Department of Computing, Imperial College London, Huxley Building, 180 Queen's Gate, South Kensington, London, UK.
Medical Faculty, University Witten Herdecke, 58455, Witten, Germany.
Int J Comput Assist Radiol Surg. 2024 Jun;19(6):1061-1073. doi: 10.1007/s11548-024-03096-0. Epub 2024 Mar 27.
Probe-based confocal laser endomicroscopy (pCLE) enables intraoperative tissue characterization with improved resection rates of brain tumours. Although a plethora of deep learning models have been developed for automating tissue characterization, their lack of transparency is a concern. To tackle this issue, techniques like Class Activation Map (CAM) and its variations highlight image regions related to model decisions. However, they often fall short of providing human-interpretable visual explanations for surgical decision support, primarily due to the shattered gradient problem or insufficient theoretical underpinning.
In this paper, we introduce XRelevanceCAM, an explanation method rooted in a better backpropagation approach, incorporating sensitivity and conservation axioms. This enhanced method offers greater theoretical foundation and effectively mitigates the shattered gradient issue when compared to other CAM variants.
Qualitative and quantitative evaluations are based on ex vivo pCLE data of brain tumours. XRelevanceCAM effectively highlights clinically relevant areas that characterize the tissue type. Specifically, it yields a remarkable 56% improvement over our closest baseline, RelevanceCAM, in the network's shallowest layer as measured by the mean Intersection over Union (mIoU) metric based on ground-truth annotations (from 18 to 28.07%). Furthermore, a 6% improvement in mIoU is observed when generating the final saliency map from all network layers.
We introduce a new CAM variation, XRelevanceCAM, for precise identification of clinically important structures in pCLE data. This can aid introperative decision support in brain tumour resection surgery, as validated in our performance study.
基于探针的共聚焦激光内窥镜检查（pCLE）能够实现术中组织特征分析，从而提高脑肿瘤的切除率。虽然已经开发出了大量用于自动组织特征分析的深度学习模型，但这些模型的不透明性令人担忧。为了解决这个问题，像 Class Activation Map（CAM）及其变体这样的技术突出了与模型决策相关的图像区域。然而，它们通常无法为手术决策支持提供可解释的可视化解释，主要是因为存在梯度碎片化问题或缺乏足够的理论基础。
在本文中，我们引入了 XRelevanceCAM，这是一种解释方法，其根源在于更好的反向传播方法，结合了敏感性和守恒公理。与其他 CAM 变体相比，这种改进的方法提供了更大的理论基础，并有效地缓解了梯度碎片化问题。
定性和定量评估是基于脑肿瘤的离体 pCLE 数据进行的。XRelevanceCAM 能够有效地突出与组织类型相关的临床相关区域。具体来说，在基于真实注释的网络最浅层的平均交并比（mIoU）指标上，与我们最接近的基线 RelevanceCAM 相比，XRelevanceCAM 提高了 56%（从 18 提高到 28.07%）。此外，当从所有网络层生成最终显着图时，mIoU 提高了 6%。
我们引入了一种新的 CAM 变体 XRelevanceCAM，用于准确识别 pCLE 数据中具有临床意义的结构。这可以在脑肿瘤切除术的术中决策支持中得到验证。
