Association of body mass index with pathologic agreement of preoperative and postoperative tumor grade in endometrial cancer.

Abstract

Objective: We aim to study association of BMI of EC patients, with the level of agreement between preoperative and postoperative tumor grade.

Methods: A retrospective study. We included patients with EC diagnosed in an outpatient clinic which had surgical staging as in our division. We categorized patients into BMI categories according to the World Health Organization; (BMI < 18.5 kg/m²), (BMI 18.5-24.9 kg/m²), (BMI 25-29.9 kg/m²), (BMI 30-34.9 kg/m²), (BMI 35-39.9 kg/m²), and (BMI ≥ 40 kg/m²). We further dichotomized the study population for obesity, defined as BMI ≥ 30.0. We analyzed agreement between preoperative and postoperative tumor grade, stratified by patient's BMI.

Results: Overall, 623 women met study inclusion criteria, with a median age of 64 [interquartile range (IQR) 57-72]. Among the study cohort, the median BMI was 30.7 [IQR 25.6-38.8], with 330 (53.0%) patients being obese. EC grade 1 was diagnosed preoperatively in 353 (56.7%), grade 3 in 148 (23.8%), and grade 2 in 122 (19.6%). Endometrioid histology was diagnosed in 463 (74.3%), serous in 78 (12.5%), mixed histotype in 51 (8.2%), clear cell in 20 (3.2%) and carcinosarcoma in 11 (1.8%). In 68.7% (n = 428), there was no change in postoperative grade, and in 24.9% (n = 155), there was upgrading of tumor, and in 6.4% (n = 40), there was a tumor downgrade. There were 3 (0.5%) cases in which no tumor was found on final pathology. The rate of no change was higher in preoperative grade 3 (89.9%) vs. grades 1 (63.5%) and grade 2 (58.2%), p < .001). There was no difference in grading agreement when obese patients were compared to non-obese, p = .248. There was no difference in grading agreement when comparing the various BMI categories, with no change proportion ranging between 58.2% in BMI 30.0-34.9 mg/kg² and 79.7% in BMI 35.0-39.9 mg/kg², p = .104. ROC analysis of BMI as predictor of no-change yielded an area under the curve of 0.466 (95% confidence interval 0.418-0.515) with a maximal performance at a BMI of 33.8 mg/kg². The agreement between preoperative and postoperative tumor grade among all patients was kappa = 0.517. The agreement did not differ when compared between obese patients (kappa = 0.456) and non-obese (kappa = 0.575).

Conclusion: Our study found no significant association between BMI and the agreement between preoperative and postoperative tumor grading in EC.