Sharif Naeha, Gilani Syed Zulqarnain, Suter David, Reid Siobhan, Szulc Pawel, Kimelman Douglas, Monchka Barret A, Jozani Mohammad Jafari, Hodgson Jonathan M, Sim Marc, Zhu Kun, Harvey Nicholas C, Kiel Douglas P, Prince Richard L, Schousboe John T, Leslie William D, Lewis Joshua R
Nutrition & Health Innovation Research Institute, Edith Cowan University, Perth, Australia; Centre for AI&ML, School of Science, Edith Cowan University, Perth, Australia; Department of Computer Science and Software Engineering, The University of Western Australia, Perth, Australia.
Nutrition & Health Innovation Research Institute, Edith Cowan University, Perth, Australia; Centre for AI&ML, School of Science, Edith Cowan University, Perth, Australia.
EBioMedicine. 2023 Aug;94:104676. doi: 10.1016/j.ebiom.2023.104676. Epub 2023 Jul 11.
Lateral spine images for vertebral fracture assessment can be easily obtained on modern bone density machines. Abdominal aortic calcification (AAC) can be scored on these images by trained imaging specialists to assess cardiovascular disease risk. However, this process is laborious and requires careful training.
Training and testing of model performance of the convolutional neural network (CNN) algorithm for automated AAC-24 scoring utilised 5012 lateral spine images (2 manufacturers, 4 models of bone density machines), with trained imaging specialist AAC scores. Validation occurred in a registry-based cohort study of 8565 older men and women with images captured as part of routine clinical practice for fracture risk assessment. Cox proportional hazards models were used to estimate the association between machine-learning AAC (ML-AAC-24) scores with future incident Major Adverse Cardiovascular Events (MACE) that including death, hospitalised acute myocardial infarction or ischemic cerebrovascular disease ascertained from linked healthcare data.
The average intraclass correlation coefficient between imaging specialist and ML-AAC-24 scores for 5012 images was 0.84 (95% CI 0.83, 0.84) with classification accuracy of 80% for established AAC groups. During a mean follow-up 4 years in the registry-based cohort, MACE outcomes were reported in 1177 people (13.7%). With increasing ML-AAC-24 scores there was an increasing proportion of people with MACE (low 7.9%, moderate 14.5%, high 21.2%), as well as individual MACE components (all p-trend <0.001). After multivariable adjustment, moderate and high ML-AAC-24 groups remained significantly associated with MACE (HR 1.54, 95% CI 1.31-1.80 & HR 2.06, 95% CI 1.75-2.42, respectively), compared to those with low ML-AAC-24.
The ML-AAC-24 scores had substantial levels of agreement with trained imaging specialists, and was associated with a substantial gradient of risk for cardiovascular events in a real-world setting. This approach could be readily implemented into these clinical settings to improve identification of people at high CVD risk.
The study was supported by a National Health and Medical Research Council of Australia Ideas grant and the Rady Innovation Fund, Rady Faculty of Health Sciences, University of Manitoba.
在现代骨密度仪上可以轻松获取用于椎体骨折评估的脊柱侧位图像。经过培训的影像专家可以根据这些图像对腹主动脉钙化（AAC）进行评分，以评估心血管疾病风险。然而，这个过程很繁琐，需要仔细培训。
利用5012张脊柱侧位图像（2个制造商，4种骨密度仪型号）对用于自动AAC-24评分的卷积神经网络（CNN）算法的模型性能进行训练和测试，并由经过培训的影像专家给出AAC评分。在一项基于注册登记的队列研究中进行验证，该研究纳入了8565名老年男性和女性，其图像是作为骨折风险评估常规临床实践的一部分采集的。使用Cox比例风险模型来估计机器学习AAC（ML-AAC-24）评分与未来发生的主要不良心血管事件（MACE）之间的关联，MACE包括死亡、住院急性心肌梗死或缺血性脑血管疾病，这些事件是通过关联的医疗保健数据确定的。
5012张图像的影像专家评分与ML-AAC-24评分之间的平均组内相关系数为0.84（95%CI 0.83, 0.84），已确定的AAC组的分类准确率为80%。在基于注册登记的队列中平均随访4年期间，1177人（13.7%）报告了MACE结局。随着ML-AAC-24评分的增加，发生MACE的人群比例增加（低评分组7.9%，中等评分组14.5%，高评分组21.2%），以及各个MACE组分（所有p趋势<0.001）。经过多变量调整后，与低ML-AAC-24组相比，中等和高ML-AAC-24组仍与MACE显著相关（HR分别为1.54，95%CI 1.31 - 1.80和HR 2.06，95%CI 1.75 - 2.42）。
ML-AAC-24评分与经过培训的影像专家评分具有高度一致性，并且在现实环境中与心血管事件风险的显著梯度相关。这种方法可以很容易地应用于这些临床环境中，以改善对心血管疾病高风险人群的识别。
该研究由澳大利亚国家卫生与医学研究委员会的创意基金以及曼尼托巴大学雷迪健康科学学院的雷迪创新基金资助。
Zotero 插件
