目的：回顧本院治療 NK/T 細胞淋巴癌之成果，並釐清預後因子對整體存活、疾病進展、與局 部控制的影響。 材料與方法：我們針對本院 2000-2014 年治療之所有 NK/T 細胞淋巴癌病人做一回溯性病歷分 析。病人至少需在本院接受一線的治療。我們從病歷資料中提取病人基本資訊、相關醫療記 錄、治療細節，與其治療成果。統計方面，我們採用 Kaplan-Meier 法估算整體存活、無進展存 活、與無局部區域復發存活，並以 Cox 比例風險模式分析預後因子對治療成果的差異。 結果：研究總計收錄 21 位病人，其中 12（57%）位分期為第 I-II 期、9（43%）位為第 IV 期。 所有病人之中位整體存活為 2 年。早期病人之中位存活（3.8 年）顯著較晚期病人為長（2.5 個 月）。在早期病人中，NK/T 淋巴癌預後指標（NKPI）與整體存活有顯著相關；有淋巴結侵犯的 病人，其局部復發的機率有較低的趨勢（P= 0.05, HR= 0.17）。 結論：與過去文獻報告相仿，NK/T 細胞淋巴癌的表現相當惡性，其預後不甚理想。提高未來治 療成效的策略包含：確實找出高風險的族群、使用較新的化學治療藥物、或提高放射線劑量。

Purpose : We aim to review treatment outcomes at our institution and evaluate prognostic factors that influence overall survival, progression, and local recurrence. Material and Methods : We conducted a retrospective chart review of all patients with NK/T-cell lymphoma treated from 2000-2014 at our institution. Patients must have received at least one line of therapy. Baseline characteristics and treatment information were extracted from medical records. We employed the Kaplan-Meier product limit method to estimate loco-regional recurrence-free survival (LRFS), progression-free survival (PFS) and overall survival (OS). Cox proportional hazards models were used to assess significance of prognostic factors on treatment outcomes. Results : A total of 21 patients were included, of which 12 (57%) were stage I-II, while the other 9 (43%) were stage IV. Median overall survival for all patients was 2 years. Median survival was significantly longer for stage I-II (3.8 years) than stage IV (2.5 months). In early stage patients, the Natural Killer/T-Cell Lymphoma Prognostic Index (NKPI) was associated with overall survival; patients with clinically positive lymph nodes had a trend towards lower risk of local recurrence (P= 0.05, HR= 0.17). Conclusions : Treatment outcomes of NK/T-cell lymphoma suggest an aggressive pattern and poor outcome, comparable with reports in the literature. Identification of high-risk subgroups and aggressive treatment with modern chemotherapeutic agents, and/or radiotherapy dose escalation may be warranted to improve treatment outcomes.