- Diagnostic Radiology
Clinical Assistant Professor, Radiology
Program Director - Neuroradiology Fellowship, Stanford Hospital (2021 - Present)
Associate Program Director - Neuroradiology Fellowship, Stanford Hospital (2018 - 2021)
Board Certification: American Board of Radiology, Neuroradiology (2015)
Fellowship: Stanford University Neuroradiology Fellowship (2015) CA
Board Certification: American Board of Radiology, Diagnostic Radiology (2013)
Residency: University of Texas Health Sciences Center at San Antonio Radiology Residency (2013) TX
Internship: Baylor College of Medicine Internal Medicine Residency (2009) TX
Medical Education: University of Texas Southwestern Medical School Registrar (2008) TX
Additional Clinical Info
Graduate and Fellowship Programs
Neuroradiology (Fellowship Program)
- Generative Editing via Convolutional Obscuring (GECO): A Generative Adversarial Network for MRI de-artifacting LIPPINCOTT WILLIAMS & WILKINS. 2023
Neuroradiologic Evaluation of MRI in High-Contact Sports.
Frontiers in neurology
2021; 12: 701948
Background and Purpose: Athletes participating in high-contact sports experience repeated head trauma. Anatomical findings, such as a cavum septum pellucidum, prominent CSF spaces, and hippocampal volume reductions, have been observed in cases of mild traumatic brain injury. The extent to which these neuroanatomical findings are associated with high-contact sports is unknown. The purpose of this study was to determine whether there are subtle neuroanatomic differences between athletes participating in high-contact sports compared to low-contact athletic controls. Materials and Methods: We performed longitudinal structural brain MRI scans in 63 football (high-contact) and 34 volleyball (low-contact control) male collegiate athletes with up to 4 years of follow-up, evaluating a total of 315 MRI scans. Board-certified neuroradiologists performed semi-quantitative visual analysis of neuroanatomic findings, including: cavum septum pellucidum type and size, extent of perivascular spaces, prominence of CSF spaces, white matter hyperintensities, arterial spin labeling perfusion asymmetries, fractional anisotropy holes, and hippocampal size. Results: At baseline, cavum septum pellucidum length was greater in football compared to volleyball controls (p = 0.02). All other comparisons were statistically equivalent after multiple comparison correction. Within football at baseline, the following trends that did not survive multiple comparison correction were observed: more years of prior football exposure exhibited a trend toward more perivascular spaces (p = 0.03 uncorrected), and lower baseline Standardized Concussion Assessment Tool scores toward more perivascular spaces (p = 0.02 uncorrected) and a smaller right hippocampal size (p = 0.02 uncorrected). Conclusion: Head impacts in high-contact sport (football) athletes may be associated with increased cavum septum pellucidum length compared to low-contact sport (volleyball) athletic controls. Other investigated neuroradiology metrics were generally equivalent between sports.
View details for DOI 10.3389/fneur.2021.701948
View details for PubMedID 34456852
View details for PubMedCentralID PMC8385770
- Neuroradiologic Evaluation of MRI in High-Contact Sports FRONTIERS IN NEUROLOGY 2021; 12
Realistic generation of diffusion-weighted magnetic resonance brain images with deep generative models.
Magnetic resonance imaging
We study two state of the art deep generative networks, the Introspective Variational Autoencoder and the Style-Based Generative Adversarial Network, for the generation of new diffusion-weighted magnetic resonance images. We show that high quality, diverse and realistic-looking images, as evaluated by external neuroradiologists blinded to the whole study, can be synthesized using these deep generative models. We evaluate diverse metrics with respect to quality and diversity of the generated synthetic brain images. These findings show that generative models could qualify as a method for data augmentation in the medical field, where access to large image database is in many aspects restricted.
View details for DOI 10.1016/j.mri.2021.06.001
View details for PubMedID 34116133
Brainstem Atrophy in Gulf War Illness.
BACKGROUND: Gulf War illness (GWI) is a condition that affects about 30% of veterans who served in the 1990-91 Persian Gulf War. Given its broad symptomatic manifestation, including chronic pain, fatigue, neurological, gastrointestinal, respiratory, and skin problems, it is of interest to examine whether GWI is associated with changes in the brain. Existing neuroimaging studies, however, have been limited by small sample sizes, inconsistent GWI diagnosis criteria, and potential comorbidity confounds.OBJECTIVES: Using a large cohort of US veterans with GWI, we assessed regional brain volumes for their associations with GWI, and quantified the relationships between any regional volumetric changes and GWI symptoms.METHODS: Structural magnetic resonance imaging (MRI) scans from 111 veterans with GWI (Age=49±6, 88% Male) and 59 healthy controls (age=51±9, 78% male) were collected at the California War Related Illness and Injury Study Center (WRIISC-CA) and from a multicenter study of the Parkinson's Progression Marker Initiative (PPMI), respectively. Individual MRI volumes were segmented and parcellated using FreeSurfer. Regional volumes of 19 subcortical, 68 cortical, and 3 brainstem structures were evaluated in the GWI cohort relative to healthy controls. The relationships between regional volumes and GWI symptoms were also assessed.RESULTS: We found significant subcortical atrophy, but no cortical differences, in the GWI group relative to controls, with the largest effect detected in the brainstem, followed by the ventral diencephalon and the thalamus. In a subsample of 58 veterans with GWI who completed the Chronic Fatigue Scale (CFS) inventory of Centers for Disease Control and Prevention (CDC), smaller brainstem volumes were significantly correlated with increased severities of fatigue and depressive symptoms.CONCLUSION: The findings suggest that brainstem volume may be selectively affected by GWI, and that the resulting atrophy could in turn mediate or moderate GWI-related symptoms such as fatigue and depression. Consequently, the brain stem should be carefully considered in future research focusing on GWI pathology.
View details for DOI 10.1016/j.neuro.2020.02.006
View details for PubMedID 32081703
MR susceptibility contrast imaging using a 2D simultaneous multi-slice gradient-echo sequence at 7T.
2019; 14 (7): e0219705
PURPOSE: To develop a 7T simultaneous multi-slice (SMS) 2D gradient-echo sequence for susceptibility contrast imaging, and to compare its quality to 3D imaging.METHODS: A frequency modulated and phase cycled RF pulse was designed to simultaneously excite multiple slices in multi-echo 2D gradient-echo imaging. The imaging parameters were chosen to generate images with susceptibility contrast, including T2*-weighted magnitude/phase images, susceptibility-weighted images and quantitative susceptibility/R2* maps. To compare their image quality with 3D gradient-echo imaging, both 2D and 3D imaging were performed on 11 healthy volunteers and 4 patients with multiple sclerosis (MS). The signal to noise ratio (SNR) in gray and white matter and their contrast to noise ratio (CNR) was simulated for the 2D and 3D magnitude images using parameters from the imaging. The experimental SNRs and CNRs were measured in gray/white matter and deep gray matter structures on magnitude, phase, R2* and QSM images from volunteers and the visibility of MS lesions on these images from patients was visually rated. All SNRs and CNRs were compared between the 2D and 3D imaging using a paired t-test.RESULTS: Although the 3D magnitude images still had significantly higher SNRs (by 13.0~17.6%), the 2D magnitude and QSM images generated significantly higher gray/white matter or globus pallidus/putamen contrast (by 13.3~87.5%) and significantly higher MS lesion contrast (by 5.9~17.3%).CONCLUSION: 2D SMS gradient-echo imaging can serve as an alternative to often used 3D imaging to obtain susceptibility-contrast-weighted images, with an advantage of providing better image contrast and MS lesion sensitivity.
View details for DOI 10.1371/journal.pone.0219705
View details for PubMedID 31314813
MR Imaging-Based Radiomic Signatures of Distinct Molecular Subgroups of Medulloblastoma.
AJNR. American journal of neuroradiology
Distinct molecular subgroups of pediatric medulloblastoma confer important differences in prognosis and therapy. Currently, tissue sampling is the only method to obtain information for classification. Our goal was to develop and validate radiomic and machine learning approaches for predicting molecular subgroups of pediatric medulloblastoma.In this multi-institutional retrospective study, we evaluated MR imaging datasets of 109 pediatric patients with medulloblastoma from 3 children's hospitals from January 2001 to January 2014. A computational framework was developed to extract MR imaging-based radiomic features from tumor segmentations, and we tested 2 predictive models: a double 10-fold cross-validation using a combined dataset consisting of all 3 patient cohorts and a 3-dataset cross-validation, in which training was performed on 2 cohorts and testing was performed on the third independent cohort. We used the Wilcoxon rank sum test for feature selection with assessment of area under the receiver operating characteristic curve to evaluate model performance.Of 590 MR imaging-derived radiomic features, including intensity-based histograms, tumor edge-sharpness, Gabor features, and local area integral invariant features, extracted from imaging-derived tumor segmentations, tumor edge-sharpness was most useful for predicting sonic hedgehog and group 4 tumors. Receiver operating characteristic analysis revealed superior performance of the double 10-fold cross-validation model for predicting sonic hedgehog, group 3, and group 4 tumors when using combined T1- and T2-weighted images (area under the curve = 0.79, 0.70, and 0.83, respectively). With the independent 3-dataset cross-validation strategy, select radiomic features were predictive of sonic hedgehog (area under the curve = 0.70-0.73) and group 4 (area under the curve = 0.76-0.80) medulloblastoma.This study provides proof-of-concept results for the application of radiomic and machine learning approaches to a multi-institutional dataset for the prediction of medulloblastoma subgroups.
View details for DOI 10.3174/ajnr.A5899
View details for PubMedID 30523141
Imaging Evaluation of the Adult Presenting With New-Onset Seizure.
AJR. American journal of roentgenology
OBJECTIVE: The purpose of this study is to discuss the evidence supporting the use of neuroimaging in adult patients presenting with new-onset seizure.CONCLUSION: Unenhanced CT should be the initial imaging examination performed for adults presenting with first unprovoked seizure in the acute setting to exclude conditions requiring urgent or emergent intervention. MRI has added benefits and should be considered for adults presenting acutely for whom the initial CT is negative and for those presenting with new-onset seizure in the nonacute setting.
View details for PubMedID 30299997
Contemporary Imaging of Cerebral Arteriovenous Malformations.
AJR. American journal of roentgenology
Brain arteriovenous malformation (AVM) rupture results in substantial morbidity and mortality. The goal of AVM treatment is eradication of the AVM, but the risk of treatment must be weighed against the risk of future hemorrhage.Imaging plays a vital role by providing the information necessary for AVM management. Here, we discuss the background, natural history, clinical presentation, and imaging of AVMs. In addition, we explain advances in techniques for imaging AVMs.
View details for DOI 10.2214/AJR.16.17306
View details for PubMedID 28267351
In Vivo 7T MR Quantitative Susceptibility Mapping Reveals Opposite Susceptibility Contrast between Cortical and White Matter Lesions in Multiple Sclerosis
AMERICAN JOURNAL OF NEURORADIOLOGY
2016; 37 (10): 1808-1815
Magnetic susceptibility measured with quantitative susceptibility mapping has been proposed as a biomarker for demyelination and inflammation in patients with MS, but investigations have mostly been on white matter lesions. A detailed characterization of cortical lesions has not been performed. The purpose of this study was to evaluate magnetic susceptibility in both cortical and WM lesions in MS by using quantitative susceptibility mapping.Fourteen patients with MS were scanned on a 7T MR imaging scanner with T1-, T2-, and T2*-weighted sequences. The T2*-weighted sequence was used to perform quantitative susceptibility mapping and generate tissue susceptibility maps. The susceptibility contrast of a lesion was quantified as the relative susceptibility between the lesion and its adjacent normal-appearing parenchyma. The susceptibility difference between cortical and WM lesions was assessed by using a t test.The mean relative susceptibility was significantly negative for cortical lesions (P < 10(-7)) but positive for WM lesions (P < 10(-22)). A similar pattern was also observed in the cortical (P = .054) and WM portions (P = .043) of mixed lesions.The negative susceptibility in cortical lesions suggests that iron loss dominates the susceptibility contrast in cortical lesions. The opposite susceptibility contrast between cortical and WM lesions may reflect both their structural (degree of myelination) and pathologic (degree of inflammation) differences, in which the latter may lead to a faster release of iron in cortical lesions.
View details for DOI 10.3174/ajnr.A4830
View details for Web of Science ID 000383984600014
View details for PubMedID 27282860
Imaging Neck Masses in the Neonate and Young Infant
SEMINARS IN ULTRASOUND CT AND MRI
2015; 36 (2): 120-137
Head and neck masses occurring in the neonatal period and early infancy consist of vascular tumors, vascular malformations, benign and malignant soft tissue tumors, and other developmental lesions. Although some lesions can be diagnosed on clinical grounds, others can only be diagnosed by imaging. Beyond diagnosis, imaging plays a significant role in evaluating the location and extent of a lesion for possible intervention. In this article, we review the clinical presentation and imaging appearance of common and rare masses that may be encountered in this age group. We also highlight current treatment strategies for specific lesions.
View details for DOI 10.1053/j.sult.2015.01.004
View details for Web of Science ID 000355575300003
View details for PubMedID 26001942