Skip to Main content Skip to Navigation
Master Thesis

Caractérisation des masses annexielles en IRM : apport de l’imagerie 3D dynamique injectée

Abstract : Purpose: to test the accuracy of a new 3D DCE MR sequence, with a lower temporal resolution in addition to ADNEX MR-SCORE for preoperative adnexal masses characterization. Material and methods: sixty-seven women (age range 25-88) with 83 adnexal masses with solid tissue (34 benign, 10 borderline, 39 invasive malignant) underwent MR scan with 3D DCE MR sequence with pharmacokinetic modeling. A senior and a junior radiologist reviewed images, scored MR results according to previously published ADNEXMR-SCORE. Then, DCE-MR imaging parameters (AUC, WASHIN, WASHOUT, PEAK, SER, KEP, KTRANS, VE, VP) were evaluated on solid tissue in average, percentiles, standard deviation, skewness and kurtosis, according to surgical histopathology or two-years follow-up findings. Statistical analysis was performed by using Fischer exact, chi-squared, Kruskal-Wallis, and Mann Whitney tests and ROC analysis. Results: time intensity curve type 1 predicts benignity with a PLR=31.7 with both ROIs technique. Time intensity curve type 3 better predicts invasiveness than other time intensity curve with an OR = 26.0 using Hotspot ROIs technique (p<0.0001, 95%IC 5.46 – 250.84) and OR= 9.1 using AreaROIs technique (p=0.002). Overall accuracy of ADNEXMR score was 87.5% (73/83) with one false negative and 9 false positives. Borderline tumors display higher 75th percentiles in KEP than benign and lower than invasive malignant tumors (p<0.05). An ADNEX MRSCORE 4 or 5 in a tumor without solid portion predicts borderline with a PLRborderline= 12.8 and in a subgroup of tumors ADNEX MR-SCORE 4 or 5 with a solid portion, a WASHOUT average > 29% predicts invasiveness with a PLRinvasiveness = 27.1. Conclusion: a low temporal resolution 3D DCE MR sequence is an accurate wide generalizable technique to improve adnexal masses characterization.
Document type :
Master Thesis
Complete list of metadata
Contributor : Université Paris Descartes - Scd Connect in order to contact the contributor
Submitted on : Tuesday, February 23, 2021 - 10:39:29 AM
Last modification on : Thursday, September 2, 2021 - 3:09:14 AM


Files produced by the author(s)


Distributed under a Creative Commons Attribution - NonCommercial - NoDerivatives 4.0 International License


  • HAL Id : dumas-03149565, version 1


Louise Gervais-André. Caractérisation des masses annexielles en IRM : apport de l’imagerie 3D dynamique injectée. Médecine humaine et pathologie. 2017. ⟨dumas-03149565⟩



Record views


Files downloads