b) Pie chart depicting the age distribution of the incorrect diagnosis group.

Figure 3

Examples of results analyzed by the AI-based lesion detection software. (a) A 17-month-old boy with pneumonia in the right upper lobe. The software detected consolidation (Csn) with an abnormality score of 91% in the right upper lobe, as marked in the grayscale map. (b) A 3-month-old boy with a cardiothoracic ratio of 50%, within normal range. The software detected cardiomegaly (Cm) with an abnormality score of 56% on the anteroposterior chest radiograph. (c) A 4-month-old girl without remarkable findings on the chest radiograph. The software detected normal thymus as consolidation (Csn) and nodule (Ndl) with an abnormality score of 88%.