Nomogram with potential clinical use to predict lymph node metastasis in endometrial cancer patients diagnosed incidentally by postoperative pathological assessment.

ABSTRACT

PURPOSE: To determine the clinical and pathological risk factors for lymph node metastasis (LNM) in patients with endometrial cancer and to create a nomogram to predict LNM in patients without surgical staging. METHODS: All patients with endometrial adenocarcinoma who were treated surgically at a university based gynecologic oncology clinic between January 2011 and December 2014 were recruited. Women with endometrial adenocarcinoma who were surgically staged including lymphadenectomy were included in the study. Data regarding clinical and pathological risk factors were recorded. The histopathologic slides from the staging surgeries were re-evaluated microscopically by a gynecologic pathologist for all parameters along with lymphovascular space invasion (LVSI). RESULTS: A total of 279 patients with endometrial cancer were analyzed. Among those, 31 (11.1%) had lymph node metastasis. According to the univariate analyses, elevated CA 125 (>35 U/mL), LVSI, myometrial invasion ≥50%, grade 3 disease, non-endometrioid type, and cervical stromal involvement were significantly associated with LNM. The multivariate logistic regression analysis showed that LVSI, non-endometrioid type, elevated CA 125, and cervical stromal involvement increased the risk of LNM. However, myometrial invasion and grade did not significantly affect the risk of LNM. A nomogram to predict LNM was constructed using these factors (concordance index 0.92). CONCLUSIONS: LVSI is the most important predictor for LNM. The present nomogram can be useful to decide if adjuvant therapy is required for patients who undergo simple hysterectomy for a benign etiology and incidentally diagnosed with endometrial cancer by pathological evaluation.