目的
肝癌是國人癌症死亡重要原因。手術是現今最具治癒性與積極性之治療，但肝癌復發仍是影響預後的重要因素，尤其是術後早期復發影響更甚。回顧肝癌患者接受肝切除手術治療，術後早期復發的經驗，加以分析其早期復發的危險因子。

方法
以中部某醫學中心接受肝切除術之原發性肝癌患者為對象，利用回顧性研究和Cox 比例風險回歸進行分析並評估術後早期復發的相關危險因子。

結果
共記錄68 位患者，平均年齡為61 歲。術後一年內復發22 位，於單變量分析發現血小板數低於100,000/μL 的族群復發機率增加5.32 倍；腫瘤每增加1 公分，復發機率增加4.00 倍；有血管侵犯者，復發機率增加2.70 倍；腫瘤分期越晚期者，復發機率增加7.50 倍；術中出血量多的族群，復發機率增加5.61 倍。多變量分析顯示，復發存活率獨立預測因素分別為：術前血小板數值低於100,000/μL（獨立預測因素：9.29；95％信賴區間為1.34-31.66）；術前C 型肝炎病毒量高於50-100 萬以上（獨立預測因素：1.02；95％信賴區間為0.78-1.82）。

結論
腫瘤愈大、有血管侵犯者、肝癌分期越晚期及術前血小板數目低及C 型肝炎病毒量高、術中出血量多的病患，術後一年內必須更加密切的追蹤。

Purposes
Hepatocellular carcinoma (HCC) is a major cause of death in Taiwan, and hepatectomy is one of the most effective curative treatments. However, the recurrence of HCC remains a significant factor that affects prognosis, especially the early postoperative recurrence of HCC. This retrospective study examined patients with early postoperative recurrence of HCC who underwent hepatectomy at a medical center in central Taiwan and analyzed the risk factors for the early recurrence of HCC.

Methods
Cox proportional-hazards regression was used to analyze and assess the risk factors for the early postoperative recurrence of HCC.

Results
A total of 68 patients, averaging 61 years old, were enrolled. Of whom, 22 experienced a postoperative recurrence of HCC. A univariate analysis showed that the recurrence rate of the group with a platelet count of <100,000/μL was 5.32 times higher than that of the group with a normal platelet count. A 1-cm increase in tumor size was associated with a 4-fold increase in the risk of recurrence. Patients with vascular invasion had a recurrence rate 2.7 times higher than that of those without vascular invasion. The risk of recurrence in patients with late-stage cancer increased by 7.5 times. HCC recurrence was 5.61 times more likely to occur in the group with a high volume of blood loss. A multivariate analysis showed that the independent predictors of recurrence rate were: a preoperative platelet count of < 100,000/μL (independent predictor: 9.29; 95% confidence interval: 1.34-31.66) and a preoperative hepatitis C virus (HCV) RNA level ranging from 500,000 to 1 million (independent predictor: 1.02; 95% confidence interval: 0.78-1.82).

Conclusions
Our results suggested that close follow-up was required for patients with large tumors, vascular invasion, advanced HCC staging, low preoperative platelet count, high HCV-RNA levels, and large volumes of blood loss.