請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/22975
標題: | 運用腹部超音波於依指標個案界定肝癌高危險群之篩檢計畫評估 Evaluation of HCC Screening Program with Abdominal Sonography for Subjects Identified from Their Family Index HCC Cases |
作者: | Lei-Ling Wang 王鐳鴒 |
指導教授: | 陳秀熙(Hsiu-Hsi Chen) |
關鍵字: | 肝癌,肝癌高危險群,腹部超音波,篩檢計畫評估, HCC,Abdominal Sonography,Evaluation of Screening Program, |
出版年 : | 2009 |
學位: | 碩士 |
摘要: | 背景 肝癌的發生除了B型肝炎病毒表面抗原陽性及C型肝癌抗體陽性等已被廣為人知的危險因子之外,家族病史也已被證實為其中一項非常重要的危險因子。因此利用肝癌病患做為指標個案(Index Case)並據此找出高危險群提供腹部超音波篩檢則被視為早期偵測肝癌個案的一項有效肝癌死亡防制政策。因此,針對此類資料進行疾病自然病史估計並進而推估與篩檢效益相關之參數則非常重要。
研究目的 包括 (1)評估利用指標個案界定高危險群的方法是否能較一般臨床個案發現有較高的疾病偵測率(high yield of detecting HCC cases); (2)釐清人口學變項、肝功能指數、甲型胎兒蛋白、具肝癌家族病史之親屬等變項與肝癌的偵測發生、死亡之間的相關性; (3)量化肝癌疾病自然病史、篩檢計畫敏感度及陽性預測值; (4)推估利用指標個案界定高危險群並提供腹部超音波之篩檢策略,在調整相關因子及前導期偏差(lead time bias)之後對肝癌死亡率降低的影響。 方法 本研究共計收集來自多家醫院所診斷之肝癌個案其一等親或二等親內之家屬,提供腹部超音波篩檢服務,計1992至1997年間共收錄20348名個案。此世代與臺灣癌症資料庫及臺灣死亡資料庫聯結以確認篩檢計畫之外因臨床症狀就醫而發現之臨床偵測個案及死亡案例。本研究利用邏吉斯迴歸(logistic regression)分析影響篩檢偵測之相關因子,利用比例危險迴歸模式(proportional hazards regression models)找出與肝癌發生或死亡相關的危險因子。在篩檢評估部份,則利用兩種統計模式:卜瓦松模式(Poisson model)及多階段隨機模式進行臨床症前篩檢可偵測期、篩檢計畫敏感度和預測值之估計。 結果 在調整已知的生化環境因子後,我們發現親等關係與篩檢偵測個案相關(指標個案之兄弟姐妹相對於其子女之勝算比(odds ratio,OR)為4.0 (95%信賴區間: 1.84-8.53),父母親則為4.16 (95%信賴區間: 1.35-12.88));就篩檢間隔個案或篩檢後個案發生的危險性而言,親等關係仍為一顯著影響因子,(指標個案之兄弟姐妹相對於其子女之危險對比值(hazard ratio,HR)為1.82 (95%信賴區間: 1.49-2.21),其父母親則為1.62 (95%信賴區間: 1.14-2.31),但此親等關係與肝癌由偵測發生至死亡的存活則無顯著相關;此外,低教育程度則被發現與較差的肝癌存活率有關。 篩檢結果發現盛行篩檢之偵測率約為每千人2例 (41/20348),在第二次及第三次後篩檢則分別降為每千人0.93例及每千人0.75例。篩檢間隔個案及篩檢後個案發生率分別為每千人年1.87及每千人 2.10例。就篩檢效益而言,在考量前導期偏差校正、親等別及其他相關生化環境因子之後,本研究發現相較於篩檢後個案,於篩檢期間偵測之個案其肝癌死亡率約降低27%(95%信賴區間: 1%-46%)。 本研究分別利用卜瓦松模式求得之臨床症前篩檢可偵測期的平均時間及敏感度分別約為0.35年及61.2%;和隨機過程模式,其估計值為0.76年及 79.4%。利用這兩種方式得到之結果針對此篩檢計畫之陽性預測值估計分別為24.2%及67.7%。 結論 本研究發現針對具肝癌家族病史之個案提供腹部超音波之篩檢,在調整與指標個案之親屬關係、相關生化因子(包括檢驗B型肝炎病毒表面抗原及C型肝炎病毒抗體的結果、肝功能指數及甲型胎兒蛋白)、人口學變項(包括年齡、性別及教育程度)及前導期偏差後,肝癌死亡率可以降低27%。因此,愈密集的肝癌篩檢可能可以導致更大的篩檢效益。 Background In addition to other risk factors such as HBV and HCV infection, family history is also demonstrated to be associated with the risk of hepatocellular carcinoma (HCC). Therefore, screening for high-risk subjects ascertained through family index case with sonography is regarded as one of approaches to identify early-detected HCC cases. Evaluation of effectiveness of ultrasonography screening and the better understanding of disease progression of HCC using such kind of data have significant implication for early detection of HCC cases among these high-risk subjects. Aims The aims of this study are therefore to (1)assess whether an index approach has a high yield of detecting HCC cases compared with the routine-finding in the absence of organized screening program; (2)elucidate the association between demographic features, ALT, AST, alpha-fetoprotein, degrees of relative relationship, and occurrence of HCC cases or deaths; (3)quantify the disease natural history of HCC disease progression, the program sensitivity, and positive predictive value; (4)project mortality reduction as a result of ultrasonography screening with adjustment for these putative risk factors and lead-time bias. Methods A total of 20,348 first- or second-degree relatives of patients with HCC diagnosed in multiple hospitals were enrolled in a multicentre hospital-based ultrasonography screening between 1992 and 1997. The cohort was linked with Taiwan Cancer Registry and Taiwan Mortality Registry for further confirmation of clinically-detected cases and death. We used logistic regression to identify factors affecting the detection in the prevalent screen, and proportional hazards regression models for the factors affecting the incidence of and mortality from HCC. Two statistical models, early indicator of the proportion of post-screening cases to the expected incidence rate conjugated with Poisson model and multi-state stochastic modeling, were used to estimate mean sojourn time (MST), sensitivity, and positive predictive values of the screening program. Results After adjusting for the well-established biological environmental factors, we found degrees of relative relationship was significantly associated with cases detected in the screening (odds ratio (OR)=4.0 (95% confidence interval : 1.84-8.53) for siblings compared to offspring, and 4.16 (95% confidence interval : 1.35-12.88) for parents compared to offspring). For the incidence of interval cancer or post-screening cases, degrees of relative relationship were also significant factors (hazard ratio (HR) of siblings versus offspring was 1.82 (95% confidence interval : 1.49-2.21), and of parents versus offspring was 1.62 (95% confidence interval : 1.14-2.31). However, it is not significantly associated with survival of HCC. Level of education was found associated with HCC survival. The lower level of education, the poorer survival was found. Regarding the screening findings, the detection rate at prevalence screen was 2 per 1000 (41/20348). It decreased to 0.93 per 1000, and 0.75 per 1000 at the second and third or further subsequent screening. The incidence rate of interval cancer and post screening cases were 1.87 and 2.10, respectively. As regards the efficacy of screening with ultrasonography, compared to post-screening cases, screening detected and interval cancers combined had a 27% lower risk of dying from HCC (95% confidence interval : 1%-46%), with adjustment for other biological environmental risk factores, degrees of relative relationship and lead-time. The estimated mean sojourn time and sensitivity were 0.35 years and 61.2% with Poisson model, and 0.76 and 79.4% with the multistate stochastic model. The MST together with the program sensitivity yielded 24.2% and 67.7% of two positive predictive values for the Poisson model and three-state stochastic model, respectively. Conclusion This study demonstrated ultrasonography screening can lead to a 27% (95% confidence interval : 1%-46%) statistically significant mortality reduction with adjustment for family history, other environmental factors, and lead time obtained from natural history. The estimated parameters of disease natural history suggests the intensive screening may even lead to a large benefit. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/22975 |
全文授權: | 未授權 |
顯示於系所單位: | 流行病學與預防醫學研究所 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-98-1.pdf 目前未授權公開取用 | 509.97 kB | Adobe PDF |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。