Andrea S Doria

SickKids, Toronto, Ontario, Canada

Are you Andrea S Doria?

Claim your profile

Publications (76)195.97 Total impact

  • [Show abstract] [Hide abstract]
    ABSTRACT: Our aim was to test the feasibility of blood oxygen level dependent magnetic resonance imaging (BOLD MRI) and dynamic contrast-enhanced (DCE) MRI to monitor periarticular hypoxic/inflammatory changes over time in a juvenile rabbit model of arthritis.
    European radiology. 09/2014;
  • [Show abstract] [Hide abstract]
    ABSTRACT: To assess the sequence and type of active joints in a cohort of newly diagnosed juvenile idiopathic arthritis (JIA) patients with full access to current treatment at first visit and during a follow-up period of 5-years, in order to identify an index joint/group of joints for magnetic resonance imaging in JIA. Patient charts of all consecutive newly diagnosed JIA patients with a follow-up duration of at least 5 years were analyzed. Patients were derived from two tertiary pediatric rheumatology centers. Patient characteristics and data concerning the presence of joints with arthritis and the use of medication were recorded. Findings from 95 JIA patients [39 (41 %) oligoarticular and 56 (59 %) polyarticular] were analyzed. At first visit, distribution of active joints among patients was as follows: knee (n = 70, 74 %), ankle (n = 55, 58 %), elbow (n = 23, 24 %), wrist (n = 23, 24 %), metacarpophalangeal (MCP) (n = 20, 21 %), proximal interphalangeal (PIP) (n = 13, 14 %), hip (n = 6, 6 %), shoulder (n = 5, 5 %), and distal interphalangeal (DIP) (n = 4, 4 %) joints. After a follow-up period of 5 years, the cumulative percentage of patients with specific joint involvement changed into: knee (n = 88, 93 %), ankle (n = 79, 83 %), elbow (n = 43, 45 %), wrist (n = 38, 40 %), MCP (n = 36, 38 %), PIP (n = 29, 31 %), shoulder (n = 20, 21 %), hip (n = 17, 19 %), and DIP (n = 9, 10 %) joints. Despite changes in treatment strategies over the years, the knee remains the most commonly involved joint at onset and during follow-up in JIA, followed by the ankle, elbow, and wrist. For the evaluation of outcome with MRI, the knee appears the most appropriate joint in JIA.
    Rheumatology International 08/2014; · 2.21 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Dual-energy absorptiometry (DXA) is the current reference standard for assessing pediatric osteoporosis; however due to its areal nature, it has limitations. Thus, quantitative ultrasound (QUS), a modality free of ionizing radiation, has been proposed as a potential surrogate for DXA.
    Pediatric Radiology 06/2014; · 1.57 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Traumatic lumbar punctures with blasts (TLP+) in children with acute lymphoblastic leukaemia (ALL) obscure central nervous system status and are associated with a poorer event-free survival (EFS). We conducted a retrospective cohort study of all lumbar punctures (LPs) for children with ALL diagnosed at our institution from 2005 to 2009. We utilised random-effects and fixed-effects repeated-measures logistic regression analyses to identify risk factors for TLPs. Fixed-effects models use each patient as his or her own control. We used survival analysis to describe outcomes after a TLP+. 264 children underwent 5267 evaluable lumbar punctures (LPs), of which 944 (17.9%) were traumatic. In the multivariable random-effects model, variables significantly associated with TLPs were age <1year (odds ratio (OR) 3.46, 95% confidence interval (CI) 2.06-5.81) or age ⩾10years (OR 2.00, CI 1.66-2.40); body mass index percentile ⩾95 (OR 1.44, CI 1.19-1.75); platelet count <100×10(3)/μL (OR 1.49, CI 1.08-20.7); fewer days since previous LP (OR 5.13, CI 2.34-11.25 for ⩾16days versus 0-3days); and a preceding TLP (OR 1.43, CI 1.19-1.73). In the fixed-effects model, image-guidance reduced the odds of TLP (OR 0.55, CI 0.32-0.95). The 5-year EFS (±SE) for children with TLP+ (77±8%) was significantly lower than for children with CNS1 status (93±2%; p=0.002). The frequency of TLP remains high. Consistent with previous studies, a TLP+ at diagnosis was associated with a poorer EFS. These risk factors can allow identifying interventions to reduce TLPs and directing interventions to those at highest risk.
    European journal of cancer (Oxford, England: 1990) 03/2014; · 4.12 Impact Factor
  • Source
  • [Show abstract] [Hide abstract]
    ABSTRACT: Recent advances in hemophilia prophylaxis have raised the need for accurate noninvasive methods for assessment of early cartilage damage in maturing joints to guide initiation of prophylaxis. Such methods can either be semiquantitative or quantitative. Whereas semiquantitative scores are less time-consuming to be performed than quantitative methods, they are prone to subjective interpretation. To test the feasibility of a manual segmentation and a quantitative methodology for cross-sectional evaluation of articular cartilage status in growing ankles of children with blood-induced arthritis, as compared with a semiquantitative scoring system and clinical-radiographic constructs. Twelve boys, 11 with hemophilia (A, n = 9; B, n = 2) and 1 with von Willebrand disease (median age: 13; range: 6-17), underwent physical examination and MRI at 1.5 T. Two radiologists semiquantitatively scored the MRIs for cartilage pathology (surface erosions, cartilage loss) with blinding to clinical information. An experienced operator applied a validated quantitative 3-D MRI method to determine the percentage area of denuded bone (dAB) and the cartilage thickness (ThCtAB) in the joints' MRIs. Quantitative and semiquantitative MRI methods and clinical-radiographic constructs (Hemophilia Joint Health Score [HJHS], Pettersson radiograph scores) were compared. Moderate correlations were noted between erosions and dAB (r = 0.62, P = 0.03) in the talus but not in the distal tibia (P > 0.05). Whereas substantial to high correlations (r range: 0.70-0.94, P < 0.05) were observed between erosions, cartilage loss, HJHS and Pettersson scores both at the distal tibia and talus levels, moderate/borderline substantial (r range: 0.55-0.61, P < 0.05) correlations were noted between dAB/ThCtAB and clinical-radiographic constructs. Whereas the semiquantitative method of assessing cartilage status is closely associated with clinical-radiographic scores in cross-sectional studies of blood-induced arthropathy, quantitative measures provide independent information and are therefore less applicable for that research design.
    Pediatric Radiology 02/2014; · 1.57 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: OBJECTIVE. Our objective was to evaluate the diagnostic accuracy and reliability of MRI and its ability to depict responsiveness to treatment for the evaluation of the axial joints (temporomandibular joint [TMJ], spinal joints, and sacroiliac joints) in juvenile idiopathic arthritis (JIA). CONCLUSION. There is fair (grade B) evidence that MRI is an accurate diagnostic method for evaluating early and intermediate changes in the TMJ in JIA and insufficient evidence to indicate MRI is an accurate diagnostic method for detecting JIA in the spinal (grade I) and sacroiliac (grade I) joints.
    American Journal of Roentgenology 01/2014; 202(1):199-210. · 2.90 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Blood-oxygen-level-dependent (BOLD) MRI has the potential to identify regions of early hypoxic and vascular joint changes in inflammatory arthritis. There is no standard protocol for analysis of BOLD MRI measurements in musculoskeletal disorders. To optimize the following BOLD MRI reading parameters: (1) statistical threshold values (low, r > 0.01 versus high, r > 0.2); (2) summary measures of BOLD contrast (percentage of activated voxels [PT%] versus percentage signal difference between on-and-off signal intensities [diff_on_off]); and (3) direction of BOLD response (positive, negative and positive + negative). Using BOLD MRI protocols at 1.5 T, arthritic (n = 21) and contralateral (n = 21) knees of 21 juvenile rabbits were imaged at baseline and on days 1, 14 and 28 after a unilateral intra-articular injection of carrageenan. Nine non-injected rabbits served as external control knees (n = 18). By comparing arthritic to contralateral knees, receiver operating characteristic curves were used to determine diagnostic accuracy. Using diff_on_off and positive + negative responses, a threshold of r > 0.01 was more accurate than r > 0.2 (P = 0.03 at day 28). Comparison of summary measures yielded no statistically significant difference (P > 0.05). Although positive + negative (AUC = 0.86 at day 28) and negative responses (AUC = 0.90 at day 28) for PT% were the most diagnostically accurate, positive + negative responses for diff_on_off (AUC = 0.78 at day 28) also had acceptable accuracy. The most clinically relevant reading parameters included a lower threshold of r > 0.01 and a positive + negative BOLD response. We propose that diff_on_off is a more clinically relevant summary measure of BOLD MRI, while PT% can be used as an ancillary measure.
    Pediatric Radiology 12/2013; · 1.57 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Recent advances in magnetic resonance imaging (MRI) techniques have substantially improved the evaluation of joint pathologies in juvenile idiopathic arthritis (JIA). Because of the current availability of highly effective antirheumatic therapies and the unique and useful features of MRI, there is a growing need for an accurate and reproducible MRI assessment scoring system for JIA, such as the rheumatoid arthritis MRI Scoring (RAMRIS) for patients with rheumatoid arthritis (RA). To effectively evaluate the efficacy of treatment in clinical research trials, we need to develop and validate scoring methods to accurately measure joint outcomes, standardize imaging protocols for data acquisition and interpretation, and create imaging atlases to differentiate physiologic and pathologic joint findings in childhood and adolescence. Such a standardized, validated, JIA-MRI scoring method could be used as an outcome measure in clinical trials.
    The Journal of Rheumatology 11/2013; · 3.26 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Our purposes were to determine: (i) whether there is direct evidence that currently available MRI techniques are accurate for early diagnosis of pathological findings in haemophilic arthropathy; (ii) whether there is an MRI scoring system that best correlates with clinical/radiological constructs for evaluation of haemophilic arthropathy; (iii) whether there is an MRI scoring system that best correlates with clinical/radiological constructs for evaluation of haemophilic arthropathy. Articles were screened using MEDLINE (n = 566), EMBASE (n = 201), and the Cochrane Library (n = 1). Two independent reviewers assessed articles for inclusion under the overarching purposes of the review by using the Standards for Reporting of Diagnostic Accuracy (STARD) tool, and the quality of the studies were graded using the Quality Assessment of Diagnostic Accuracy Studies 2 (QUADAS-2) tool. The electronic literature search retrieved 777 references (after duplicates were removed). A total of 32 studies were chosen for inclusion from the results of the search and review of bibliographical references. Using the STARD tool, seven studies were of excellent quality of reporting, and using the QUADAS-2 tool, 10 studies were judged to be of adequate quality. There is 'fair' evidence to recommend MRI as an accurate test for detecting evidence of haemophilic arthropathy and the use of second or third generation MRI scales for assessing haemophilic arthropathy. However, there is no evidence that screening of early intra-articular soft tissue bleed with MRI improves the functional status of joints over time.
    Haemophilia 08/2013; · 3.17 Impact Factor
  • Source
    Clinics (São Paulo, Brazil) 05/2013; 68(5). · 1.59 Impact Factor
  • Andrea S Doria
    Evidence-based medicine 03/2013;
  • [Show abstract] [Hide abstract]
    ABSTRACT: Evaluation of prophylactic treatment of haemophilia requires sensitive methods. To design and test a new magnetic resonance imaging (MRI) scale for haemophilic arthropathy, two scales of a combined MRI scoring scheme were merged into a single scale which includes soft tissue and osteochondral subscores. Sixty-one joint MRI's of 46 patients with haemophilia were evaluated by four radiologists using the new and older scales. Forty-six of the joints were evaluated using two X-ray scales. For all MRI scores, interreader agreement and correlations with X-ray scores and lifetime number of haemarthroses were analysed. The interreader agreement intraclass correlation coefficient was 0.82, 0.89 and 0.88 for the soft tissue and osteochondral subscores and the total score, as evaluated according to the new MRI scale, compared to 0.80 and 0.89 as for the older scales. The total score and osteochondral subscore according to the new scale, as well as scores according to the older scales were correlated (P < 0.01) with number of haemarthroses (Spearman correlation 0.35-0.68) and with the X-ray scores (Spearman correlation 0.40-0.76), but no correlation (P > 0.05) was found between the soft tissue subscore of the new MRI scale and the X-ray scores. The new MRI scale is simpler to apply than the older and has similar reader reliability and correlation with lifetime number of haemarthroses, and by separating soft tissue and osteochondral changes it gives additional information. The new scale is useful for analyses of early and moderate stages of arthropathy, and may help to evaluate prophylactic haemophilia treatment.
    Haemophilia 07/2012; · 3.17 Impact Factor
  • Source
    Pediatric Rheumatology 07/2012; 10(1). · 1.47 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Color Doppler US (CDUS) has been used for evaluation of cerebral venous sinuses in neonates. However, there is very limited information available regarding the appearance of superficial and deep normal cerebral venous sinuses using CDUS and the specificity of the technique to rule out disease. To determine the specificity, inter-modality and inter-reader agreement of color Doppler US (CDUS). To evaluate normal cerebral venous sinuses in neonates in comparison to MR venography (MRV). Newborns undergoing a clinically indicated brain MRI were prospectively evaluated. All underwent a dedicated CDUS of the cerebral venous sinuses within 10 h (mean, 3.5 h, range, and 2-7.6 h) of the MRI study using a standard protocol. Fifty consecutive neonates participated in the study (30 males [60%]; 25-41 weeks old; mean, 37 weeks). The mean time interval between the date of birth and the CDUS study was 19.1 days. No cases showed evidence of thrombosis. Overall agreement for US reading was 97% (range, 82-100%), for MRV reading, 99% (range, 96-100%) and for intermodality, 100% (range, 96-100%). Excellent US-MRI agreement was noted for superior sagittal sinus, cerebral veins, straight sinus, torcular Herophili, sigmoid sinus, superior jugular veins (94-98%) and transverse sinuses (82-86%). In 10 cases (20%), MRV showed flow gaps whereas normal flow was demonstrated with US. Visualization of the inferior sagittal sinus was limited with both imaging techniques. Excellent reading agreement was noted for US, MRV and intermodality. CDUS is highly specific to rule out cerebral venous thrombosis in neonates and holds potential for clinical application as part of clinical-laboratory-imaging algorithms of pre/post-test probabilities of disease.
    Pediatric Radiology 04/2012; 42(9):1070-9. · 1.57 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Therapeutic options for the treatment of plaque morphea are limited. We explored the efficacy and safety of imiquimod cream in children with plaque morphea. Prospective, open-label, double-baseline study, using imiquimod 5% cream topically for 9 months. The primary outcome measure was improvement in the thickening of morphea plaques using a visual analog scale (VAS) and the DIET score (dyspigmentation/induration/erythema/telangiectasia). Secondary outcome measures were clinicoradiographic correlations and frequency of adverse events. Nine patients, 89% females, with a mean age of 11.33 years (SD = 3.52) were enrolled. At 36 weeks, the mean VAS had decreased from 48.08 (SD = 18.85) to 22.7 (SD = 12.9) (p < 0.0001), and the mean DIET score from 4.38 (SD = 1.2) to 3.06 (SD = 1.39) (p = 0.23). There was very good interrater reliability between DIET score assessments (intraclass correlation coefficient, ICC = 0.75) and VAS (ICC = 0.59) and moderate agreement between parent and investigator VAS (ICC = 0.5). Ultrasonographically measured dermis thickness changed from 1.05 (SD = 0.34) to 0.95 (SD = 0.19) (p = 0.001). One patient experienced ulceration that required temporary discontinuation of intervention. This proof of concept study revealed that imiquimod 5% cream is effective in decreasing the thickening of plaque morphea and safe for pediatric use. Further prospective studies are warranted.
    Dermatology 02/2012; 223(4):363-9. · 2.02 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Because of the ability of blood-oxygen-level-dependent (BOLD) MRI to assess blood oxygenation changes within the microvasculature, this technique holds potential for evaluating early perisynovial changes in inflammatory arthritis. To evaluate the feasibility of BOLD MRI to detect interval perisynovial changes in knees of rabbits with inflammatory arthritis. Rabbit knees were injected with albumin (n=9) or saline (n=6) intra-articularly, or were not injected (control knees, n=9). Except for two rabbits (albumin-injected, n=2 knees; saline-injected, n=2 knees) that unexpectedly died on days 7 and 21 of the experiment, respectively, all other animals were scanned with BOLD MRI on days 0, 1, 7, 14, 21 and 28 after induction of arthritis. T2*-weighted gradient-echo MRI was performed during alternate 30 s of normoxia/hyperoxia. BOLD MRI measurements were compared with clinical, laboratory and histological markers. Percentage of activated voxels was significantly greater in albumin-injected knees than in contralateral saline-injected knees (P=0.04). For albumin-injected knees (P<0.05) and among different categories of knees (P=0.009), the percentage of activated BOLD voxels varied over time. A quadratic curve for on-and-off BOLD difference was delineated for albumin- and saline-injected knees over time (albumin-injected, P=0.047; saline-injected, P=0.009). A trend toward a significant difference in synovial histological scores between albumin-injected and saline-injected knees was noted only for acute scores (P=0.07). As a proof of concept, BOLD MRI can depict perisynovial changes during progression of experimental arthritis.
    Pediatric Radiology 08/2011; 42(1):63-75. · 1.57 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: The clinimetric properties of blood oxygen level‒dependent (BOLD) magnetic resonance imaging (MRI) for assessment of musculoskeletal changes have been poorly investigated. The study objectives were to assess the interframework reliability of data acquisition of BOLD MRI and to test its convergent validity in chronic arthritis in a rabbit model of inflammatory arthritis as compared with corresponding clinical and laboratory measures. One of the knees of 12 New Zealand male white rabbits was injected with a 1% carrageenin solution, and the contralateral (control) one was not. Twelve rabbits were euthanized on day 28 of arthritis (chronic arthritis). Clinical (joint diameters), laboratory (serum amyloid A concentration), and BOLD MRI measurements were obtained on days 0, 1, and 28 of arthritis. Twenty paradigms of data acquisition and analysis were applied. The most reliable MRI parameters set, regardless of threshold values used for data analysis, was spiral technique (level 1), 40 ms of echo time (level 2), 60 seconds of on_ and off_ paradigm (level 3) and carbogen mixture of gases (95% O2 + 5% CO2) (level 4). With regard to construct validity, BOLD imaging correlated moderately (r = -.54, P < .0001) with knee diameters, and weakly (r = -.35, P = .01) with laboratory indices (high threshold for analysis). BOLD MRI has a substantial or excellent interframework reliability for assessment of arthritic rabbit knees; however, it correlates only moderately or poorly with clinical and laboratory measures. Nevertheless, this study supports further validation of BOLD MRI for assessment of soft tissue changes in a rabbit model of arthritis.
    Academic radiology 03/2011; 18(5):615-25. · 2.09 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Osteonecrosis (ON) is a major complication after treatment of developmental dysplasia of the hip (DDH). Several studies have explored the absence of the femoral head ossific nucleus at the time of hip reduction as a risk factor for the development of ON, but findings have been inconsistent. We therefore determined the incidence of ON in children who underwent reduction of a dislocated hip in the presence or absence of the ossific nucleus. We retrospectively reviewed the radiographs of 105 hips in 89 patients treated for DDH at the age of 18 months or younger. Radiographs were graded for the presence of the ossific nucleus at the time of hip reduction and for the presence of ON, as graded by the Bucholz and Ogden classification, for patients at a mean age of 10 years. We used log-binomial regression to estimate if the presence of the ossific nucleus was associated with a lower incidence of ON. We identified ON in 37 of the 105 hips (35%). The incidence of ON at 10 years was 40% in the absence of the ossific nucleus and 32% in the presence of the ossific nucleus (adjusted relative risk, 0.86; 95% confidence interval, 0.36-1.81). When only radiographic changes of Grade II and greater were considered ON, the risk was still not increased (relative risk, 1.26; 95% confidence interval, 0.62-2.56). Patients with an ossific nucleus at the time of hip reduction showed a slight tendency toward better outcomes. The ossific nucleus did not protect for ON. Level III, prognostic study. See Guidelines for Authors for a complete description of levels of evidence.
    Clinical Orthopaedics and Related Research 02/2011; 469(10):2838-45. · 2.79 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Obesity constitutes an independent predictor for a nondiagnostic screening ultrasound (US) in suspected appendicitis. Furthermore, the imaging route consisting of the screening US with a second US if necessary has a significantly lower diagnostic accuracy in the obese children (83%) than their lean counterparts (93%). However, the effect of this association on clinical outcomes in the obese population is unknown. In children presenting to the emergency department (ED) undergoing US for suspected appendicitis, the objective was to examine if there is a significant difference in the proportions of those with a "desirable outcome" in the obese group versus their normal-weight counterparts. This was a reanalysis of a recent prospective cohort study conducted at a tertiary care pediatric ED. A consecutive sample of 263 previously healthy children 4-17 years of age undergoing screening US for suspected appendicitis was enrolled. Study patients were divided into obese (body mass index-for-age percentiles [BMI-FAP] ≥ 85%) and normal-weight groups. The main outcome was the proportion of children with the desirable outcome in the obese and normal-weight groups. Desirable outcomes for patients with a nondiagnostic screening US were defined as either 1) nonperforated appendicitis or alternate surgical diagnosis undergoing appropriate surgical procedure within 24 hours of assessment without a computed tomography (CT) scan or 2) no appendicitis, discharged from the ED within 24 hours without a CT scan. Secondary outcomes included hospitalization, CT imaging for any reason within 30 days of initial presentation, length of hospital stay, return visits, time to and length of surgery, and complications. Of the 263 patients, 76 (28.9%) were obese, and 187 had a normal weight. Ninety-nine children (37.6%) had appendicitis (obese = 24, normal-weight = 75). The desirable outcome was achieved in 46 (64.5%) of the obese patients versus 137 (73.3%) of the normal-weight children (95% confidence interval [CI] for the difference = -0.21 to 0.04). Compared to the normal-weight children, the obese children had similar hospitalization rates (44.7% vs. 45.5%), median length of hospital stay (20.5 hours vs. 23.0 hours), and return ED visit rates (11.8% vs. 9.1%). The median time to surgery (obese = 10.5 hours, normal-weight = 9.0 hours), complications (obese = 25%, normal-weight = 21.3%), perforations (obese = 33%, normal-weight = 31%), and length of surgery (obese 80 [±37] minutes, normal-weight 67 [±25] minutes) were also similar. The rate of CT within 30 days of initial presentation was higher in the obese compared to the normal-weight group (13/76 or 17.1% vs. 13/187 or 6.9%, 95% CI = 0.02 to 0.23). Previously healthy obese children undergoing screening US for suspected appendicitis experienced comparable clinical outcomes to their normal-weight counterparts. However, they were three times more likely to receive a CT scan during their investigation. These results argue for continuing current practice of using the screening US in this population, followed by a CT if clinically warranted upon reexamination.
    Academic Emergency Medicine 02/2011; 18(2):167-73. · 1.76 Impact Factor

Publication Stats

959 Citations
195.97 Total Impact Points


  • 2002–2013
    • SickKids
      • • Department of Diagnostic Imaging
      • • Division of Paediatric Emergency Medicine
      • • Division of Rheumatology
      Toronto, Ontario, Canada
  • 2005–2012
    • Lund University
      • • Department of Clinical Sciences
      • • Department of Radiology
      Lund, Skane, Sweden
  • 2010
    • Hackensack University Medical Center
      Hackensack, New Jersey, United States
  • 2001–2010
    • University of Toronto
      • Hospital for Sick Children
      Toronto, Ontario, Canada
  • 2009
    • Christian Medical College Vellore
      • Department of Radiology
      Vellore, State of Tamil Nadu, India
    • McMaster University
      Hamilton, Ontario, Canada
  • 2007
    • University College London
      • Institute of Child Health
      London, ENG, United Kingdom
  • 2003
    • Hospital das Clínicas da Faculdade de Medicina da Universidade de São Paulo
      San Paulo, São Paulo, Brazil
  • 2001–2002
    • University of São Paulo
      • Hospital das Clínicas (FMUSP)
      San Paulo, São Paulo, Brazil