Endometrial Cancer - Studies on recurrences, complications and preoperative diagnostics

Abstract

Introduction: The most common gynecological cancer is Endometrial Cancer (EC). The prognosis is generally favorable, mainly due to an early diagnosis. However, there are subgroups of EC with a higher risk for metastases and recurrences resulting in poorer survival. Primary treatment for EC is surgical, with hysterectomy and bilateral salpingo-oophorectomy and in higher risk groups adding surgical staging with lymph node assessment for the adjuvant treatment planning. Aim: The overall aim of this thesis was to study recurrence, survival and surgical complications in a population-based cohort and to assess the introduction of the first national guidelines (NGEC), which recommended pelvic and para-aortic lymphadenectomy (PPLND) in high-risk EC. A second aim was to evaluate preoperative risk classification assessment with transvaginal ultrasound (TVUS) and magnetic resonance imaging (MRI) in low-grade endometrioid EC. Methods: Paper I-III were regional population-based studies in the Western Sweden Health Care Region (WSHCR). Data was retrieved from the Swedish Quality Register for Gynecological Cancer (SQRGC) for all EC patients in the WSHCR 2010-2017. Medical records were reviewed for details of recurrence, complications, and patient characteristics, such as BMI and comorbidities. Patients with primary surgical treatment for pre-operative early- stage EC were included in the studies. Paper I encompassed patients with endometrioid EC and Paper II non-endometrioid EC. In Paper III, patients who underwent surgery at the tertiary center were included and complications 30 days postoperatively were recorded and graded according to the Clavien-Dindo (CD) classification system. Overall (OS), net (NS) and disease- free survival (DFS) were calculated using the Kaplan-Meier method. The Cox proportional hazards regression model was used in Paper I-III to evaluate the effect of identified variables on DFS and OS. Uni- and multivariable logistic regression analyses were performed with complications as outcome in Paper III. Paper IV was a prospective multicenter study in the WSHCR including patients with low-grade EC planned for primary surgery during 2017-2019. The patients were examined preoperatively with both TVUS and MRI to assess deep myometrial infiltration (MI) and cervical stroma invasion (CSI) for the decision on surgery with or without PPLND. The TVUS was performed by gynecologists, and the MRI was performed according to a standardized protocol. The methods were analyzed for sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy. The methods were compared using McNemar’s test and Cohen’s kappa (k).

Results: In the endometrioid EC cohort in Paper I, 8.3% (136/1630) experienced a recurrence. In the non-endometrioid EC cohort in Paper II, the recurrence rate was 29% (67/228). The total 5-year DFS was 83.9% for the endometrioid EC cohort (Paper I) and 61.9% for the non- endometrioid EC cohort (Paper II). If no recurrence occurred, the 5-year OS was 91.9% in the endometrioid EC cohort (Paper I) and 88.5% in the non-endometrioid EC cohort (Paper II). When a recurrence occurred the 5-year OS for the endometrioid EC cohort was 77.0% for isolated vaginal recurrences compared to 36.1% for all other recurrences (Paper I). The 5-year OS was 13.4% when a recurrence occurred in the non-endometrioid EC cohort (Paper II). In Paper I, age, FIGO stage and primary treatment were found independent risk factors for recurrence. In Paper II, the OS before the implementation of NGEC was 57.3% compared to 72.0% after. Age, FIGO stage and lymph node dissection were found significant factors for DFS, where having a lymph node dissection decreased the risk of recurrence or death. In Paper III, 19.7% (108/549) had a surgical complications of CD grade II-V. Surgical technique, BMI and lymph node dissection, were found to be risk factors for complications CD. In Paper IV (n=259), MRI and TVUS were compared for the assessment of deep MI and CSI and there was a statistically significant difference in specificity, with MRI having a higher specificity. No difference in sensitivity was found. Conclusions: For endometrioid EC, the recurrence rate was overall low in contrast to non- endometrioid EC where the recurrence rate was rather high. The survival was excellent when no recurrence occurred, in both endometrioid and non-endometrioid EC. However, in cases of recurrence, survival was poor, with the exception of isolated vaginal recurrence, where the prognosis was favorable. A significant improvement in survival was seen in non-endometrioid EC after the NGEC implementation with lymph node staging tailoring adjuvant radiotherapy. However, in Paper III we show that surgical staging with lymphadenectomy is a risk factor for surgical complications. This may be taken into consideration in treatment guidelines for EC, where steps moving towards a less extensive lymph node assessment surgery with the sentinel node procedure may be advocated. For the assessment of deep MI, MRI had a higher accuracy than TVUS. Nevertheless, the sensitivity of TVUS performed by gynecologists was evaluated as acceptable and did not differ from MRI. TVUS is readily available, and Paper IV supports this method for first-hand use in similar settings.