使用辨別分析建立肝癌手術預後模式

Establishing Prognosis model of Hepatoma for Opeation by Discriminant Analysis

曾映滋(Ying-Tzu Tseng)

指導教授 ： 蔡崇弘

中山醫學大學/醫學院/醫學研究所/碩士(2015年)

https://doi.org/10.6834/CSMU.2015.00041

摘要

目的：肝癌(Hepatocellular Carcinoma)是世界上第五大常見的癌症，也是癌症死亡的第三位。肝癌患者最常見有慢性B型肝炎及C型肝炎慢性感染。外科手術包括肝切除及肝臟移植是目前被視為最有效治癒的治療方式。因部分肝癌患者本身存在合併症，包括肝硬化，肝代謝功能不良，心血管疾病，呼吸系統疾病，糖尿病，腎功能不全，藥物的藥代動力學改變，因合併症使得肝癌患者只有12 % 輔合手術切除或者是換肝手術進而影響患者術後的預後及存活率。因此，本研究將探討肝癌手術的患者的公認的實驗室值，觀察一年後存活情況，並通過判別分析，建立肝癌患者預後模式。
材料與方法：本研究收集中山醫學大學附設醫院原發性肝癌病人接受肝切除術治療的病例。時間自西元2009年11月至2014年12月，共77案例以回溯記錄的方法，數據分析以SPSS14.0版本，設計肝癌手術判別方析，將術前抽血報告套入辨別方程式中，從判別方析中預測肝癌手術後一年的存活率。
結果：本研究中的患者進行了肝癌辨別分析建立肝癌手術預後模式來確定第一年死亡風險因素。經過分析，患者的平均年齡為61.0±5.10歲。單變項因素分析中，術前平均值低血紅素實驗室（死亡組11.87±2.48 VS存活組13.69±1.80（P = 0.02）），術前低白蛋白實驗室值（死亡組正常範圍56.25％，異常範圍43.75％後，存活組正常範圍90.16％與異常範圍9.83％（P = 0.004）），高血清尿素氮實驗室值（死亡組正常範圍56.25％，異常範圍43.75％與生存組正常範圍：83.61％，異常範圍16.39％，（P =0.037））。 以判別函數係數得到方程Fisher的線性判別函數，-0.02=-0.238血紅素+0.029白蛋白+0.792血清尿素氮，經交叉驗證正確率70.1％。
結論與建議：研究發現，術前實驗室檢驗值低血紅素，低白蛋白和高血清尿素氮，影響存活率。

關鍵字

判別分析 ； 肝癌 ； 肝切除術

並列摘要

Objective: Hepatocellular carcinoma (HCC) is the fifth-most common cancer in the world, and the third, also of cancer death in surgery including liver resection and liver transplantation is currently considered the most effective treatment to cure. It is worth noting that only 5% of patients with secondary liver cancer resection combined or liver transplants, including the presence of comorbidities part of liver cancer patients, including cardiovascular disease, respiratory disease, diabetes, renal insufficiency, drug pharmacokinetics change, thereby affecting the prognosis and survival of patients after surgery. Therefore, this study will examine the accepted laboratory values of liver cancer surgery patients. One year after death experience, and the establishment of Prognosis model of HCC patients by discriminant analysis.
Methods and Materials: This study collected underwent liver resection Medical University Hospital in the treatment of primary liver cancer. BC since November 2009 to December 2014 years, a total of 77 people have been collecting case studies backtracking method of recording, after data collection to data analysis SPSS version 14.0, liver blood test to assess the value of using discriminant analysis. Discriminant analysis from side can confirm that after a year survival rate of liver cancer surgery.
Results: This study was conducted in patients with hepatocellular carcinoma before liver surgery prognosis establish discriminant analysis to determine the risk factors set one year after the first death. After closing the case analysis, the average age of the patients was 61.0 ± 5.10 years. After univariate analysis, the mean preoperative low heme Laboratory (death group 11.87 ± 2.48 VS surviving group 13.69 ± 1.80 ( P = 0.02) , preoperative albumin measurement before ( death group normal range 56.25% abnormal range 43.75%; VS mal survival group normal range 90.16%, abnormal range 9.83%( P = 0.004) High BUN count ( death group normal range 56.25%, abnormal range 43.75% VS survival group normal range: 83.61%, abnormal range 16.39%, (P) . Classification Function Coeffcients discriminant function coefficients obtained equation Fisher’s linear discriminant functions-0.02=-0.238 Hb+0.029Alb+0.792BUN, Cross-validation of grouping 70.1% is correct.Conclusions and Recommendations: Study found that preoperative laboratory trace hemoglobin, albumin and BUN of the measure, and forces the patient to maintain normal metrology laboratory, surgery can help prolong survival.

並列關鍵字

Discriminant analysis ； Hepatocellular carcinoma ； Hepatectomy

參考文獻

2. 劉鐘軒、蔡正中、陳海雄．肝癌的診斷及治療最新發展．內科學誌．2013; (24)，85-94。

3. 林錦城．目前台灣地區肝癌流行與醫療趨勢．北市中醫會刊，2010; (61)，75-96。

14.林志陵、高嘉宏．肝癌流行病學．中華癌醫會誌，2008;24(5)，277-281。

49.蔡玉純、楊福麟、邱艷芬、張芙美、李茹萍．消化道重症手術患者營養狀況疾病嚴重度及術後合併症與預後之關係．台灣醫學，2006;(10)1，1-9。

52.潘恆之、鄭昌綺、陳永昌．第一和第二型肝腎症候群．腎臟與透析，2011;(25)2，129-135。