AI-integrated software solutions have already morphed the diagnostic dimensions for adults. Now, this advanced technology also got a green signal for pediatric practices. The medical health care sector has already used AI-enabled software for adult chest X-rays. However, a recent experiment also affirmed that AI-integrated software solutions can be equally helpful for pediatric diagnostic services.
The Study
The study considered children belonging to the age group of 2 to 18. There were almost 2273 children in this group. Surprisingly, the accuracy of the diagnostic outcomes of the AI-integrated software came out to be around 86% to 96.6%.
Eun-Kyung Kim works at Seoul’s Yonsei University’s Radiology Department in South Korea. Kim and other associates mentioned that pediatric radiology could not use the advanced technologies as severely as adult laterals do. However, a chest radiograph is more significant for children than adults. Therefore, AI-integrated diagnostic methods can be as critical for children as they are for adults.
The software used in this study measured critical pediatric radiograph aspects like nodules, consolidation, cardiomegaly, fibrosis, atelectasis, pleural effusion, pneumoperitoneum and pneumothorax. The team of researchers took the reference of interpretations of an expert pediatric radiologist to verify the accuracy level of the AI-based software’s results.
The results showed a strange trend. Age had a lot to do with the outcomes thus derived. Most of the errors regarding diagnosis (80%) occurred in the case of children under the age of two years.
The Expert’s Opinion
When the researchers excluded the outcomes belonging to young children under the age of two years, the accuracy level of the outcomes increased gradually and steadily. The accuracy percentage varied from 86.4% to 96.9%. The researchers concluded that pneumothorax had the highest accuracy for children over the age of two years.
By accuracy, the experts meant parameters like sensitivity, NPV, PPV, specificity, etc. even though the team of researchers gave a convincing account of AI’s relevance in pediatric radiology practices, they did not recommend its inclusion right away. The team explained that for introducing AI to pediatric practices, they need more data and more detailed research.
Children often respond to external stimuli unusually. An adult’s body responds differently. Therefore, the same technology that is secure for an adult might not be equally safe for a child. Hence, the experts deemed further validation necessary.
The team also affirmed that after further investigation and validation pediatric radiographs can use lesion detection software solutions powered by AI.
The database supporting the impact of such methods on children should be wider. Hence, more introspection into the matter is a necessity, according to the team of researchers. Technology can benefit pediatric radiographs in many ways.
However, a superficial understanding of the influences of such methods on children’s health may lead to undesired outcomes in the long run. Hence, the researchers felt the necessity of further research before concluding.
At sepStream®, we ensure pocket-friendly, accurate, and reliable software solutions for diagnostic practices. Our intuitive software solutions ensure flawless detection of diseases to help radiologists perform efficiently.