Scholarlink

Purpose <p>To construct a model to predict the early risk of recurrence in pancreatic ductal adenocarcinoma (PDAC) patients following curative R0 resection based on preoperative magnetic resonance imaging (MRI) features and serum CA19-9 levels.</p> Materials and methods <p>From January 2017 to June 2021, 131 patients with radically&#xa0;resected PDAC were reviewed and divided into a training cohort (n = 91) and a validation cohort (n = 40). In the training cohort, preoperative MRI findings and clinical variables associated with early recurrence were identified by using univariate and multivariate Logistic analyses. The nomogram was constructed using multivariate Logistic regression models. The calibration curves and receiver operating characteristics (ROC) curve were used to assess the calibration and predictive performance of the model respectively. Survival curves were plotted to verify the preoperative risk stratification performance of the proposed model.</p> Results <p>Rim enhancement (OR, 18.93; 95% CI: 2.79, 128.50; <i>P</i> = 0.003), adjacent organs invasion (OR, 3.84; 95% CI: 1.07, 13.69; <i>P</i> = 0.038) and preoperative serum CA19-9 levels &gt; 180 U/mL (OR, 7.67; 95% CI: 2.20, 26.73; <i>P</i> = 0.001) were independently associated with early recurrence after curative resection. The AUC values of the proposed nomogram for predicting early recurrence in the development and validation cohorts were 0.87 (95%CI: 0.80, 0.94) and 0.83 (95% CI: 95%CI; 0.69, 0.97), respectively. Patients in the high-risk of early recurrence exhibited significantly worse RFS compared to those in low-risk group (Log-rank <i>P</i> &lt; 0.001 in both the training and validation cohorts).</p> Conclusion <p>Nomogram based on preoperative MRI features and serum CA19-9 levels may predict early recurrence for PDAC patients after curative resection and facilitate further risk stratification and decision-making for individualized treatment.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Predicting early recurrence risk in patients with pancreatic ductal adenocarcinoma patients after curative resection based on preoperative MRI features and CA19-9

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月14日

查看全文
  • Dengfeng Li,
  • Wei Cai,
  • Xinming Zhao

摘要

Purpose

To construct a model to predict the early risk of recurrence in pancreatic ductal adenocarcinoma (PDAC) patients following curative R0 resection based on preoperative magnetic resonance imaging (MRI) features and serum CA19-9 levels.

Materials and methods

From January 2017 to June 2021, 131 patients with radically resected PDAC were reviewed and divided into a training cohort (n = 91) and a validation cohort (n = 40). In the training cohort, preoperative MRI findings and clinical variables associated with early recurrence were identified by using univariate and multivariate Logistic analyses. The nomogram was constructed using multivariate Logistic regression models. The calibration curves and receiver operating characteristics (ROC) curve were used to assess the calibration and predictive performance of the model respectively. Survival curves were plotted to verify the preoperative risk stratification performance of the proposed model.

Results

Rim enhancement (OR, 18.93; 95% CI: 2.79, 128.50; P = 0.003), adjacent organs invasion (OR, 3.84; 95% CI: 1.07, 13.69; P = 0.038) and preoperative serum CA19-9 levels > 180 U/mL (OR, 7.67; 95% CI: 2.20, 26.73; P = 0.001) were independently associated with early recurrence after curative resection. The AUC values of the proposed nomogram for predicting early recurrence in the development and validation cohorts were 0.87 (95%CI: 0.80, 0.94) and 0.83 (95% CI: 95%CI; 0.69, 0.97), respectively. Patients in the high-risk of early recurrence exhibited significantly worse RFS compared to those in low-risk group (Log-rank P < 0.001 in both the training and validation cohorts).

Conclusion

Nomogram based on preoperative MRI features and serum CA19-9 levels may predict early recurrence for PDAC patients after curative resection and facilitate further risk stratification and decision-making for individualized treatment.