Department of Computer Science and Information Engineering, National Taiwan University, Taipei, Taiwan, Republic of China.
Department of Medical Imaging, China Medical University Hsinchu Hospital, Hsinchu, Taiwan, Republic of China.
Eur Radiol. 2022 Aug;32(8):5371-5381. doi: 10.1007/s00330-022-08633-6. Epub 2022 Feb 24.
To examine the role of ADC threshold on agreement across observers and deep learning models (DLMs) plus segmentation performance of DLMs for acute ischemic stroke (AIS).
Twelve DLMs, which were trained on DWI-ADC-ADC combination from 76 patients with AIS using 6 different ADC thresholds with ground truth manually contoured by 2 observers, were tested by additional 67 patients in the same hospital and another 78 patients in another hospital. Agreement between observers and DLMs were evaluated by Bland-Altman plot and intraclass correlation coefficient (ICC). The similarity between ground truth (GT) defined by observers and between automatic segmentation performed by DLMs was evaluated by Dice similarity coefficient (DSC). Group comparison was performed using the Mann-Whitney U test. The relationship between the DSC and ADC threshold as well as AIS lesion size was evaluated by linear regression analysis. A p < .05 was considered statistically significant.
Excellent interobserver agreement and intraobserver repeatability in the manual segmentation (all ICC > 0.98, p < .001) were achieved. The 95% limit of agreement was reduced from 11.23 cm for GT on DWI to 0.59 cm for prediction at an ADC threshold of 0.6 × 10 mm/s combined with DWI. The segmentation performance of DLMs was improved with an overall DSC from 0.738 ± 0.214 on DWI to 0.971 ± 0.021 on an ADC threshold of 0.6 × 10 mm/s combined with DWI.
Combining an ADC threshold of 0.6 × 10 mm/s with DWI reduces interobserver and inter-DLM difference and achieves best segmentation performance of AIS lesions using DLMs.
• Higher Dice similarity coefficient (DSC) in predicting acute ischemic stroke lesions was achieved by ADC thresholds combined with DWI than by DWI alone (all p < .05). • DSC had a negative association with the ADC threshold in most sizes, both hospitals, and both observers (most p < .05) and a positive association with the stroke size in all ADC thresholds, both hospitals, and both observers (all p < .001). • An ADC threshold of 0.6 × 10 mm/s eliminated the difference of DSC at any stroke size between observers or between hospitals (p = .07 to > .99).
探讨 ADC 阈值在观察者和深度学习模型（DLM）之间的一致性，以及 DLM 对急性缺血性脑卒中（AIS）的分割性能的作用。
使用来自 76 例 AIS 患者的 DWI-ADC-ADC 组合，对 12 个 DLM 进行训练，使用 6 个不同的 ADC 阈值，由 2 位观察者手动勾画真值，在同一家医院的 67 例患者和另一家医院的 78 例患者中进行测试。通过 Bland-Altman 图和组内相关系数（ICC）评估观察者与 DLM 之间的一致性。通过 Dice 相似系数（DSC）评估观察者定义的真值（GT）和 DLM 自动分割之间的相似性。使用 Mann-Whitney U 检验进行组间比较。通过线性回归分析评估 DSC 与 ADC 阈值以及 AIS 病变大小的关系。p<0.05 被认为具有统计学意义。
在手动分割中实现了优秀的观察者间一致性和观察者内重复性（所有 ICC>0.98，p<0.001）。GT 在 DWI 上的 95%一致性限从 11.23cm 降低到 ADC 阈值为 0.6×10mm/s 时的 0.59cm。结合 DWI，DLM 的分割性能整体 DSC 从 DWI 上的 0.738±0.214 提高到 ADC 阈值为 0.6×10mm/s 时的 0.971±0.021。
将 ADC 阈值 0.6×10mm/s 与 DWI 相结合，减少了观察者间和 DLM 间的差异，并且使用 DLM 实现了 AIS 病变的最佳分割性能。
与单独使用 DWI 相比，ADC 阈值结合 DWI 可实现更高的 DSC（所有 p<0.05）。
在大多数大小、两个医院和两个观察者中，DSC 与 ADC 阈值呈负相关（多数 p<0.05），而与中风大小呈正相关（所有 p<0.001）。
在任何中风大小下，ADC 阈值 0.6×10mm/s 消除了观察者之间或医院之间的 DSC 差异（p=0.07 至>0.99）。
