Quantitative Sciences Unit, Department of Medicine, Stanford University, Palo Alto, California.
Center for Biomedical Informatics, University of Tennessee Health Science Center, Memphis.
JAMA Netw Open. 2023 Jun 1;6(6):e2319420. doi: 10.1001/jamanetworkopen.2023.19420.
Abusive head trauma (AHT) in children is often missed in medical encounters, and retinal hemorrhage (RH) is considered strong evidence for AHT. Although head computed tomography (CT) is obtained routinely, all but exceptionally large RHs are undetectable on CT images in children.
To examine whether deep learning-based image analysis can detect RH on pediatric head CT.
DESIGN, SETTING, AND PARTICIPANTS: This diagnostic study included 301 patients diagnosed with AHT who underwent head CT and dilated fundoscopic examinations at a quaternary care children's hospital. The study assessed a deep learning model using axial slices from 218 segmented globes with RH and 384 globes without RH between May 1, 2007, and March 31, 2021. Two additional light gradient boosting machine (GBM) models were assessed: one that used demographic characteristics and common brain findings in AHT and another that combined the deep learning model's risk prediction plus the same demographic characteristics and brain findings.
Sensitivity (recall), specificity, precision, accuracy, F1 score, and area under the curve (AUC) for each model predicting the presence or absence of RH in globes were assessed. Globe regions that influenced the deep learning model predictions were visualized in saliency maps. The contributions of demographic and standard CT features were assessed by Shapley additive explanation.
The final study population included 301 patients (187 [62.1%] male; median [range] age, 4.6 [0.1-35.8] months). A total of 120 patients (39.9%) had RH on fundoscopic examinations. The deep learning model performed as follows: sensitivity, 79.6%; specificity, 79.2%; positive predictive value (precision), 68.6%; negative predictive value, 87.1%; accuracy, 79.3%; F1 score, 73.7%; and AUC, 0.83 (95% CI, 0.75-0.91). The AUCs were 0.80 (95% CI, 0.69-0.91) for the general light GBM model and 0.86 (95% CI, 0.79-0.93) for the combined light GBM model. Sensitivities of all models were similar, whereas the specificities of the deep learning and combined light GBM models were higher than those of the light GBM model.
The findings of this diagnostic study indicate that a deep learning-based image analysis of globes on pediatric head CTs can predict the presence of RH. After prospective external validation, a deep learning model incorporated into CT image analysis software could calibrate clinical suspicion for AHT and provide decision support for which patients urgently need fundoscopic examinations.
儿童虐待性头部外伤 (AHT) 在医疗接触中经常被忽视，视网膜出血 (RH) 被认为是 AHT 的有力证据。尽管头部计算机断层扫描 (CT) 通常是常规进行的，但在儿童中，除了异常大的 RH 外，所有 RH 在 CT 图像上都无法检测到。
检查基于深度学习的图像分析是否可以在儿科头部 CT 上检测到 RH。
设计、地点和参与者：这项诊断研究纳入了在一家四级儿童医院接受 AHT 诊断并接受头部 CT 和散瞳检查的 301 名患者。该研究评估了一个深度学习模型，该模型使用了 2007 年 5 月 1 日至 2021 年 3 月 31 日期间 218 个分段球体的轴向切片，这些球体存在 RH 和 384 个不存在 RH。还评估了另外两个轻梯度提升机 (GBM) 模型：一个使用 AHT 中的人口统计学特征和常见脑发现，另一个结合了深度学习模型的风险预测以及相同的人口统计学特征和脑发现。
评估了每个模型预测球体中 RH 存在或不存在的敏感性 (召回率)、特异性、精度、准确性、F1 分数和曲线下面积 (AUC)。通过显着性映射可视化影响深度学习模型预测的球体区域。通过 Shapley 加法解释评估人口统计学和标准 CT 特征的贡献。
最终的研究人群包括 301 名患者（187 [62.1%] 名男性；中位[范围]年龄，4.6 [0.1-35.8] 个月）。共有 120 名患者（39.9%）在眼底检查中存在 RH。深度学习模型的表现如下：敏感性为 79.6%；特异性，79.2%；阳性预测值（精度），68.6%；阴性预测值，87.1%；准确率，79.3%；F1 分数，73.7%；AUC，0.83（95%CI，0.75-0.91）。通用轻 GBM 模型的 AUC 为 0.80（95%CI，0.69-0.91），组合轻 GBM 模型的 AUC 为 0.86（95%CI，0.79-0.93）。所有模型的敏感性相似，而深度学习和组合轻 GBM 模型的特异性高于轻 GBM 模型。
这项诊断研究的结果表明，基于深度学习的儿科头部 CT 球图像分析可以预测 RH 的存在。经过前瞻性外部验证后，将深度学习模型纳入 CT 图像分析软件中，可以校准对 AHT 的临床怀疑，并为哪些患者迫切需要眼底检查提供决策支持。
