第一部分  局部晚期或转移性非小细胞肺癌接受抗PD-1/PD-L1免疫治疗的预后分析及疗效预测因素的回顾性研究

【研究目的】 本研究旨在评估真实世界局部晚期或转移性非小细胞肺癌（non-small cell lung cancer，NSCLC）患者抗程序性细胞死亡受体-1（programmed cell death-1，PD-1）/ 程序性细胞死亡-配体1（programmed cell death-ligand 1，PD-L1）免疫单药或联合治疗的疗效及预后，并挖掘与一线抗PD-1免疫联合化疗治疗疗效及长期获益（long term response，LTR）相关的预测因素。

    【方法】 本研究收集了2018年1月1日至2022年2月28日期间在中国医学科学院肿瘤医院接受抗PD-1/PD-L1免疫单药或联合治疗的局部晚期或转移性NSCLC患者（队列A）的临床资料，进行治疗疗效和生存结局的分析。在此基础上，全面收集接受一线抗PD-1免疫联合化疗治疗的NSCLC患者（队列B）的临床特征、基线实验室指标、外周血淋巴细胞亚群及常见基因突变状态，通过Logistic回归分析挖掘与LTR相关的特征，并通过Cox回归分析探究与无进展生存期（progression-free survival，PFS）和总生存期（overall survival，OS）相关的预测指标。LTR定义为PFS超过24个月。

【结果】 队列A共纳入319例患者，截止到2024年5月21日，患者的中位随访时间为32.1个月 [95%置信区间（confidence interval，CI）29.2-38.0个月]，队列A的中位PFS为7.7个月（95%CI 6.2-9.7个月），中位OS为29.9个月（95%CI 23.6-37.5个月），接受一线和二线抗PD-1/PD-L1免疫单药或联合治疗患者的5年OS率分别为47.2%（95%CI 38.7%-57.6%）和18.6%（95%CI 9.8%-35.6%）。队列B共纳入167例患者，其中25.1%（n = 42）的患者获得LTR。多因素Logistic分析结果显示，年龄＜65岁 [优势比（odds ratio，OR）= 3.22，P = 0.024]、超重或肥胖 [身体质量指数（body mass index, BMI）≥24 kg/m², OR = 3.26，P = 0.020]以及C反应蛋白/白蛋白比值（C-reactive protein /albumin ratio，CAR）＜0.07（OR = 9.94，P = 0.039）是LTR的独立预测因素。此外，在调整年龄、BMI和TNM分期后，外周血B细胞占比≥14.5%（OR = 9.23，P = 0.045）也被确认为LTR的预测指标。多因素Cox回归分析结果显示，TNM分期为IV期 [风险比（hazard ratio, HR）= 1.86，P = 0.03]和肝转移（HR = 2.80，P = 0.01）是PFS的独立不良预后因素。BMI＜18.5 kg/m²（HR = 4.52，P = 0.005）、基线白蛋白/胆红素评分（albumin-bilirubin，ALBI）评分＞-3.0（HR = 2.45，P = 0.042）、肺免疫预后指数（lung immune prognostic index，LIPI）评分为2分（HR = 27.5，P = 0.015）、CAR评分≥0.07（HR = 2.83，P = 0.016）以及肝转移（HR = 3.51，P = 0.02）是OS的独立不良预后因素。基因突变方面，携带TP53突变患者的PFS （HR = 2.09，P = 0.028）与OS（HR = 2.45，P = 0.023）较野生型患者显著缩短。

    【结论】 BMI＜18.5kg/m²、基线ALBI评分＞-3.0、LIPI评分为2分、CAR评分≥0.07和肝转移是接受一线抗PD-1免疫联合化疗治疗的NSCLC患者OS的独立不良预后因素。年龄＜65岁、BMI≥24 kg/m²及基线CAR评分＜0.07的患者更可能实现长期获益。

【关键词】 NSCLC；免疫治疗；疗效；长期获益

第二部分  多组学揭示CTSL作为非小细胞肺癌抗PD-1/PD-L1免疫治疗的疗效预测生物标志物

【研究目的】 抗原提呈在肿瘤免疫中扮演重要角色，本研究旨在从抗原提呈相关基因中挖掘可预测NSCLC患者抗PD-1/PD-L1免疫治疗疗效和预后的潜在生物标志物。

【方法】 本研究收集了两个接受抗PD-1/PD-L1免疫单药治疗的晚期NSCLC数据集（GSE126044和GSE135222）的RNA测序数据，并将其整合作为GEO发现队列。基于京都基因与基因组百科全书（Kyoto Encyclopedia of Genes and Genomes，KEGG）数据库收集的抗原加工与呈递基因集（hsa04612），利用GEO发现队列数据进行PFS的单因素Cox回归分析，挖掘关键疗效预测基因。利用GSE117570和GSE131907单细胞测序数据集分析关键基因在不同免疫细胞亚型的分布特征及其细胞间通讯特点。基因本体生物过程（Gene ontology biological process，GOBP）、KEGG、单样本基因集富集分析（single sample gene set enrichment analysis，ssGSEA）用于关键基因的功能富集分析和免疫微环境分析。收集来自中国医学科学院肿瘤医院接受抗PD-1/PD-L1免疫治疗的局部晚期或转移性NSCLC患者的组织样本进行多重免疫荧光染色，收集疗前血浆样本通过酶联免疫吸附试验（enzyme-linked immune sorbent assay，ELISA）测定关键基因的血浆蛋白浓度，评估和验证其免疫治疗疗效预测价值。最后，与第一部分研究发现的OS独立预后因素相整合，建立可预测接受抗PD-1/PD-L1免疫联合化疗治疗的局部晚期或转移性NSCLC患者预后的列线图模型。
    【结果】 基于GEO发现队列的43例接受抗PD-1/PD-L1免疫单药治疗的晚期NSCLC转录组测序数据，本研究筛选出17个抗原提呈相关基因的表达水平与PFS显著相关，并识别出其中的关键基因：组织蛋白酶L（Cathepsin L，CTSL）。根据中位表达值分界，CTSL高表达组呈现更活跃的免疫富集通路及免疫细胞浸润特征。单细胞RNA测序分析结果显示，CTSL主要在单核/巨噬细胞中表达，与CTSL阴性的单核/巨噬细胞相比，CTSL阳性的单核/巨噬细胞与其他免疫和基质细胞之间有更丰富的细胞互作网络。多项免疫治疗队列验证了CTSL的RNA及组织蛋白高表达与NSCLC患者接受抗PD-1/PD-L1治疗更佳疗效显著相关。在包含19例NSCLC肿瘤样本的多重免疫荧光分析队列中，CTSL+细胞和CD8+细胞的数量呈显著正相关关系（r = 0.60，P = 0.0068）。在84例接受抗PD-1免疫联合化疗治疗的局部晚期或转移性NSCLC疗前血浆样本中，CTSL血浆蛋白浓度高（≥4180 pg/mL）的患者PFS和OS显著延长（中位PFS: 26.3 vs 7.6个月，P = 0.009; 中位OS: 未达到 vs 21.2个月，P = 0.012）。基于BMI、基线ALBI评分、CAR评分、肝转移状态及CTSL血浆蛋白浓度这5个指标构建的列线图模型预测接受抗PD-1免疫联合化疗治疗的局部晚期或转移性NSCLC患者1年、2年和3年OS率的性能良好 [训练集的C-指数：0.761（95%CI 0.642-0.880）；验证集的C-指数：0.862（95%CI 0.739-0.985）]。

【结论】 本研究首次揭示了CTSL可作为局部晚期或转移性NSCLC患者抗PD-1/PD-L1免疫治疗疗效预测的潜在生物标志物。基于BMI、基线ALBI评分、CAR评分、肝转移状态及CTSL血浆蛋白浓度构建的列线图模型可较好预测接受抗PD-1免疫联合化疗治疗的局部晚期或转移性NSCLC患者预后。

【关键词】 NSCLC；免疫治疗；分子标志物

Part 1.  Prognostic analysis and investigation of predictive factors for anti-PD-1/PD-L1 immunotherapy in locally advanced or metastatic non-small cell lung cancer: a retrospective study

Objective: In this study, we aimed to evaluate the real-world efficacy and survival outcome of anti-programmed cell death-1 (PD-1) / programmed cell death-ligand 1 (PD-L1) monotherapy or combination therapy in locally advanced or metastatic non-small cell lung cancer (NSCLC), as well as to investigate factors predicting efficacy and LTR in patients receiving first-line treatment of anti-PD-1 plus chemotherapy.

Methods: In this study, we retrospectively collected clinical data of patients with locally advanced or metastatic NSCLC who received anti-PD-1/PD-L1 monotherapy or combination therapy from January 1st, 2018 to February 28th, 2022 at the Cancer Hospital, Chinese Academy of Medical Sciences & Peking Union Medical College (cohort A), the efficacy and survival outcomes were analyzed. Then various characteristics including clinical parameters, baseline laboratory examination, peripheral blood lymphocyte subsets and common gene mutation profile in NSCLC patients who received first-line anti-PD-1 plus chemotherapy treatment (cohort B) were collected comprehensively. The predictive factors associated with LTR were explored via Logistic regression analysis, the predictive factors associated with progression-free survival (PFS) and overall survival (OS) were investigated using Cox regression analysis. LTR was defined as PFS exceeding 24 months.

Results: Cohort A included 319 patients. As of May 21th, 2024, with a median follow-up time of 32.1 months (95% confidence interval [CI] 29.2-38.0), a median PFS of 7.7 months (95%CI 6.2-9.7) and a median OS of 29.9 months (95%CI 23.6-37.5) were observed in patients in cohort A. The 5 year-OS rates among patients receiving the first-line and second-line anti-PD-1/PD-L1 monotherapy or combination therapy were 47.2% (95%CI 38.7% -57.6%) and 18.6% (95%CI 9.8%-35.6%), respectively. Cohort B included 167 patients, and 25.1% (n = 42) of patients achieved LTR. The results of multivariate Logistic analysis showed that independent baseline predictors of LTR included age＜65 years (odds ratio [OR] = 3.22, P = 0.024), overweight or obesity (body mass index [BMI]≥24 kg/m², OR = 3.26, P =0.020), and baseline C-reactive protein/albumin ratio (CAR) score＜0.07 (OR = 9.94, P = 0.039). A peripheral B cell percentage≥14.5% was found to be an indicator for LTR (OR = 9.23, P = 0.045) after adjusting for age, BMI and TNM stage. The results of multivariate Cox analysis showed that TNM stage IV (hazard ratio [HR] = 1.86, P = 0.03) and liver metastasis (HR = 2.80, P = 0.01) were independent adverse predictors of PFS. Besides, BMI＜18.5kg/m² (HR = 4.52, P = 0.005), baseline albumin-bilirubin (ALBI) score＞-3.0 (HR = 2.45, P = 0.042), lung immune prognostic index (LIPI) score of 2 (HR = 27.5, P = 0.015), a CAR score≥0.07 (HR = 2.83, P = 0.016) and liver metastasis (HR = 3.51, P = 0.02) were independent adverse predictors of OS. In terms of genetic mutations, patients harboring TP53 mutations showed significantly shorter PFS (HR = 2.09, P = 0.028) and OS (HR = 2.45, P = 0.023) compared to those with wild-type TP53.

Conclusion:  BMI＜18.5kg/m², baseline ALBI score＞-3.0, LIPI score of 2, CAR score≥0.07 and liver metastasis were independent adverse predictors of OS in patients with NSCLC who received the first-line treatment of anti-PD-1 plus chemotherapy. Patients whose age＜65 years, BMI≥24 kg/m², and baseline CAR score＜0.07 were more likely to achieve LTR.

Keywords: NSCLC, efficacy, long-term response, immunotherapy

Part 2. Multi-omics revealed CTSL as a potential biomarker in predicting the efficacy of anti-PD-1/PD-L1 immunotherapy in non-small cell lung cancer

Objective: Antigen presentation plays a pivotal role in tumor immunity. This study aimed to identify potential antigen presenting-related biomarkers for predicting the efficacy and prognosis of anti-PD-1/PD-L1 immunotherapy in patients with NSCLC.

Methods: The RNA-sequencing data from two advanced NSCLC cohorts receiving anti-PD-1/PD-L1 immunotherapy (GSE126044 and GSE135222) were integrated to form the GEO discovery cohort. Based on an antigen process and presentation-related gene set (hsa04612) collected from the Kyoto Encyclopedia of Genes and Genomes (KEGG) database, univariate Cox regression analysis of PFS was performed on the GEO discovery cohort to identify key genes. The immune cellular distribution of the key gene and cell-cell communication network were investigated using single-cell RNA-seq data from the GSE117570 and GSE131907 datasets. Gene ontology biological process (GOBP), KEGG, and single sample gene set enrichment analysis (ssGSEA) analyses were performed to explore functional enrichment and immune cell infiltration characteristics associated with the key gene. To validate the efficacy and prognostic predictive value of the key gene, baseline tissue samples from patients with locally advanced or metastatic NSCLC who received anti-PD-1/PD-L1 immunotherapy at the Cancer Hospital, Chinese Academy of Medical Sciences were collected to perform the multiplex immunofluorescence (mIF) staining, baseline plasma sample were also collected to examine the plasma concentration of the key gene via enzyme-linked immune sorbent assay (ELISA). Finally, the finding was integrated with independent prognostic indicators of OS (as identified in Part One) to construct a nomogram model to predict survival of locally advanced or metastatic NSCLC patients who received anti-PD-1/PD-L1 plus chemotherapy.

Results: Based on the RNA sequencing data of 43 advanced NSCLC patients who received anti-PD-1/PD-L1 monotherapy (GEO discovery cohort), we initially discovered 17 antigen presenting-related genes that were associated with PFS. Among them, Cathepsin L (CTSL) was identified as a key gene. Based on the median cutoff, the high CTSL expression group exhibited more active immune cell infiltration and enrichment of immune-related pathways. Single-cell RNA sequencing analysis revealed that CTSL was primarily expressed in monocytes/macrophages. Compared to CTSL- monocytes/macrophages, CTSL+ monocytes/macrophages exhibited a more enriched cell-cell communication network with other immune and stromal cells. Several immunotherapy cohorts confirmed that both higher RNA and tissue protein expression levels of CTSL were associated with better efficacy of anti-PD-1/PD-L1 immunotherapy in patients with NSCLC. In a cohort of 19 tumor specimens with mIF staining, a significant positive correlation was observed between the abundance of CTSL+ cells and CD8+ T cells (r = 0.60, P = 0.0068). Notably, among baseline plasma samples of 84 patients with locally advanced or metastatic NSCLC who received anti-PD-1 plus chemotherapy, those with higher plasma CTSL concentration (≥4180 pg/mL) exhibited significantly improved prognosis (median PFS: 26.3 vs 7.6 months, P = 0.009; median OS: not reached vs 21.2 months, P = 0.012). Finally, a nomogram predictive model incorporating BMI, ALBI score, CAR score, liver metastasis status, and plasma CTSL concentration was developed to estimate the likelihood of achieving 1-year, 2-year, and 3-year OS rates from the treatment of anti-PD-1 plus chemotherapy in patients with locally advanced or metastatic NSCLC, with robust performance (C-index of training set: 0.761 [95%CI 0.642-0.880], C-index of validation set: 0.862 [95%CI 0.739-0.985]).

Conclusion: This study first revealed CTSL as a potential biomarker in predicting the efficacy of anti-PD-1/PD-L1 immunotherapy in patients with locally advanced or metastatic NSCLC. The nomogram model based on BMI, baseline ALBI score, CAR score, liver metastasis status, and plasma CTSL concentration was established with favorable potential in predicting survival of patients with locally advanced and metastatic NSCLC who received anti-PD-1 plus chemotherapy treatment.

Keywords: NSCLC, immunotherapy, biomarkers.