Texture Analysis in Preoperative Magnetic Resonance Imaging for Assessing Recurrence Risk in Endometrial Cancer After Hysterectomy
- Author:
Juhun LEE
1
;
Hyun Jung LEE
Author Information
- Publication Type:Original Article
- From:Investigative Magnetic Resonance Imaging 2024;28(2):76-86
- CountryRepublic of Korea
- Language:English
-
Abstract:
Purpose:This study aimed to assess the feasibility of texture analysis using preoperative magnetic resonance imaging (MRI) for assessing the recurrence risk of endometrial cancer after hysterectomy.
Materials and Methods:Eighty-five patients who underwent surgery and had pathologically confirmed endometrial cancer were considered in this study. Histographic parameters (perfusion ratio, integrated density, skewness, and kurtosis) and gray-level co-occurrence matrix textural parameters (angular second moment, contrast, correlation, entropy, and inverse difference moment) obtained from normalized perfusion mapping and apparent diffusion coefficient (ADC) maps of pelvic MRI were correlated with pathological features, including tumor type, International Federation of Gynecology and Obstetrics staging, tumor-involved lymph node metastasis, and recurrence after hysterectomy.
Results:In distinctions between type I and II endometrial cancers, the histogram analysis of the perfusion map revealed significant differences in tumor area (p = 0.014), perfusion ratio (p = 0.001), integrated density (p = 0.042), entropy (p = 0.001) on the perfusion map, and energy (p = 0.004) and entropy (p = 0.007) on the ADC map. The assessment of the relationship with recurrence revealed significant differences in contrast (p = 0.013), and entropy (p < 0.001) on the perfusion map, and energy (p < 0.001) and entropy (p = 0.003) on the ADC map. Entropy obtained via texture analysis demonstrated associations between integrated density and linear correlation, with notable differences observed between type I (R 2 = 0.363) and type II (R 2 = 0.471) endometrial cancer subtypes (p = 0.010).
Conclusion:The parameters obtained through a texture analysis on preoperative MRI could be employed as potential quantitative predictors for the assessment of the recurrence risk in endometrial cancer after hysterectomy.
