medwireNews: An artificial intelligence (AI)-assisted tool can speed up anthropometric calculations for paediatric endocrinologists monitoring growth in children with achondroplasia, research indicates.
“This AI-assisted tool provides a user-friendly, accessible, and highly accurate method for automated growth assessment in pediatric achondroplasia”, say Yael Lebenthal (Tel Aviv Sourasky Medical Center, Israel) and colleagues.
“It facilitates efficient clinical and research applications, with potential for future integration into electronic health records and web-based platforms”, they write in the European Journal of Pediatrics.
The Excel-based tool integrates the European Lambda-Mu-Sigma (LMS) growth reference information for height, weight, BMI, and head circumference at 6-month intervals for patients from birth to age 4 years and at 1-year intervals thereafter up to age 20 years. In addition, data at 1-year intervals were included for sitting height, leg length, arm span and relative sitting height for ages 2–20 years and for foot length from ages 3-20 years.
Large language models were used to extract data from the LMS tables and create complex formulas so that a patient’s individual measurements could be inputted and used to generate their sex- and age-specific z-scores and percentiles in real time.
To test the tool, 10 paediatric endocrinologists were asked to calculate z-scores for three patients with achondroplasia using the conventional manual growth charts and then with the automated tool.
“The tool demonstrated complete agreement with the LMS-based calculations across all anthropometric parameters, including at interpolation points and percentile extremes”, Lebenthal et al report.
The calculations took significantly less time to perform with the AI tool than manually, at an average of 10.1 versus 23.4 minutes.
Moreover, the median absolute z-score deviation across the paediatric endocrinologists in each patient case was significantly lower when they used the tool than when they did the calculations manually. For example, for case 1, the median z-score deviation between the paediatric endocrinologists performing the calculations with the tool was 0.0 versus 0.17 when they did so manually.
“These findings demonstrate consistent improvements in accuracy and precision across all evaluators and test cases”, the researchers emphasise.
Lebenthal and co-workers highlight that their tool uses European-specific growth charts rather than the local or US-based growth references used by other growth calculators that are not specific to children with achondroplasia. They also note that the tool works offline and therefore has “minimal” privacy considerations as “no personal health information is transmitted or stored externally.”
The authors conclude: “[T]his AI-assisted Excel tool automates z-score and percentile calculations for pediatric achondroplasia, ensuring accuracy, efficiency, and accessibility.
“It eliminates manual errors and provides real-time assessments, making it a potentially invaluable resource for clinicians and researchers.”
By Lynda Williams
medwireNews is an independent medical news service provided by Springer Healthcare Ltd. © 2025 Springer Healthcare Ltd, part of Springer Nature
