利用神经影像数据的机器学习识别孤独症谱系障碍:系统综述与Meta分析
该系统综述与Meta分析纳入44项研究(5697例孤独症谱系障碍患者和6013例典型发育个体),评估机器学习联合神经影像数据分类孤独症谱系障碍的汇总敏感度为86.25%(95%CI 81.24, 90.08)、汇总特异度为83.31%(95%CI 78.12, 87.48),HSROC曲线下面积(AUC)为0.889;异质性高(Higgins I²>90%,Cochran's Q p<0.0001);作者认为该方向有前景,但需进一步研究提高可靠性以用于临床实践。
要点
- 机器学习联合神经影像数据区分孤独症谱系障碍与典型发育个体的汇总敏感度为86.25%、汇总特异度为83.31%,AUC为0.889。
- 未使用脑图谱的研究汇总特异度更高(90.91%,95%CI 80.67, 96.00,p=0.032)。
- 同时招募男性和女性的研究汇总敏感度更高(89.04%,95%CI 83.84, 92.72,p=0.021),联合多种影像模态的研究汇总敏感度也更高(94.12%,95%CI 85.43, 97.76,p=0.036)。
核心结论
- 机器学习联合神经影像数据在识别孤独症谱系障碍个体方面是一个有前景的方向,但需要进一步研究以提高其可靠性以便用于临床实践。
研究设计与样本
- 系统综述与Meta分析,检索PubMed、Scopus和Embase,纳入44项研究(5697例孤独症谱系障碍患者和6013例典型发育个体),采用QUADAS-2评估研究质量,使用双变量随机效应模型进行Meta分析并通过HSROC曲线评估诊断性能,同时进行Meta回归。
诊断标准与关键数值
- 机器学习联合神经影像数据区分孤独症谱系障碍与典型发育个体的汇总敏感度为86.25%(95%CI 81.24, 90.08)。
- 汇总特异度为83.31%(95%CI 78.12, 87.48),HSROC曲线下面积(AUC)为0.889。
- 异质性高:Higgins I²>90%,Cochran's Q p<0.0001。
- 未使用脑图谱的研究汇总特异度为90.91%(95%CI 80.67, 96.00,p=0.032)。
- 同时招募男性和女性的研究汇总敏感度为89.04%(95%CI 83.84, 92.72,p=0.021);联合多种影像模态的研究汇总敏感度为94.12%(95%CI 85.43, 97.76,p=0.036)。
临床应用与阅片要点
- 作者认为机器学习联合神经影像数据在检测孤独症谱系障碍个体方面是一个有前景的方向,但需进一步研究提高可靠性以用于临床实践。
局限性
- 异质性高:Higgins I²>90%,Cochran's Q p<0.0001。
- 作者指出需要进一步研究以提高机器学习联合神经影像数据识别孤独症谱系障碍的可靠性以便用于临床实践。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
英文原文(摘要)· 点击展开
PURPOSE: Autism Spectrum Disorder (ASD) is diagnosed through observation or interview assessments, which is time-consuming, subjective, and with questionable validity and reliability. Thus, we aimed to evaluate the role of machine learning (ML) with neuroimaging data to provide a reliable classification of ASD.
METHODS: A systematic search of PubMed, Scopus, and Embase was conducted to identify relevant publications. Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) was used to assess the studies' quality. A bivariate random-effects model meta-analysis was employed to evaluate the pooled sensitivity, the pooled specificity, and the diagnostic performance through the hierarchical summary receiver operating characteristic (HSROC) curve of ML with neuroimaging data in classifying ASD. Meta-regression was also performed.
RESULTS: Forty-four studies (5697 ASD and 6013 typically developing individuals [TD] in total) were included in the quantitative analysis. The pooled sensitivity for differentiating ASD from TD individuals was 86.25 95% confidence interval [CI] (81.24, 90.08), while the pooled specificity was 83.31 95% CI (78.12, 87.48) with a combined area under the HSROC (AUC) of 0.889. Higgins I2 (> 90%) and Cochran's Q (p < 0.0001) suggest a high degree of heterogeneity. In the bivariate model meta-regression, a higher pooled specificity was observed in studies not using a brain atlas (90.91 95% CI [80.67, 96.00], p = 0.032). In addition, a greater pooled sensitivity was seen in studies recruiting both males and females (89.04 95% CI [83.84, 92.72], p = 0.021), and combining imaging modalities (94.12 95% [85.43, 97.76], p = 0.036).
CONCLUSION: ML with neuroimaging data is an exciting prospect in detecting individuals with ASD but further studies are required to improve its reliability for usage in clinical practice.
引用
- Song DY, Topriceanu CC, Ilie-Ablachim DC, et al. Machine learning with neuroimaging data to identify autism spectrum disorder: a systematic review and meta-analysis. Neuroradiology. 2021. PMID: 34420058. doi:10.1007/s00234-021-02774-z
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。