胆囊息肉样病变性质的术前预测分析

Preoperative prediction of the nature of gallbladder polypoid

目的分析胆囊肿瘤性息肉的危险因素，建立预测胆囊息肉性质的Logistic回归方程。方法回顾性收集580例因胆囊息肉行胆囊切除术患者的临床病理资料，通过单因素分析和多因素分析对比肿瘤性息肉与非肿瘤性息肉以及腺瘤与胆囊恶性肿瘤的临床病理资料特点，拟合Logistic回归方程，并做其诊断试验。结果单因素分析显示，超声测量息肉基底宽度〉0．72cm（P〈0．05）、CEA〉1．51ng／ml（P〈0．05）、CA199〉12．92U／ml分别对诊断胆囊肿瘤性息肉差异有统计学意义（P〈0．05）。多因素回归分析可得回归方程l，以P〉0．256诊断肿瘤性息肉其灵敏度为72．4％，特异度为80．0％，总的正确比例为77．8％，AUC为0．828。当患者存在肿瘤性息肉风险时，平扫CT值〉32．50Hu对诊断恶性肿瘤差异有统计学意义（P〈0．05）。多因素回归分析可得回归方程2，以P〉0．496诊断肿瘤性息肉其灵敏度为86．0％，特异度为90．5％，总的正确比例为89．0％，AUC为0．938。结论在术前，根据患者年龄、症状、超声参数得到的回归方程可以比较准确地预测胆囊息肉性质。

Objective To evaluate the risk factors of gallbladder neoplastic polyps and establish a Logistic regression equation. Methods We retrospectively collected clinicopathologieal data of 580 patients who underwent choleeysteetomy for gallbladder polyps. Compared to the clinicopathological parameters between non-neoplastic polys and neoplastic polys, and adenomas and gallbladder cancer by univariate and multivariate analysis, we developed a Logistic regression equations to conduct diagnostic test. Results The univariate analyses showed that ultrasonic measuring gallbladder basal 〉 0. 72 cm, CEA 〉 1.51 ng/ml, CA199 〉 12. 92 U/ml for the diagnosis of gallbladder adenomatous polyps were statistically significant respectively （P 〈 0. 05 ）. Multivariate regression analysis showed regression equations 1. When P 〉 0. 256, the sensitivity of the diagnosis of adenomatous polyps was 72. 4% , the specificity was 80. 0% , and the whole correct ratio was 77. 8%. The positive likelihood ratio of adenomatous polyps was 10. 44, AUC was 0. 828. As for neoplasm risk factors, the CT value 〉 32, 50 Hu for the diagnosis of malignancy was statistically significant （ P 〈 0. 05 ）. Multivariate regression analysis displayed regression equations 2. When P 〉 0. 496, the sensitivity for the diagnosis of adenomatous polyps was 86. 0%, the specificity was 90. 5%, and the whole correct ratio was 89. 0%. The positive likelihood ratio for malignant tumor was 9. 08, AUC was 0. 938. Conclusion Preoperatively, patients＇ age, symptoms, ultrasonic parameters, can be used to develop a regression equation to predict accurately the nature of gallbladder polyps.