Deep Reinforcement Learning for Detection of Abnormal Anatomies

Abstract

Automatic detection of abnormal anatomies or malformations of different structures of the human body is a challenging task that could provide support for clinicians in their daily practice. Compared to normative anatomies, there is a low presence of anatomical abnormalities in patients, and the great variation within malformations make it challenging to design deep learning frameworks for automatic detection. We propose a framework for anatomical abnormality detection, which benefits from using a deep reinforcement learning model for landmark detection trained in normative data. We detect the abnormalities using the variability between the predicted landmarks configurations in a subspace based on a point distribution model of landmarks using Procrustes shape alignment and principal component analysis projection from normative data. We demonstrate the performance of this implementation on clinical CT scans of the inner ear, and show how synthetically created abnormal cochlea anatomy can be detected using the prediction of five landmarks around the cochlea. Our approach shows a Receiver Operating Characteristics (ROC) Area Under The Curve (AUC) of 0.97, and 96% accuracy for the detection of abnormal anatomy on synthetic data.

A Web-Based Automated Image Processing Research Platform for Cochlear Implantation-Related Studies

“…Post-operatively, Nautilus makes possible the exploration of anatomo-physiologically-tuned fitting [78,79] or the exploration of the relationship between electrode geometrical configuration within the cochlea and clinical outcomes, including perhaps residual hearing. For all its utility, Nautilus could in the future be extended with additional features to address a broader spectrum of investigations, such as these related to the prediction of insertion difficulties during surgical planning, including for abnormal anatomies [80,81]. The delineation of other structures, including the facial nerve, chorda tympani, or RW would then be required.…”

Deep Reinforcement Learning for Detection of Inner Ear Abnormal Anatomy in Computed Tomography

“…To test our approach, we synthetically generated abnormal inner ear CT scans from the original images by removing the cochlea (simulating cochlear aplasia) from the images, thus generating corresponding pairs of normal and abnormal CT scans with the same surrounding structures. The cochlea was segmented using ITK-SNAP software [20] and then replaced by Gaussian noise with mean and standard deviation estimated from the intensities of the tissue surrounding the segmentation [12]. An example of the transformation process as well as the location of the anatomical landmarks we use are shown in Figure 1.…”

“…We use both the communicative multiple agent reinforcement learning (C-MARL [11]) model and the standard multiple agent reinforcement learning (MARL [19]) model to locate a set of landmarks in the inner ear. We extract two pieces of critical information from these models: First, the variability of the predicted location of a certain landmark across different runs/agents which we evaluate in a subspace defined by the normative data landmarks after they are all aligned using Procrustes, and a principal component analysis (PCA) of the shape variation is performed to define the subspace as presented by López Diez et al in [12]. Second, as a measurement of abnormality, we use the distribution of the predicted Q-values for each agent over the last ten states, including the final position where the landmark is placed.…”