Totally 11639 adults were diagnosed with EGJ cancer from 2004 to 2010, and 2859 patients were excluded due to multiple primary tumors. Patients with other histology (N=1810) except adenocarcinoma, without histological confirmation (N=4206), with distant metastasis (N=407), and with unknown examined LNs (N=209) were also excluded. Besides, primary AEG with follow-up less than 3 months (N=103), unknown size (N=265), unknown invasion depth (N=11), and cause of unknown death (N=14) were also excluded. Finally, 1755 cases were included in further analysis (Figure 1). A sum of 373 females (21.2%) and 1382 males (78.8%) were included in the analysis. Sixty years was used as a cut-off of elder people. T stage ranged from T1 to T4 (N= 355, 227, 768, 405, respectively), N stage from N0 to N3 (N=716, 391, 340, 308, respectively). The details of the baseline characteristics of participants were shown in Table 1.
AEG survival prediction model
In the final multivariate proportional sub-distribution hazard model of clinical characteristics for prognosis of AEG, age >60 ys. (SHR 1.389, P < 0.001), high T stage (T1-4, SHR 1.592, 2.167, 2.555, respectively, P < 0.001), high N stage (N0-3, SHR 1.814, 2.505, 3.335, P < 0.001), high Grade (SHR 1.281, P < 0.001) were associated to the poor prognosis. Number of examined LN presented to be a protective factor (<10, <15,>=15, SHR 0.751, 0.635, 0.540, P<0.001). The detailed results of multivariate analysis were listed in Table 2.
Construction of competing risk nomogram
AEG cause-specific death predicting model of the nomogram was established based on selected prognostic factors (Figure 2). The nomogram showed that N stage contributed to be the most impact factor of prognosis, followed by T stage, and age. Amount of examined LN and grade had a modest effect on survival. Each subtype of the variables was assigned a score. A straight line to determine the estimated probability of survival can be drawn at each time point on the total point scale, according to the total point.
Internal Validation and Evaluation of the nomogram
In the analysis of specificity, we used both receiver operating characteristic (ROC) curve and Brier score to evaluate the diagnostic value and accuracy of the nomogram model. With respect to the ROC curve, the nomogram model was greater than traditional TNM staging in the cohort (1-year AUC:0.747 vs. 0.641, 3-year AUC: 0.761 vs. 0.679, 5-year AUC: 0.759 vs. 0.682, 7-year AUC: 0.749 vs. 0.673, respectively, P<0.001, Figure 3). The Brier score is a measure of overall performance and captures aspects of both calibration and discrimination. It is a representation of the difference between the predicted probability and the actual outcome. The score ranges from 0 to 1, with values closer to 0 indicating better predictive ability. Concerning Brier score, the accuracy of the nomogram was also better than traditional TNM stage at 3-year (0.198 vs. 0.217, P=0.012), 5-year (0.198 vs. 0.216, P=0.008), 7-year (0.199 vs. 0.215, P=0.014) (Figure 3). The calibration curves showed the dots close to a 45° diagonal line, meaning the nomograms were well calibrated (Figure 4).
Then we used DCA to compare the clinical usability between the nomogram and traditional TNM staging. Based on a continuum of potential thresholds for death (x-axis) and the net benefit of using the model to risk stratify patients (y-axis) relative to assuming all patients will be alive, the DCA graphically presented that the nomogram was better than traditional TNM staging in clinical conditions (Figure 5). Compared with traditional TNM staging, the nomogram showed a larger net benefit across the range of death risk in the analysis.