OBJECTIVE: A SVM predictive model consisting of preoperative tumor markers and inflammatory factors was established to explore its significance in evaluating the prognosis of patients with ESCC. METHODS: Clinical data of 311 patients with ESCC who underwent surgery were collected and followed up until October 2019. Statistical software SPSS version 22.0, and R (version 3.6.1) were used to analyze the data. RESULTS: In the Test, Val1 and Val2 groups, the sensitivity of preoperative optimal combination (SVM5) to predict the prognosis of patients with ESCC was 88.89%, 76.92%, and 73.68%, respectively. The specificity was 92.00%, 74.42%, and 78.00%, respectively. The sensitivity and specificity were not statistically different from those of SVM9 (P > 0.05), while the sensitivity of SVM9+5 for predicting the prognosis of patients with ESCC was 91.84%, 82.26%, and 80.36%, respectively. The specificity was 97.44%, 75.93%, and 78.00%, respectively. Its sensitivity and specificity were higher than those of SVM9 (P < 0.001). CONCLUSIONS: We used a nomogram to input the indicators in the SVM5 into the artificial intelligence program for patients with ESCC who have not yet developed an individualized plan. It can predict and evaluate the postoperative outcome of patients with ESCC with a sensitivity of 79.04%, specificity of 81.82%, PPV of 83.54%, NPV of 76.97%, and accuracy of 80.32%. For patients who have undergone surgery, we can enter the indicators in SVM9+5 into the artificial intelligence program. AJTR
OBJECTIVE: A SVM predictive model consisting of preoperative tumor markers and inflammatory factors was established to explore its significance in evaluating the prognosis of patients with ESCC. METHODS: Clinical data of 311 patients with ESCC who underwent surgery were collected and followed up until October 2019. Statistical software SPSS version 22.0, and R (version 3.6.1) were used to analyze the data. RESULTS: In the Test, Val1 and Val2 groups, the sensitivity of preoperative optimal combination (SVM5) to predict the prognosis of patients with ESCC was 88.89%, 76.92%, and 73.68%, respectively. The specificity was 92.00%, 74.42%, and 78.00%, respectively. The sensitivity and specificity were not statistically different from those of SVM9 (P > 0.05), while the sensitivity of SVM9+5 for predicting the prognosis of patients with ESCC was 91.84%, 82.26%, and 80.36%, respectively. The specificity was 97.44%, 75.93%, and 78.00%, respectively. Its sensitivity and specificity were higher than those of SVM9 (P < 0.001). CONCLUSIONS: We used a nomogram to input the indicators in the SVM5 into the artificial intelligence program for patients with ESCC who have not yet developed an individualized plan. It can predict and evaluate the postoperative outcome of patients with ESCC with a sensitivity of 79.04%, specificity of 81.82%, PPV of 83.54%, NPV of 76.97%, and accuracy of 80.32%. For patients who have undergone surgery, we can enter the indicators in SVM9+5 into the artificial intelligence program. AJTR