We included a total of 5037 young patients and 60,612 older EC patients from the SEER database. In terms of race, both groups had a high proportion of white patients, 73.5% and 81.3%, respectively. The histological type of adenocarcinoma accounted for 80.4% of the younger patients. Of the histological grades, a higher proportion of younger patients were in grade I (50.8%) and a higher number of older patients were in grade II or above (63.5%). T1 stage was more frequent in both AJCC stages, accounting for 82.2% and 77.6% of patients, respectively, with fewer patients with lymph nodes and distant metastases. The number of young patients and older patients receiving adjuvant treatment were both low. Survival analysis showed that overall survival and cancer-specific survival were higher in younger patients than in older patients (P-value < 0.001) (Fig. 2).
Younger Patient Characteristics
Confirmation of the cut-off age X-tile software analysis result displayed that the optimal cut-off age in early-onset EC patients was 33 years; the optimal cut-off tumor size in early-onset EC patients was 4 cm.
For all younger patients, there were 838 (16.6%) patients ≤ 33 years old and 4199 (83.4%) patients 33ཞ44 years old. For the race group, 3703 (75.5%) patients were white, whereas 492 (9.8%) patients were black and 842(16.7%) patients were of other races. Further, 2478 (49.2%) were identified as married and 2558 (50.8%) patients were in other marital status. Regarding the pathological type, adenocarcinoma was the most prevalent, accounting for 80.4% of all tumors (4052). With respect to TMN stage, the majority of patients were classified as T1 (4138; 82.2%), N0 (4707; 93.4%), and M0 (4897;97.2%) according to laboratory examinations and postoperative pathological results. In the SEER stage group,4037(80.1%) patients were found to have localized disease. There were 1497 (29.7%) patients with tumor size (cm) < 4. The treatment protocol for the patients included chemotherapy (591;11.7%) and radio therapy (798;15.8%) (Table 2).
Table 2
Baseline demographic and clinicopathological characteristics with early-onset EC patients.
Variables | The training cohort | The validation cohort | Total | Cardinality | P |
| No. (%) | No. (%) | No. (%) | | |
Total | 3526 | 1511 | 5037 | | |
Age (year) | | | | 2.17 | 0.15 |
≤ 33 | 569 (16.1) | 269(17.8) | 838(16.6) | | |
33–44 | 2957(83.9) | 1242(82.2) | 4199(83.4) | | |
Marital status | | | | 1.05 | 0.31 |
Married | 1752(49.7) | 727 (48.1) | 2478(49.2) | | |
Other | 1774(50.3) | 784(51.9) | 2558(50.8) | | |
Race | | | | 0.21 | 0.90 |
White | 2586(73.3) | 1117(73.9) | 3703 (73.5) | | |
Black | 348(9.9) | 144(9.5) | 492(9.8) | | |
Other | 592(16.8) | 250(16.5) | 842(16.7) | | |
Histology | | | | 0.002 | 0.97 |
Adenocarcinoma | 2837(80.5) | 1215(80.4) | 4052(80.4) | | |
Other | 689(19.5) | 296(19.6) | 985(19.6) | | |
Tumor size(cm) | | | | 3.02 | 0.22 |
< 4 | 1038(29.4) | 459(30.4) | 1497(29.7) | | |
≥ 4 | 985(27.9) | 447(29.6) | 1432(28.4) | | |
Unknown | 1503(42.6) | 605(40.0) | 2108(41.9) | | |
Grade | | | | 1.07 | 0.90 |
Grade I | 1798(51.0) | 762(50.4) | 2560(50.8) | | |
Grade II | 799(22.7) | 335(22.2) | 1134(22.5) | | |
Grade III | 296(8.4) | 126(8.3) | 422(8.4) | | |
Grade IV | 42(1.2) | 17(1.1) | 59(1.2) | | |
Unknown | 591(16.8) | 271(17.9) | 862(17.1) | | |
SEER stage | | | | 1.78 | 0.41 |
Localized | 2809(79.7) | 1228(81.3) | 4037(80.1) | | |
Regional | 598(17.0) | 238(15.8) | 836(16.6) | | |
Distant | 119(3.4) | 45(3.0) | 164(3.3) | | |
AJCC T stage | | | | 2.77 | 0.43 |
T1 | 2881(81.7) | 1257(83.2) | 4138(82.2) | | |
T2 | 285(8.1) | 102(6.8) | 387(7.7) | | |
T3 | 328(9.3) | 139(9.2) | 467(9.3) | | |
T4 | 32(0.9) | 13(0.9) | 45(0.9) | | |
AJCC N stage | | | | 0.138 | 0.710 |
N0 | 3292(93.4) | 1415(93.6) | 4707(93.4) | | |
N1 | 234(6.6) | 96(6.4) | 330(6.6) | | |
AJCC M stage | | | | 0.56 | 0.46 |
M0 | 3424(97.1) | 1472(97.5) | 4897(97.2) | | |
M1 | 102(2.9) | 38(2.5) | 140(2.8) | | |
Radiotherapy | | | | 0.62 | 0.43 |
Yes | 568(16.1) | 230(15.2) | 798(15.8) | | |
No/Unknown | 2958(83.9) | 1281(84.8) | 4239(84.2) | | |
Chemotherapy | | | | 0.79 | 0.38 |
Yes | 423(12.0) | 168(11.1) | 591(11.7) | | |
No/Unknown | 3103(88.0) | 1343(88.9) | 4446(88.3) | | |
Univariate And Multivariate Analysis
Cox analysis univariate analysis identified age, marital status, race, histology, tumor size, grade, SEER stage, T stage, N stage, M stage, radiotherapy and chemotherapy as risk factors significantly associated with overall survival. These risk factors were included in a multivariate analysis (P-value < 0.05) and age, marital status, race, SEER stage, grade and T stage were found to be independent risk factors for poor overall survival (Table 3).
Table 3
Univariate and multivariate analysis of variables related to OS in the training cohort. (n = 3526)
Variables | Univariable analysis | Multivariable analysis |
| HR | 95%Cl | p-value | HR | 95%Cl | p-value |
Age | | | 0.010 | | | |
≤ 33 | Ref | | | Ref | | |
33–44 | 0.61 | 0.42–0.89 | 0.011 | 1.70 | 1.15–2.49 | 0.008 |
Marital status | | | < 0.001 | | | |
Married | Ref | | | Ref | | |
Other | 0.62 | 0.49–0.79 | < 0.001 | 1.33 | 1.04–1.71 | 0.024 |
Race | | | < 0.001 | | | |
White | Ref | | | Ref | | |
Black | 0.95 | 0.68–1.31 | 0.740 | 1.65 | 1.19–2.30 | 0.003 |
Other | 1.84 | 1.23–2.76 | 0.003 | 1.07 | 0.77–1.49 | 0.702 |
Histology | | | < 0.001 | | | |
Adenocarcinoma | Ref | | | Ref | | |
Other | 0.61 | 0.47–0.78 | < 0.001 | 1.02 | 0.78–1.52 | 0.863 |
Tumor size(cm) | | | < 0.001 | | | |
< 4 | Ref | | | Ref | | |
≥ 4 | 1.01 | 0.73–1.40 | 0.972 | 1.09 | 0.79–1.52 | 0.595 |
Unknown | 2.56 | 1.96–3.35 | < 0.001 | | | 0.780 |
Grade | | | < 0.001 | | | |
Grade I | Ref | | | Ref | | |
Grade II | 0.40 | 0.27–0.60 | < 0.001 | 1.75 | 1.21–2.51 | 0.003 |
Grade III | 1.01 | 0.68–1.50 | 0.965 | 5.01 | 3.44–7.29 | < 0.001 |
Grade IV | 5.12 | 3.55–7.39 | < 0.001 | 5.43 | 2.98–9.90 | < 0.001 |
Unknown | 6.61 | 3.75–11.66 | < 0.001 | 1.87 | 1.24–9.91 | 0.003 |
SEER stage | | | < 0.001 | | | |
Localized | Ref | | | Ref | | |
Regional | 0.04 | 0.03–0.05 | < 0.001 | 2.33 | 1.24–4.39 | 0.009 |
Distant | 0.20 | 0.15–0.27 | < 0.001 | 4.94 | 1.60-15.23 | 0.005 |
AJCC T stage | | | < 0.001 | | | |
T1 | Ref | | | Ref | | |
T2 | 0.03 | 0.02–0.05 | < 0.001 | 1.08 | 0.57–2.03 | 0.823 |
T3 | 0.10 | 0.06–0.18 | < 0.001 | 2.00 | 1.13–3.53 | 0.017 |
T4 | 0.33 | 0.20–0.53 | | 3.20 | 1.36–7.54 | 0.008 |
AJCC N stage | | | < 0.001 | | | |
N0 | Ref | | | Ref | | |
N1 | 0.13 | 0.10–0.17 | < 0.001 | 1.39 | 0.99–1.94 | 0.059 |
AJCC M stage | | | < 0.001 | | | |
M0 | Ref | | | Ref | | |
M1 | 0.07 | 0.05–0.09 | < 0.001 | 1.31 | 0.52–3.29 | 0.567 |
Radiotherapy | | | < 0.001 | | | |
Yes | Ref | | | Ref | | |
No/Unknown | 2.65 | 2.07–3.40 | < 0.001 | 1.30 | 0.97–1.73 | 0.080 |
Chemotherapy | | | < 0.001 | | | |
Yes | Ref | | | Ref | | |
No/Unknown | 8.10 | 6.40-10.25 | < 0.001 | 1.05 | 0.72–1.52 | 0.819 |
Nomogram Construction
Different subtypes of each independent prognostic variable were projected onto a scale to obtain a score for each item. The scores corresponding to the independent prognostic factors were summed to obtain a total score. The higher the total score, the worse the prognosis. For example, for a married 40-year-old EC patient in our hospital with Grade II classification, distant metastases and stage T2, adding up the points for each prognostic predictor, she scored 147.99 in the overall survival line chart respectively, and the 5-year overall survival was estimated to be 83% according to the nomogram. The nomogram showed that the factor that has the greatest impact on prognosis was SEER stage (Fig. 3).
Internal Validation
In the training cohort, the concordance index for the nomogram of overall survival was 0.839 (0.814–0.864), which was greater than the concordance index for a single independent prognostic risk factor. In the validation cohort, the concordance index for the nomogram of overall survival was 0.826(0.785–0.867). The result was also superior to its independent prognostic risk factors. In addition, the concordance index for overall survival was calculated to be 0.834 (0.618–0.856) for the total data. The AUC values for overall survival at 1-, 3-, 5- and 8-year in the training cohort (0.91, 0.865, 0.837 and 0.828) were significantly higher than for SEER stage (0.869, 0.821, 0.94 and 0.773). The results for the validation cohort also showed significantly higher AUC values for the nomogram (0.888,0.817, 0.803 and 0.781) than for SEER stage (0.835, 0.793, 0.765 and 0.742) (Fig. 4a, b). The calibration curves for the training and validation cohorts were close to a 45degree diagonal, indicating that the probabilities of the predicted values were generally consistent with the actual probabilities (Fig. 4c, d). In addition, the DCA curves confirmed the validity of the nomogram (Fig. 4e, f).
External Validation
In seventy early-onset EC patients, there were 3 cases aged < 33; Among T stages, 85.7%, 12.9% and 1.4% were T1, T2 and T3, respectively. Among differentiation, 61.4%, 32.9% and 5.7% were high, moderate and low differentiation, respectively. In SEER stage, lesions confined to the uterus account for 84.3% of cases (Supplementary Table 1). The AUCs at 5-, 8-and 10-year for the nomogram (0.896, 0.89 and 0.852) were higher than those for SEER stage (0.778, 0.776 and 0.675) (Supplementary Fig. 2).
Risk Stratification
We divided the internal validation cohort and the external validation cohort into high- and low-risk groups based on critical values. Comparison of overall survival between groups using Kaplan-Meier curves showed that overall survival rates were higher in all low-risk groups than in the high-risk group (P-value < 0.05) (Fig. 5).