Abstract:Dear Editor, Currently, the association of patients’ sex with survival benefit of immune checkpoint inhibitor (ICI) is being actively investigated, but inconsistent results have been produced. A meta-analysis1 reported that male patients had significantly lager overall survival (OS) benefits from ICI versus control treatment than did female patients. However, an updated meta-analysis2 found no significant difference in OS between sexes. This conflict indicated that the sex-related effects on ICI efficacy involved complex and unknown elements of tumor microenvironment. The phase III KEYNOTE-024 trial3 enrolled cancer patients with PD-L1 expression on at least 50% tumor cells and found a strikingly improved OS benefit with ICI compared with chemotherapy in male but only a minimal improvement in female. However, neither male nor female had significant OS benefit from ICI over chemotherapy in some trials recruiting cancer patients with a lower PD-L1 expression threshold (≥1%).4, 5 Therefore, we hypothesize that PD-L1 expression has essential impact on the clinical usefulness of sex. This study, based on a post hoc analysis of prospective individual patient data from five clinical trials including OAK, POPLAR, IMvigor210, KEYNOTE-001, and CheckMate-012, and a meta-analysis of nine randomized controlled trials (RCTs), is the first to clarify the sex-related difference in ICI efficacy by using PD-L1 expression (Figure 1A). We further evaluated the landscape of tumor immune microenvironment to explore potential factors underpinning this difference. Full methods were described in the Supporting Information Methods. Characteristics and references of included trials and patients were described in the Supporting Information Result 1 and Tables S1 and S2. The individual-patient level analysis of OAK, POPLAR, and IMvigor210 trials showed that for patients receiving ICI, there was no significant difference in OS between male and female (HR 0.85, 95% CI 0.72-1.02; P = .073; Figure S2A). However, stratified by PD-L1 expression of 1%, OS for patients with PD-L1 expression <1% was significantly longer for female compared with male (HR 0.58, 95% CI 0.42-0.80; P < .001), but OS for patients with PD-L1 expression≥1% did not significantly differed between the sexes (HR 1.01, 95% CI 0.79-1.29; P = .910) (Figure 1B,C and Table S3). Results were consistent stratified by PD-L1 expression of 50%. Overall survival was significantly longer for female than for male at PD-L1 threshold of <50% (HR 0.76, 95% CI 0.61-0.95; P = .013), but OS did not significantly differ between the sexes at PD-L1 threshold of 1–49% (HR 0.76, 95% CI 0.52-1.13; P = .173) or ≥50% (HR 0.64, 95% CI 0.31-1.33; P = .231) (Figure S2B-D and Table S3). When examining chemotherapy, sex was not associated with OS regardless of PD-L1 expression (see details in Supporting Information Results 2, Figure 1D-E, and Figure S3). The individual-patient level analysis of POPLAR and OAK trials found that ICI improved OS compared with chemotherapy in both female (HR 0.67, 95% CI 0.53-0.84; P < .001) and male (HR 0.74, 95% CI 0.62-0.88; P < .001) (Figure S4A,B). However, among patients with PD-L1 expression <1%, the benefit from ICI versus chemotherapy was significantly different in female (HR 0.57, 95% CI 0.38-0.85; P = .006; Figure 2A), but there was no significant difference between the two treatments in male (HR 0.93, 95% CI 0.68-1.26; P = .621; Figure 2B). Conversely, among patients with PD-L1 expression ≥1%, there was no significant difference between the two treatments in female (HR 0.69, 95% CI 0.47-1.01; P = .057; Figure 2C), whereas the benefit from ICI over chemotherapy was significantly different in male (HR 0.71, 95% CI 0.54-0.95; P = .022; Figure 2D). Further, we improved the evidence for patients with PD-L1 expression ≥1% by pooling the individual patient-level result with the meta-analysis result of eight other trials, which showed that although female patients showed significant OS benefit from ICI over chemotherapy (eight RCTs, 1602 patients; HR 0.86, 95% CI 0.74-1.00; P = .05), this benefit only remained significant in subsequent line (four RCTs, 850 patients; HR 0.77, 95% CI 0.62-0.96; P = .02), without significant benefits for patients in first-line setting (four RCTs, 752 patients; HR 0.96, 95% CI 0.81-1.15; P = .67) or in subgroups by regimen and ICI class (Figure 2E, Figures S5 and S6 and Table S4). Male patients with PD-L1 expression ≥1% could derive significant OS benefit from ICI over chemotherapy (nine RCTs, 3166 patients; HR 0.76, 95% CI 0.67-0.85; P < .01; Figure 2F). This benefit remained significant in both first line (four RCTs, 1393 patients; HR 0.79, 95% CI 0.65-0.97; P = .02) and subsequent line (five RCTs, 1723 patients; HR 0.73, 95% CI 0.62-0.85; P < .01), and in other tested subgroups (Figure 2F, Figures S7 and S8 and Table S4). There were no significant differences in effects on OS between subgroups. Similar findings were observed stratified by PD-L1 expression of 50% (see details in Supporting Information Results 3 and Figures S9 and S10). Finally, central memory T cells (Tcm), rather than tumor mutation burden (TMB) and neoantigen burden (NAB), was found to potentially correlated with the OS differences between the sexes (see details in Supporting Information Results 4; Figures S11-S24 and Tables S5-S7). To our knowledge, this is the first study to reveal that PD-L1 expression has decisive effect on sex-associated differences in ICI efficacy. Previous meta-analyses included patients across varieties of PD-L1 thresholds, which might explain their inconsistency.1, 2 Previous research indicated that sex differences in mutational landscape might explain ICI efficacy differentially associated with sex,6 but neither TMB nor NAB was significantly differed between male and female at any PD-L1 expression thresholds in our study. Instead, Tcm probably has potential impact on sex differences in ICI efficacy. However, with only bladder cancer patients from IMvigor210 trial having RNA sequencing data, we were unable to evaluate the role of Tcm in other cancer types. Additionally, multi-omics have been shown to mutually predict ICI efficacy7, 8; therefore, future studies are warranted to comprehensively investigate sex difference in immune landscape using multi-omics across different cancer types. This study found the survival benefits of ICI in male and female were greatly influenced by PD-L1 expression, especially in NSCLC. At PD-L1 expression <1%, ICI should be recommended for female but not for male, and Tcm might be essential to drive this recommendation. We suggest that sex and PD-L1 expression should be jointly taken into account in the clinical decision making for ICI in cancer. We sincerely thank all of the investigators and patients who participated in the included trials. Li, Chen, Zhang, Zhong, Ou, Hu, Yu, and Yao jointly designed the study and drafted of the manuscript. Hu, Yu, and Yao provided supervision. Li and Yao obtained funding. All authors contributed to the data collection, data analysis and interpretation, manuscript revision, and approval of the final version. National Science and Technology Major Project; Grant Number: 2020ZX09201021; Medical Artificial Intelligence Project of Sun Yat-Sen Memorial Hospital; Grant Number: YXRGZN201902; National Natural Science Foundation of China; Grant Numbers: 81572596, 81972471, and U1601223; Natural Science Foundation of Guangdong Province; Grant Number: 2017A030313828; Guangzhou Science and Technology Major Program; Grant Number: 201704020131; Guangdong Science and Technology Department; Grant Number: 2017B030314026; Sun Yat-Sen University Clinical Research 5010 Program; Grant Number: 2018007; Sun Yat-Sen Clinical Research Cultivating Program; Grant Number: SYS-C-201801; Special Funds for the Cultivation of Guangdong College Students’ Scientific and Technological Innovation; Grant Number: pdjh2019a0212; National Students’ Innovation and Entrepreneurship Training Program; Grant Number: 201910571001. The study protocol was approved by the ethics committee of the Sun Yat-sen Memorial Hospital of Sun Yat-sen University. The requirement for informed consent of study participants was waived because the human data were obtained from publicly available datasets. The authors declare that they have no conflict of interest. The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request (yaoherui@mail.sysu.edu.cn). Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Abstract:Dear Editor, We identified an immunomodulatory signature that was significantly associated with clinical improvement of immune checkpoint blockade (ICB) therapy response and better prognosis in urothelial carcinoma1 and melanoma.2 Our finding is helpful for identifying cancer patients who may benefit from ICB therapy. ICB therapy is an effective treatment regimen for cancer patients. Tumor mutation burden (TMB),3 expression of PD-1, PD-L1,4 and CTLA-4, are known factors associated with ICB therapy response. However, our understanding of biomarkers for response to ICB therapy remains incomplete. This necessitates further investigation on identification of reliable biomarkers for ICB therapy response. A flowchart depicting procedures of the study was shown in Figure 1A, which consisted of feature representation learning of gene expression from single cells and association of expression signatures with ICB therapy response in two clinical trials. We collected gene expression data from 71,494 single cells encompassing tumor cells, stroma cells, dendritic cells, CD4+, and CD8+ T cells from seven cancer types (Table S1). A deep learning model5, 6 was developed by iteratively training on gene expression data of these 71,494 single cells. Subsequently, it was applied to extract 258 expression signatures for single cells and samples from two clinical trials. We ranked the extracted expression signatures of single cells with a gene set involved in immunomodulation7 (see Supporting Information Methods, Table S2) and examined association of the top five ranking signatures with clinical improvement of ICB therapy and prognosis. The t-SNE embedded scatter plot shown in Figure 1B depicted single cell clusters unveiled from expression signatures learned by deep learning model. The same cell types (such as CD4+ or CD8+ T cell) from different data set were clustered together, whereas different cell types were separated distinctively (Figures S1A and B). Notably, a cluster of unsorted melanoma single cells were interspersed with different T-cell types (Figure S1C). We observed that 63.6% (164/258) of expression signatures extracted from single cells were significantly associated with immunomodulation7 (Figure 1C), out of which the receptor subcategory was the most significantly enriched among seven groups. The top five ranking immunomodulatory signatures were characterized by upregulation of genes related to immunomodulatory ligands (e.g., TNFSF4, CCL5, and IFNG), receptors (e.g., HAVCR2, TNFRSF18, PDCD1, and LAG3), and antigen presentation (e.g., HLA-DRB5, HLA-DRA, and HLA-DRB1) (Figure 1D). We categorized ICB therapy response into clinical and no clinical improvement (see Supporting Information Methods). The baseline characteristics are shown in Table S3. The urothelial carcinoma clinical trial1 consisted of 298 urothelial cancer patients all treated with PD-L1 inhibitor, 68 of which were categorized as clinical improvement and the rest 230 were categorized as no clinical improvement group. The melanoma clinical trial2 consisted of 119 melanoma cancer patients, 47 of which achieved clinical improvement and the rest did not. In melanoma clinical trial, 72 patients received PD-1 inhibitor, whereas 47 patients received CTLA4 in conjunction with PD-1 inhibitor. The other available molecular markers related to ICB therapy response or prognosis such as TMB, the degree of cytotoxic lymphocytes infiltration (CTL),8 expression levels of PD-1, PD-L1, and CTLA4, and clinical features are summarized in Table S2. TMB data were available for 78.5% (234/298) and 100% (119/119) of patients in these two clinical trials. Among these top five ranking signature, one of the expression signature scores was significantly different between clinical improvement group and no clinical improvement group for the urothelial carcinoma (Figure 2A; −0.72 versus −0.09, 95% CI = −0.82 to −0.62 versus −0.30 to 0.12; Wilcoxon test, adjusted P < .001) and melanoma (Figure 2B; −0.69 versus −0.24, 95% CI = −0.88 to −0.50 versus −0.50 to 0.02; Wilcoxon test, adjusted P = .03) clinical trials. This immunomodulatory signature was significantly associated with clinical improvement of ICB therapy across two clinical trials (OR = 4.34 and 3.32, 95% CI = 2.30-8.23 and 1.35-8.47; Fisher's exact test, P < 0.001 and P = 0.006). Association of this immunomodulatory signature with clinical improvement of ICB therapy remained statistically significant after controlling for TMB, CTL, expression levels of PD-1, PD-L1, and CTLA4 across the urothelial carcinoma clinical trial (OR = 1.30, 95% CI = 1.14-1.47; multivariate logistic model, P < 0.001; Figure 2C) and melanoma clinical trial (OR = 1.32, 95% CI = 1.09-1.59; multivariate logistic model, P = 0.005; Figure 2D). We further examined the significance of identified immunomodulatory signature with respect to anti-PD1 and anti-PD-L1 inhibitors. In urothelial carcinoma clinical trial, all patients were treated with anti-PD-L1 inhibitors. In melanoma clinical trial, the identified immunomodulatory signature remained significant in patients treated with anti-PD-1 treatment (OR = 1.32, 95% CI = 1.05-1.67, P = 0.02; Figure S2A), while it also trended toward better association in patients treated with anti-PD-1 in conjunction with anti-CTLA4 (Figure S2B). In addition, the identified immunomodulatory signature was significantly associated with improved overall survival in urothelial carcinoma clinical trial (log-rank test, P = 0.004; Figure 3A) and marginally significant in melanoma trial (log-rank test, P = 0.07; Figure 3B). This identified signature remained significant in urothelial cohort (HR = 0.61, 95% CI = 0.39-0.96; multivariate Cox model, P = 0.03; Figure 3C) and marginal significant in melanoma cohort (HR = 0.57, 95% CI = 0.30-1.07; multivariate Cox model, P = 0.08; Figure 3D) after controlling for confounding factors. Gene set enrichment analysis revealed that immune-related signaling circuits9 (Table S4) were overrepresented in patients with activated immunomodulatory signature, while signatures featuring cancer cell proliferation and invasiveness such as epithelial-to-mesenchymal transition and angiogenesis were underrepresented (Figure S3). In addition, tumor immune microenvironment signatures obtained from CIBERSORT algorithm10 showed that there were significant differences among infiltration of CD8+ T cells, activated CD4+ memory T cells, M1 macrophages, and dendritic cells in patients stratified by this signature (Figure S4). In summary, we reported that an immunomodulatory signature dissected from large-scale single cell expression data was significantly associated with clinical improvement of ICB therapy and better prognosis in the two independent ICB therapy clinical trials. However, further investigation in prospective randomized clinical trial is warranted. We want to thank Prof. Wei Zhang from Wake Forest Cancer Center for his constructive comments and editing the manuscript and Liangtao Zheng from Peking University for his comments on data interpretation as well as researchers for their generosity to made their data publicly available. This study was approved by the institutional review board (IRB) of Tianjin Cancer Hospital. Informed consent was exempted by the IRB given that data were obtained from public, open access database. The authors declare that they have no conflict of interest. National Natural Science Foundation of China; Grant No. 31801117; Program for Changjiang Scholars and Innovative Research Team in University in China; Grant No.: IRT_14R40; Tianjin Science and Technology Committee Foundation; Grant No.: 17JCYBJC25300; Chinese National Key Research and Development Project; Grant No. 2018YFC1315600. Xiangchun Li and Kexin Chen designed and supervised the study; Hongru Shen and Xiangchun Li performed data collection, analysis, and wrote the manuscript; Dan Wu, Xilin Shen, Mengyao Feng, Yichen Yang, Yang Li, and Meng Yang collected data; Wei Wang and Qiang Zhang interpretated the data; Xiangchun Li, Kexin Chen, and Hongru Shen revised the manuscript. Data are available in open access database. All data relevant to the study are included in the article or uploaded as supplementary information. All data generated or analyzed during this study are included in this manuscript. A website running the developed deep learning model to extract gene expression signatures is freely available at https://lixiangchun.github.io. Figure S1. t-SNE visualization of (A) CD4+ and CD8+ T cells, (B) astrocytoma, melanoma, and head and neck carcinoma, and (C) CD4+, CD8+ T cells, and unsorted admixed melanoma cells. Figure S2. Association between the identified immunomodulatory signature and clinical improvement in (A) anti-PD-1 and (B) anti-PD-1 in conjunction with anti-CTL4 from the melanoma ICB therapy clinical trial. Confounding factors included TMB, CTL and expression levels of PD-1, PD-L1 and CTLA-4. Figure S3. Gene set enrichment analysis of the identified immunomodulatory signature. Signaling pathways statistically significant (adjusted p < 0.10) in both two datasets were displayed in the heatmap. Figure S4. Tumor immune microenvironment signatures unveiled by CIBERTSORT. The infiltration difference of each immune signature was evaluated by t-test. The p-values were subjected for multiple hypothesis test. Adjusted p-value < 0.05 was considered to be significant. (*** represented adjusted p ≤ 0.001 and ** 0.05 < adjusted p < 0.001) Table S1. Data source information. Table S2. Immunomodulatory gene sets. Table S3. Clinical characteristic of the patients with ICB therapy in two clinical trials. Table S4. Immune related signaling gene sets. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Joint Association of Patients’ Sex and PD‐L1 Expression with Overall Survival Benefits and Tumor‐immune Microenvironment in Immune Checkpoint Inhibitors for Cancers

1036P Patients’ Sex and PD-L1 Expression Jointly Associated with Overall Survival Benefits of Immune Checkpoint Inhibitors in Cancer

Abstract 4056: Correlation Analysis Between PD-L1 Expression, TMB and Clinical Characteristics in Chinese Non-Small Cell Lung Cancer

Impact of Clinicopathological Features on the Efficacy of Immune Checkpoint Inhibitors Plus Conventional Treatment in Patients with Advanced Lung Cancer

Sex‐based immune microenvironmental feature heterogeneity in response to PD‐1 blockade in combination with chemotherapy for patients with untreated advanced non‐small‐cell lung cancer

163POX40/OX40L protein expression in non-small cell lung cancer and its role in clinical outcome and relationships with other immune biomarkers

Abstract 620: Associations of Genomic Alteration, Tumor Mutational Burden with PD-L1 Expression and Response to Immune Checkpoint Inhibitors in Chinese Lung Patients

Effect of Sex on the Efficacy of Patients Receiving Immune Checkpoint Inhibitors in Advanced Non‐small Cell Lung Cancer

Impact of Clinicopathologic Features on the Efficacy of PD-1/PD-L1 Inhibitors in Patients With Previously Treated Non-small-cell Lung Cancer.

An Immunomodulatory Signature of Responsiveness to Immune Checkpoint Blockade Therapy

Immunotherapy Efficacy and Safety of Programmed Cell Death 1 (PD-1) Versus Programmed Death Ligand-1 (PD-L1) Inhibitors in Pan-Cancer Patients: A Systematic Review and Meta-Analysis

Association of Immune Checkpoint Inhibitor with Survival in Patients with Cancers with Protein Tyrosine Phosphatase Receptor T Mutation

Association between sex and efficacy of immune checkpoint inhibitors: a systematic review and meta-analysis

Association Between Efficacy of Immune Checkpoint Inhibitors and Sex: An Updated Meta-Analysis on 21 Trials and 12,675 Non-Small Cell Lung Cancer Patients

Clinical and Molecular Characteristics Associated with Survival among Cancer Patients Receiving First-Line Anti-Pd-1/pd-l1-based Therapies.

Relationship Between Patients' Baseline Characteristics and Survival Benefits in Immunotherapy-Treated Non-Small-Cell Lung Cancer: A Systematic Review and Meta-Analysis

Effects of Clinicopathological Characteristics on the Survival of Patients Treated with PD-1/PD-L1 Inhibitor Monotherapy or Combination Therapy for Advanced Cancer: A Systemic Review and Meta-Analysis

Association of PD-L 1 expression with clinicopathologic features and overall survival in patients with non-small-cell lung cancer

The Association of PD-L1 Expression with the Efficacy of Anti-Pd-1/pd-l1 Immunotherapy and Survival of Non-Small Cell Lung Cancer Patients: a Meta-Analysis of Randomized Controlled Trials

The Effect of Sex on the Therapeutic Efficiency of Immune Checkpoint Inhibitors: A Systematic Review and Meta-Analysis Based on Randomized Controlled Trials

Identification of TMB, CD8 T-cell Abundance, and Homologous Repair Pathway Mutation Frequency As Predictors of the Benefit-Toxicity Ratio of Anti-Pd-1/pd-l1 Therapy