You are here

Article Text

Article menu

Download PDFPDF

ePoster Viewing
AS04. Endometrial/Uterine corpus cancers
EP136/#225  Evaluation of uterine endometrial carcinoma histological grades using magnetic resonance imaging texture analyses
Free
  1. Eito Kozawa1,
  2. Kaiji Inoue1,
  3. Saki Tuchihashi1,
  4. Hirokazu Shimizu1,
  5. Yasutaka Baba2,
  6. Kosei Hasegawa3 and
  7. Masanori Yasuda4
  1. 1Saitama Medical University, Radiology, Saitama, Japan
  2. 2Saitama Medical Univetsity, International Medical Center, Imaging Diagnosis, Saitama, Japan
  3. 3Saitama Medical university, International Medical Center, Gynecologic Oncology, Saitama, Japan
  4. 4Saitama Medical University, International Medical Center, Pathologic Diagnosis, Saitama, Japan

Abstract

Introduction The histological tumor grade of uterine endometrial carcinoma (UEC) is one factor that can determine the prognosis. However, studies have shown that some histological grades assigned by preoperative biopsy results did not correspond to the final grades of the surgical specimens. This study evaluated the possibility of predicting the UEC histological grade using magnetic resonance imaging texture features (TFs).

Methods This retrospective study included 70 patients with UEC. We evaluated axial T2-weighted imaging (T2WI) TFs, axial apparent diffusion coefficient (ADC) TFs, sagittal T2WI TFs, and their combinations to determine histological class 1 (Grade 1: n=33) and class 2 (Grade 2 and Grade 3: n=37) using texture analyses. The least absolute shrinkage and selection operator was used to select four TFs for each model and construct a discriminative model. A binary logistic regression analysis and receiver-operating characteristic analysis of the axial T2WI TFs, axial ADC TFs, sagittal T2WI TFs, and combined TFs models were performed to compare the two histological class.

Results Four models were constructed from each of the four selected features. The area under the curve (AUC) values of the discriminative model using these features were 0.71, 0.70, 0.77, and 0.82 for the sagittal T2WI TFs, axial T2WI TFs, axial ADC TFs, and combined TFs models, respectively . The AUC value of the combined TFs model was the highest.

Conclusion/Implications A combined TFs model may help distinguish UEC histological grades.

View this table:
Abstract EP136/#225 Table 1

Diagnostic performance of the model for differentiating histological class of uterus endometrial carcinoma

Abstract EP136/#225 Figure 1
Abstract EP136/#225 Figure 1

The receiver operating characteristic curves analysis of the diagnostic model

https://doi.org/10.1136/ijgc-2023-IGCS.232

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.