All Publications


  • Age-Dependent White Matter Characteristics of the Cerebellar Peduncles from Infancy Through Adolescence. Cerebellum (London, England) Bruckert, L., Shpanskaya, K., McKenna, E. S., Borchers, L. R., Yablonski, M., Blecher, T., Ben-Shachar, M., Travis, K. E., Feldman, H. M., Yeom, K. W. 2019

    Abstract

    Cerebellum-cerebrum connections are essential for many motor and cognitive functions and cerebellar disorders are prevalent in childhood. The middle (MCP), inferior (ICP), and superior cerebellar peduncles (SCP) are the major white matter pathways that permit communication between the cerebellum and the cerebrum. Knowledge about the microstructural properties of these cerebellar peduncles across childhood is limited. Here, we report on a diffusion magnetic resonance imaging tractography study to describe age-dependent characteristics of the cerebellar peduncles in a cross-sectional sample of infants, children, and adolescents from newborn to 17years of age (N=113). Scans were collected as part of clinical care; participants were restricted to those whose scans showed no abnormal findings and whose history and exam had no risk factors for cerebellar abnormalities. A novel automated tractography protocol was applied. Results showed that mean tract-FA increased, while mean tract-MD decreased from infancy to adolescence in all peduncles. Rapid changes were observed in both diffusion measures in the first 24months of life, followed by gradual change at older ages. The shape of the tract profiles was similar across ages for all peduncles. These data are the first to characterize the variability of diffusion properties both across and within cerebellar white matter pathways that occur from birth through later adolescence. The data represent a rich normative data set against which white matter alterations seen in children with posterior fossa conditions can be compared. Ultimately, the data will facilitate the identification of sensitive biomarkers of cerebellar abnormalities.

    View details for PubMedID 30637673

  • Deep-learning-assisted diagnosis for knee magnetic resonance imaging: Development and retrospective validation of MRNet. PLoS medicine Bien, N., Rajpurkar, P., Ball, R. L., Irvin, J., Park, A., Jones, E., Bereket, M., Patel, B. N., Yeom, K. W., Shpanskaya, K., Halabi, S., Zucker, E., Fanton, G., Amanatullah, D. F., Beaulieu, C. F., Riley, G. M., Stewart, R. J., Blankenberg, F. G., Larson, D. B., Jones, R. H., Langlotz, C. P., Ng, A. Y., Lungren, M. P. 2018; 15 (11): e1002699

    Abstract

    BACKGROUND: Magnetic resonance imaging (MRI) of the knee is the preferred method for diagnosing knee injuries. However, interpretation of knee MRI is time-intensive and subject to diagnostic error and variability. An automated system for interpreting knee MRI could prioritize high-risk patients and assist clinicians in making diagnoses. Deep learning methods, in being able to automatically learn layers of features, are well suited for modeling the complex relationships between medical images and their interpretations. In this study we developed a deep learning model for detecting general abnormalities and specific diagnoses (anterior cruciate ligament [ACL] tears and meniscal tears) on knee MRI exams. We then measured the effect of providing the model's predictions to clinical experts during interpretation.METHODS AND FINDINGS: Our dataset consisted of 1,370 knee MRI exams performed at Stanford University Medical Center between January 1, 2001, and December 31, 2012 (mean age 38.0 years; 569 [41.5%] female patients). The majority vote of 3 musculoskeletal radiologists established reference standard labels on an internal validation set of 120 exams. We developed MRNet, a convolutional neural network for classifying MRI series and combined predictions from 3 series per exam using logistic regression. In detecting abnormalities, ACL tears, and meniscal tears, this model achieved area under the receiver operating characteristic curve (AUC) values of 0.937 (95% CI 0.895, 0.980), 0.965 (95% CI 0.938, 0.993), and 0.847 (95% CI 0.780, 0.914), respectively, on the internal validation set. We also obtained a public dataset of 917 exams with sagittal T1-weighted series and labels for ACL injury from Clinical Hospital Centre Rijeka, Croatia. On the external validation set of 183 exams, the MRNet trained on Stanford sagittal T2-weighted series achieved an AUC of 0.824 (95% CI 0.757, 0.892) in the detection of ACL injuries with no additional training, while an MRNet trained on the rest of the external data achieved an AUC of 0.911 (95% CI 0.864, 0.958). We additionally measured the specificity, sensitivity, and accuracy of 9 clinical experts (7 board-certified general radiologists and 2 orthopedic surgeons) on the internal validation set both with and without model assistance. Using a 2-sided Pearson's chi-squared test with adjustment for multiple comparisons, we found no significant differences between the performance of the model and that of unassisted general radiologists in detecting abnormalities. General radiologists achieved significantly higher sensitivity in detecting ACL tears (p-value = 0.002; q-value = 0.019) and significantly higher specificity in detecting meniscal tears (p-value = 0.003; q-value = 0.019). Using a 1-tailed t test on the change in performance metrics, we found that providing model predictions significantly increased clinical experts' specificity in identifying ACL tears (p-value < 0.001; q-value = 0.006). The primary limitations of our study include lack of surgical ground truth and the small size of the panel of clinical experts.CONCLUSIONS: Our deep learning model can rapidly generate accurate clinical pathology classifications of knee MRI exams from both internal and external datasets. Moreover, our results support the assertion that deep learning models can improve the performance of clinical experts during medical imaging interpretation. Further research is needed to validate the model prospectively and to determine its utility in the clinical setting.

    View details for PubMedID 30481176

  • Deep learning for chest radiograph diagnosis: A retrospective comparison of the CheXNeXt algorithm to practicing radiologists. PLoS medicine Rajpurkar, P., Irvin, J., Ball, R. L., Zhu, K., Yang, B., Mehta, H., Duan, T., Ding, D., Bagul, A., Langlotz, C. P., Patel, B. N., Yeom, K. W., Shpanskaya, K., Blankenberg, F. G., Seekins, J., Amrhein, T. J., Mong, D. A., Halabi, S. S., Zucker, E. J., Ng, A. Y., Lungren, M. P. 2018; 15 (11): e1002686

    Abstract

    BACKGROUND: Chest radiograph interpretation is critical for the detection of thoracic diseases, including tuberculosis and lung cancer, which affect millions of people worldwide each year. This time-consuming task typically requires expert radiologists to read the images, leading to fatigue-based diagnostic error and lack of diagnostic expertise in areas of the world where radiologists are not available. Recently, deep learning approaches have been able to achieve expert-level performance in medical image interpretation tasks, powered by large network architectures and fueled by the emergence of large labeled datasets. The purpose of this study is to investigate the performance of a deep learning algorithm on the detection of pathologies in chest radiographs compared with practicing radiologists.METHODS AND FINDINGS: We developed CheXNeXt, a convolutional neural network to concurrently detect the presence of 14 different pathologies, including pneumonia, pleural effusion, pulmonary masses, and nodules in frontal-view chest radiographs. CheXNeXt was trained and internally validated on the ChestX-ray8 dataset, with a held-out validation set consisting of 420 images, sampled to contain at least 50 cases of each of the original pathology labels. On this validation set, the majority vote of a panel of 3 board-certified cardiothoracic specialist radiologists served as reference standard. We compared CheXNeXt's discriminative performance on the validation set to the performance of 9 radiologists using the area under the receiver operating characteristic curve (AUC). The radiologists included 6 board-certified radiologists (average experience 12 years, range 4-28 years) and 3 senior radiology residents, from 3 academic institutions. We found that CheXNeXt achieved radiologist-level performance on 11 pathologies and did not achieve radiologist-level performance on 3 pathologies. The radiologists achieved statistically significantly higher AUC performance on cardiomegaly, emphysema, and hiatal hernia, with AUCs of 0.888 (95% confidence interval [CI] 0.863-0.910), 0.911 (95% CI 0.866-0.947), and 0.985 (95% CI 0.974-0.991), respectively, whereas CheXNeXt's AUCs were 0.831 (95% CI 0.790-0.870), 0.704 (95% CI 0.567-0.833), and 0.851 (95% CI 0.785-0.909), respectively. CheXNeXt performed better than radiologists in detecting atelectasis, with an AUC of 0.862 (95% CI 0.825-0.895), statistically significantly higher than radiologists' AUC of 0.808 (95% CI 0.777-0.838); there were no statistically significant differences in AUCs for the other 10 pathologies. The average time to interpret the 420 images in the validation set was substantially longer for the radiologists (240 minutes) than for CheXNeXt (1.5 minutes). The main limitations of our study are that neither CheXNeXt nor the radiologists were permitted to use patient history or review prior examinations and that evaluation was limited to a dataset from a single institution.CONCLUSIONS: In this study, we developed and validated a deep learning algorithm that classified clinically important abnormalities in chest radiographs at a performance level comparable to practicing radiologists. Once tested prospectively in clinical settings, the algorithm could have the potential to expand patient access to chest radiograph diagnostics.

    View details for PubMedID 30457988

  • ASL PERFUSION IMAGING OF THE FRONTAL LOBES PREDICTS THE OCCURRENCE AND RESOLUTION OF POSTERIOR FOSSA SYNDROME Yecies, D., Shpanskaya, K., Grant, G., Cheshier, S., Hong, D., Edwards, M., Yeom, K. OXFORD UNIV PRESS INC. 2018: 170
  • Sex Differences in Cognitive Decline in Subjects with High Likelihood of Mild Cognitive Impairment due to Alzheimer's disease SCIENTIFIC REPORTS Sohn, D., Shpanskaya, K., Lucas, J. E., Petrella, J. R., Saykin, A. J., Tanzi, R. E., Samatova, N. F., Doraiswamy, P. 2018; 8: 7490

    Abstract

    Sex differences in Alzheimer's disease (AD) biology and progression are not yet fully characterized. The goal of this study is to examine the effect of sex on cognitive progression in subjects with high likelihood of mild cognitive impairment (MCI) due to Alzheimer's and followed up to 10 years in the Alzheimer's Disease Neuroimaging Initiative (ADNI). Cerebrospinal fluid total-tau and amyloid-beta (Aβ42) ratio values were used to sub-classify 559 MCI subjects (216 females, 343 males) as having "high" or "low" likelihood for MCI due to Alzheimer's. Data were analyzed using mixed-effects models incorporating all follow-ups. The worsening from baseline in Alzheimer's Disease Assessment Scale-Cognitive score (mean, SD) (9 ± 12) in subjects with high likelihood of MCI due to Alzheimer's was markedly greater than that in subjects with low likelihood (1 ± 6, p < 0.0001). Among MCI due to AD subjects, the mean worsening in cognitive score was significantly greater in females (11.58 ± 14) than in males (6.87 ± 11, p = 0.006). Our findings highlight the need to further investigate these findings in other populations and develop sex specific timelines for Alzheimer's disease progression.

    View details for PubMedID 29748598

  • Toward Personalized Network Biomarkers in Alzheimer's Disease: Computing Individualized Genomic and Protein Crosstalk Maps FRONTIERS IN AGING NEUROSCIENCE Padmanabhan, K., Shpanskaya, K., Bello, G., Doraiswamy, P., Samatova, N. F. 2017; 9: 315

    View details for PubMedID 29085293

  • Characterizing Gene and Protein Crosstalks in Subjects at Risk of Developing Alzheimer's Disease: A New Computational Approach PROCESSES Padmanabhan, K., Nudelman, K., Harenberg, S., Bello, G., Sohn, D., Shpanskaya, K., Dikshit, P., Yerramsetty, P. S., Tanzi, R. E., Saykin, A. J., Petrella, J. R., Doraiswamy, P., Samatova, N. F., Alzheimer's Dis Neuroimaging 2017; 5 (3)

    View details for DOI 10.3390/pr5030047

    View details for Web of Science ID 000412051700015

  • Intracranial pressure management Handbook of Neurosurgery, Neurology, and Spinal Medicine for Nurses and Advanced Practice Health Professionals Johnson, E., Yi-ren, C., Shpanskaya, K., Harris, O. 2017
  • Educational attainment and hippocampal atrophy in the Alzheimer's disease neuroimaging initiative cohort JOURNAL OF NEURORADIOLOGY Shpanskaya, K. S., Choudhury, K., Hostage, C., Murphy, K. R., Petrella, J. R., Doraiswamy, P., Alzheimer's Dis Neuroimaging Initi 2014; 41 (5): 350–57

    Abstract

    Subjects with higher cognitive reserve (CR) may be at a lower risk for Alzheimer's disease (AD), but the neural mechanisms underlying this are not known. Hippocampal volume loss is an early event in AD that triggers cognitive decline.Regression analyses of the effects of education on MRI-measured baseline HV in 675 subjects (201 normal, 329 with mild cognitive impairment (MCI), and 146 subjects with mild AD), adjusting for age, gender, APOE ɛ4 status and intracranial volume (ICV). Subjects were derived from the Alzheimer's Disease Neuroimaging Initiative (ADNI), a large US national biomarker study.The association between higher education and larger HV was significant in AD (P=0.014) but not in cognitively normal or MCI subjects. In AD, HV was about 8% larger in a person with 20 years of education relative to someone with 6 years of education. There was also a trend for the interaction between education and APOE ɛ4 to be significant in AD (P=0.056).A potential protective association between higher education and lower hippocampal atrophy in patients with AD appears consistent with prior epidemiologic data linking higher education levels with lower rates of incident dementia. Longitudinal studies are warranted to confirm these findings.

    View details for PubMedID 24485897

  • Impact of F-18-florbetapir PET imaging of beta-amyloid neuritic plaque density on clinical decision-making NEUROCASE Zannas, A. S., Doraiswamy, P., Shpanskaya, K. S., Murphy, K. R., Petrella, J. R., Burke, J. R., Wong, T. Z. 2014; 20 (4): 466–73

    Abstract

    ¹⁸F-florbetapir positron emission tomography (PET) imaging of the brain is now approved by the Food and Drug Administration (FDA) approved for estimation of β -amyloid neuritic plaque density when evaluating patients with cognitive impairment. However, its impact on clinical decision-making is not known. We present 11 cases (age range 67-84) of cognitively impaired subjects in whom clinician surveys were done before and after PET scanning to document the theoretical impact of amyloid imaging on the diagnosis and treatment plan of cognitively impaired subjects. Subjects have been clinically followed for about 5 months after the PET scan. Negative scans occurred in five cases, leading to a change in diagnosis for four patients and a change in treatment plan for two of these cases. Positive scans occurred in six cases, leading to a change in diagnosis for four patients and a change in treatment plan for three of these cases. Following the scan, only one case had indeterminate diagnosis. Our series suggests that both positive and negative florbetapir PET scans may enhance diagnostic certainty and impact clinical decision-making. Controlled longitudinal studies are needed to confirm our data and determine best practices.

    View details for PubMedID 23672654

  • Mapping the effects of ApoE4, age and cognitive status on 18F-florbetapir PET measured regional cortical patterns of beta-amyloid density and growth NEUROIMAGE Murphy, K. R., Landau, S. M., Choudhury, K., Hostage, C. A., Shpanskaya, K. S., Sair, H. I., Petrella, J. R., Wong, T. Z., Doraiswamy, P., Alzheimer's Dis Neuroimaging 2013; 78: 474–80

    Abstract

    Although it is well known that many clinical and genetic factors have been associated with beta-amyloid deposition, few studies have examined the interactions of such factors across different stages of Alzheimer's pathogenesis.We used 18F-florbetapir F18 PET imaging to quantify neuritic beta-amyloid plaque density across four cortical regions in 602 elderly (55-94 years) subjects from the national ADNI biomarker study. The group comprised of 194 normal elderly, 212 early mild cognitive impairment [EMCI], 132 late mild cognitive impairment [LMCI], and 64 mild Alzheimer's (AD).In a model incorporating multiple predictive factors, the effect of apolipoprotein E ε4 and diagnosis was significant on all four cortical regions. The highest signals were seen in cingulate followed by frontal and parietal with lowest signals in temporal lobe (p<0.0001). The effect of apolipoprotein E ε4 (Cohen's D 0.96) on beta-amyloid plaque density was approximately twice as large as the effect of a diagnosis of AD (Cohen's D 0.51) and thrice as large as the effect of a diagnosis of LMCI (Cohen's D 0.34) (p<0.0001). Surprisingly, ApoE ε4+ normal controls had greater mean plaque density across all cortical regions than ε4- EMCI and ε4- LMCI (p<0.0001, p=0.0009) and showed higher, though non-significant, mean value than ε4- AD patients (p<0.27). ApoE ε4+ EMCI and LMCI subjects had significantly greater mean plaque density across all cortical regions than ε4- AD patients (p<0.027, p<0.0001).Neuritic amyloid plaque load across progressive clinical stages of AD varies strongly by ApoE4 genotype. These findings support the need for better pathology-based and supported diagnosis in routine practice. Our data also provides additional evidence for a temporal offset between amyloid deposition and clinically relevant symptoms.

    View details for PubMedID 23624169

    View details for PubMedCentralID PMC3749874

  • SPICE: discovery of phenotype-determining component interplays BMC SYSTEMS BIOLOGY Chen, Z., Padmanabhan, K., Rocha, A. M., Shpanskaya, Y., Mihelcic, J. R., Scott, K., Samatova, N. F. 2012; 6: 40

    Abstract

    A latent behavior of a biological cell is complex. Deriving the underlying simplicity, or the fundamental rules governing this behavior has been the Holy Grail of systems biology. Data-driven prediction of the system components and their component interplays that are responsible for the target system's phenotype is a key and challenging step in this endeavor.The proposed approach, which we call System Phenotype-related Interplaying Components Enumerator (SPICE), iteratively enumerates statistically significant system components that are hypothesized (1) to play an important role in defining the specificity of the target system's phenotype(s); (2) to exhibit a functionally coherent behavior, namely, act in a coordinated manner to perform the phenotype-specific function; and (3) to improve the predictive skill of the system's phenotype(s) when used collectively in the ensemble of predictive models. SPICE can be applied to both instance-based data and network-based data. When validated, SPICE effectively identified system components related to three target phenotypes: biohydrogen production, motility, and cancer. Manual results curation agreed with the known phenotype-related system components reported in literature. Additionally, using the identified system components as discriminatory features improved the prediction accuracy by 10% on the phenotype-classification task when compared to a number of state-of-the-art methods applied to eight benchmark microarray data sets.We formulate a problem--enumeration of phenotype-determining system component interplays--and propose an effective methodology (SPICE) to address this problem. SPICE improved identification of cancer-related groups of genes from various microarray data sets and detected groups of genes associated with microbial biohydrogen production and motility, many of which were reported in literature. SPICE also improved the predictive skill of the system's phenotype determination compared to individual classifiers and/or other ensemble methods, such as bagging, boosting, random forest, nearest shrunken centroid, and random forest variable selection method.

    View details for PubMedID 22583800

  • NIBBS-Search for Fast and Accurate Prediction of Phenotype-Biased Metabolic Systems PLOS COMPUTATIONAL BIOLOGY Schmidt, M. C., Rocha, A. M., Padmanabhan, K., Shpanskaya, Y., Banfield, J., Scott, K., Mihelcic, J. R., Samatova, N. F. 2012; 8 (5): e1002490

    Abstract

    Understanding of genotype-phenotype associations is important not only for furthering our knowledge on internal cellular processes, but also essential for providing the foundation necessary for genetic engineering of microorganisms for industrial use (e.g., production of bioenergy or biofuels). However, genotype-phenotype associations alone do not provide enough information to alter an organism's genome to either suppress or exhibit a phenotype. It is important to look at the phenotype-related genes in the context of the genome-scale network to understand how the genes interact with other genes in the organism. Identification of metabolic subsystems involved in the expression of the phenotype is one way of placing the phenotype-related genes in the context of the entire network. A metabolic system refers to a metabolic network subgraph; nodes are compounds and edges labels are the enzymes that catalyze the reaction. The metabolic subsystem could be part of a single metabolic pathway or span parts of multiple pathways. Arguably, comparative genome-scale metabolic network analysis is a promising strategy to identify these phenotype-related metabolic subsystems. Network Instance-Based Biased Subgraph Search (NIBBS) is a graph-theoretic method for genome-scale metabolic network comparative analysis that can identify metabolic systems that are statistically biased toward phenotype-expressing organismal networks. We set up experiments with target phenotypes like hydrogen production, TCA expression, and acid-tolerance. We show via extensive literature search that some of the resulting metabolic subsystems are indeed phenotype-related and formulate hypotheses for other systems in terms of their role in phenotype expression. NIBBS is also orders of magnitude faster than MULE, one of the most efficient maximal frequent subgraph mining algorithms that could be adjusted for this problem. Also, the set of phenotype-biased metabolic systems output by NIBBS comes very close to the set of phenotype-biased subgraphs output by an exact maximally-biased subgraph enumeration algorithm ( MBS-Enum ). The code (NIBBS and the module to visualize the identified subsystems) is available at http://freescience.org/cs/NIBBS.

    View details for PubMedID 22589706

    View details for PubMedCentralID PMC3349732