LifeSemantics 'AI-based Blood Pressure Prediction' Confirmed Accurate in Validation Clinical Trial

LifeSemantics announced on the 4th that it has successfully completed a confirmatory clinical trial demonstrating the blood pressure prediction accuracy of its blood pressure prediction artificial intelligence (AI). The blood pressure prediction AI is a medical AI developed through the Doctor Answer 2.0 project, led by the Ministry of Science and ICT and the National IT Industry Promotion Agency (NIPA).

LifeSemantics Headquarters Exterior
[Photo by LifeSemantics]

In July of last year, LifeSemantics conducted a multicenter, prospective confirmatory clinical trial involving 158 hypertension patients in collaboration with Chonnam National University Hospital, Bucheon St. Mary's Hospital, and Hanyang University Hospital to verify the clinical efficacy of the medical AI. This study measured the accuracy of the software’s predicted average blood pressure over four weeks and the average blood pressure for each week. The acceptance criteria were an average difference between the software’s predicted values and actual measurements of ±3 mmHg or less, a standard deviation of 5 mmHg or less, and a rate of 85% or higher for differences between predicted and measured values within ±10 mmHg.

The results showed that the difference between predicted and measured average blood pressure was an average of 0.48 mmHg with a standard deviation of 3.09 mmHg. The rate of differences within ±10 mmHg between predicted and measured values was 99.39%. For diastolic blood pressure, the performance was also superior, with values of 0.31 mmHg, 2.1 mmHg, and 100%, respectively. The accuracy for weekly average blood pressure met clinical standards across all indicators.

LifeSemantics’ blood pressure prediction AI is being developed as a medical device that predicts blood pressure 1 to 4 weeks ahead based on clinical data when patients record their blood pressure at home via a mobile application for 8 weeks. Beyond simple blood pressure measurement and recording, it is designed to provide information on future blood pressure trends, risk levels, and management appropriateness evaluations, helping patients receive necessary treatment at more appropriate times and encouraging them to continue self-monitoring and managing their blood pressure.

Based on this clinical trial, LifeSemantics plans to quickly complete the medical device approval process for the blood pressure prediction AI and sequentially proceed with commercialization.

Hong Seung-yong, head of LifeSemantics’ Medical Device Business Division, stated, “Medical AI technology for blood pressure prediction and management appropriateness evaluation can be utilized in software medical devices used in clinical settings or digital therapeutics that assist patient self-management. We will advance this technology toward normalizing hypertension self-management by linking it with devices for biometric data measurement and telemedicine platforms such as DoctorCall.”