Dow-Mu Koh

Institute of Cancer Research, Sutton, ENG, United Kingdom

Are you Dow-Mu Koh?

Claim your profile

Publications (61)207.98 Total impact

  • [Show abstract] [Hide abstract]
    ABSTRACT: Liver perfusion magnetic resonance (MR) imaging is currently being actively investigated as a functional imaging technique that provides physiologic information on the microcirculation and microenvironment of liver tumors and the underlying liver. It has gained importance in light of antiangiogenic therapy for hepatocellular carcinoma and colorectal liver metastases. This article explains the various model-free and model-based approaches for liver perfusion MR imaging and their relative clinical utility. Relevant published works are summarized for each approach so that the reader can understand their relative strengths and weaknesses, to make an informed choice when performing liver perfusion MR imaging studies.
    Magnetic Resonance Imaging Clinics of North America. 08/2014; 22(3):417–432.
  • [Show abstract] [Hide abstract]
    ABSTRACT: Purpose To compare revised Choi criteria that incorporate concurrent size and attenuation changes at early follow-up imaging with Response Evaluation Criteria in Solid Tumors (RECIST) 1.1 and original Choi criteria in stratification of clinical outcomes in patients with metastatic renal cell carcinoma (mRCC) treated with sunitinib. Materials and Methods Institutional review board approved this retrospective study and waived informed consent. Baseline and first follow-up computed tomographic scans in 69 patients (50 men, 19 women; mean age, 60.3 years; range, 19-83 years) with mRCC treated with sunitinib from October 1, 2008, to March 1, 2013, were evaluated for tumor response by using RECIST 1.1, original Choi criteria, and revised Choi criteria. Correlations with overall survival (OS) and progression-free survival (PFS) were compared and stratified according to each radiologic criteria with Kaplan-Meier and multivariate Cox regression analysis. Results Median follow-up time was 29.7 months (95% confidence interval [CI]: 18.9, 45.9). Response according to revised Choi criteria was independently correlated with OS (hazard ratio, 0.47 [95% CI: 0.23, 0.99]; P = .046) and PFS (hazard ratio, 0.53 [95% CI: 0.29, 0.99]; P = .047). Response according to RECIST was not significantly correlated with OS (hazard ratio, 0.65 [95% CI: 0.27, 1.58]; P = .344) or PFS (hazard ratio, 0.89 [95% CI: 0.42, 1.91]; P = .768). Response according to original Choi criteria was not significantly correlated with OS (hazard ratio, 0.60 [95% CI: 0.32, 1.11]; P = .106) or PFS (hazard ratio, 0.59 [95% CI: 0.34, 1.02]; P = .060). Median OS and PFS in responders according to revised Choi criteria was 39.4 months (95% CI: 9.1, upper limit not estimated) and 13.7 months (95% CI: 6.4, 24.6), respectively, compared with 12.8 months (95% CI: 8.7, 18.0) and 5.3 months (95% CI: 3.9, 8.4), respectively, in nonresponders. Conclusion Contemporaneous reduction in tumor size and attenuation were correlated with favorable clinical outcomes. Response according to revised Choi criteria showed better correlation with clinical outcomes compared with that according to RECIST or original Choi criteria in patients with mRCC treated with sunitinib. © RSNA, 2014.
    Radiology 05/2014; · 6.34 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: The intra-voxel incoherent motion (IVIM) model relates signal decay on diffusion-weighted MRI (DWI) to tissue characteristics of perfusion and diffusivity. We explored the effects of altering the diffusion time of DWI sequences on IVIM perfusion (f,D*) and diffusion (D) related parameters in abdominal organs of healthy volunteers. At longer diffusion times, the calculated perfusion fraction (f) was significantly increased while the effect on the D and D* parameters was more variable, depending on the organ under study. The diffusion time has a significant impact on measured IVIM parameters in abdominal organs and should be reported when employing the IVIM model.
    Joint Annual Meeting ISMRM-ESMRMB, Milan, Italy; 05/2014
  • [Show abstract] [Hide abstract]
    ABSTRACT: We investigated how the technical differences between ss-EPI(R=2), FLASH(R=2) and FLASH(R=3) ACS acquisition methods visually manifest themselves in the context of WBDWI imaging. We analyzed: geometric distortion, SNR, overall image quality, and inter-station registration. FLASH(R=3) had the highest geometric fidelity, fewer image artifacts, and good station-station alignment. These advantages are offset by a lower SNR of ~17%: this may be partially responsible for the tendency of FLASH(R=2) ACS data to be rated higher than FLASH(R = 3) ACS data in terms of overall MIP quality. There was no situation in this study in which ss-EPI outperformed FLASH.
    Joint Annual Meeting ISMRM-ESMRMB, Milan, Italy; 05/2014
  • [Show abstract] [Hide abstract]
    ABSTRACT: Respiratory motion commonly confounds abdominal DWI, and motion minimisation strategies adversely affect scan efficiency and comfort. Blurring is due to post-acquisition combination of images from separate signal averages and diffusion-gradient directions, and is not inherent to the images. In a volunteer cohort where all images were stored separately, taking a (voxel-by-voxel) median image instead of a mean at each b-value yields parameter maps with much improved sharpness while still retaining tissue features. ADCs from ROIs in liver and kidneys were 108±18 vs 120±26 (p=0.007) and 182±17 vs 188±13 (p=0.04) x10-5 mm2s-1 for median and mean, respectively.
    Joint Annual Meeting ISMRM-ESMRMB, Milan, Italy; 05/2014
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The early identification of children presenting ALKF1174L-mutated neuroblastoma, which are associated with resistance to the promising ALK inhibitor crizotinib and a marked poorer prognosis, has become a clinical priority. In comparing the radiology of the novel Th-ALKF1174L/Th-MYCN and the well-established Th-MYCN genetically-engineered murine models of neuroblastoma using MRI, we have identified a marked ALKF1174L-driven vascular phenotype. We demonstrate that quantitation of the transverse relaxation rate R2* (s-1) using intrinsic susceptibility-MRI under baseline conditions and during hyperoxia, can robustly discriminate this differential vascular phenotype, and identify MYCN-driven tumors harboring the ALKF1174L mutation with high specificity and selectivity. Intrinsic susceptibility-MRI could thus potentially provide a non-invasive and clinically-exploitable method to help identifying children with MYCN-driven neuroblastoma harboring the ALKF1174L mutation at the time of diagnosis.
    PLoS ONE 01/2014; 9(3):e92886. · 3.73 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Autophagy is a highly regulated, energy dependent cellular process where proteins, organelles and cytoplasm are sequestered in autophagosomes and digested to sustain cellular homeostasis. We hypothesized that during autophagy induced in cancer cells by i) starvation through serum and amino acid deprivation or ii) treatment with PI-103, a class I PI3K/mTOR inhibitor, glycolytic metabolism would be affected, reducing flux to lactate, and that this effect may be reversible. We probed metabolism during autophagy in colorectal HT29 and HCT116 Bax knock-out cells using hyperpolarized 13C-magnetic resonance spectroscopy (MRS) and steady-state 1H-MRS. 24 hr PI103-treatment or starvation caused significant reduction in the apparent forward rate constant (kPL) for pyruvate to lactate exchange compared with controls in HT29 (100 μM PI-103: 82%, p = 0.05) and HCT116 Bax-ko cells (10 μM PI-103: 53%, p = 0.05; 20 μM PI-103: 42%, p<0.0001; starvation: 52%, p<0.001), associated with reduced lactate excretion and intracellular lactate in all cases, and unchanged lactate dehydrogenase (LDH) activity and increased NAD+/NADH ratio following PI103 treatment or decreased LDH activity and unchanged NAD+/NADH ratio following starvation. After 48 hr recovery from PI103 treatment, kPL remained below control levels in HT29 cells (74%, p = 0.02), and increased above treated values, but remained below 24 hr vehicle-treated control levels in HCT116 Bax-ko cells (65%, p = 0.004) both were accompanied by sustained reduction in lactate excretion, recovery of NAD+/NADH ratio and intracellular lactate. Following recovery from starvation, kPL was significantly higher than 24 hr vehicle-treated controls (140%, p = 0.05), associated with increased LDH activity and total cellular NAD(H). Changes in kPL and cellular and excreted lactate provided measureable indicators of the major metabolic processes accompanying starvation- and drug-induced autophagy. The changes are reversible, returning towards and exceeding control values on cellular recovery, which potentially identifies resistance. kPL (hyperpolarized 13C-MRS) and lactate (1H-MRS) provide useful biomarkers for the autophagic process, enabling non-invasive monitoring of the Warburg effect.
    PLoS ONE 01/2014; 9(3):e92645. · 3.73 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: BACKGROUND: Dichloroacetate (DCA) has been found to have antitumour properties. METHODS: We investigated the cellular and metabolic responses to DCA treatment and recovery in human colorectal (HT29, HCT116 WT and HCT116 Bax-ko), prostate carcinoma cells (PC3) and HT29 xenografts by flow cytometry, western blotting, electron microscopy, (1)H and hyperpolarised (13)C-magnetic resonance spectroscopy. RESULTS: Increased expression of the autophagy markers LC3B II was observed following DCA treatment both in vitro and in vivo. We observed increased production of reactive oxygen species (ROS) and mTOR inhibition (decreased pS6 ribosomal protein and p4E-BP1 expression) as well as increased expression of MCT1 following DCA treatment. Steady-state lactate excretion and the apparent hyperpolarised [1-(13)C] pyruvate-to-lactate exchange rate (kPL) were decreased in DCA-treated cells, along with increased NAD(+)/NADH ratios and NAD(+). Steady-state lactate excretion and kPL returned to, or exceeded, control levels in cells recovered from DCA treatment, accompanied by increased NAD(+) and NADH. Reduced kPL with DCA treatment was found in HT29 tumour xenografts in vivo. CONCLUSIONS: DCA induces autophagy in cancer cells accompanied by ROS production and mTOR inhibition, reduced lactate excretion, reduced kPL and increased NAD(+)/NADH ratio. The observed cellular and metabolic changes recover on cessation of treatment.
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: We describe our semi-automatic segmentation of whole-body diffusion-weighted MRI (WBDWI) using a Markov random field (MRF) model to derive tumor total diffusion volume (tDV) and associated global apparent diffusion coefficient (gADC); and demonstrate the feasibility of using these indices for assessing tumor burden and response to treatment in patients with bone metastases. WBDWI was performed on eleven patients diagnosed with bone metastases from breast and prostate cancers before and after anti-cancer therapies. Semi-automatic segmentation incorporating a MRF model was performed in all patients below the C4 vertebra by an experienced radiologist with over eight years of clinical experience in body DWI. Changes in tDV and gADC distributions were compared with overall response determined by all imaging, tumor markers and clinical findings at serial follow up. The segmentation technique was possible in all patients although erroneous volumes of interest were generated in one patient because of poor fat suppression in the pelvis, requiring manual correction. Responding patients showed a larger increase in gADC (median change = +0.18, range = -0.07 to +0.78×10-3 mm2/s) after treatment compared to non-responding patients (median change = -0.02, range = -0.10 to +0.05×10-3 mm2/s, p = 0.05, Mann-Whitney test), whereas non-responding patients showed a significantly larger increase in tDV (median change = +26%, range = +3 to +284%) compared to responding patients (median change = -50%, range = -85 to +27%, p = 0.02, Mann-Whitney test). Semi-automatic segmentation of WBDWI is feasible for metastatic bone disease in this pilot cohort of 11 patients, and could be used to quantify tumor total diffusion volume and median global ADC for assessing response to treatment.
    PLoS ONE 01/2014; 9(4):e91779. · 3.73 Impact Factor
  • The British journal of radiology 08/2013; · 2.11 Impact Factor
  • PET Clinics 07/2013; 8(3):259–277.
  • [Show abstract] [Hide abstract]
    ABSTRACT: Diffusion-weighted MRI in the body must account for a microcirculation fraction, separate to self-diffusion, within imaging voxels. Explicit control of diffusion pulse length and delay allows reproducible application of diffusion gradients with varying echo times; calculation of mono-exponential T2 estimates with applied gradients of b=0 and b=200 s/mm2 shows significant changes observed in liver, kidney and spleen. This suggests that the microcirculation component, with its own distinct T2, is being removed, allowing the generation of flow-free T2 maps more robustly estimating tissue T2s. This approach enables appropriate b-value choices when considering diffusion models that include or exclude microcirculation contribution.
    ISMRM Annual Meeting & Exhibition, Salt Lake City, Utah, USA; 04/2013
  • [Show abstract] [Hide abstract]
    ABSTRACT: Purpose:To determine whether perfusion computed tomography (CT)-derived vascular parameters-namely, blood flow, mean transit time (MTT), volume transfer constant (K(trans)), permeability-surface area product (PS), extracellular extravascular space volume, and vascular volume-correlate with the immunohistologic markers of angiogenesis in colorectal tumors.Materials and Methods:This prospective study was approved by the Regional Ethics and Research and Development Committees. The perfusion CT protocol was incorporated in the staging CT after informed consent in 29 patients (14 men, 15 women; mean age, 70 years; age range, 55-94 years). The perfusion parameters were calculated over two regions of interest (ROIs), at the invasive and luminal site defined by two radiologists independently. Accurate representative data were captured manually by correcting for motion artifacts and were analyzed by using Matlab software. The vascular heterogeneity between ROIs was assessed by using the Wilcoxon signed rank test. Perfusion CT parameters were correlated with the microvessel density (MVD) count at both corresponding sites obtained by means of immunohistochemical staining of the selected histologic slide with factor VIII and CD105 antigens by using Spearmen rank coefficient.Results:There was no statistically significant difference found between perfusion CT vascular parameters at the two ROIs by either of the radiologists. The Pearson coefficient for blood flow, MTT, K(trans), and PS at the two ROIs demonstrated good to moderate interobserver variability (for the two ROIs, 0.46 and 0.44; 0.67 and 0.64; 0.41 and 0.72; and 0.86 and 0.56, respectively). None of these parameters correlated with MVD count at the invasive or the luminal site for either of the two antigens.Conclusion:Perfusion CT measurements may measure vascularity of colorectal tumors, however, correlation with MVD, which is a morphologic measure, appears inappropriate.© RSNA, 2013.
    Radiology 04/2013; · 6.34 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: PURPOSE: To evaluate the effect on diffusion-weighted image-derived parameters in the apparent diffusion coefficient (ADC) and intra-voxel incoherent motion (IVIM) models from choice of either free-breathing or navigator-controlled acquisition. MATERIALS AND METHODS: Imaging was performed with consent from healthy volunteers (n = 10) on a 1.5T Siemens Avanto scanner. Parameter-matched free-breathing and navigator-controlled diffusion-weighted images were acquired, without averaging in the console, for a total scan time of ∼10 minutes. Regions of interest were drawn for renal cortex, renal pyramid, whole kidney, liver, spleen, and paraspinal muscle. An ADC diffusion model for these regions was fitted for b-values ≥ 250 s/mm(2) , using a Levenberg-Marquardt algorithm, and an IVIM model was fitted for all images using a Bayesian method. RESULTS: ADC and IVIM parameters from the two acquisition regimes show no significant differences for the cohort; individual cases show occasional discrepancies, with outliers in parameter estimates arising more commonly from navigator-controlled scans. The navigator-controlled acquisitions showed, on average, a smaller range of movement for the kidneys (6.0 ± 1.4 vs. 10.0 ± 1.7 mm, P = 0.03), but also a smaller number of averages collected (3.9 ± 0.1 vs. 5.5 ± 0.2, P < 0.01) in the allocated time. CONCLUSION: Navigator triggering offers no advantage in fitted diffusion parameters, whereas free-breathing appears to offer greater confidence in fitted diffusion parameters, with fewer outliers, for matched acquisition periods.J. Magn. Reson. Imaging 2013. © 2013 Wiley Periodicals, Inc.
    Journal of Magnetic Resonance Imaging 04/2013; · 2.57 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: In addition to the diffusion coefficient, fitting the intravoxel incoherent motion model to multiple b-value diffusion-weighted MR data gives pseudo-diffusion measures associated with rapid signal attenuation at low b-values that are of use in the assessment of a number of pathologies. When summary measures are required, such as the average parameter for a region of interest, least-squares based methods give adequate estimation accuracy. However, using least-squares methods for pixel-wise fitting typically gives noisy estimates, especially for the pseudo-diffusion parameters, which limits the applicability of the approach for assessing spatial features and heterogeneity. In this article, a Bayesian approach using a shrinkage prior model is proposed and is shown to substantially reduce estimation uncertainty so that spatial features in the parameters maps are more clearly apparent. The Bayesian approach has no user-defined parameters, so measures of parameter variation (heterogeneity) over regions of interest are determined by the data alone, whereas it is shown that for the least-squares estimates, measures of variation are essentially determined by user-defined constraints on the parameters. Use of a Bayesian shrinkage prior approach is, therefore, recommended for intravoxel incoherent motion modeling. Magn Reson Med, 2013. © 2013 Wiley Periodicals, Inc.
    Magnetic Resonance in Medicine 02/2013; · 3.27 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Purpose:To evaluate noninvasive and clinically translatable magnetic resonance (MR) imaging biomarkers of therapeutic response in the TH-MYCN transgenic mouse model of aggressive, MYCN-amplified neuroblastoma.Materials and Methods:All experiments were performed in accordance with the local ethical review panel and the UK Home Office Animals Scientific Procedures Act 1986 and with the UK National Cancer Research Institute guidelines for the welfare of animals in cancer research. Multiparametric MR imaging was performed of abdominal tumors found in the TH-MYCN model. T2-weighted MR imaging, quantitation of native relaxation times T1 and T2, the relaxation rate R2*, and dynamic contrast-enhanced MR imaging were used to monitor tumor response to cyclophosphamide (25 mg/kg), the vascular disrupting agent ZD6126 (200 mg/kg), or the antiangiogenic agent cediranib (6 mg/kg, daily). Any significant changes in the measured parameters, and in the magnitude of the changes after treatment between treated and control cohorts, were identified by using Student two-tailed paired and unpaired t test, respectively, with a 5% level of significance.Results:Treatment with cyclophosphamide or cediranib induced a 54% or 20% reduction in tumor volume at 48 hours, respectively (P < .005 and P < .005, respectively; P < .005 and P < .005 versus control, respectively). Treatment with ZD6126 induced a 45% reduction in mean tumor volume 24 hours after treatment (P < .005; P < .005 versus control). The antitumor activity of cyclophosphamide, cediranib, and ZD6126 was consistently associated with a decrease in tumor T1 (P < .005, P < .005, and P < .005, respectively; P < .005, P < .005, and P < .005 versus control, respectively) and with a correlation between therapy-induced changes in native T1 and changes in tumor volume (r = 0.56; P < .005). Tumor response to cediranib was also associated with a decrease in the dynamic contrast-enhanced MR imaging-derived volume transfer constant (P = .07; P < .05 versus control) and enhancing fraction (P < .05; P < .01 versus control), and an increase in R2* (P < .005; P < .05 versus control).Conclusion:The T1 relaxation time is a robust noninvasive imaging biomarker of response to therapy in tumors in TH-MYCN mice, which emulate high-risk neuroblastoma in children. T1 measurements can be readily implemented on clinical MR systems and should be investigated in translational clinical trials of new targeted therapies for pediatric neuroblastoma.© RSNA, 2012Supplemental material:
    Radiology 11/2012; · 6.34 Impact Factor
  • Dow-Mu Koh, Aslam Sohaib
    [Show abstract] [Hide abstract]
    ABSTRACT: Diffusion-weighted magnetic resonance (MR) imaging (DWI) is now widely incorporated as a standard MR imaging sequence for the assessment of the male pelvis. DWI can improve the detection, characterization, and staging of pelvic malignancies, such as prostate, bladder, and rectal cancers. There is growing interest in applying quantitative DWI for the assessment of tumor treatment response. In addition, the technique seems promising for the evaluation of metastatic nodal and bone disease in the pelvis.
    Radiologic Clinics of North America 11/2012; 50(6):1127-44. · 1.95 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Picture archiving and communication systems (PACS) provide limited flexibility for the development of novel research methods. By contrast, the research model of data access is more flexible but has vulnerabilities in numerous areas. No single monolithic application can fulfill the diverse and rapidly changing needs of the clinical imaging research community. Instead, the focus should be on the interoperability of preexisting systems. To a large extent, this can be achieved by means of a unified interface for storing and retrieving data. The concept of a research PACS combines the advantages of the clinical and research models of data access while eliminating the disadvantages. A research PACS streamlines the data management process. Instead of a single software program, it consists of a confederation of independent applications brought together by the ability to store and retrieve data in a common database. A prototype research PACS has been developed that is based on the Extensible Neuroimaging Archive Toolkit (XNAT) in association with two new in-house tools: a data selection tool and a data archiving tool. By taking as an example the comparison of regions of interest in multifunctional liver data, it was demonstrated that this framework allows a number of in-house and open-source applications originally designed to work on a stand-alone basis to be integrated into a unified workflow, with minimal redevelopment effort.© RSNA, 2012.
    Radiographics 08/2012; · 2.79 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: OBJECTIVE: We examine the clinical impetus for whole-body diffusion-weighted MRI and discuss how to implement the technique with clinical MRI systems. We include practical tips and tricks to optimize image quality and reduce artifacts. The interpretative pitfalls are enumerated, and potential challenges are highlighted. CONCLUSION: Whole-body diffusion-weighted MRI can be used for tumor staging and assessment of treatment response. Meticulous technique and knowledge of potential interpretive pitfalls will help to avoid mistakes and establish this modality in radiologic practice.
    American Journal of Roentgenology 08/2012; 199(2):252-62. · 2.90 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: OBJECTIVES: Silicone breast prostheses prove technically challenging when performing diffusion-weighted MR imaging in the breasts. We describe a combined fat and chemical suppression scheme to achieve dual suppression of fat and silicone, thereby improving the quality of diffusion-weighted images in women with breast implants. METHODS: MR imaging was performed at 3.0 and 1.5 T in women with silicone breast implants using short-tau inversion recovery (STIR) fat-suppressed echo-planar (EPI) diffusion-weighted MR imaging (DWI) on its own and combined with the slice-select gradient-reversal (SSGR) technique. Imaging was performed using dedicated breast imaging coils. RESULTS: Complete suppression of the fat and silicone signal was possible at 3.0 T using EPI DWI with STIR and SSGR, evaluated with dedicated breast coils. However, a residual silicone signal was still perceptible at 1.5 T using this combined approach. Nevertheless, a further reduction in silicone signal at 1.5 T could be achieved by employing thinner slice partitions and the addition of the chemical-selective fat-suppression (CHESS) technique. CONCLUSIONS: DWI using combined STIR and SSGR chemical suppression techniques is feasible to eliminate or reduce silicone signal from prosthetic breast implants. KEY POINTS : • Breast magnetic resonance imaging (MRI) is frequently needed following breast implants • Unsuppressed signal from silicone creates artefacts on diffusion-weighted MR sequences • Dual fat/chemical suppression can eliminate signal from fat and silicone • STIR with slice selective gradient reversal can suppress fat and silicone signal.
    European Radiology 06/2012; · 4.34 Impact Factor

Publication Stats

2k Citations
207.98 Total Impact Points


  • 2004–2013
    • Institute of Cancer Research
      • CR-UK and EPSRC Cancer Imaging Centre
      Sutton, ENG, United Kingdom
  • 2008–2012
    • The Royal Marsden NHS Foundation Trust
      Londinium, England, United Kingdom
  • 2011
    • École Polytechnique Fédérale de Lausanne
      Lausanne, Vaud, Switzerland
  • 2010
    • State University of New York Downstate Medical Center
      • Department of Radiology
      Brooklyn, NY, United States
  • 2009
    • University of Michigan
      Ann Arbor, Michigan, United States
    • Johns Hopkins Medicine
      • Department of Radiology and Radiological Science
      Baltimore, MD, United States
    • Inselspital, Universitätsspital Bern
      • University Institute of Diagnostic, Interventional and Pediatric Radiology
      Bern, BE, Switzerland
    • NYU Langone Medical Center
      • Department of Radiology
      New York City, NY, United States
    • University of Toronto
      Toronto, Ontario, Canada