ARRS 2022 Abstracts

RETURN TO ABSTRACT LISTING


E2163. Diagnostic Accuracy of MRI in Predicting Depth of Invasion in Endometrial Cancer
Authors
  1. Anuja Liyanage ; Middlemore Hospital
  2. Supriya Cardoza ; Middlemore Hospital
  3. Helen Moore; Auckland City Hospital
  4. Darshna Kasabia; Middlemore Hospital
Objective:
Depth of myometrial invasion in endometrial cancer staging is the most important prognostic factor, correlating with tumor grade, presence of lymph node metastases, and overall patient survival. Accurate preoperative MRI assessment of depth of invasion leads to improved preoperative assessment, triage, and treatment. The objective of this research project was to investigate the accuracy of MRI in predicting the depth of myometrial invasion in preoperative assessment of women with endometrial cancer, using the 2009 International Federation of Gynaecology and Obstetrics (FIGO) stage as 1A versus 1B. Although the standard surgical intervention for both stages typically includes a total hysterectomy, bilateral salpingo-oophorectomy, and peritoneal cytology, the difference lies in an added step of nodal sampling if greater than 50% myometrial invasion (1B staging) is demonstrated due to high risk of nodal metastases. Therefore, the goal of this exhibit is to become aware of predictive capacities and raise possible factors influencing inaccuracy to potentially improve surgical planning and patient outcomes.

Materials and Methods:
We conducted a retrospective study of early endometrial cancer using collected data from the regional weekly gynecology multidisciplinary meeting held at Auckland City Hospital between January 2020 and January 2021. NHI and basic demographic details were collected, noting staging initially mentioned on the original report (if available) and the final stage assigned by a specialist radiologist at the MDM. The originally reported stage and stage discussed at MDM were compared to evaluate for discrepancy, with subsequent staging from MDM compared with final histopathology. Further data were collected regarding the presence of background uterine disease such as fibroids and adenomyosis. Exclusion criteria included: advanced stage; exact FIGO not clearly stated at MDM discussion; MRI not performed; no surgical intervention performed; and if histology tumour type was unsuitable for standard FIGO staging (adenosarcoma). Data were evaluated for sensitivity, specificity, NPV, PPV, accuracy, and overall statistical significance of MRI staging and histology outcomes.

Results:
Sensitivity was 52% (95% CI: 0.38–0.65); specificity was 86% (95% CI: 0.80–0.91); PPV was 54% (95% CI: 0.40–0.67); NPV was 85% (95% CI: 0.79–0.90); agreement rate was 78% (95% CI: 0.72–0.83); false-positive rate was 14%; false-negative rate was 48%; McNemar’s Test with Yate’s correction was p = 0.89); out of false negatives and false positives, 13/52 were background adenomyosis and 7/52 were background leiomyoma.

Conclusion:
Our study showed high specificity and moderate sensitivity; however, the difference between 1B and 1A groups was not statistically significant. In the presence of confounding factors (adenomyoma and leiomyoma) MRI prediction of depth of invasion is less sensitive.