台灣懸浮微粒減量與肝癌存活率提升之關聯

Abstract

背景 流行病學研究已證實大氣懸浮微粒暴露為引發肝癌發生與死亡率的顯著風險因子之一。然而，目前關於懸浮微粒暴露及其減量是否能實質影響肝癌患者診斷後的存活率與預後表現，現存文獻仍相當有限且缺乏針對不同肝炎病毒感染背景及懸浮微粒交互作用的系統性評估。因此，本研究旨在探討台灣粗懸浮微粒(PM10)暴露和肝細胞癌 (HCC) 以及不同肝炎病毒型態之肝癌患者肝癌發生率及發生之後存活率之間的關聯性。 材料與方法 本研究之材料來自台灣癌症登記資料庫(Taiwan Cancer Registry, TCR)與環境部空氣品質監測資料。在癌症方面，納入2000 年至 2022 年間台灣本島診斷為肝癌之患者(ICD-O-3 : C22)，並且依據病毒感染狀態（B型肝炎表面抗原、C型肝炎抗體檢驗紀錄）將患者分為B 型肝炎病毒（HBV）陽性組、C 型肝炎病毒（HCV）陽性組、B 肝與 C 肝皆為陰性組，以及B 肝與 C 肝皆陽性組等四組。此外，根據形態學編碼81703、81713、81723、81733、81743、81753另行分出肝細胞癌（Hepatocellular carcinoma, HCC）進行分析。PM10數據源自環境部75個台灣本島空氣品質監測站逐時資料，並採用反距離加權法（Inverse Distance Weighting, IDW）進行空間內插，以估算全國各鄉鎮行政區之PM10濃度及癌症患者的暴露量。資料分析分群體層次與個人層次進行。在群體層次以WHO標準化人口計算2000年起到2022年的逐年肝癌標準化發生率（ASIR）與5和15年的PM10移動平均濃度進行簡單線性回歸及廣義相加模型分析。在個體層次的存活分析，包含 2011 年至 2018 年診斷為肝癌的92,331名患者，其中肝癌診斷患者含88,290 名肝細胞癌，及依肝炎病毒感染狀態可分為：B 型肝炎病毒（HBV）陽性組、22,058 名 C 型肝炎病毒（HCV）陽性組、15,055 名B 肝與 C 肝皆為陰性，以及3,690名B 肝與 C 肝皆為陽性之肝癌患者。肝癌患者的PM10 之暴露量分為自診斷日起和自1994 年1月1日起計算至死亡或追蹤截止日(2023年12月31日) 兩個時間區段，採用 Cox 比例風險模型評估 PM10 暴露量減少對肝癌之死亡之存活風險的影響，並探討PM10 暴露與至少接受手術、化學治療、免疫治療、標靶治療或放射治療中之一種治療方式對存活率的交互作用，所有模式都校正了年齡、性別、身體質量指數（BMI）、吸菸、飲酒、肝纖維化等級、癌症分期。 研究結果 本研究發現，2000-2009年台灣肝癌年齡標準化發生率約40/105左右的水平，2010年後逐步下降至2022年的23.61/105，同期PM10五年年移動平均濃度從58.51µg/m3至下滑至33.07µg/m3。簡單線性回歸顯示 PM₁₀ 與肝癌五年及十年年齡標準化發生率呈顯著正相關，且在肝細胞癌及各肝炎病毒亞組分析中亦呈現相似結果。在82,290 名肝細胞癌（HCC）患者中，男性佔 70.6%，女性佔 29.4%。臨床處置方面，接受治療者74.5%，未接受治療者9.1%，另有16.4%治療狀態不明。至追蹤截止時存活個案為 18,402 人（22.4%），死亡個案為 63,888 人（77.6%）。統計結果顯示，死亡組的平均年齡（66.96 ± 12.70 歲）高於存活組（60.99 ± 10.99 歲）。全因死亡組的平均PM₁₀ 暴露濃度（48.41 ± 15.49 µg/m³）均高於存活組的PM₁₀（37.45 ± 9.15 µg/m³）。Cox 比例風險模型分析未納入 PM10的模型中，結果顯示接受臨床治療可大幅降低全因死亡風險（HR =0.35, 95% CI: 0.34-0.36）。Cox 比例風險模型加入PM10 暴露變相後發現治療與PM10 暴露改善對於肝癌病人的存活率具有顯著的交互作用（p < 0.01）。以診斷後的PM10暴露模型中為例，當肝細胞癌患者PM10濃度平均暴露（45.96 µg/m³）時，接受臨床治療患者相對於無治療患者的全死因風險比為 0.33。當肝細胞癌患者PM10暴露減少 1 個標準差 (15.02 µg/m³)，接受臨床治療患者相對於無治療患者的全死因風險比為 HR = 0.25。在各病毒性肝癌亞組中，雙陰性組 (B–C–)患者PM10濃度平均暴露量（45.42 µg/m³）時，接受臨床治療患者相對於無治療患者的全死因風險比為 0.36。當B–C–患者PM10暴露減少 1 個標準差 (15.06 µg/m³)，接受臨床治療患者相對於無治療患者的全死因風險比為 HR = 0.29。B 肝陽性組 (B+)患者PM10濃度平均暴露量（44.23 µg/m³）時，接受臨床治療患者相對於無治療患者的全死因風險比為 0.24。當B+患者PM10暴露減少 1 個標準差 (14.37 µg/m³)，接受臨床治療患者相對於無治療患者的全死因風險比為 HR = 0.18。C 肝陽性組 (C+)患者PM10濃度平均暴露（46.60 µg/m³）時，接受臨床治療患者相對於無治療患者的全死因風險比為 0.31。當C+患者PM10暴露減少 1 個標準差 (14.33 µg/m³)，接受臨床治療患者相對於無治療患者的全死因風險比為 HR = 0.24。雙陽性組 (B+C+)患者診斷後的PM10濃度平均暴露（47.06 µg/m³）時，接受臨床治療患者相對於無治療患者的全死因風險比為 0.24。當B+C+患者PM10暴露減少 1 個標準差 (14.56 µg/m³)，接受臨床治療患者相對於無治療患者的全死因風險比為 HR = 0.17。模式進一步針對手術以及化學治療兩項臨床介入處置，採用 Cox 比例風險模型評估 PM10 暴露量減少對肝癌之死亡之存活風險的影響，結果顯示在肝細胞癌患者、有無病毒感染的肝癌患者的PM₁₀ 暴露量都與他們接受手術及化學治療在癌症存活率存在交互作用。針對 15,055 名 B、C 肝皆陰性（B-C-）的世代效應（Cohort Analysis）分析中，在校正了四個出生年代（1945 年前、1945–1954 年、1955–1964 年及 1965 年以後）後，診斷後的PM10濃度平均暴露量（45.42 µg/m³）時，接受化療患者相對於無治療患者的全死因風險比為 0.35。當B+C+患者PM10暴露減少 1 個標準差 (15.06 µg/m³)，接受化療患者相對於無治療患者的全死因風險比為 HR = 0.25。 結論 本研究證實PM10暴露的削減與台灣肝癌發生率下降及患者生存預後提升具有顯著相關，治療與PM10暴露量下降對對肝細胞癌病理分型及肝炎病毒感染狀態皆有貢獻且兩者之間有交互作用的關係。

Background Epidemiological studies have confirmed that exposure to atmospheric particulate matter is a significant risk factor for the incidence of and mortality from liver cancer. However, existing literature regarding whether particulate matter exposure—or its reduction—substantially affects post-diagnosis survival and prognosis in liver cancer patients remains limited; furthermore, there is a lack of systematic assessment concerning the interactions between particulate matter exposure and various hepatitis virus infection backgrounds. Therefore, this study aims to investigate the associations between exposure to coarse particulate matter (PM10) and hepatocellular carcinoma (HCC), specifically examining the incidence of HCC and post-diagnosis survival rates among patients with different hepatitis virus infection profiles in Taiwan. Materials and Methods Data for this study were obtained from the Taiwan Cancer Registry (TCR) and air quality monitoring records from the Ministry of Environment. For cancer cases, patients diagnosed with primary liver cancer (ICD-O-3: C22) in mainland Taiwan between 2000 and 2022 were included. Based on viral infection status (HBsAg and anti-HCV laboratory records), patients were categorized into four groups: hepatitis B virus (HBV)-positive, hepatitis C virus (HCV)-positive, HBV/HCV dual-negative, and HBV/HCV dual-positive. Additionally, hepatocellular carcinoma (HCC) cases were specifically isolated for analysis using morphology codes 81703, 81713, 81723, 81733, 81743, and 81753. Hourly PM₁₀ data were collected from 75 ambient air quality monitoring stations operated by the Ministry of Environment in mainland Taiwan. Inverse Distance Weighting (IDW) spatial interpolation was applied to estimate township-level PM₁₀ concentrations and individual exposure levels for cancer patients. Data analysis was performed at both population and individual levels. At the population level, the WHO standard population was used to calculate the annual age-standardized incidence rate (ASIR) of liver cancer from 2000 to 2022, and simple linear regression alongside generalized additive models were utilized to evaluate its relationship with 5-year and 15-year moving average PM₁₀ concentrations. For the individual-level survival analysis, a cohort of 92,331 patients diagnosed with liver cancer between 2011 and 2018 was included, comprising 88,290 HCC patients. Stratified by viral infection status, this cohort included HBV-positive patients, 22,058 HCV-positive patients, 15,055 HBV/HCV dual-negative patients, and 3,690 HBV/HCV dual-positive liver cancer patients. Individual PM₁₀ exposure for liver cancer patients was calculated across two time windows: starting from the date of diagnosis and starting from January 1, 1994, both calculated through death or the follow-up end date of December 31, 2023. Cox proportional hazards models were employed to evaluate the impact of reduced PM₁₀ exposure on mortality risk in liver cancer patients. The models further investigated the interactive effect on survival rates between PM₁₀ exposure and receiving at least one therapeutic modality—including surgery, chemotherapy, immunotherapy, targeted therapy, or radiotherapy. All models were adjusted for age, sex, body mass index (BMI), smoking status, alcohol consumption, liver fibrosis category, and cancer stage. Results The study found that the age-standardized incidence rate (ASIR) of liver cancer in Taiwan remained at a level of around 40/105 from 2000 to 2009, then gradually declined after 2010 to 23.61/105 in 2022. During the same period, the 5-year moving average concentration of ambient PM₁₀ dropped from 58.51 μg/m³ to 33.07 μg/m³. Simple linear regression demonstrated that PM₁₀ was significantly and positively correlated with the 5-year and 10-year age-standardized incidence rates of liver cancer, with similar results observed in subgroup analyses for hepatocellular carcinoma and across various viral hepatitis profiles. Among 82,290 patients with hepatocellular carcinoma (HCC), 70.6% were male and 29.4% were female. Regarding clinical management, 74.5% received treatment, 9.1% did not, and the treatment status of the remaining 16.4% was unknown. By the end of the follow-up period, 18,402 patients (22.4%) were alive, while 63,888 (77.6%) had died. Statistical results indicated that the mean age of the deceased group (66.96 ± 12.70 years) was higher than that of the surviving group (60.99 ± 10.99 years). The mean PM₁₀ exposure concentration for the all-cause mortality group (48.41 ± 15.49 µg/m³) was higher than that for the surviving group (37.45 ± 9.15 µg/m³). A Cox proportional hazards model analysis that did not include PM₁₀ showed that receiving clinical treatment significantly reduced the risk of all-cause mortality (HR = 0.35, 95% CI: 0.34–0.36). When PM₁₀ exposure was included as a variable in the Cox proportional hazards model, a significant interaction (p < 0.01) was observed between treatment and PM₁₀ exposure regarding the survival rate of HCC patients. Taking the post-diagnosis PM₁₀ exposure model as an example, when hepatocellular carcinoma (HCC) patients were exposed to the mean PM₁₀ concentration (45.96 μg/m³), the all-cause mortality hazard ratio (HR) for treated patients relative to untreated patients was 0.33. When PM₁₀ exposure was reduced by 1 standard deviation (15.02 μg/m³), the all-cause mortality hazard ratio for treated patients relative to untreated patients was 0.25. Across viral hepatitis subgroups, when dual-negative (B–C–) patients were exposed to the mean PM₁₀ concentration (45.42 μg/m³), the all-cause mortality HR for treated versus untreated patients was 0.36. When PM₁₀ exposure was reduced by 1 standard deviation (15.06 μg/m³), the HR was 0.29. For hepatitis B virus–positive (B+) patients at mean exposure (44.23 μg/m³), the HR was 0.24, dropping to 0.18 with a 1 SD reduction (14.37 μg/m³). For hepatitis C virus–positive (C+) patients at mean exposure (46.60 μg/m³), the HR was 0.31, dropping to 0.24 with a 1 SD reduction (14.33 μg/m³). For dual-positive (B+C+) patients at mean exposure (47.06 μg/m³), the HR was 0.24, dropping to 0.17 with a 1 SD reduction (14.56 μg/m³). The study further examined two clinical interventions—surgery and chemotherapy—using the Cox proportional hazards model to assess how reduced PM10 exposure affects survival risk regarding liver cancer mortality. The results indicate that for patients with hepatocellular carcinoma—regardless of whether they had a viral infection—PM10 exposure levels interacted with the receipt of surgery and chemotherapy to influence cancer survival rates. In a birth cohort analysis of 15,055 hepatitis B and C dual-negative (B–C–) patients, after adjusting for four birth cohorts (pre-1945, 1945–1954, 1955–1964, and post-1965), the all-cause mortality HR for chemotherapy-treated patients relative to untreated patients was 0.35 at the mean post-diagnosis PM₁₀ concentration (45.42 μg/m³). When PM₁₀ exposure was reduced by 1 standard deviation (15.06 μg/m³), the HR for chemotherapy-treated patients relative to untreated patients was 0.25. Conclusion This study demonstrates that reductions in PM₁₀ exposure are significantly associated with declining liver cancer incidence and improved survival outcomes among patients in Taiwan. Clinical treatment and reduced PM₁₀ exposure contribute favorably across both hepatocellular carcinoma pathological subtypes and viral hepatitis infection statuses, with a significant interactive effect observed between treatment and environmental exposure reduction.