Ashok Kumar Devaraj, Anburajan Mariamicheal, Snekhalatha Umapathy
Department of Biomedical Engineering, SRM Institute of Science and Technology, Chennai, Tamil Nadu, India.
Andhra Pradesh Medtech Zone (AMTZ), Visakhapatnam, India.
Int J Rheum Dis. 2018 Jul;21(7):1350-1371. doi: 10.1111/1756-185X.13326.
(i) To predict the future risk of osteoporotic fracture in women using a simple forearm radiograph. (ii) To assess osteoporosis in southern Indian women by using radiogrammetric technique in comparison with dual-energy X-ray absorptiometry (DXA) and X-ray phantom study.
The bone mineral density (BMD) of the right proximal femur by DXA and the X-ray measurements were acquired from the right forearm. The combined cortical thickness at the second to fourth metacarpal region (M-CCT), radius (R-CCT) and ulna (U-CCT) were derived in all the studied population. The aluminium phantom study was conducted by varying the X-ray source to film distance at 100 cm and 150 cm, respectively. The feed forward back propagation neural network was used for classification of low bone mass group and normal.
The combined cortical thickness of M-CCT, R-CCT and U-CCT of the total studied population was strongly correlated with DXA femur Th.BMD measurements (r = 0.77, r = 0.61 and r = 0.59 [P < 0.01]). The predicted future osteoporotic fracture risk for the low bone mass group, post-menopausal women and old-aged women population was found to be 92%, 62.8%, and 64.7%, respectively. The accuracy of neural network classifier for training set, testing set was found to be 97.5% and 87.5% in the studied population.
The results suggested that M-CCT and M-CCT (%) at the second metacarpal region are useful in predicting the future risk of osteoporotic fracture in women. The aluminium phantom study with an X-ray tube to film distance of 100 cm mimics an exact condition of forearm radiogrammetry.
（i）使用简单的前臂X线片预测女性骨质疏松性骨折的未来风险。（ii）通过放射计量技术评估印度南部女性的骨质疏松症，并与双能X线吸收法（DXA）和X线体模研究进行比较。
通过DXA测量右近端股骨的骨密度（BMD），并从前臂获取X线测量值。在所有研究人群中得出第二至第四掌骨区域（M-CCT）、桡骨（R-CCT）和尺骨（U-CCT）的皮质厚度总和。分别在100 cm和150 cm的X线源到胶片距离下进行铝体模研究。使用前馈反向传播神经网络对低骨量组和正常组进行分类。
所有研究人群的M-CCT、R-CCT和U-CCT的皮质厚度总和与DXA测量的股骨骨小梁BMD（Th.BMD）密切相关（r = 0.77、r = 0.61和r = 0.59 [P < 0.01]）。发现低骨量组、绝经后女性和老年女性人群未来骨质疏松性骨折的预测风险分别为92%、62.8%和64.7%。在研究人群中，发现神经网络分类器对训练集和测试集的准确率分别为97.5%和87.5%。
结果表明，第二掌骨区域的M-CCT和M-CCT（%）有助于预测女性未来骨质疏松性骨折的风险。X线管到胶片距离为100 cm的铝体模研究模拟了前臂放射计量的精确条件。
