VA RR& D Center for Limb Loss and MoBility, Seattle, WA 98108, USA; Department of Mechanical Engineering, University of Washington, Seattle, WA 98195, USA.
Department of Mechanical Engineering, University of Washington, Seattle, WA 98195, USA.
Foot (Edinb). 2023 Sep;56:101989. doi: 10.1016/j.foot.2023.101989. Epub 2023 Feb 25.
Plantar ulceration is a serious complication of diabetes. However, the mechanism of injury initiating ulceration remains unclear. The unique structure of the plantar soft tissue includes superficial and deep layers of adipocytes contained in septal chambers, however, the size of these chambers has not been quantified in diabetic or non-diabetic tissue. Computer-aided methods can be leveraged to guide microstructural measurements and differences with disease status.
Adipose chambers in whole slide images of diabetic and non-diabetic plantar soft tissue were segmented with a pre-trained U-Net and area, perimeter, and minimum and maximum diameter of adipose chambers were measured. Whole slide images were classified as diabetic or non-diabetic using the Axial-DeepLab network, and the attention layer was overlaid on the input image for interpretation.
Non-diabetic deep chambers were 90 %, 41 %, 34 %, and 39 % larger in area (26,954 ± 2428 µm vs 14,157 ± 1153 µm), maximum (277 ± 13 µm vs 197 ± 8 µm) and minimum (140 ± 6 µm vs 104 ± 4 µm) diameter, and perimeter (405 ± 19 µm vs 291 ± 12 µm), respectively, than the superficial (p < 0.001). However, there was no significant difference in these parameters in diabetic specimens (area 18,695 ± 2576 µm vs 16627 ± 130 µm, maximum diameter 221 ± 16 µm vs 210 ± 14 µm, minimum diameter 121 ± 8 µm vs 114 ± 7 µm, perimeter 341 ± 24 µm vs 320 ± 21 µm). Between diabetic and non-diabetic chambers, only the maximum diameter of the deep chambers differed (221 ± 16 µm vs 277 ± 13 µm). The attention network achieved 82 % accuracy on validation, but the attention resolution was too coarse to identify meaningful additional measurements.
Adipose chamber size differences may provide a basis for plantar soft tissue mechanical changes with diabetes. Attention networks are promising tools for classification, but additional care is required when designing networks for identifying novel features.
All images, analysis code, data, and/or other resources required to replicate this work are available from the corresponding author upon reasonable request.
足底溃疡是糖尿病的一种严重并发症。然而，导致溃疡形成的损伤机制仍不清楚。足底软组织的独特结构包括位于隔室中的浅层和深层脂肪细胞，然而，这些隔室的大小在糖尿病或非糖尿病组织中尚未被量化。计算机辅助方法可用于指导微观结构测量和疾病状态差异。
使用预先训练的 U-Net 对糖尿病和非糖尿病足底软组织全幻灯片图像中的脂肪隔室进行分割，并测量脂肪隔室的面积、周长以及最小和最大直径。使用轴向深度实验室网络对全幻灯片图像进行分类，并在输入图像上叠加注意力层进行解释。
非糖尿病深层隔室的面积（26954 ± 2428 µm 比 14157 ± 1153 µm）、最大直径（277 ± 13 µm 比 197 ± 8 µm）和最小直径（140 ± 6 µm 比 104 ± 4 µm）以及周长（405 ± 19 µm 比 291 ± 12 µm）分别比浅层隔室大 90%、41%、34%和 39%（p<0.001）。然而，在糖尿病标本中，这些参数没有显著差异（面积 18695 ± 2576 µm 比 16627 ± 130 µm，最大直径 221 ± 16 µm 比 210 ± 14 µm，最小直径 121 ± 8 µm 比 114 ± 7 µm，周长 341 ± 24 µm 比 320 ± 21 µm）。在糖尿病和非糖尿病隔室之间，只有深层隔室的最大直径不同（221 ± 16 µm 比 277 ± 13 µm）。注意力网络在验证中的准确率达到 82%，但注意力分辨率太粗糙，无法识别有意义的附加测量值。
脂肪隔室大小的差异可能为糖尿病足底软组织力学变化提供基础。注意力网络是分类的有前途的工具，但在设计用于识别新特征的网络时需要额外的关注。
复制本工作所需的所有图像、分析代码、数据和/或其他资源均可从相应作者处按合理要求获得。
