Quantitative imaging in medicine and surgery · 2025
Deep learning for automated grading of radiographic sacroiliitis.
СтатьяНазвание и аннотация: Перевод в очереди
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.
Meng Xinyi, Du Yongku, Jia Rongrong, Zhou Qing, Xia Yuwei, Shi Feng, Zhao Fanhui, Gao Yanjun
Аннотация
BACKGROUND: Grading assessment of sacroiliitis via X-ray is widely used in clinical evaluation. The aim of this study was to develop and validate an artificial intelligence (AI) system to help physicians in assessing and diagnosing sacroiliitis from standard X-ray images.
METHODS: In this retrospective study, a deep learning model for the automated grading assessment of radiographic sacroiliitis was developed using pelvic X-ray images from a training set of 465 individuals (930 single sacroiliac joints) and a validation set of 195 individuals (390 single sacroiliac joints). The algorithm was tested using an external test set of 223 individuals (446 single sacroiliac joints). The receiver operating characteristic (ROC) curve was used to calculate the area under the curve (AUC), sensitivity, and specificity to assess the model's performance. The findings of the model were used as a reference to determine its utility in aiding radiologists in the diagnosis and grading assessment of sacroiliitis.
RESULTS: The neural network model demonstrated proficiency in assessing grading of sacroiliitis. In the external test set, the model achieved a grading accuracy rate of 63.90% for radiographic sacroiliitis, and its diagnostic accuracy for determining the presence of radiographic sacroiliitis reached 90.13%. With the assistance of the model, the diagnostic accuracy of radiological sacroiliac arthritis by two junior imaging physicians improved significantly, increasing from 92.45% and 91.10% to 97.17% and 95.29%, respectively. Furthermore, the accuracy of image grading (grades 0 to 4) also showed notable improvement, rising from 75.00% and 74.08% to 88.89% and 80.90%, respectively.
CONCLUSIONS: The AI model demonstrated high diagnostic accuracy and can greatly enhance the precision of radiographic sacroiliitis grading.
METHODS: In this retrospective study, a deep learning model for the automated grading assessment of radiographic sacroiliitis was developed using pelvic X-ray images from a training set of 465 individuals (930 single sacroiliac joints) and a validation set of 195 individuals (390 single sacroiliac joints). The algorithm was tested using an external test set of 223 individuals (446 single sacroiliac joints). The receiver operating characteristic (ROC) curve was used to calculate the area under the curve (AUC), sensitivity, and specificity to assess the model's performance. The findings of the model were used as a reference to determine its utility in aiding radiologists in the diagnosis and grading assessment of sacroiliitis.
RESULTS: The neural network model demonstrated proficiency in assessing grading of sacroiliitis. In the external test set, the model achieved a grading accuracy rate of 63.90% for radiographic sacroiliitis, and its diagnostic accuracy for determining the presence of radiographic sacroiliitis reached 90.13%. With the assistance of the model, the diagnostic accuracy of radiological sacroiliac arthritis by two junior imaging physicians improved significantly, increasing from 92.45% and 91.10% to 97.17% and 95.29%, respectively. Furthermore, the accuracy of image grading (grades 0 to 4) also showed notable improvement, rising from 75.00% and 74.08% to 88.89% and 80.90%, respectively.
CONCLUSIONS: The AI model demonstrated high diagnostic accuracy and can greatly enhance the precision of radiographic sacroiliitis grading.