PD-(L)1抑制劑併用化學治療用於晚期食道鱗狀細胞癌作為一線治療：系統性回顧、統合分析及臨床試驗計畫書

Abstract

背景 多項隨機對照第三期臨床試驗顯示，對於晚期食道鱗狀細胞癌（ESCC）患者，PD-(L)1抑制劑併用化療作為第一線全身性藥物治療的療效顯著優於單獨化療。雖然在不同研究中，PD-(L)1抑制劑併用化療療效顯著優於單獨化療在PD-L1高表現度的患者中，但在PD-L1低表現度患者，是否應該使用PD-(L)1抑制劑併用化療仍未有共識。此外，缺乏直接比較PD-(L)1抑制劑與platinum與fluoropyrimidine (PF)或platinum 與paclitaxel (TP)化療併用組合。藉由系統性回顧和統合分析，我們間接比較了PD-(L)1抑制劑與PF或TP化療組合作為晚期食道鱗狀細胞癌患者一線治療的療效。 方法 我們在Cochrane、PubMed和Embase數據庫中進行了檢索。符合納入標準的研究文獻是根據以下PICO篩選的：(1) Population-未接受過治療的不可切除局部晚期或轉移性食道鱗狀細胞癌患者。(2) Intervention-PD-(L)1抑制劑併用化療。(3) Control-化療單用。(4) Outcome-整體存活期。採用研究文獻中整體存活期（overall survival, OS）和疾病無惡化存活期（progression free survival, PFS）的危險比（hazard ratio, HR）及95%信賴區間（95% Confidence Interval, 95%CI），進行了統合分析及網絡統合分析比較不同治療方案。我們採用了隨機效應模型。在所選的研究中採用了三種不同PD-L1表現度評分標準，包括腫瘤比率分數（TPS）、綜合陽性分數（CPS）和腫瘤陽性區域（TAP）；在本篇論文中，TPS ≥1%、TAP ≥10% 和 CPS ≥10 被歸類為PD-L1高表現度次族群；而 TPS <1%、TAP <10% 和 CPS <10 被歸類為PD-L1低表現度次族群。此外，我們還根據PD-L1表現度，評估了PD-(L)1抑制劑搭配兩種不同化療組合—TP（paclitaxel併用platinum agent）和PF（5-fluorouracil或capecitabine併用platinum agent）的療效。 结果 在本研究中納入了八個符合納入標準的對照試驗。這八個試驗的統合分析結果顯示PD-(L)1抑制劑併用化療效果顯著優於化療單用，PD-(L)1抑制劑併用化療顯著改善了疾病無惡化存活期（PFS）和整體存活期（OS），其危險比（HR）分別為0.63（95% CI，0.58–0.69）和0.69（95% CI，0.64–0.75）。 進一步將納入病人依照PD-L1表現度區分為兩組次族群，PD-(L)1抑制劑併用化療的效果在兩組病人均呈現顯著優於化療單獨使用。並且發現併用治療的效果在PD-L1高表現度病人群的又比在低表現度病人組更具優勢：PFS的HR在高表現度為0.54（95% CI，0.49–0.60），在低表現度為0.71（95% CI，0.62–0.81），兩者差異達到統計顯著 (組間異質性檢定p值< 0.01)；OS的HR在高表現度為0.60（95% CI，0.53–0.67），在低表現度為0.78（95% CI，0.70–0.87），兩者差異達到統計顯著 (組間異質性檢定p值 < 0.01)。 進一步分析了在不同PD-L1表現度次族群中，化療組合對PD-(L)1抑制劑併用化療帶來的療效影響。在PD-L1高表現度族群中，PD-(L)1抑制劑併用化療的PFS危險比分別為PF組的0.53和TP組的0.56以及OS危險比分別為PF組的0.59和TP組的0.60，兩種化療方案組的PFS和OS危險比之間沒有顯著差異(組間異質性檢定p值分別為p=0.75和p=0.90)。然而，在PD-L1低表現度族群中，PD-(L)1抑制劑併用化療中的PFS危險比分別為PF組的0.82和TP組的0.59；OS的危險比分別為PF組的0.84和TP組的0.72，PFS的危險比在兩種化療組合次族群之間存在顯著差異，TP組更具優勢（p=0.01），而OS的危險比在兩種化療組合次族群之間未達到統計顯著，但趨向於TP組合（p=0.32）。 結論 本研究說明在晚期食道鱗狀細胞癌患者且PD-L1低表現度族群中，相較於化療單用，PD-(L)1抑制劑與化療併用的疾病無惡化存活期和整體存活期都具有顯著優勢。且透過本研究的系統性整理，發現在晚期食道鱗狀細胞癌患者且PD-L1低表現度族群中，PD-(L)1抑制劑併用TP比併用PF帶來更顯著的生存益處，但直接比較PD-(L)1抑制劑併用不同化療方案（如PF與TP組合）的資料較為缺乏，進一步的前瞻性研究將有助於確定化療選擇作為PD-(L)1抑制劑合併用藥。

Background In patients diagnosed with advanced esophageal squamous cell carcinoma (ESCC), multiple randomized controlled phase III studies have demonstrated that the efficacy of PD-(L)1 inhibitors combined with chemotherapy as first-line systemic therapy is significantly better than chemotherapy alone. While the benefit of PD-(L)1 inhibitors combined with chemotherapy over chemotherapy alone has been consistently seen in patient with high PD-L1 expression across different studies, there is no consensus on this benefit among patients with low PD-L1 expression. Moreover, direct comparison between PD-(L)1 inhibitors combined with platinum plus fluoropyrimidine (PF) or platinum plus paclitaxel (TP) chemotherapy regimens is lacking. By a systematic review and meta-analysis, we conducted direct and indirect comparisons of the efficacy of combinations of PD-(L)1 inhibitors and PF or TP chemotherapy as initial treatment for patients with advanced ESCC. Methods We conducted searches in the Cochrane, PubMed, and Embase databases. Articles were selected following the PICO framework if they met the inclusion criteria outlined below: (1) Population: untreated patients with unresectable locally advanced, or metastatic ESCC. (2) Intervention: PD-(L)1 inhibitors plus chemotherapy. (3) Control: chemotherapy alone. (4) Outcome: overall survival (OS). For this meta-analysis and network meta-analysis, we used study-level pooled analyses of hazard ratios (HR) for progression-free survival (PFS) and overall survival (OS), along with 95% confidence intervals (95% CI), to compare the different therapeutic regimens. A random-effects model was employed. Three different scoring systems of PD-L1 expression had been used in the selected clinical trials including tumor proportion score (TPS), tumor area positivity (TAP), and combined positive score (CPS). In the current thesis work, TPS ≥1%, TAP ≥10%, and CPS ≥10 were classified as high PD-L1 expression subgroup; and TPS <1%, TAP <10%, and CPS <10 were classified as low PD-L1 expression subgroup. Additionally, the efficacy of two different chemotherapy regimens—TP (paclitaxel plus a platinum agent) and PF (a platinum agent plus 5-fluorouracil or capecitabine)—in combination with anti-PD-(L)1 inhibitors was also evaluated, stratified by PD-L1 expression levels. Results Eight controlled trials meeting the criteria were included in the current work. A meta-analysis of these 8 trials showed that PD-(L)1 inhibitors combined with chemotherapy were significantly more effective than chemotherapy alone. The addition of PD-(L)1 inhibitors to chemotherapy significantly improved progression-free survival (PFS) and overall survival (OS), with hazard ratios (HRs) of 0.63 (95% CI, 0.58–0.69) and 0.69 (95% CI, 0.64–0.75), respectively. Further analysis divided patients into two subgroups based on PD-L1 expression levels. The combined therapy of PD-(L)1 inhibitors and chemotherapy was significantly more effective than chemotherapy alone in both subgroups and a more significant beneficial effect was seen in high PD-L1 expression group than low PD-L1 expression group. The HRs of adding PD-(L)1 inhibitors to chemotherapy for PFS were 0.54 (95% CI, 0.49-0.60) and 0.71 (95%CI, 0.62-0.81) for high-expression and low-expression subgroups, respectively (p< 0.01 for subgroup differences). The HRs of adding PD-(L)1 inhibitors to chemotherapy for OS were 0.60 (95% CI, 0.53-0.67) and 0.78 (95%CI, 0.70-0.87) for high-expression and low-expression subgroups, respectively (p< 0.001 for subgroup differences). The impact of chemotherapy regimens on the beneficial effect conferred by the addition of PD-(L)1 inhibitors to chemotherapy was further examined in different PD-L1 expression subgroups. In the high PD-L1 expression subgroup, the HRs of adding PD-(L)1 inhibitors to chemotherapy for PFS were 0.53 and 0.56 for PF group and TP group, respectively; and the HRs for OS were 0.59 and 0.60 for PF group and TP group, respectively. The HRs for PFS and OS were not significantly different between two chemotherapy regimen groups (p= 0.75 and 0.90, respectively). In the low PD-L1 expression subgroup, the HRs of adding PD-(L)1 inhibitors to chemotherapy for PFS were 0.82 and 0.59 for PF group and TP group, respectively; and the HRs for OS were 0.84 and 0.72 for PF group and TP group, respectively. The HRs for PFS were significantly different between two chemotherapy regimen groups in favor of TP group (p= 0.01), and the HRs for OS did not differ significantly between two chemotherapy regimen groups but trended to be in favor of TP regimen (p= 0.32). Conclusions The study revealed that, in patients with advanced ESCC and low PD-L1 expression, PD-(L)1 inhibitors combined with chemotherapy offers notable benefits for both PFS and OS compared with chemotherapy alone. Through systematic analysis, it was found that in this patient subgroup, the combination of PD-(L)1 inhibitors with TP regimen was associated with more significant survival benefits than the combination with PF regimen. However, data directly comparing the benefit of adding PD-(L)1 inhibitors to different chemotherapy regimens such as PF versus TP regimens are currently lacking. Further prospective studies are needed to determine the optimal chemotherapy regimen in combination with PD-(L)1 inhibitors in advanced ESCC patients with low PD-L1 expression.