Not all high-intermediate risk endometrial cancers are created equal: recurrence-free survival and cause-specific survival after observation or vaginal brachytherapy in all possible subgroups of early-stage high-intermediate risk endometrial cancer

Abstract

<h3>Objectives:</h3> Historically, patients with early-stage high-intermediate risk (HIR) endometrioid endometrial cancer (EC) have been known to benefit from improved recurrence-free survival (RFS) with the use of adjuvant therapy. However, specific subgroups of HIR patients may benefit from observation ± vaginal brachytherapy (VB) only. We sought to evaluate RFS and cause-specific survival (CSS) after observation ± VB in the subgroups of early-stage HIR endometrioid EC. <h3>Methods:</h3> We identified patients with early-stage HIR endometrioid EC, who underwent a total hysterectomy, bilateral salpingo-oopho-rectomy, and sentinel lymph node biopsy or pelvic ± paraaortic lymphadenectomy at Mayo Clinic and Cleveland Clinic between 1999 and 2016. HIR group was defined based on revised GOG99 criteria used in the GOG249 trial: age ≥70 with 1 risk factor, age ≥50 with 2 risk factors, or all the 3 risk factors (grade 2-3, ≥50% myometrial invasion, or positive lymphovascular space invasion [LVSI]). HIR subgroups were determined by all possible combinations of risk factors, and 3-year RFS, site-specific RFS, and CSS were estimated. <h3>Results:</h3> Among 1,507 women with stage I or II EC, 447 stage I endometrioid EC patients were identified who fulfilled HIR criteria, and 349 (78.1%) were managed with observation ± VB. Eleven subgroups were defined and evaluated (Table). Among stage IA patients (1-5 subgroups), subgroup 5 (G3 ≥50y LVSI+) had the lowest RFS and CSS of 68.9% (46.6-100.0%) and 86.7% (71.1-100.0%), respectively, whereas the remaining 4 subgroups had 3-year RFS and CSS >90%. In all stage IA subgroups, 3-year lymphatic RFS was >90%; 3-year distant RFS was lowest in subgroup 5 (79.6% [56.9-100.0%] and 97.1% [91.8-100.0%] in subgroup 3 [G2 ≥50y LVSI+]). Additionally, 3-year CSS was <100% in these two subgroups (vs. 100% in subgroups 1, 2, and 4). Among the 6 stage IB subgroups (6-11), 3-year RFS ranged from 68.2% (48.6-95.7%) to 93.6% (87.7-99.9%) and 3-year CSS from 80.0% (58.7-100.0%) to 100.0% (100.0-100.0%). Lymphatic recurrence was observed in subgroup 8 (IB G2 ≥50y LVSI -), but site-specific RFS was >90%. Conversely, all stage IB subgroups experienced distant recurrence; 3-year distant RFS ranged from 73.1% (53.5-99.7%) to 96.5% (91.8-100.0%). 3-year distant RFS lower than 90% was observed in those subgroups with grade 3 or positive LVSI (7, 9, 10, and 11). <h3>Conclusions:</h3> Our results suggest that observation ± VB can be sufficient in some subgroups of stage I endometrioid HIR patients, specifically stage IA endometrioid HIR EC without the concomitant presence of grade 3 and LVSI. Among stage IB endometrioid HIR EC subgroups, particularly those with grade 3 or LVSI, observation ± VB is associated with lower RFS and CSS. Stratifying HIR patients into risk subgroups would alleviate the risk of overtreatment and undertreatment among patients with early-stage endometrioid EC; in this regard, the molecular characterization will provide a further instrument to achieve this objective.