結直腸癌篩檢狀態和發生率、死亡率及存活率之相關性：台灣全國性結直腸癌篩檢計畫的分析

Abstract

背景：結直腸癌是全球重大的健康議題，而多項研究證實透過篩檢能有效預防或早期發現結直腸癌。然而同為參與篩檢，每個個體篩檢後之狀態不盡相同：接受篩檢或未接受篩檢，篩檢結果陽性或陰性，參與篩檢的次數等，是一個包含多個面向之複雜因子。然而，過去的研究多限於特定族群或只考慮篩檢狀態的某些方面，缺乏對不同篩檢狀態對結直腸癌發生率和死亡率全面的探討。此外，對於個人層次而言，篩檢似乎也可能是結直腸癌患者的一個預後因子。然而，過去的研究在探究篩檢對預後的影響仍有爭議。 目的：本研究將利用台灣的戶政資料、結直腸癌篩檢計劃資料、癌症登記資料和死亡登記資料，建立兩個世代研究：一般民眾世代研究及結直腸癌患者世代研究來充分探討不同篩檢狀態對結直腸癌發生、死亡率及患者預後之相關性。 方法：一般民眾世代：從台灣戶政資料庫隨機抽取民眾並進行四年（2010-2013）觀察以確定他們的篩檢狀態。然後所有人從2014年起進行隨訪以評估結直腸癌發病率和死亡率。結腸癌患者世代：利用台灣癌症登記資料庫選取 2013至2015年首次診斷為結直腸癌患者。並連結他們與台灣結直腸癌篩檢資料庫來定義患者之篩檢狀態，並對所有患者進行追蹤至2019年年底。主要分析皆使用Cox比例風險模型進行相關性評估。 結果：在一般民眾世代中，與無篩檢者相比，糞便免疫化學檢測 (FIT) 篩檢可將結直腸癌發生及死亡率分別降低21%及58%。經歷一次、兩次、三次以上之FIT 陰性顯示結直腸癌發生及死亡風險對比值降低（0.66、0.58 和 0.45；0.37、0.19 和 0.12），反之一次、兩次以上之 FIT 陽則風險對比值增加（發生風險：2.09 和 3.30；死亡風險：1.42 和 2.25）。在FIT陽性經歷次數相同下，結直腸癌發生及死亡風險由高至低：不依從後續檢查>部分依從>全依從、而最後糞便血紅蛋白濃度（μg Hb/ g糞便）為「100+」>「20-99」>「<20」。 在風險最低的兩組中，結直腸癌死亡風險甚至低於無篩檢者。而在患者世代研究中，與有進行後續檢查之FIT陽性患者相比，未進行檢查之FIT 陽性患者、FIT 陰性患者和無篩檢患者其結直腸癌死亡之風險對比值依序為 1.40、1.63和1.76，均達統計顯著意義。當FIT陽性後雖有後續診斷性檢查，但此診斷性檢查檢出結直腸癌已拖超過 12 個月時，其死亡風險仍增加。 結論：FIT篩檢結果呈陽性，罹患結直腸癌的風險會提高，但完全依從後續檢查，即使檢查出是結直腸癌，仍然提升後續存活率。反之，FIT篩檢結果呈陰性，雖然罹患結直腸癌的風險較低，但一旦被診斷有結直腸癌，也就是所謂的間隔癌，其預後表現將比FIT陽性者更差。建議參與篩檢的人，FIT陽性則配合後續診斷性檢查；FIT陰性者則繼續參與篩檢，因多次FIT陰性結果能持續降低結直腸癌相關風險。本研究闡明了篩檢狀態與結直腸癌發生率、死亡率風險及與患者預後之間複雜的關係，並提供篩檢作為結直腸癌風險預測的一獨立因子之相關有力證據。呈現各種篩檢狀態與結直腸癌風險關係之全貌，有助於當局針對不同篩檢狀態人群調整未來的篩檢策略。

Background: Colorectal cancer (CRC) is a significant global health issue, and multiple studies have confirmed the effectiveness of screening in preventing or detecting CRC at an early stage. However, for the same participation in screening, each individual's screening status can vary, including participation in screening, fecal immunochemical testing (FIT) results (positive or negative), and the number of screening rounds. This complexity involves multiple aspects of screening status. Previous studies often focused on specific populations or considered only single aspect of screening status, lacking a comprehensive exploration of the impact of different screening statuses on CRC incidence and mortality rates. Additionally, from an individual perspective, screening appears to be a prognostic factor for CRC patients. However, there is still controversy regarding the impact of screening on prognosis in previous studies. Objective: This study aims to utilize Taiwan population-based data resources, including household registration database, CRC screening program, cancer registration, and death registration dataset, to establish two cohort studies: a general population cohort and a CRC patient cohort. This study will comprehensively investigate the correlation between different screening statuses and CRC incidence, mortality rates, and patient prognosis. Methods: General population cohort: randomly selected individuals will be sampled from the Taiwan household registration database, and their screening statuses will be determined based on an equal four-year observation period (2010-2013).Then, Follow-up will be conducted from 2014 onwards to assess CRC incidence and mortality rates. On the other hand, CRC patient cohort: patients first-time diagnosed with CRC between 2013 and 2015 will be identified from the Taiwan Cancer Registry dataset. Their screening statuses will be defined by linking them with the Taiwan CRC screening dataset, and all patients will be followed up until the end of 2019. Cox proportional hazard models will be used for the major analyses to evaluate any association. Results: In the general population cohort, compared to individuals who did not undergo screening, FIT screening reduced CRC incidence and mortality rates by 21% and 58%, respectively. Among individuals with FIT-negative, the hazard ratios for CRC incidence and mortality decreased with each additional negative screening rounds (0.66, 0.58, and 0.45; 0.37, 0.19, and 0.12), while the hazard ratios increased for individuals with one, two or more rounds of FIT-positive (incidence: 2.09 and 3.30; mortality: 1.42 and 2.25). Among FIT-positive individuals with the same number of FIT experienced round, the risk of CRC incidence and mortality decreased in the following order: non-compliance > partial compliance > full compliance with follow-up examinations. Moreover, the last fecal hemoglobin concentration (μg Hb/g feces) also reveals the similar tendency (categories: "100+" > "20-99" > "<20"). In the lowest-risk two groups (full compliance, "<20"), the risk of CRC mortality was even lower than that of individuals who did not undergo screening. On the other hand, in CRC patient cohort, compared to FIT-positive patients who underwent follow-up examinations, the risk odds ratios for CRC death were 1.40, 1.63, and 1.76 for FIT-positive patients without follow-up examinations, FIT-negative patients, and patients who did not undergo screening, respectively, all of which were statistically significant. Notably, the risk of CRC mortality increased when the diagnostic examination for FIT-positive individuals took more than 12 months to diagnose CRC. Conclusion: A positive FIT screening result increases the risk of developing CRC, but complete compliance with follow-up examinations can significantly improve survival rates after CRC diagnosis. On the other hand, a negative FIT screening result indicates a lower risk of developing CRC, but if diagnosed with CRC, for instance, the interval cancers, the prognosis is worse than FIT-positive individuals. This study recommends that individuals participating in screening well compliance with follow-up diagnostic examinations if their FIT results are positive, while FIT-negative individuals should continue participating in screening as multiple negative results can continuously reduce the risk of CRC. This study elucidates the complex relationship between screening status and CRC incidence, mortality rates, and patient prognosis, providing convinced evidence for screening status as an independent factor in predicting CRC risk. By presenting the comprehensive landscape of the relationship between various screening statuses and CRC risk, it will help authorities in adjusting future screening strategies for different populations based on their screening status.