Seoul St. Mary's Hospital Develops AI to Predict Blood Carbon Dioxide Levels in Pediatric Surgeries
Estimating Arterial Carbon Dioxide Without Blood Sampling; Published in International Journal
Mean Error Improved by 23% Compared to Conventional End-Tidal CO2 Measurement
A Promising Auxiliary Tool to Reduce Repeated Blood Draws in Neonates
An artificial intelligence (AI) model has been developed that can predict the arterial partial pressure of carbon dioxide (PaCO₂) in pediatric patients during surgery without the need for blood sampling. The model demonstrated higher accuracy than traditional noninvasive tests while reducing the burden of repeated arterial blood draws, confirming its clinical applicability.
(From left) Hyunho Kim, Professor of Pediatrics at Seoul St. Mary's Hospital, Juhyun Park, Resident in Anesthesiology and Pain Medicine at Asan Medical Center, and Chaeun Cho, student at Korea University College of Medicine. Seoul St. Mary's Hospital
View original imageOn July 21, Seoul St. Mary's Hospital announced that a research team led by Professor Hyunho Kim of the Department of Pediatrics at St. Nicholas Children's Hospital (within Seoul St. Mary's) has developed an AI model that estimates arterial carbon dioxide partial pressure in pediatric patients by utilizing biosignal data collected during surgery. The first authors of the study are Resident Juhyun Park from the Department of Anesthesiology and Pain Medicine at Asan Medical Center and medical student Chaeun Cho at Korea University College of Medicine.
Arterial carbon dioxide partial pressure is a critical indicator for evaluating ventilation status and acid-base balance during surgery. Currently, the standard method involves inserting an arterial catheter, drawing blood, and measuring it with a blood gas analyzer. However, due to the smaller blood vessels in pediatric patients, there is a higher risk of complications from arterial catheterization and the potential for anemia resulting from repeated blood draws.
To address this, end-tidal carbon dioxide (EtCO₂), a noninvasive measurement, is often used. However, there are limitations since differences can occur between the actual arterial CO₂ levels and EtCO₂. The research team developed a model that uses AI to analyze EtCO₂ and clinical data in order to estimate arterial carbon dioxide partial pressure more accurately.
The researchers analyzed 8,853 data pairs obtained from 3,586 pediatric patients using the publicly available operating room biosignal database (VitalDB) at Seoul National University Hospital. They then applied SHAP (Shapley Additive Explanations) analysis to identify ten key variables: end-tidal carbon dioxide, body temperature, minute ventilation, age, fraction of inspired oxygen (FiO₂), presence of heart disease, lung compliance, preoperative hemoglobin level, body mass index (BMI), and mean arterial pressure. These variables were incorporated into a machine learning model.
As a result, the developed AI model achieved a mean absolute error (MAE) of 2.73 mmHg, which is approximately 23% lower than the error of EtCO₂ alone (3.56 mmHg).
The team also verified the model’s reproducibility through external validation: for 50 pediatric patients (92 cases) at Chungnam National University Hospital and 499 patients (2,138 cases) at Seoul National University Hospital between January 2024 and June 2025, the MAEs were 3.65 mmHg and 3.67 mmHg, respectively.
Hot Picks Today
"Two Spend 500,000 Won in a Flash, Yet Wait Two Hours?" ... 'Dreamland' Shunned by Kids and Foreign Tourists
- The Era of Getting 200 Million Won After 30 Years Is Over... A Japanese Bank's "Perpetual Department Head" Who Just Clocked In Cries Out
- Spain Hit by Its Worst Tragedy... Lee Kangin, Moreno and Real Madrid Join Soccer World in Mourning
- They Said, “Make It Your Home for Life” - Then Suddenly Kicked 200 Passengers Off... What Happened to the Around-the-World Cruise?
- "130 Rice Bowls Sold Out in a Day"... 'Unfiltered Face' Adds 8 Million Won to Night Market Sales: China's Bungeunwoo
Professor Hyunho Kim, who led the research, stated, "While this model cannot completely replace arterial blood gas testing, it can serve as a helpful auxiliary tool in situations where arterial catheterization is difficult or frequent blood gas analysis is not feasible." He added, "Its potential is especially promising for pediatric neurosurgery patients, where precision carbon dioxide management is critical." He further emphasized, "Further validation through prospective clinical trials is necessary for real-world clinical application." The findings were recently published in the international journal 'Anesthesia and Pain Medicine.'
© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.