The Angle orthodontist · 2025
What amount of data is required to develop artificial intelligence that can accurately predict soft tissue changes after orthognathic surgery?
СтатьяНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Kim Jong-Hak, Kwon Naeun, Park Ji-Ae, Youn Sung Bin, Seo Byoung-Moo, Lee Shin-Jae
Аннотация
OBJECTIVES: To suggest a sample size calculation method to develop artificial intelligence (AI) that can predict soft tissue changes after orthognathic surgery with clinically acceptable accuracy.
MATERIALS AND METHODS: From data collected from 705 patients who had undergone combined surgical-orthodontic treatment, 10 subsets of the data were generated through random resampling procedures, specifically with reduced data sizes of 75, 100, 150, 200, 300, 400, 450, 500, 600, and 700. Resampling was repeated four times, and each subset was used to create a total of 40 AI models using a deep-learning algorithm. The prediction results for soft tissue change after orthognathic surgery were compared across all 40 AI models based on their sample sizes. Clinically acceptable accuracy was set as a 1.5-mm prediction error. The predictive performance of AI models was evaluated on the lower lip, which was selected as a primary outcome variable and a benchmark landmark. Linear regression analysis was conducted to estimate the relationship between sample size and prediction error.
RESULTS: The prediction error decreased with increasing sample size. A sample size greater than 1700 datasets was estimated as being required for the development of an AI model with a prediction error < 1.5 mm at the lower lip area.
CONCLUSIONS: A fairly large quantity of orthognathic surgery data seemed to be necessary to develop software programs for visualizing surgical treatment objectives with clinically acceptable accuracy.
MATERIALS AND METHODS: From data collected from 705 patients who had undergone combined surgical-orthodontic treatment, 10 subsets of the data were generated through random resampling procedures, specifically with reduced data sizes of 75, 100, 150, 200, 300, 400, 450, 500, 600, and 700. Resampling was repeated four times, and each subset was used to create a total of 40 AI models using a deep-learning algorithm. The prediction results for soft tissue change after orthognathic surgery were compared across all 40 AI models based on their sample sizes. Clinically acceptable accuracy was set as a 1.5-mm prediction error. The predictive performance of AI models was evaluated on the lower lip, which was selected as a primary outcome variable and a benchmark landmark. Linear regression analysis was conducted to estimate the relationship between sample size and prediction error.
RESULTS: The prediction error decreased with increasing sample size. A sample size greater than 1700 datasets was estimated as being required for the development of an AI model with a prediction error < 1.5 mm at the lower lip area.
CONCLUSIONS: A fairly large quantity of orthognathic surgery data seemed to be necessary to develop software programs for visualizing surgical treatment objectives with clinically acceptable accuracy.