Bilwaj Gaonkar

Computerized Assessment of Superior Semicircular Canal Dehiscence Size Using Advanced Morphological Imaging Operators

Abstract Superior semicircular canal dehiscence (SSCD) describes a pathological aperture at the level of the arcuate eminence. Techniques for quantifying defect size are described with most studies using two‐dimensional lengths that underestimate the pathology. The objective of this study is to describe a novel method of measurement that combines manual segmentation of high‐resolution computed tomography (HRCT) images of the temporal bone and a morphological skeletonization transform to calculate dehiscence volume. Images were imported into a freely available image segmentation tool: ITK‐SNAP (version 3.4.0; available at: http://www.itksnap.org/) software. Coronal and sagittal planes were used to outline the dehiscence in all slices demonstrating the defect using the paintbrush tool. A morphological skeletonization transform derived a single‐pixel thick representation of the original delineation. This “sheet” of voxels overlaid the dehiscence. Volume was calculated by counting the number of nonzero image voxels within this “sheet” and multiplying this number by the volume (mm3) of each voxel. A total of 70 cases of SSCD were identified. Overall, mean volume was 0.88 mm3 (standard deviation: 0.57, range: 0.11‐2.27). We present a novel technique for measuring SSCD, which we believe provides a more accurate representation of the pathology, and has the potential to standardize measurement of SSCD.

Exploring intentions of physician-scientist trainees: factors influencing MD and MD/PhD interest in research careers

BackgroundPrior studies have described the career paths of physician-scientist candidates after graduation, but the factors that influence career choices at the candidate stage remain unclear. Additionally, previous work has focused on MD/PhDs, despite many physician-scientists being MDs. This study sought to identify career sector intentions, important factors in career selection, and experienced and predicted obstacles to career success that influence the career choices of MD candidates, MD candidates with research-intense career intentions (MD-RI), and MD/PhD candidates.MethodsA 70-question survey was administered to students at 5 academic medical centers with Medical Scientist Training Programs (MSTPs) and Clinical and Translational Science Awards (CTSA) from the NIH. Data were analyzed using bivariate or multivariate analyses.ResultsMore MD/PhD and MD-RI candidates anticipated or had experienced obstacles related to balancing academic and family responsibilities and to balancing clinical, research, and education responsibilities, whereas more MD candidates indicated experienced and predicted obstacles related to loan repayment. MD/PhD candidates expressed higher interest in basic and translational research compared to MD-RI candidates, who indicated more interest in clinical research. Overall, MD-RI candidates displayed a profile distinct from both MD/PhD and MD candidates.ConclusionsMD/PhD and MD-RI candidates experience obstacles that influence their intentions to pursue academic medical careers from the earliest training stage, obstacles which differ from those of their MD peers. The differences between the aspirations of and challenges facing MD, MD-RI and MD/PhD candidates present opportunities for training programs to target curricula and support services to ensure the career development of successful physician-scientists.

Multi-Parameter Ensemble Learning for Automated Vertebral Body Segmentation in Heterogeneously Acquired Clinical MR Images

The development of quantitative imaging biomarkers in medicine requires automatic delineation of relevant anatomical structures using available imaging data. However, this task is complicated in clinical medicine due to the variation in scanning parameters and protocols, even within a single medical center. Existing literature on automatic image segmentation using MR data is based on the analysis of highly homogenous images obtained using a fixed set of pulse sequence parameters (TR/TE). Unfortunately, algorithms that operate on fixed scanning parameters do not avail themselves to real-world daily clinical use due to the existing variation in scanning parameters and protocols. Thus, it is necessary to develop algorithmic techniques that can address the challenge of MR image segmentation using real clinical data. Toward this goal, we developed a multi-parametric ensemble learning technique to automatically detect and segment lumbar vertebral bodies using MR images of the spine. We use spine imaging data to illustrate our techniques since low back pain is an extremely common condition and a typical spine clinic evaluates patients that have been referred with a wide range of scanning parameters. This method was designed with special emphasis on robustness so that it can perform well despite the inherent variation in scanning protocols. Specifically, we show how a single multi-parameter ensemble model trained with manually labeled T2 scans can autonomously segment vertebral bodies on scans with echo times varying between 24 and 147 ms and relaxation times varying between 1500 and 7810 ms. Furthermore, even though the model was trained using T2-MR imaging data, it can accurately segment vertebral bodies on T1-MR and CT, further demonstrating the robustness and versatility of our methodology. We believe that robust segmentation techniques, such as the one presented here, are necessary for translating computer assisted diagnosis into everyday clinical practice.

Deep learning for medical image segmentation – using the IBM TrueNorth neurosynaptic system

Deep convolutional neural networks have found success in semantic image segmentation tasks in computer vision and medical imaging. These algorithms are executed on conventional von Neumann processor architectures or GPUs. This is suboptimal. Neuromorphic processors that replicate the structure of the brain are better-suited to train and execute deep learning models for image segmentation by relying on massively-parallel processing. However, given that they closely emulate the human brain, on-chip hardware and digital memory limitations also constrain them. Adapting deep learning models to execute image segmentation tasks on such chips, requires specialized training and validation. In this work, we demonstrate for the first-time, spinal image segmentation performed using a deep learning network implemented on neuromorphic hardware of the IBM TrueNorth Neurosynaptic System and validate the performance of our network by comparing it to human-generated segmentations of spinal vertebrae and disks. To achieve this on neuromorphic hardware, the training model constrains the coefficients of individual neurons to {-1,0,1} using the Energy Efficient Deep Neuromorphic (EEDN)1 networks training algorithm. Given the ∼1 million neurons and 256 million synapses, the scale and size of the neural network implemented by the IBM TrueNorth allows us to execute the requisite mapping between segmented images and non-uniform intensity MR images >20 times faster than on a GPU-accelerated network and using <0.1 W. This speed and efficiency implies that a trained neuromorphic chip can be deployed in intra-operative environments where real-time medical image segmentation is necessary.

Novel Method of Measuring Canal Dehiscence and Evaluation of its Potential as a Predictor of Symptom Outcomes After Middle Fossa Craniotomy

BACKGROUND Superior semicircular canal dehiscence (SSCD) is an osseous defect of the arcuate eminence of the petrous temporal bone. Strategies for measuring dehiscence size are variable, and the usefulness of such parameters remains in clinical equipoise. OBJECTIVE To present a novel method of measuring dehiscence volume and to evaluate its potential as a predictor of symptom outcomes after surgical repair of SSCD. METHODS High-resolution computed tomographic temporal bone images were imported into a freely available segmentation software. Dehiscence lengths and volumes were ascertained by independent authors. Inter-rater observer reliability was assessed using Cronbachs alpha. Correlation and regression analyses were performed to evaluate for relationships between dehiscence size and symptoms (pre- and post-operative). RESULTS Thirty-seven dehiscences were segmented using the novel volumetric assessment. Cronbachs alpha for dehiscence lengths and volumes were 0.97 and 0.95, respectively. Dehiscence lengths were more variable as compared to dehiscence volumes (σ2 8.92 vs σ2 0.55, F = 1.74). The mean dehiscence volume was 2.22 mm3 (0.74, 0.64-0.53 mm3). Dehiscence volume and headache at presentation were positively correlated (Rpb = 0.67, P = .03). Dehiscence volume and vertigo improvement after surgery were positively correlated, although this did not reach statistical significance (Rpb = 0.46, P = .21). CONCLUSION SSCD volumetry is a novel method of measuring dehiscence size that has excellent inter-rater reliability and is less variable compared to dehiscence length, but its potential as a predictor of symptom outcomes is not substantiated. However, the study is limited by low power.

Automatic segmentation of lumbar vertebrae in CT images

Lower back pain is one of the most prevalent disorders in the developed/developing world. However, its etiology is poorly understood and treatment is often determined subjectively. In order to quantitatively study the emergence and evolution of back pain, it is necessary to develop consistently measurable markers for pathology. Imaging based measures offer one solution to this problem. The development of imaging based on quantitative biomarkers for the lower back necessitates automated techniques to acquire this data. While the problem of segmenting lumbar vertebrae has been addressed repeatedly in literature, the associated problem of computing relevant biomarkers on the basis of the segmentation has not been addressed thoroughly. In this paper, we propose a Random-Forest based approach that learns to segment vertebral bodies in CT images followed by a biomarker evaluation framework that extracts vertebral heights and widths from the segmentations obtained. Our dataset consists of 15 CT sagittal scans obtained from General Electric Healthcare. Our main approach is divided into three parts: the first stage is image pre-processing which is used to correct for variations in illumination across all the images followed by preparing the foreground and background objects from images; the next stage is Machine Learning using Random-Forests, which distinguishes the interest-point vectors between foreground or background; and the last step is image post-processing, which is crucial to refine the results of classifier. The Dice coefficient was used as a statistical validation metric to evaluate the performance of our segmentations with an average value of 0.725 for our dataset.

Automatic vertebral bodies detection of x-ray images using invariant multiscale template matching

Lower back pain and pathologies related to it are one of the most common results for a referral to a neurosurgical clinic in the developed and the developing world. Quantitative evaluation of these pathologies is a challenge. Image based measurements of angles/vertebral heights and disks could provide a potential quantitative biomarker for tracking and measuring these pathologies. Detection of vertebral bodies is a key element and is the focus of the current work. From the variety of medical imaging techniques, MRI and CT scans have been typically used for developing image segmentation methods. However, CT scans are known to give a large dose of x-rays, increasing cancer risk [8]. MRI can be substituted for CTs when the risk is high [8] but are difficult to obtain in smaller facilities due to cost and lack of expertise in the field [2]. X-rays provide another option with its ability to control the x-ray dosage, especially for young people, and its accessibility for smaller facilities. Hence, the ability to create quantitative biomarkers from x-ray data is especially valuable. Here, we develop a multiscale template matching, inspired by [9], to detect centers of vertebral bodies from x-ray data. The immediate application of such detection lies in developing quantitative biomarkers and in querying similar images in a database. Previously, shape similarity classification methods have been used to address this problem, but these are challenging to use in the presence of variation due to gross pathology and even subtle effects [1].