Sarah J Drouin

Hôpitaux Universitaires La Pitié salpêtrière - Charles Foix, Lutetia Parisorum, Île-de-France, France

Are you Sarah J Drouin?

Claim your profile

Publications (69)141.28 Total impact

  • [Show abstract] [Hide abstract]
    ABSTRACT: Controversy exists regarding the propensity of hypogonadism and metabolic disorders to promote the development of high-risk prostate cancer (PCa). Our aim was to prospectively test whether preoperative circulating testosterone levels, obesity, and metabolic syndrome (MetS) were correlated with aggressive pathological features after radical prostatectomy (RP). Overall, 354 patients undergoing robot-assisted RP at our academic institution, between 2010 and 2013, to treat clinically localized PCa were included in this prospective study. Pelvic lymphadenectomy was performed in 116 (32.8%) patients and confirmed the absence of nodal metastases in all of them. Cardiovascular risk factors and body-mass index (BMI) were used to define MetS and obesity, respectively. Total testosterone (TT) levels were assessed using an immunoassay method, whereas bioavailable testosterone (BT) and free testosterone (FT) levels were estimated using Vermeulen׳s formula. Multivariate logistic regression analyses assessed independent predictors for postoperative aggressive pathological features (i.e., a pathological Gleason score [GS]≥7, extracapsular extension [ECE], seminal vesicle invasion [SVI], and positive surgical margins [PSM]) and GS upgrading. Low TT, BT, and FT levels were found in 54 (15.2%), 70 (19.8%), and 62 (17.5%) patients, respectively. Median BMI was 26.3kg/m(2) (range: 17.4-43.9), and prevalence of MetS was 18.9%. Significantly higher rates of pathological GS≥7 were observed in groups with a low TT level (46.3% vs. 33.3%; P = 0.01), low BT level (44.3% vs. 33.1%; P<0.001), and low FT level (46.8% vs. 32.9%; P = 0.001). Multivariate analyses demonstrated that only low BT and FT levels were independent predictors of pathological GS≥7 (odds ratio [OR] = 1.76; P<0.001 and OR = 1.39; P<0.001, respectively) and GS upgrading (OR = 2.82; P<0.001 and OR = 1.71; P<0.001, respectively), but there was no significant correlation between low circulating testosterone levels and ECE, SVI, or PSM. Furthermore, BMI (OR = 1.28; P = 0.04) and MetS (OR = 1.19; P = 0.01) were only correlated with PSM. Hypogonadism, obesity, and MetS were not independent predictors of pathological GS ≥7, ECE, or SVI after RP. Our data suggest that only low BT and FT levels, which might logically result in an active androgen-depleted environment, were linked with high-grade PCa. Copyright © 2014 Elsevier Inc. All rights reserved.
    Urologic Oncology 01/2015; DOI:10.1016/j.urolonc.2014.11.010
  • [Show abstract] [Hide abstract]
    ABSTRACT: Aim To describe the main prognostic factors with an impact on survival of patients diagnosed with upper tract urothelial carcinomas (UTUC). Material and methods A systematic review of the literature has been performed using Pubmed without timeline restriction with the following keywords (MeSH): urothelial carcinoma; ureter; renal pelvis; prognosis; recurrence; survival; predictive models; nomogram. Results The level of evidence was low (3) in every available studies. There were 4 categories of prognostic factors in UTUCs: clinical (patient and tumor characteristics); surgical; pathological and molecular. The most important pre-operative prognostic factors were: size > 3 cm, grade (biopsy and cytology); multifocality; important hydronephrosis; co-morbidity (ASA), ECOG status, and a surgical delay of no more than 3 months. After surgery, the most important prognostic factors are: stage, grade, carcinoma in situ, lymphovascular invasion and lymph node involvement. Serum markers from inflammation (CRP) could be useful for the prediction of advanced stages. Molecular markers are still under evaluation. Conclusion The identification of prognostic factors in UTUC has improved over the past years. These prognostic factors can be considered alone but also as a panel or inside predictive tools to predict accurately patient's survival.
    Progrès en Urologie 11/2014; DOI:10.1016/j.purol.2014.07.013
  • [Show abstract] [Hide abstract]
    ABSTRACT: To describe the main prognostic factors with an impact on survival of patients diagnosed with upper tract urothelial carcinomas (UTUC).
  • European Urology Supplements 04/2014; 13(1):e14–e14a. DOI:10.1016/S1569-9056(14)60016-6
  • [Show abstract] [Hide abstract]
    ABSTRACT: To study the prognostic value of extent, number, and location of positive surgical margins (PSM). A total of 1,504 consecutive adjuvant treatment naive and node-negative radical prostatectomy men were included in a prospective database including extent, number, and location of PSM. Mean follow-up was 33 months. Endpoint was biochemical progression-free (bPFS) survival. The impact of margin status and characteristics was assessed in time-dependent analyses using Cox regression and Kaplan-Meier methods. PSM was reported in 26.7 % of patients. The predominant PSM locations were apex and posterior locations. Median PSM length was 4.0 mm. The 2-year bPFS was 73.7 % in PSM patients as compared to 93.0 % in NSM patients (p < 0.001). The rate and extent of PSM increased significantly with pathologic stage (p < 0.001). The extent of PSM length was linearly correlated with bPFS (p = 0.017, coefficient: -0.122). In univariable analysis, extent and number of PSM were significantly linked to outcomes. None of PSM subclassifications significantly influenced the bPFS rates in the subgroup of pT2 disease patients. Conversely, stratification by PSM location (apex vs. other locations, p = 0.008), by PSM number (p = 0.006), and by PSM length (p < 0.001) showed significant differences in pT3-4 cancer patients. In that subgroup, PSM length also added to bPFS prediction using PSM status only in multivariable models (p = 0.005). PSM subclassifications do not improve the biochemical recurrence prediction in organ-confined disease. In non-organ-confined disease, PSM length (≥3 mm), multifocality (≥3 sites), and apical location are significantly linked to poorer outcomes and could justify a more aggressive adjuvant treatment approach.
    World Journal of Urology 01/2014; 32(6). DOI:10.1007/s00345-014-1243-3
  • Progrès en Urologie 11/2013; 23(13):1022. DOI:10.1016/j.purol.2013.08.013
  • Progrès en Urologie 11/2013; 23(13):1023-1024. DOI:10.1016/j.purol.2013.08.016
  • [Show abstract] [Hide abstract]
    ABSTRACT: To investigate the impact of 3-month androgen deprivation therapy (st-ADT) a secondary chemoprevention of indolent-localized prostate cancer (PCa). A prospective phase II study enrolled men over 4 years with low-risk PCa and the following characteristics: PSA < 10 ng/mL, Gleason score of 6 (3 + 3) or less, three positive cores or less, and tumor stage T2a or less. Patients received a single sub-cutaneous injection of 22.5 mg of leuprolide acetate with Atrigel 3-month depot associated with a daily oral intake of bicalutamide 50 mg/day during 15 days around the injection. Follow-up included PSA and bioavailable testosterone blood tests every 3 months and yearly surveillance biopsies. Primary end point was the presence of PCa on biopsy at last follow-up. Secondary end points were detailed pathological features and adverse events. Overall, 98 men were included and 45 of them (45.9 %) had a negative biopsy after a median follow-up of 13 months [11-19.5]. Of the 53 patients with positive biopsy, 17 had pathologic progression because of upgraded Gleason score (11 patients), four or more positive cores (three patients) or both (three patients). The only significant predictive factor biopsy outcome was the number of positive cores at diagnosis. Secondary chemoprevention by st-ADT for localized PCa could be useful to pinpoint indolent tumors suitable for AS. Indeed, after st-ADT nearly one patient out of two had negative biopsies and 17 % had pathological progression. This is an innovative option to consider as an alternative to current AS protocols contingent upon confirmation in subsequent studies.
    World Journal of Urology 10/2013; DOI:10.1007/s00345-013-1196-y
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The gold standard treatment of muscle invasive bladder cancer is radical cystectomy. This population is known to be at risk of malnutrition due to age and disease factors. Current evidence has established the nutritional and immunological benefits of immune-enhancing nutritional supplements in upper gastrointestinal surgery. There is no present guideline for immunonutrition use in urology and bladder cancer specifically. We did a systematic review of available literature on the MEDLINE/Embase database. We assessed the rates of malnutrition in cystectomy cohorts and analysed the clinical impacts of nutritional deficiency. The impact of immune enhancing supplements was also investigated in cystectomy cohorts with regards to postoperative outcomes. The prevalence of severe malnutrition was found to be between 16% and 22%. There was a consistent association of malnourished patients with adverse postoperative outcomes in terms of mortality and morbidity. There is a paucity of data regarding immunonutrition in urological cohorts. Postoperative immunonutrition in cystectomy was not found to have significant benefits beyond early return to a normal diet. There is not enough evidence in malnourished urological cohorts to establish a consensus on immunonutrition. Until there is more well controlled comparative effective studies or randomized trials, the role of immunonutrition should be considered investigational in the bladder cancer cohort.
    BJU International 10/2013; 114(2). DOI:10.1111/bju.12529
  • [Show abstract] [Hide abstract]
    ABSTRACT: The aim of the study was to assess the outcome after nephron-sparing surgery (NSS) of patients with small renal masses (SRMs) who would have been eligible for active surveillance (AS). Data were collected retrospectively for 758 patients who underwent NSS over a 5-year period. Outcomes were assessed in two groups of patients who were eligible for AS according to different criteria. Group 1 criteria were as follows: age >75 years, renal mass ≤4 cm, significant comorbidities [Charlson Comorbidity Index (CCI) >2]. Group 2 criteria were as follows: any SRM ≤ 4 cm regardless of age, severe comorbidities with a 10-year mortality risk >50 % (CCI > 4). The two groups were not compared statistically because some patients were included in both. Fifty-five patients (7.3 %) were included in Group 1 and 62 (8.2 %) in Group 2. There was a significant proportion of benign tumours in Group 1 (N = 6; 11 %) and Group 2 (N = 6; 10 %). Six (11 %) positive margins were observed in Group 1 and 8 (13 %) in Group 2. The 2- and 5-year recurrence-free survival rates were 100 and 77.4 %, respectively, in Group 1, and 88.5 and 79.6 % in Group 2. The 2- and 5-year overall survival rates were 100 and 74.7 % in Group 1, and 96.7 and 78.1 % in Group 2. The majority of patients with SRMs who would have been eligible for AS had no recurrence after initial tumour removal. In these patients, a CCI > 4 appeared to be a pertinent criterion to identify those patients less likely to benefit from immediate surgery.
    World Journal of Urology 07/2013; DOI:10.1007/s00345-013-1131-2
  • [Show abstract] [Hide abstract]
    ABSTRACT: To assess functional outcomes obtained after surgical management of post-radiation urinary incontinence after prostate cancer. A retrospective review of the data from patients treated in our centre between September 2004 and February 2012 by surgery for vesicosphincteric injuries after prostatic external beam radiation therapy was performed. A total of seven men with a median age of 70years±4.1 were included. Mean follow-up was 32.3months±29 (3-86). All patients underwent a partial cystectomy and augmentation enterocystoplasty. The vesicocutaneous fistula rate was 33% occurring within a mean time of 18.5days±2.1 (17-20). The mean length of urinary catheter and hospital stay were respectively 16days±8.4 (12-35) and 18days±7.8 (13-37). Five patients underwent asynchronous insertion of artificial urinary sphincter. The success rate of partial cystectomy and augmentation enterocystoplasty with asynchronous implantation of artificial urinary sphincter to treat post-radiation urinary incontinence was 71.5%. The median time between partial cystectomy and augmentation enterocystoplasty and artificial urinary sphincter implantation was 27.6months±26.4 (7-72). Surgical management of post-radiation urinary cystitis offers good functional outcomes albeit its morbidity is not negligible. It should be proposed only in expert surgical centres.
    Cancer/Radiothérapie 06/2013; DOI:10.1016/j.canrad.2013.04.003
  • [Show abstract] [Hide abstract]
    ABSTRACT: Introduction Prostate cancer (PCa) is the most common disease in male patients and it has the particularity to be androgen dependent. The aim of the current study was to provide an overview about the interest of testosterone dosage during the management of PCa regardless of the stage of the disease. Patients et methods A systematic review of the literature was done from the PubMed database by searching the following key words alone or in combination: prostate cancer; testosterone; risk; aggressiveness; hormonotherapy; active surveillance; prognosis; androgen; cardiovascular risk; biochemical recurrence. Results The level of plasmatic testosterone depends on the moment of the day with a peak between the end of the night and in the morning. We can test either the whole testosterone level, the free testosterone level or the bioavailable testosterone. The bioavailable testosterone is more representative of the presence of androgen in tissues but a specialized laboratory is mandatory. The testosterone plasmatic rate is potentially useful during several steps of the PCa management: in localized prostate cancer cases, men with low testosterone levels are more likely to have an aggressive disease and are therefore not good candidates for active surveillance. An extensive radical prostatectomy should be considered in case of young men since these patients are more likely to recur subsequently; in advanced prostate cancer cases, a testosterone level has to be less or equal to 0.2 ng/mL to guarantee an appropriate castration when a patient is undergoing an androgen deprivation treatment. A dissociation between the trend of PSA and testosterone levels can be the starting point of the castration-resistant period of the disease. Conclusion The testosterone level can bring useful information regarding the profile of PCa and its ability to evolve during the whole natural history of the disease.
    Progrès en Urologie 06/2013; 23(7):438–443. DOI:10.1016/j.purol.2013.01.007
  • [Show abstract] [Hide abstract]
    ABSTRACT: PURPOSE: To assess the surgical approach using the pathological specimen obtained after open radical prostatectomy (ORP) or robot-assisted radical prostatectomy (RALRP). METHODS: A prospective study has been performed in patients who underwent either ORP or RALRP for localized prostate cancer. Two dedicated uro-pathologists, blinded to the surgeons and the operating rooms' schedules, analyzed the pathological specimens according to the Stanford protocol. Both pathologists also determined the surgical approach used based on several criteria pertaining to the pathological specimen. RESULTS: Overall, 117 patients with a median age of 63 years were included. The main characteristics (i.e., Gleason score, pTNM stage, preoperative PSA and margin) were comparable in both groups (p > 0.05). Pathologists 1 and 2 were able to significantly assess the surgical procedure from the pathological specimen provided (in 76.1 and 69.2 % of cases, respectively). Pathologist 1 had a better performance than pathologist 2 (AUC 0.75, IC 95 % [0.67-0.83] vs. AUC = 0.68 IC 95 % [0.59-0.77]) (p = 0.017). The κ index of the inter-observer agreement was satisfactory (0.76). In a univariate analysis, the criteria linked to the pathologist's assessment were as follows: macroscopic integrity of the specimen (p = 0.04), presence of periprostatic fat (p = 0.04), width of periprostatic tissue (p < 0.001) and nerve-sparing status (p < 0.001). CONCLUSION: It was possible to determine the surgical procedure from the analysis of the specimen obtained after a radical prostatectomy. In view of these data and from this perspective, one could infer that there are indeed oncological differences between the robotic and open approaches to radical prostatectomy.
    World Journal of Urology 05/2013; DOI:10.1007/s00345-013-1107-2
  • [Show abstract] [Hide abstract]
    ABSTRACT: PURPOSE: To assess whether the PSA level (threshold 4 ng/mL) is a prognostic factor in biochemical recurrence-free survival in men with prostate cancer (PCa) with an initial PSA level <10 ng/mL who underwent robotic-assisted laparoscopic radical prostatectomy (RARLP). METHODS: We prospectively recruited data for consecutive patients treated by RARLP for PCa with an initial PSA level below 10 ng/mL between 2003 and 2011 at our institution. We divided the population into two groups: patients with a PSA level below 4 ng/mL (G1; n = 53) and patients with a PSA level between 4 and 10 ng/mL (G2; n = 371). Biochemical recurrence was defined as a single increase in PSA greater than 0.2 ng/mL after surgery. Multivariate analysis was used to assess prognostic factors of recurrence-free survival. RESULTS: Overall, 424 patients were included, and the median age was 62 (58-67) years. The median PSA was 5.8 ng/mL (4.8-7.7 ng/mL). Overall, 6 patients from G1 and 34 patients from G2 experienced a biochemical recurrence. Overall, the 5-year recurrence-free survival rate was 86.6 %. The PSA level at diagnosis (under or over 4 ng/mL) was not significantly linked to recurrence-free survival (HR = 0.59, p = 0.25). However, positive margins and a Gleason score >7 on the specimen were significantly linked to recurrence-free survival with respective hazard ratios of 4.30 (p < 0.0001) and 6.18 (p < 0.0001), respectively. CONCLUSION: A PSA level <4 ng/mL alone appears to be obsolete as a cut-off to define a population of men likely to have indolent disease.
    World Journal of Urology 04/2013; DOI:10.1007/s00345-013-1089-0
  • [Show abstract] [Hide abstract]
    ABSTRACT: OBJECTIVE: Our aim was to assess the effect of surgical wait time on the survival of patients with urological neoplasms, including prostate, bladder, penile, and testicular cancers and upper tract tumours (UTUC). MATERIALS AND METHODS: Current, relevant studies were identified from the literature. Keywords used for article retrieval were as follows: delay; surgery; prostate cancer; urothelial carcinoma; renal cell carcinoma; testicular cancer; bladder; renal pelvis; ureter; and survival. RESULTS: Regarding the length of surgical wait time, it does not matter in cases of incidental T1a renal cell carcinomas. In other cases of renal cell carcinomas, surgery should be considered within <1 month; it is of crucial importance in bladder cancer and should be <1 month for a TURBT in cases of non-muscle-invasive bladder cancer and <1 month for a radical cystectomy in cases of muscle-invasive bladder cancer; it is important in invasive UTUC and should be <1 month for a radical nephroureterectomy; it is not crucial in cases of low-risk prostate cancer. In any other case, radical prostatectomy should be considered within <2 months; it is important in testicular cancer and should be fewer than 10 days for an orchiectomy. CONCLUSION: Prolonged surgical wait times have an impact on the overall quality of life and anxiety of the patient. Extending the wait time beyond a given threshold can also have a negative impact on the patient's clinical outcomes, but this threshold differs between urological neoplasms.
    World Journal of Urology 03/2013; 32(2). DOI:10.1007/s00345-013-1045-z
  • [Show abstract] [Hide abstract]
    ABSTRACT: Cystoprostatectomy (CPT) is the gold standard surgical treatment for muscle invasive bladder cancer (MIBC). In certain cases, MIBC can invade the prostate gland and/or a prostate cancer (PCa) can be discovered fortuitously on the pathologic specimen. The aim of the current study was to report the prognostic influence of PCa in patients who underwent a CPT for MIBC. A systematic review of the scientific literature was achieved in the Pubmed database, using the following keywords: prostatic neoplasm; urinary bladder neoplasm; cystectomy; surgery; recurrence; prognosis; survival. Clinical cases and series of less than five cases were deliberately excluded herein. Overall, ten studies published between 2004 and 2011 and involving 2196 patients were selected. Only retrospective studies of low level of evidence (NP 4) were available. The incidence of neoplastic invasion of the prostate gland by MIBC ranged from 25 to 48%. Preoperative predictors were multiple BC, recurrent, location in the trigone and existence of CIS. Overall survival at 3years was significantly affected by the invasion of the prostate gland (pT4a) in these patients. The incidence of PCa discovered incidentally pathologic specimen CPT ranged from 14 to 49%. Only age was found as a positive predictor. The diagnosis of PCa did not influence survival of patients with MIBC and no specific PCa adjuvant treatment was systematically advocated. Fortuitous diagnosis of PCa and/or neoplastic invasion of the prostatic gland by BC on CPT specimen is not uncommon but this is variable across studies, depending on the quality of the pathological analysis. The invasion of the prostate gland by MIBC is a serious situation (pT4a) and linked with a poor prognosis. In case of concomitant PCa and MIBC, the prognosis is much more related to the natural history of the bladder tumour.
    Progrès en Urologie 03/2013; 23(3):165-70. DOI:10.1016/j.purol.2012.11.003
  • [Show abstract] [Hide abstract]
    ABSTRACT: OBJECTIVES: To describe the most recent data from phase I and II clinical trials of stereotactic body radiation therapy (SBRT) using image-guided robotic radiosurgery, specifically the Cyberknife(®) (Accuracy Incorporated, Sunnyvale, CA, USA). To better determine thecurrent role of this type of radiosurgery in prostate cancer (PCa) management. MATERIALS AND METHODS: Current clinical trials and relevant retrospective studies were identified from the literature, clinical trial databases, websites and conference abstracts. The indications, technical aspects, efficacy and toxicity of SBRT using the Cyberknife(®) system were summarized. RESULTS: The Cyberknife(®) system is an experimental treatment mostly used for localized PCa in stage cT1/T2a-b N0 M0 with a Gleason score ≤7 and PSA level ≤20 ng/mL. Hypofractionated radiation therapy was delivered in five fractions of 7-7.25 Gy for a total dose of 35-36.25 Gy. After treatment, the median PSA levelfell from 4.9-8.3 ng/mL to 0.1-1.6 ng/mL at a median follow-up of 4-60 months. The biochemical progression-free survival rates ranged from 78.3 to 100%. Acute and late toxicities were mostly grade 1/2 rectal or urinary complications. Few grade 3 and no grade 4 toxicities occurred during follow-up; however, erectile dysfunction and testes toxicity were also reported. CONCLUSIONS: The use of the Cyberknife(®) system is limited mainly by its pretreatment and maintenance costs. Despite encouraging preliminary results, longer-term follow-up and randomized controlled phase III clinical trials are necessary before the Cyberknife(®) system becomes a standard treatment method.
    BJU International 01/2013; 111(5). DOI:10.1111/bju.12000
  • [Show abstract] [Hide abstract]
    ABSTRACT: PURPOSE: To analyze the results of the bibliometric system (SIGAPS score) of scientific publications in the Assistance publique-Hôpitaux de Paris (AP-HP) and to compare the scientific production among the various medical and surgical specialties of the academic hospitals of Paris. METHODS: All the publications imported from Pubmed between 2006 and 2008 were included. The following data were taken into account and analysed: the hospital department of origin, the number of articles published, the number of full-time physicians, the SIGAPS score. RESULTS: Thirty-eight thousand, seven hundred and nine publications were included. The departments were consisted of 747 full-time practitioners 5719 (1895 Professors [33.1%], 2772 Assistant Professors [48.4%] and 1052 fellows [18.4%]). The average number of full-time practitioner by department was 7.7±6.7 (range 1-69). The average total number of articles published in a department was 51.8±49.4 (range 1-453). The average SIGAPS score was more important in medicine than in surgery (621.2±670.1 vs. 401±382.2; P=0.01) but not the average number of article per practitioner (8.1±8.3 vs. 6.6±6.2; P=0.0797). The mean number of publication by full-time practitioner was 7.9±7.8 (1-45), or an average of 2.7±2.6 for each full-time practitioner each year. CONCLUSION: Academic hospitals in Paris have a reasonably scientific output but with a mean of 2.7 articles per full-time practitioner per year. No major differences between medical and surgical disciplines were observed.
    La Revue de Médecine Interne 12/2012; 34(6). DOI:10.1016/j.revmed.2012.08.006
  • [Show abstract] [Hide abstract]
    ABSTRACT: OBJECTIVES: To identify the predictive tools which have emerged recently in the field of urothelial carcinomas. MATERIALS AND METHODS: We performed a thorough MEDLINE literature review using a combination of the following keywords: urothelial carcinoma, transitional cell carcinoma, bladder, renal pelvis, ureter, predictive tools, predictive models and nomograms. We found 117 articles, but only the relevant reports were selected. RESULTS: The majority of available tools are prediction models, particularly nomograms. These models combine good performance accuracy with ease of use. They appear to be more accurate than risk grouping or tree modeling and are more suitable for clinicians than artificial intelligence. The most recent nomograms have been designed to be used in daily clinical practice and are even available as computer or smartphone applications. They focus on pathological outcomes or more frequently on survival statistics or recurrence risk after surgery. They provide an accurate prediction of disease evolution and may help clinicians to choose the most appropriate treatment option. However, these prediction tools still need to be validated and regularly utilized. CONCLUSION: Predictive tools represent very helpful clinical decision-making aids but need to be validated in larger populations.
    World Journal of Urology 12/2012; 31(1). DOI:10.1007/s00345-012-1008-9
  • Progrès en Urologie 11/2012; 22(13):744–745. DOI:10.1016/j.purol.2012.08.015

Publication Stats

228 Citations
141.28 Total Impact Points


  • 2014–2015
    • Hôpitaux Universitaires La Pitié salpêtrière - Charles Foix
      Lutetia Parisorum, Île-de-France, France
  • 2013
    • Polytech Paris-UPMC
      Lutetia Parisorum, Île-de-France, France
  • 2010–2013
    • Assistance Publique – Hôpitaux de Paris
      • Department of Radiology
      Lutetia Parisorum, Île-de-France, France
  • 2011–2012
    • Pierre and Marie Curie University - Paris 6
      • Faculté de médecine Pierre et Marie Curie
      Lutetia Parisorum, Île-de-France, France