表皮生長因子受體酪胺酸酶抑制劑產生抗藥性的非小細胞肺癌患者使用免疫治療合併化學治療加上有無抗血管新生抑制劑與單獨化學治療之療效比較：系統性文獻回顧及統合分析

Abstract

1. 研究背景 表皮生長因子受體（EGFR）突變為非小細胞肺癌（NSCLC）中常見且重要的致癌驅動因子，特別常見於不吸菸的亞洲族群。雖然EGFR酪胺酸激酶抑制劑（TKI）為第一線標準治療，但幾乎無可避免地會出現後續抗藥性導致疾病進展。目前疾病進展後的主要治療方式為化學治療，然而免疫檢查點抑制劑合併化療或抗血管新生療法的效益仍未明確。鑒於過去研究顯示免疫合併化療於EGFR突變患者中的效果有限，本研究進行系統性回顧與統合分析，以探討免疫合併治療在此族群中的療效與安全性，並進一步分析不同次族群患者是否可能從中獲益。

2. 研究方法 本研究依循PRISMA 2020指引進行，系統性搜尋PubMed、Embase、Cochrane Library及ClinicalTrials.gov等資料庫中截至2024年12月31日前發表之隨機對照試驗，納入對象為接受表皮生長因子受體突變且標靶藥物治療後發生疾病進展之非小細胞肺癌患者，評估免疫檢查點抑制劑為基礎之合併治療（免疫治療+化療±抗血管新生療法）對照化療的隨機對照試驗。主要觀察指標為無惡化存活期（PFS），次要臨床指標包括整體存活期（OS）、治療相關不良事件（AEs）、次族群分析、出版偏倚。本研究使用 RoB 2.0 工具進行偏倚風險評估，並以漏斗圖進行發表偏倚之視覺化檢視。資料統整與統計分析皆使用 RStudio 軟體（版本 2024.12.1+563）進行。 鑑於各治療比較組中納入研究數量有限，本研究以漏斗圖作為探索性工具，評估潛在的出版偏倚。針對無惡化生存期與整體生存期所繪製的漏斗圖經視覺檢查後，未見明顯不對稱，顯示出版偏倚風險較低。此外，本研究亦繪製網絡結構圖，呈現四種不同治療之比較關係，共納入六項隨機對照試驗，其中兩項為三臂設計並共用對照組，最終納入八組治療效果估計值。考量各比較組資料有限，且間接比較證據迴路受限，本研究未進行 Egger’s 迴歸檢定、節點分裂分析與迴圈不一致性檢定等統計評估，以避免產生統計不穩定或誤導性結果。然而，根據整體網絡結構與治療效益趨勢，整體一致性仍具可接受程度。

3. 研究結果 本研究納入六篇符合條件的第三期隨機對照試驗，共計1,777名非小細胞肺癌患者，免疫治療相關合併療法在疾病無惡化存活期方面顯示具臨床意義之改善效果。IO合併化療相較於單獨化療，其合併HR為0.77（95% CI: 0.67–0.88）；三合一治療（免疫治療+化療+抗血管新生）效果最佳，HR為0.54（95% CI: 0.44–0.67）。不同治療策略（免疫治療+化療與免疫治療+化療+抗血管新生）之間的分層比較顯示PFS療效存在統計學上顯著差異（p = 0.0080）。整體存活期方面，免疫治療+化療組與化療組相比雖呈改善趨勢（HR = 0.86；95% CI: 0.75–1.00），但未達統計顯著性。三合一組合則未展現明顯存活效益（HR = 0.98；95% CI: 0.77–1.25）。與抗血管新生合併療法相比，免疫治療合併組亦未顯現統計上優勢。在安全性方面，免疫合併療法整體與化療相比，任何等級不良事件相對風險比為1.03（95% CI: 0.96–1.10）；而三合一組合的Grade ≥3不良事件風險最高（RR = 1.55；95% CI: 0.01–161.03），但異質性較高，解釋上需謹慎。但整體安全仍具可接受性。 次族群分析顯示，免疫治療合併方案在多項臨床與分子特徵中皆展現相對化療更一致的無疾病惡化存活期改善效果。在年齡方面，無論是 <65 歲（HR = 0.76，95% CI: 0.64–0.91）或 ≥65 歲族群（HR = 0.81，95% CI: 0.65–1.02），皆觀察到療效趨勢，且無顯著異質性（交互作用 p = 0.63）。性別方面，男性（HR = 0.76，95% CI: 0.61–0.95）與女性（HR = 0.80，95% CI: 0.66–0.98）同樣展現PFS改善，差異無統計顯著交互作用。亞洲族群具統計顯著的療效改善（HR = 0.80，95% CI: 0.68–0.94），而非亞洲族群則無一致趨勢（HR = 1.13，95% CI: 0.52–2.47），具明顯異質性。功能狀態分析中，ECOG PS為1的患者顯示明確療效（HR = 0.75，95% CI: 0.63–0.88），而ECOG 0患者雖呈現改善趨勢但未達統計顯著（HR = 0.84，95% CI: 0.65–1.09）。吸菸史方面，不論是現/曾吸菸者（HR = 0.75，95% CI: 0.60–0.95）或未吸菸者（HR = 0.81，95% CI: 0.66–0.99）皆有益處，無顯著交互作用。值得注意的是，在有腦轉移患者中，三合一治療（免疫治療+抗血管新生 + 化療）展現顯著療效（HR = 0.32，95% CI: 0.19–0.53），而免疫治療+化療則未顯著（HR = 0.92，95% CI: 0.71–1.19；次族群交互作用 p = 0.0029）。分子標誌方面，無T790M突變者（HR = 0.69，95% CI: 0.52–0.92）與 PD-L1表現 ≥1%者（HR = 0.79，95% CI: 0.63–0.99）明顯獲益；反之，T790M陽性患者（HR = 0.97，95% CI: 0.74–1.27）與 PD-L1表現<1%者（HR = 0.84，95% CI: 0.66–1.07）則未達統計顯著。此結果支持臨床上依據患者特徵進行個別化治療選擇之重要性。

4. 研究結論 統合分析結果顯示，對於EGFR突變且TKI治療失效之非小細胞肺癌患者，免疫合併化學治療，特別是同時合併抗血管新生藥物者，能在特定族群（如L858R突變、T790M陰性與合併腦轉移）中提供明顯的PFS益處。此結果支持個別化治療策略的重要性，亦突顯未來需進行具生物標記分層設計的前瞻性臨床試驗，以驗證此類療法於不同患者族群中的適用性與安全性。

Background Epidermal growth factor receptor (EGFR) mutations are key oncogenic drivers in non-small cell lung cancer (NSCLC), especially among never-smoking Asians population. While EGFR tyrosine kinase inhibitors (TKIs) are standard first-line therapy, resistance inevitably emerges. Following progression, chemotherapy remains the mainstay; however, the efficacy of immunotherapy-based regimens with or without anti-angiogenic agents remains uncertain. We aim to conduct a systematic review and meta-analysis to evaluate the efficacy and safety of immunotherapy-based regimens compared with chemotherapy alone in EGFR TKI resistant NSCLC, with subgroup analyses to identify potential beneficiaries of combination therapy, alongside evaluations of publication bias and network consistency.

Methods A systematic literature review and meta-analysis were conducted in accordance with PRISMA 2020 guidelines. PubMed, Embase, Cochrane Library, and ClinicalTrials.gov were searched up to December 31, 2024, for randomized controlled trials (RCTs) evaluating immunotherapy-based combinations (IO + chemotherapy ± anti-angiogenesis) versus chemotherapy in patients with EGFR-mutant NSCLC following TKI progression.

The primary outcome was progression-free survival (PFS); secondary outcomes included overall survival (OS), grade ≥3 treatment-related adverse events (AEs), subgroup analyses, publication bias. Risk of bias was assessed using the RoB 2.0 tool. Publication bias was evaluated using funnel plots inspection. Data synthesis was performed using RStudio (version 2024.12.1+563).

Due to the limited number of studies per treatment comparison, funnel plots were employed as an exploratory tool to assess potential publication bias. Visual inspection of funnel plots for both PFS and OS revealed no obvious asymmetry, suggesting a low risk of publication bias. Additionally, a network plot was generated to illustrate the treatment comparisons, including a total of four distinct therapeutic comparisons across six randomized controlled trials, two of which included three-arm designs with shared control arms, resulting in eight treatment effect estimates. Given the sparsity of data within each comparison and the limited indirect evidence loops, formal statistical assessments such as Egger’s regression test, node-splitting analysis, and loop-specific inconsistency testing were not conducted, in order to avoid generating unstable or potentially misleading results. Nonetheless, the consistency of treatment effects appeared acceptable based on the structure and coherence of the network plot.

Results In this meta-analysis of six phase III RCTs involving 1,777 patients met the inclusion criteria were included. For PFS, the pooled HR favored immunotherapy plus chemotherapy (IO+Chemo) over chemotherapy alone (HR = 0.77; 95% CI: 0.67–0.88), and triplet regimens combining immunotherapy, chemotherapy, and anti-angiogenic agents (IO+Chemo+Anti-angiogenic) demonstrated the greatest PFS benefit (HR = 0.54; 95% CI: 0.44–0.67). Stratified analysis by treatment regimen (IO+Chemo vs. IO+Chemo+Anti-angiogenic) demonstrated significant differences in PFS benefits (p = 0.0080). For OS, the benefit was less pronounced. IO+Chemo showed a trend toward improved OS versus chemotherapy (HR = 0.86; 95% CI: 0.75–1.00), while the triplet regimen did not significantly improve OS (HR = 0.98; 95% CI: 0.77–1.25). No significant differences were observed when comparing immunotherapy-based strategies to anti-angiogenesis-based combinations. In terms of safety, immunotherapy-based combinations were associated with a mild increase in any-grade adverse events (RR = 1.03; 95% CI: 0.96–1.10), and triplet regimens showed the highest risk increase in grade ≥3 adverse events (RR = 1.55; 95% CI: 0.01–161.03), though with substantial heterogeneity. However, across comparisons, most differences were not statistically significant, and toxicity profiles remained within a clinically manageable range.

Subgroup analyses revealed consistent PFS benefits favoring immunotherapy-based combinations over chemotherapy across multiple clinical and molecular characteristics. Statistically significant improvements in PFS were observed in both age groups (<65 years: HR 0.76, 95% CI 0.64–0.91; ≥65 years: HR 0.81, 95% CI 0.65–1.02), with no evidence of heterogeneity (p for interaction = 0.63). Similarly, both males (HR 0.76, 95% CI 0.61–0.95) and females (HR 0.80, 95% CI 0.66–0.98) derived comparable benefit, with no significant interaction effect. Among geographic subgroups, Asian patients experienced a significant PFS benefit (HR 0.80, 95% CI 0.68–0.94), while non-Asian populations did not show a consistent trend (HR 1.13, 95% CI 0.52–2.47), with notable heterogeneity. Patients with ECOG performance status 1 showed significant improvement (HR 0.75, 95% CI 0.63–0.88), whereas ECOG 0 patients demonstrated a numerically favorable but nonsignificant trend (HR 0.84, 95% CI 0.65–1.09). With respect to smoking history, both current/former smokers (HR 0.75, 95% CI 0.60–0.95) and never-smokers (HR 0.81, 95% CI 0.66–0.99) benefited from immunotherapy combinations, though no interaction was found. Particularly, in patients with brain metastases, IO + anti-angiogenic combinations yielded substantial benefit (HR 0.32, 95% CI 0.19–0.53), whereas IO + chemo alone did not (HR 0.92, 95% CI 0.71–1.19; p for subgroup interaction = 0.0029).Finally, patients without T790M mutations (HR 0.69, 95% CI 0.52–0.92) and those with PD-L1 ≥1% expression (HR 0.79, 95% CI 0.63–0.99) exhibited improved PFS, while those with T790M-positive status (HR 0.97, 95% CI 0.74–1.27) or PD-L1 <1% (HR 0.84, 95% CI 0.66–1.07) did not derive statistically significant benefit. These finding underscore the clinical value of stratified treatment selection based on demographic and molecular profiles.

Conclusion Compared with chemotherapy, immunotherapy-based combinations, particularly with anti-angiogenic agents, provide meaningful PFS benefit in EGFR-mutant NSCLC after TKI resistance. While OS benefits were not statistically significant, favorable trends in selected subgroups with EGFR L858R mutations and absence of T790M, highlighting the potential role of immunotherapy-based strategies. These findings warrant prospective validation in biomarker-stratified clinical trials.