• 2019-10
  • 2019-11
  • 2020-03
  • 2020-07
  • 2020-08
  • NSC 232003 br Results br Cohort selection


    3. Results
    3.1. Cohort selection
    Of the 5795 women identified with FIGO Stage III-IVA Type 1 (endometrioid histology, grades 1 or 2) uterine cancer with known type of adjuvant therapy who underwent hysterectomy and received ei-ther RT and/or CT, 1260 (21.7%) received RT only, 2465 (42.5%) received CT only, 564 (9.7%) received RT before CT, and 1506 (26.0%) received RT after CT. Clinicopathologic and demographic characteristics of women within each treatment subgroup are shown in Table S1. The most com-mon fractionation scheme for RT was 45 Gy in 25 fractions, while the second most common was 50.4 Gy in 28 fractions. Clinicopathologic and treatment characteristics were well-balanced for all matched co-horts (Tables S2–S9).
    3.2. Kaplan-Meier estimates and multivariable survival analyses
    0.001). Older age, more advanced tumor or nodal stage, higher grade, receipt of omentectomy, positive peritoneal cytology, positive margin status, black race, Medicare insurance or no insurance, higher comor-bidity scores, a history of prior cancer, and treatment other than at an academic facility were also associated with significantly shorter OS in this model, while Hispanic ethnicity was associated with significantly longer OS (Table 1).
    Table 1
    Kaplan-Meier estimates of survival and multivariable parametric accelerated failure time models of all women who received adjuvant radiotherapy and/or chemotherapy.
    Variable Univariate analysis of all patients
    Multivariable AFT model
    Adjuvant therapy
    Tumor stage
    Nodal stage
    Nodal surgery
    Grade r> LVSI
    Margin status
    Insurance status
    Comorbidity score
    Prior cancer
    Income quartile
    Education quartile
    Treatment facility
    (continued on next page)
    Table 1 (continued)
    Variable Univariate analysis of all patients
    Multivariable AFT model
    Year diagnosed
    Significance determined by Log-Rank test (univariate analyses) or Wald test (multivariable analyses). OS: overall survival; mo: months; y: years; N: number; RMST: restricted mean sur-
    vival time; TR (95% CI): time ratio (95% confidence interval); P: P-value; RT: radiotherapy; CT: chemotherapy; LND: NSC 232003 node dissection; PA: para-aortic; LVSI: lymphovascular space
    invasion; NOS: not otherwise specified; Gov: other government agency; ref.: reference level.
    = 1.03, 95% CI, 0.94–1.13, P = 0.49; Table 3, Fig. 2F). Multivariable models of unadjusted analyses are presented in Tables S10–S11. When the number of days between RT and chemotherapy start dates for the patients who received RT after CT was plotted on a histogram there appeared to be a bimodal distribution (Fig. S2). One subset of pa-tients started RT between 50 and 100 d following the start of CT (peak ~65 d), while a second subset started RT between 110 and 200 d follow-ing the start of CT (peak ~145 d). To explore whether the start date of RT
    following CT was associated with survival benefit, a multivariable anal-ysis was performed of propensity-score matched cohorts of women whose RT start date was 126–200 d after her CT start date (and there-fore thought to have received 5 or 6 cycles of CT) and women whose RT start date was 61–125 d after her CT start date (and therefore thought to have received 3 or 4 cycles of CT). There was no association of this variable with OS (TR = 0.87, 95% CI = 0.73–1.04, P = 0.13). When an expanded cohort including women with Grade 3 endometrioid histology was utilized, results were consistent those of the original cohort of Type 1 endometrioid histology alone. In this ex-panded cohort of women, treatment with RT after CT was associated with significantly longer OS compared to treatment with RT before CT (5 y OS: 72.6% vs 65.8%; time-ratio (TR) = 1.22, 95% CI = 1.05–1.42, P
    4. Discussion
    In this analysis of a large cohort of women from the NCDB with Stage III-IVA Type 1 EC following hysterectomy, treatment with RT after CT was associated with longer OS compared to treatment with RT before CT, or either treatment alone. Women who received RT before CT, how-ever, did not experience significantly longer OS than women who re-ceived CT or RT alone, nor did women who received CT alone experience significantly different survival than women who received RT alone. Treatment with multi-agent CT followed by RT may therefore 
    be the optimal adjuvant multimodality regimen for selected women with EC whose disease characteristics are prognostic for a high risk of LRR but a lower risk of distant recurrence. Historical randomized trials addressing women with locally ad-vanced EC have included heterogeneous cohorts and yielded results with limited applicability. An Italian trial of women with primarily Stage III non-serous or clear cell EC randomized after hysterectomy to RT to the pelvis ± para-aortic (PA) lymph node chain versus multi-agent CT found no difference in PFS or OS [16]. Similarly, a Japanese trial of women with a mix of stages and histological subtypes random-ized following hysterectomy and lymph node dissection (LND) to RT or to cisplatin/doxorubicin CT demonstrated no differences in PFS or OS between arms, although subset analyses suggested superior survival for those patients deemed “high risk” treated with CT [17]. Subset anal-yses of matched cohorts of patients treated with either RT or CT reflect these data (Fig. 2F). The role of adjuvant CT in this setting was cemented following GOG-122, in which women with Stage III-IVA EC randomized after hysterectomy and LND to cisplatin/doxorubicin experienced sig-nificantly longer PFS and OS compared to women who received whole abdominal RT with a pelvic boost. Recurrences in the pelvis, however, were more common in the CT arm, comprising 21% of all first failures [18]. It seemed logical to next explore treatment with a combination of CT and RT, addressing the potential for both local and systemic relapse.