No abstract

An accurate preoperative measurement of glenoid orientation is crucial for evaluating pathologies and successful total shoulder arthroplasty. Existing methods may be labor-intensive, observer-dependent, and sensitive to the misalignment between the scapula plane and CT scanning direction. In this study, we proposed a computation framework and performed an automated analysis of the glenoid orientation based on 3D surface data. Three-dimensional models of 12 scapulae were analyzed. The glenoid cavity and external anatomical features were automatically extracted from these 3D models. Glenoid version was calculated using the scapula plane and the fulcrum axis alternatively. Glenoid inclination was measured both relative to transverse axis of the scapula and the medial pole-inferior tip axis. The mean (AESD) of the fulcrum-based glenoid version was À0.55˚(AE4.17˚), while the scapular-plane-based glenoid version was À5.05˚(AE3.50˚). The mean (AESD) of glenoid inclinations based on the medial pole and inferior tip was 12.75( AE5.03˚) while the mean (AESD) of the glenoid inclination based on the medial pole and glenoid center was 4.63˚(AE4.86˚). Our computational framework was able to extract the reproducible morphological measures free of inter-and intra-observer variability. For the first time in 3D, we showed that the fulcrum axis was practically perpendicular to the glenoid plane normal (radial line), and thus extended the fulcrum-based glenoid version for quantifying 3D glenoid orientation. Keywords: glenoid orientation; fulcrum axis; version; inclination; morphometry Assessment of glenoid orientation is essential in evaluating pathologies such as degenerative wear and shoulder instability, 1 and for planning shoulder surgeries.2,3 Glenoid orientation characterizes the geometrical relationship between the glenoid cavity and the scapular body. It could be used for differentiating normal and pathological shoulders since shoulder pathologies are likely to damage the glenoid rim and alter the orientation of the glenoid cavity with respect to the scapular body.4 It has been shown that endstage glenohumeral arthritis may increase wearing of the posterior part of glenoid, thus causing increased retroversion. 5,6 In total shoulder arthroplasty, restoration of glenoid orientation can improve the chances of longer-term implant survival by balancing the forces across prosthetic glenoid components, 7,8 while failure in restoring glenoid orientation accurately can cause posterior displacement and glenoid implant loosening.6,9,10 Restoration of glenoid orientation close to a patient's native glenoid orientation is a goal at the time of surgery. 9,11 Unfortunately this goal is difficult to attain due to the extremely diverse morphology of glenoid. 8,9,[12][13][14][15] An accurate and reliable preoperative assessment of glenoid orientation is therefore vital for a successful intra-operative restoration of the shoulder joint. 8Glenoid orientation is usually quantified by glenoid version and inclination. 16,17 These two an...

No abstract

It is common for pathologists to annotate specific regions of the tissue, such as tumor, directly on the glass slide with markers. Although this practice was helpful prior to the advent of histology whole slide digitization, it often occludes important details which are increasingly relevant to immunooncology due to recent advancements in digital pathology imaging techniques. The current work uses a generative adversarial network with cycle loss to remove these annotations while still maintaining the underlying structure of the tissue by solving an image-to-image translation problem. We train our network on up to 300 whole slide images with marker inks and show that 70% of the corrected image patches are indistinguishable from originally uncontaminated image tissue to a human expert. This portion increases 97% when we replace the human expert with a deep residual network. We demonstrated the fidelity of the method to the original image by calculating the correlation between image gradient magnitudes. We observed a revival of up to 94,000 nuclei per slide in our dataset, the majority of which were located on tissue border.

No abstract