Modern imaging systems far exceed the human eye in spatial and spectral resolution and in dynamic range, thus potentially allowing machine-based image pattern analysis systems to outperform manual image interpretation. In fact, recent work in pattern recognition has demonstrated that computers can equal or even surpass image classification and pattern analysis by human experts. We have previously published work investigating the progression of osteoarthritis (OA) in the human population comprising the Baltimore Longitudinal Study of Aging (BLSA). We were able to show that WND-CHARM (see AG000671-13) is able to diagnose the existence of OA in knee X-Rays with accuracies approaching that of a panel of highly trained radiologists. We have subsequently published work that WND-CHARM can predict the future onset of radiologically detectable osteoarthritis in X-Rays that were scored as radiologically clear. We were able to show that the development of moderate OA two decades in the future can be predicted with > 70% accuracy from X-Rays scored as free of OA by a panel of three radiologists. Subsequently, we were able to further characterize OA progression and identify an early, slow period of change followed by rapid degeneration. We are following up our knee X-ray studies with an MRI dataset obtained from the Osteoarthritis Initiative as well as experimental MRI samples imaged here at NIA through a collaboration with Dr. Richard Spencer (NIA/LCI). In a recently published study, we developed a technique using multivariate linear regression of image features derived from several types of MRI scans to construct a continuously variable cartilage quality score similar to an OARSI grade. The OARSI grade is determined histologically, and involves an invasive procedure that is not amenable to early screening or tracking disease progress. While MRI methods are non-invasive, they must first be correlated with histological grading schemes before they can be used in diagnosis or evaluating cartilage quality. We analyzed samples from cadavers that were imaged using several MRI modalities as well as scored histologically using the OARSI scale. Standard classification experiments using single MRI scans of degraded vs. non-degraded cartilage as determined by OARSI had accuracies > 80%, which are substantially higher than published reports correlating scores derived from MRI images with OARSI grades. Our multivariate regression of image features from multimodal MRI scans produced a continuous score that was well correlated with the OARSI grade of the same samples (r > 0.65, p < 10-5). Defining a continuous grading system based on a non-invasive procedure is a key element in evaluating osteoarthritis treatment strategies. A recently completed study being prepared for publication involves abdominal CT scans. The viscera, bone, subcutaneous and visceral fat in these scans has been segmented into separate image masks using the characteristic densities of these tissues on the Hounsfield scale. Our analysis of these image masks indicates that a strong aging signal is present in adipose tissues as well as in the unsegmented whole CT scans. These results are based on cross-validation of classifiers trained on a middle-age group (56-70) and an older group (81-99). This is by far the largest image-based aging study done in humans, and it clearly shows that adipose tissue is one of the major factors in age-related changes occuring in the abdomen. Another radiology project is in collaboration with Dr. Maria Knoll at the Johns Hopkins Bloomberg School of Public Health. Here the goal is to diagnose viral vs., bacterial pneumonia in children using chest X-rays. A rapid accurate diagnosis of the nature of this disease will dramatically improve outcomes for children in developing countries. A standard set of chest X-rays is available from the World Health Organization that is well annotated, with each X-ray having beenn read by mmultiple expert radiologists forming a solid ground truth for training machine classifiers. The major challenge posed by this set is the extreme variation in the physical size of the subjects due to the variation in their ages. We have developed a strategy to compensate for this size variation by working with a student at JHSPH to manually annotate the X-rays with a set of fiducial marks that are anatomically comparable across the X-rays regardless of subject size. Using these manually aligned regions of the lung, we have preliminary indications that such a diagnosis may be possible. We are currently assembling a completed manually annotated set from the WHO, as well as a different set of X-rays that has recently become available through JHSPH. A project recently begun involves analyzing fundus images collected by the National Eye Institute. The goal is two-fold: 1) Develop an automated scoring system for age-related macular degeneration (AMD), and use fundus images from SardiNIA for testing and validation using manual reads by our NEI collaborators. 2) Develop a set of numerical descriptors for segmented vasculature in fundus images, correlate them with cardiovascular traits measured in SardiNIA, and use them to determine genome-wide associations to these traits. Currently, preliminary experiments have been able to automatically detect an AMD signal in NEI fundus images, and thuse preliminary findings are being refined. Additionally, software has been identified to automatically segment vasculature from fundus images. Our work in developing tools and expertise in analyzing images to obtain physiological insights that are not directly observable has recently been translated to non-image-based clinical data. In this project, we used machine learning and the totality of the quantitative trait data collected in the SardiNIA study to ascertain each participant's age. This predicted age was excpected to closely correlate to chronological age, but as it was based on broad physiological measurements, be more closely related to each participant's physiological age. The ratio between physiological and chronological age can be viewed as a measure of aging rate, and we determined that this rate is largely preserved for each participant across multiple visits. Moreover, we determined that 40% of the variation in aging rates is heritable, and by performing a GWAS showed that it is significantly associated with two genes related to telomere function. This study is currently being prepared for publication. The use of machine learning and pattern recognition to mine new insights from numerical clinical data has great potential, and we are actively pursuing this strategy with new projects. In collaboration with Madhav Thambisetty and Kevin Becker, we are extending this approach to examine traits in BLSA participants that may be predictive of later diagnosis of Alzheimer's disease. In a collaboration with Matt Oberdier and Majd AlGhatrif (LCS) we are applying these tools to analyze data from blood flow and pressure sensors to act as markers of cardiovascular state in rats.
|Ashinsky, B G; Coletta, C E; Bouhrara, M et al. (2015) Machine learning classification of OARSI-scored human articular cartilage using magnetic resonance imaging. Osteoarthritis Cartilage 23:1704-12|
|Orlov, Nikita V; Eckley, D Mark; Shamir, Lior et al. (2012) Improving class separability using extended pixel planes: a comparative study. Mach Vis Appl 23:1047-1058|
|Orlov, Nikita V; Weeraratna, Ashani T; Hewitt, Stephen M et al. (2012) Automatic detection of melanoma progression by histological analysis of secondary sites. Cytometry A 81:364-73|