Explaining Clinical Decision Support Systems in Medical Imaging using Cycle-Consistent Activation Maximization

This includes applications in microscopy and histopathology [1, 2], time-continuous biosignal analysis [3, 4], and, quite prominently, medical image analysis for volumetric imaging data as generated by computed tomography [5, 6], positron emission tomography [7, 8] or magnetic resonance imaging [9, 10, 11]. In the field of medical imaging, recent work has demonstrated a variety of applications for DNNs, such as organ segmentation [12], anomaly detection [13], lesion detection [14], segmentation [15] and assessment [16], providing major advantages and even repeatedly outperforming gold-standard human assessment [17]. A nearby field of similarly growing research interest established with the publications of Kumar et al. and Aerts et al. [18, 19] is,,Radiomics" using traditional machine learning (ML) techniques. Compared to deep learning techniques, traditional ML methods like random forests and support vector machines have a largely transparent decision-making process, which is generally easier to comprehend and/or depict - a clear argument for their preference in clinical practice. Many publications have shown the advantages of DNNs in comparison to traditional machine learning techniques, such as the ability to learn descriptive features from data instead of a complex and expensive handcrafted feature design, as well as an improved classification performance on medical imaging tasks [20, 21], with some architectures being on par with gold-standard human assessment [17]. However, as DNNs learn features from the given data, the semantic of these features is in general not immediately evident. Thus, clinicians understandably approach these methods with a high degree of skepticism.