Validation of ultrasound strategies to assess tumor extension and to predict high-risk endometrial cancer in women from the prospective IETA (International Endometrial Tumor Analysis)-4 cohort
Abstract
OBJECTIVES: To validate ultrasound measurements and subjective ultrasound assessment (SA) to detect deep myometrial invasion (MI), and cervical stromal invasion (CSI) in patients with endometrial cancer and to compare their performance between low and high-grade endometrial cancer, and to validate published prediction models to identify high-risk endometrial cancer (grade 3 endometrioid or non-endometrioid cancer and/or deep MI and/or CSI). METHODS: The study comprises 1538 patients from the prospective IETA4 multicenter study with endometrial cancer undergoing standardized expert transvaginal ultrasound examination.
SA and ultrasound measurements were used to predict deep MI and CSI. We assessed the diagnostic accuracy of the Tumor/Uterine anteroposterior (AP) ratio to detect deep MI and the distance from the lower margin of the tumor to outer cervical os (Dist-OCO) to detect CSI, and validated two 2-step strategies to predict high-risk cancer.
In the 2-step strategies the first step consists of biopsy grade 3/non-endometrioid cancers were classified as high-risk cancer, and the second step encompasses application of a mathematical model on the remaining tumors. The "subjective model" included biopsy grade (1 versus 2) and subjective assessment of deep MI/CSI (deep MI or CSI: yes or no), the "objective model" includes biopsy grade (1 versus 2) and minimal tumor-free margin.
The two 2-step strategies were compared to simply classifying patients as high-risk if either deep MI or CSI was suspected based on SA or if biopsy showed grade 3/non-endometroid histotype (combining SA with biopsy grade). Histological assessment from hysterectomy was considered the reference standard.
RESULTS: Among patients with measurable lesions (n=1275), SA had a sensitivity and specificity of 70% and 80% to detect deep MI in grade 1-2 tumors versus 76% and 64% in grade 3/non endometrioid tumors. The corresponding percentages for detection of CSI were 51% and 94% versus 50% and 91%.
Tumor AP diameter and Tumor/Uterine AP ratio were the best ultrasound measurements to predict deep MI, and Dist-OCO was best to predict CSI (area under receiver operating characteristics curve(AUC) of 0.77 and 0.72). The proportion of patients correctly classified as having high-risk cancer was 80% for simply combining SA with biopsy grade versus 80% and 74% for the subjective and objective 2-step strategies, respectively.
The subjective and objective models had AUC of 0.76 and 0.75 when applied to grade 1-2 endometrioid tumors. CONCLUSIONS: In the hands of experienced ultrasound examiners SA was superior to taking measurements for prediction of deep MI and CSI of endometrial cancer especially in grade 1-2 tumors.
The mathematical models for prediction of high-risk cancer performed as expected. The best strategy to predict high-risk endometrial cancer was either to simply combine SA with biopsy grade or to use the subjective 2-step strategy, both having an accuracy of 80%.