基于智能计算的自动骨龄评估及其与TW3法比较

Automatic bone age assessment based on intelligent algorithms and comparison with TW3 method

目的寻找新的算法以提高自动化骨龄评估(ABAA)的准确性和实用性。方法建立基于目标的兴趣区(ROD。按照Tanner-Whitehouse(TW3)法将ROI分为RUS(包括尺桡骨及掌指骨)ROI及腕骨ROI。按离子群优化(PSO)法,每个兴趣区提取5项特征(包括大小、形态及融合或比邻状态)输入人工神经网络(ANN)分类器,ANN建立在前馈的多层网络基础上,并以反向传播算法规则训练ANN以分别处理RUS及腕骨特征。约1046份左手及腕的数字X线片被随机分成两部分,一半用以训练ANN,另一半用以ABAA,而之前全部采用TW3法有两名小儿内分泌专家人工判读骨龄。结果不同专家判读的骨龄间比较提示:RUS骨龄的标准差大于腕骨骨龄(4.40和2.42),但二者的变异系数(CV)均为4.0,且均有很高的一致率(95.5%及94.2%),不同判读者间RUS及腕骨骨龄均无显著性差异(P>0.05)。通过比较ABAA与人工判读骨龄的比较发现,RUS骨龄的标准差大于腕骨骨龄。但腕骨骨龄<9岁及RUS骨龄≥9岁者CV很接近,分别为3.0和3.1,而对RUS骨龄<9岁者CV较大,为3.5。本研究中不管是RUS骨龄还是腕骨骨龄,ABAA与人工判读相比均有很高的一致率(97.0%、93.8%与96.5%)并且无显著性差异(P>0.05)。结论PSO对图像分割与特征的提取更为有效和准确。该ANN经训练后能更全面地处理影像特征信息,准确判断骨龄。基于智能算法的ABAA系统成功地应用于骨龄0～18岁所有病例。

Objective To improve the validity, accuracy and practicality of automatic bone age assessment （ABAA） with new algorithms. Methods The concept of object-based ROI was proposed. Thirteen RUS （including radius, ulna and short finger bones） ROIs and seven Carpal ROIs were appointed respectively according to Tanner-Whitehouse （TW3） method. Five features including size, morphologie features and fusional/adjacent stage of each ROI were extracted based on particle swarm optimization （PSO） and input into ANN classifiers. ANNs were built upon feed-forward muhilayer networks and trained with back-propagation algorithm rules to process RUS and Carpal features respectively. About 1046 digital left hand- wrist radiographs were randomly utilized half for training ANNs and the rest for ABAA after manual reading by TW3 method, Results BA comparison between observers indicated that the SD of RUS BA was larger than that of Carpal BA （SD= 4.40, 2.42 respectively）, but interestingly, both CVs were 4.0, and both concordance rates were very high （95.5% and 94.2%）, and hoth differences between observers were not significant （both P）〉0.05 ）. It was found by comparison between results of ABAA and manual readings that RUS BA had larger SDs than Carpal BA between two methods, but the CVs were very similar in the case of Carpal BA〈9 years and RUS BA≥9 years （CV=3.0, 3. 1 respectively）, apart from a comparatively larger CV for RUS BA〈9 years （CV=3.5）. Both parts of ABAA system, RUS and Carpal, had very high concordance rates （97.0%, 93.8% and 96.5%） and no significant difference compared with manual method （all P〉0.05）. Conclusion PSO method made image segmentation and feature extraction more valid and accurate, and the ANN models were sophisticated in processing image information. ABAA system based on intelligent algorithms had been successfully applied to all cases from 0 to 18 years of bone age.