评估深度学习技术在检测动脉瘤性蛛网膜下腔出血中的效果:卷积神经网络与迁移学习模型的对比分析
本研究使用伊斯坦布尔Ümraniye培训与研究医院的影像数据,比较Inception-V3、EfficientNetB4、单层CNN和三层CNN在区分动脉瘤性与非动脉瘤性蛛网膜下腔出血中的表现,EfficientNetB4准确率最高(99.92%),作者认为基于CNN的迁移学习模型可辅助诊断并可能减少数字减影血管造影等侵入性检查的需求。
要点
- EfficientNetB4准确率99.92%,F值99.82%,召回率99.92%,精确率99.90%。
- 单层与三层CNN及迁移学习模型结果相当。
- 未观察到过拟合,模型稳健。
核心结论
- 基于CNN的迁移学习模型能通过CT图像准确诊断蛛网膜下腔出血的病因,是临床医生的有价值工具。
- 该方法可能减少数字减影血管造影等侵入性操作的需求,从而更有效利用医疗资源并改善患者结局。
研究设计与样本
- 研究使用伊斯坦布尔Ümraniye培训与研究医院的医学数字成像和通信图像,包括123例动脉瘤性蛛网膜下腔出血患者的15600张图像和80例非动脉瘤性蛛网膜下腔出血患者的7793张图像。
- 研究采用4种模型:Inception-V3、EfficientNetB4、单层CNN和三层CNN。
- 迁移学习模型通过修改最后3层并使用Adam优化器进行定制,模型在Google Collaboratory上训练,基于F值、精确率、召回率和准确率等指标评估。
主要发现
- EfficientNetB4准确率最高(99.92%),F值(99.82%)、召回率(99.92%)和精确率(99.90%)均优于其他方法。
- 单层和三层CNN与迁移学习模型产生可比的结果。
- 未观察到过拟合,开发了稳健的模型。
临床怎么用
- 作者认为基于CNN的迁移学习模型能通过CT图像准确诊断蛛网膜下腔出血的病因,是临床医生的有价值工具。
- 作者认为该方法可能减少数字减影血管造影等侵入性操作的需求,从而更有效利用医疗资源并改善患者结局。
局限性
- 原文未报告局限性。
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)。
原文
OBJECTIVE: Machine learning and deep learning techniques offer a promising multidisciplinary solution for subarachnoid hemorrhage (SAH) detection. The novel transfer learning approach mitigates the time constraints associated with the traditional techniques and demonstrates a superior performance. This study aims to evaluate the effectiveness of convolutional neural networks (CNNs) and CNN-based transfer learning models in differentiating between aneurysmal SAH and nonaneurysmal SAH. METHODS: Data from Istanbul Ümraniye Training and Research Hospital, which included 15,600 digital imaging and communications in medicine images from 123 patients with aneurysmal SAH and 7793 images from 80 patients with nonaneurysmal SAH, were used. The study employed 4 models: Inception-V3, EfficientNetB4, single-layer CNN, and three-layer CNN. Transfer learning models were customized by modifying the last 3 layers and using the Adam optimizer. The models were trained on Google Collaboratory and evaluated based on metrics such as F-score, precision, recall, and accuracy. RESULTS: EfficientNetB4 demonstrated the highest accuracy (99.92%), with a better F-score (99.82%), recall (99.92%), and precision (99.90%) than the other methods. The single- and three-layer CNNs and the transfer learning models produced comparable results. No overfitting was observed, and robust models were developed. CONCLUSIONS: CNN-based transfer learning models can accurately diagnose the etiology of SAH from computed tomography images and is a valuable tool for clinicians. This approach could reduce the need for invasive procedures such as digital subtraction angiography, leading to more efficient medical resource utilization and improved patient outcomes.
引用
- Etli MU, Başarslan MS, Varol E, et al. Evaluating Deep Learning Techniques for Detecting Aneurysmal Subarachnoid Hemorrhage: A Comparative Analysis of Convolutional Neural Network and Transfer Learning Models. World Neurosurg. 2024. PMID: 38710407. doi:10.1016/j.wneu.2024.04.168
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。