Analysis of metastasis and prognostic risk factors in T 4 pancreatic cancer and construction of a prognostic nomogram
10.3760/cma.j.cn115396-20241012-00310
- VernacularTitle:T 4期胰腺癌转移与预后风险因素分析及预后列线图的构建
- Author:
Xiaohao ZHENG
1
;
Jingyu ZHANG
;
Xiaojie CHEN
;
Zhen HAO
;
Jing LIU
;
Zewen ZHANG
;
Yun YANG
Author Information
1. 首都医科大学附属北京友谊医院普通外科中心 消化健康全国重点实验室 国家消化系统疾病临床医学研究中心，北京 100050
- Keywords:
Pancreatic cancer;
Arterial involvement;
T 4;
Nomogram;
Prognosis
- From:
International Journal of Surgery
2024;51(12):819-828
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To preliminarily investigate the risk factors for distant metastasis and prognosis, and construct a prognostic nomogram in T 4 stage pancreatic cancer. Methods:A retrospective case-control study was conducted using data from the Surveillance, Epidemiology, and End Results (SEER) database for pancreatic patients from January 1, 2010, to December 31, 2021. Based on whether the tumor invaded the celiac axis, superior mesenteric artery, and/or common hepatic artery, 38 759 patients were divided into an arterial invasion group (T 4 stage, n=7 471) and a non-arterial invasion group (non-T 4 stage, n=31 288). Clinical and pathological data, including demographic characteristics, treatment information, and tumor data were collected. The primary outcome was overall survival. Categorical data were expressed as numbers (percentages), and intergroup comparisons were made using the chi-square test. Survival benefits were measured using the Log-Rank test. A multivariate logistic model was used to identify high-risk factors for metastasis in T 4 stage pancreatic cancer. Patients were randomly divided into training ( n=5 232) and validation ( n=2 239) sets at a 7∶3 ratio. A nomogram model was created based on independent prognostic factors from the multivariate Cox regression analysis, and the model′s predictive ability was evaluated using the C-index and calibration curves. Results:The overall metastasis rate in the arterial invasion group was higher than that in the non-arterial invasion group (32.8% vs 29.0%, P<0.001), with fewer patients showing no metastasis or single-organ metastasis (86.3% vs 89.7%, P<0.001) and higher rates of lung metastasis ( P<0.001), distant lymph node metastasis ( P<0.001), and other metastases excluding liver, lung, brain, bone, and distant lymph node metastases ( P<0.001). However, no significant difference was found between groups for liver, brain, or bone metastasis rates ( P>0.05). Surgical rates for T 4 stage patients were significantly lower than for non-T 4 stage patients (all patients: 10.7% vs 38.4%, P<0.001; M 0 stage patients: 15.0% vs 52.4%, P<0.001; M 1 stage patients: 2.1% vs 4.1%, P<0.001). Additionally, significant differences were observed in age, race, radiotherapy, chemotherapy, tumor location, tumor size, and tumor stage ( P<0.05). The median survival for patients with arterial invasion was 8 months, significantly lower than the 10-month median survival for non-arterial invasion patients ( P<0.001). The median survival for surgical patients with arterial invasion was 22 months, lower than the 24-month median for non-T 4 stage patients underwent surgery ( P<0.001) but significantly higher than for patients without surgery (T 4 stage patients without surgery: 8 months, P<0.001; non-T 4 stage patients without surgery: 6 months, P<0.001). For lymph node metastasis, patients with or without positive local lymph node metastasis had similar overall survival ( P>0.05). However, Patients with distant lymph node metastasis had significantly lower overall survival than that in patients without distant lymph node metastasis ( P<0.001). The multivariate logistic model indicated that tumor location in the body and tail ( OR=2.591, 95% CI: 2.343-2.867), positive regional lymph nodes ( OR=2.033, 95% CI: 1.836-2.252), and age <70 years old ( OR=1.183, 95% CI: 1.067-1.312) were risk factors for distant metastasis in arterial invasion patients. The multivariate Cox model showed that surgery ( HR=0.451, 95% CI: 0.405-0.503), radiotherapy ( HR=0.729, 95% CI: 0.677-0.784), chemotherapy ( HR=0.277, 95% CI: 0.258-0.297), tumor location in the body and tail ( HR=0.928, 95% CI: 0.874-0.985), and household income ≥$80, 000 ( HR=0.908, 95% CI: 0.853-0.968) were independent protective factors for prognosis in arterial invasion patients. Living in areas with a population ≤1 million ( HR=1.109, 95% CI: 1.044-1.178), age ≥70 years old ( HR=1.220, 95% CI: 1.150-1.296), larger tumor size (>2 cm but ≤4 cm: HR=1.124, 95% CI: 0.954-1.323; >4 cm: HR=1.310, 95% CI: 1.114-1.541), and having a metastatic burden (lung metastasis: HR=1.049, 95% CI: 0.869-1.267; distant lymph node metastasis: HR=1.179, 95% CI: 0.910-1.527; bone metastasis: HR=1.419, 95% CI: 0.854-2.359; brain or other metastasis: HR=1.519, 95% CI: 1.350-1.709; liver metastasis: HR=1.737, 95% CI: 1.600-1.886; two types of metastasis: HR=1.913, 95% CI: 1.689-2.168; three or more types: HR=2.436, 95% CI: 1.947-3.048) were independent risk factors for prognosis. The nomogram based on these prognostic factors had a C-index of 0.749 in the training set and 0.745 in the validation set; calibration curves in both sets were near the 45° line. Conclusions:High metastasis rates and low surgery rates are characteristic of pancreatic cancer with arterial invasion. Investigating the risk factors for distant metastasis and developing a prognostic nomogram incorporating metastatic burden hold significant clinical value for T 4 stage pancreatic cancer.
