Immune-related adverse event prediction and influential factor identification.

Immune-Related Adverse Events and Outcomes in Patients with Advanced Non–Small Cell Lung Cancer: A Predictive Marker of Efficacy?

2032 Background: 30-50% of patients with non-early NSCLC will eventually develop BM, with a median survival of less than one year from BM diagnosis. There are no widely accepted clinical risk models for development of BM in patients without them at baseline. We predicted the binary risk of BM using clinical and genetic factors from a large multi-institutional cohort. Methods: Stage II-IV NSCLC patients from the AACR Project GENIE Biopharma Consortium dataset were eligible. This consisted of 4 academic institutions who curated clinical data of patients who had somatic next-generation tumor sequencing (NGS) between 2015-2017. We excluded patients who had BM at baseline, died within 30 days of NSCLC diagnosis, or did not undergo brain imaging. Covariates included demographics, anticancer therapies (received up to 90 days prior to BM development and within 5 years from NSCLC diagnosis), and NGS data; radiotherapy (RT) data were not available. NGS features included mutations and copy number alterations. These features were restricted to those classified as oncogenic by OncoKB. Univariate feature selection with Fisher’s test (p<.1) was performed on medication and genetic features. We compared 5 different machine learning models for prediction: random forest (RF), support vector machine (SVM), lasso regression, ridge regression, and an ensemble classifier. We split our data into training and test sets. 10-fold cross-validation was done on the training set for parameter tuning. The area under the receiver-operating curve (AUC) is reported on the test set. Results: 956 patients were included, 192 (20%) in the test set. Univariate features associated with BM were treatment with etoposide, Asian race, presence of bone metastases at NSCLC diagnosis, mutations in TP53 and EGFR, amplifications of ERBB2 and EGFR, and deletions of RB1, CDKN2A and CDKN2B. Univariate features inversely associated with BM were older age, treatment with nivolumab, vinorelbine, alectinib, pembrolizumab, atezolizumab, and gemcitabine, as well as mutations in NOTCH1 and KRAS. Ridge regression had the best AUC, 0.73 (Table). Conclusions: We achieved reasonable prediction performance using commonly obtained clinical and genomic information in non-early NSCLC. The biologic role of the associated alterations deserves further scrutiny; this study replicates similar findings for EGFR and KRAS in a much smaller cohort. Certain subsets of NSCLC patients may benefit from increased surveillance for BM and transition to drug therapies known to effectively cross the blood-brain barrier, e.g., nivolumab and alectinib. Inclusion of additional covariates, e.g., brain RT, may further improve model performance.[Table: see text]

Background: Cancer immunotherapy, especially immune checkpoint inhibitors (ICIs), has provided a remarkable antitumor effect in non-small cell lung cancer (NSCLC), but only a limited number of patients can derive durable benefit. Therefore, more accurate biomarkers are highly needed. The fibroblast growth factor receptor (FGFR) family of receptor tyrosine kinases is comprised of 4 members (FGFR1-4) that mediate the function of the FGFR ligand family. Although the genomic alterations (GAs) of FGFRs were frequently detected by targeted next-generation sequencing (NGS) and considered as potent oncogenes, their correlation with recognized immune biomarkers and predictive value for ICIs response in NSCLC remains unclear. Methods: Formalin-fixed, paraffin-embedded (FFPE) tumor and matched blood samples of 3433 NSCLC patients from OrigiMed were collected for targeted NGS panel sequencing from December 2017 to January 2019. GAs including single nucleotide variations, short and long insertions/deletions, copy number variations, and gene rearrangements were assessed. Genomic data and ICIs treatment outcome of 2 cohorts of NSCLC patients were derived from The Cancer Genome Atlas (TCGA) and Memorial Sloan Kettering Cancer Center (MSKCC) databases. Results: GAs of FGFR gene family were detected in 5% (172/3433) of patients with NSCLC in the OrigiMed database, including 84 FGFR1 GAs, 34 FGFR4 GAs, 28 FGFR2 GAs, and 26 FGFR3 GAs. No significant difference was observed in the frequency of FGFR4 GAs within the histologic subtypes. Compared with wild-type (WT), FGFR4 GAs were associated with a significantly higher tumor mutational burden (TMB) (6.55 vs. 4.6 muts/Mb, respectively, p<0.001). Survival analysis from 2 independent cohorts confirmed that patients with FGFR4 GAs had remarkable clinical benefit to ICIs compared with WT patients in both progression free survival (PFS) (13.17 months vs 3.17 months, respectively, p=0.025) and durable clinical benefit (DCB) (80% vs 29.3%, respectively, p=0.03). Furthermore, FGFR4 GAs were an independent risk factor of PFS (HR: 0.22, 95%CI: 0.055-0.92, p=0.037) and overall survival (OS) (HR: 0.23, 95%CI: 0.057-0.92, p=0.038). Conclusion: Our results supported that the FGFR4-mutated NSCLC patients were associated with an increased TMB and favorable response to ICIs. This suggests that GAs of FGFR4 may have implications as a biomarker for guiding ICI treatment. Citation Format: Xiangfeng Jin, Xinglong Fan, Kaihua Tian, Hanlin Xu, Nan Ge, Yafei Wang, Shiyue Zhang, Hui Chen, Shaohua Yuan, Kai Wang. FGFR4 mutationas a novel biomarker for immunotherapy in patients with non-small cell lung cancer [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 3193.

e21049 Background: Sarcopenia has been receiving attention due to its association with prognoses in patients with cancer. We investigated the association between evolving sarcopenia assessed with computed tomography (CT) and outcome of non-small cell lung cancer (NSCLC) patients receiving immune checkpoint inhibitors (ICIs). Methods: In this retrospective study, 47 NSCLC patients who received ICIs at our Hirosaki University Hospital were enrolled. Skeletal muscle area at the level of the third lumbar vertebra (L3-SMA) was measured from CT images before and 90 days after the ICI treatment. The factor related to the poor efficacy of ICIs was determined by logistic regression analysis. The prognostic cut-off value was estimated by a receiver operating characteristic (ROC) curve. As covariate factors, age, gender, smoking history, histology, performance status (PS), PD-L1 tumor proportion score (TPS), status of driver mutations, body mass index (BMI), serum protein and albumin levels, and lymphocyte count were included. Results: In the univariate analysis, reduction rate of L3-SMA, PS, BMI, serum protein and albumin levels, and lymphocyte count were significantly associated with efficacy. In the multivariate analysis, only reduction rate of L3-SMA was significant. The cut-off value was estimated to be 6%. The study subjects were divided into two groups by the cut-off value. In both of progression free and overall survival analysis, those with more than 6% of reduction had significantly shorter period being free of events than that of the other. Conclusions: Evolving sarcopenia assessed with CT images is promising as a prognostic factor in Japanese NSCLC patients receiving ICIs.

Immune checkpoint inhibitors (ICIs) have changed the treatment landscape for non-small cell lung cancer, but the role of race/ethnicity is not well understood. This is a retrospective study of non-small cell lung cancer (NSCLC) patients receiving ICI. We evaluated the role of race/ethnicity in ICI response and immune-related adverse events (irAEs) in NSCLC patients. NSCLC patients treated with ICIs from 2014 to 2022 at Los Angeles General Medical Center and Norris Comprehensive Cancer Center were included. Primary endpoints were irAE incidence, time on ICI treatment (TOT), overall survival (OS), and progression-free survival (PFS). TOT, OS, and PFS were evaluated using the Kaplan-Meier method. Landmark analysis was performed of patients receiving >6 month ICIs. Fisher's exact test was performed for analysis of variables between groups. In total, 211 NSCLC patients receiving ICIs were analyzed, including 86 (40.8%) Asian American/Pacific Islander (AAPI), 65 (30.8%) non-Hispanic White/Caucasian (NHW), 37 (17.5%) Hispanic/Latino (HIS), and 23 (10.9%) African American/Black (AA). Among stage IV patients, median OS for AAPI was 23.2 months (95% confidence interval (CI) 16.0-not reached (NR)), 23.6 months (95% CI 6.9-38.8) for NHW, 12.7 months for AA (95% CI 2.2-37.0), and 11.0 months (95% CI 4.8-14.8) for HIS (p = 0.008). Median TOT and PFS were similar among race/ethnicities. 48.3% of AAPI developed irAE versus 27.3% for AA, 27.0% for HIS, and 39.4% for NHW. Landmark analysis showed that AAPI had longer median OS 60.6 months (95% CI 20.6-NR), while HIS had shorter median OS 15.4 months (95% CI 8.9-NR; p < 0.01). Among patients who experienced irAEs, median OS for AAPI was 60.6 months (95% CI 20.6-NR), compared to NHW at 48.4 months (95% CI 20.5-NR), AA at 35.7 months (95% CI 2.2-NR), and HIS at 15.2 months (95% CI 6.3-31.0; p = 0.0211). There was no difference in median OS among patients who did not experience irAEs. Multivariate analysis for OS in stage IV patients showed that HIS versus AAPI (hazard ratio: 2.27; 95% CI 1.30-3.95) was a significant variable. Our study demonstrates that AAPIs and NHW had higher irAE incidence and longer OS. Validation studies evaluating the role of race/ethnicity in ICI response and toxicity are merited.

The impact of red cell distribution width (RDW) on outcomes in elderly patients after surgery for non-small cell lung cancer (NSCLC) is not fully understood. We retrospectively analyzed 992 NSCLC patients who underwent curative resection between 1998 and 2012. The following variables were included in the analyses to evaluate the role of RDW: age, gender, smoking index, leukocyte count, neutrophil to lymphocyte ratio (NLR), hemoglobin, platelet count, albumin, C-reactive protein, carcinoembryonic antigen, respiratory function, histology, T factor, N factor, surgical approach, surgical procedures, complications and prognosis. High RDW (>13.8) was an independent risk factor for morbidity [hazard ratio (HR) 2.1; P<0.01], recurrence (HR 2.0; P=0.01), overall survival (OS) (HR 2.1; P<0.01) and disease-free survival (DFS) (HR 2.0; P<0.01) in elderly patients (age ≥75 years, n=275), whereas it was not in younger patients (age <75 years, n=717). The surgical outcome was extremely poor in those older than 80 years with a RDW greater than 15% (morbidity, 56%; postoperative stay, 23 days; OS, 24%; DFS, 0%). RDW was unaffected by age (R =0.01; P=0.86) and elevated RDW without anemia was more prognostic than high RDW due to anemia in elderly patients. High RDW was significantly associated with high morbidity and reduced survival in elderly patients who underwent resection for NSCLC. Therefore, this parameter should be taken into account when surgery is considered in the elderly.

172 Background: Two revolutions recently occur in the treatment of advanced NSCLC: the development of targeted therapies and of ICI). Both strategies have been shown to outperform chemotherapy. Patients with molecular alterations are usually considered as poor candidate for immunotherapy. Here, we aimed to analyze the efficacy of immunotherapy in NSCLC patients with oncogenic addiction. Methods: We conducted a retrospective multicentric study, on patients treated with ICI and carrying an activating molecular abnormality: EGFR, ALK, KRAS, ROS1, HER2, BRAF, MET and RET. We have collected anonymized data, which were evaluated for clinical characteristics and outcome (progression free survival duration of ICI treatment and overall survival since initiation of ICI). Results: 209 patients were registered from 7 centers in France and Switzerland. 198 patients had adenocarcinoma (95.2%), 6 large cell carcinoma (2.9%), 161 were former or current smokers (81.3%), 99 were female (47.4%), median age at diagnosis was 59 yrs. (range 30-79). 33 patients had EGFR mutations (15.8%), 132 KRAS mutation (63.2%), 6 ALK rearrangement (2.9 %), 2 ROS1 rearrangement (1 %), 5 HER2 mutation (2.4 %), 12 BRAF mutations (5.7 %), 7 MET alteration (3.3 %), 3 RET rearrangements (1.4 %), and 9 had concomitant multiple molecular alterations (4.3%). After validated treatment 200 patients were treated with PD1 inhibitors (nivolumab, 97.5% and pembrolizumab, 2.5%), 4 with anti CTLA4 (tremelimumab), 5 with anti PDL1 (4 atezolizumab). The median PFS was 2.8 m. [95%CI 2.3;3.5] for the whole population 2.1 for EGFR [1.7;2.8], 3.5 for KRAS [2.5;4.9], 2.7 for BRAF [1.5;NR] and not estimable for other alterations. Median exposure to ICI was 1.89 m. [0;18.5]. The median overall survival for the whole population was 13.0 m. [9.4; 15.6], for EGFR 13.3 m. [4.7; NR], KRAS 11.3 m. [8.2; 16], BRAF 10.7 m. [1.5; NR] and not yet estimable for the other alterations. Conclusions: PFS (2.8 m.) and OS (13 m.) are very similar from the ones observed in pretreated unselected NSCLC patients. ICI is associated with better PFS in patients with KRAS mutation than in EGFR mutated patients. Analysis of outcome in other molecular subgroups is ongoing.

Serum markers associated with treatment response and survival in non-small cell lung cancer patients treated with anti-PD-1 therapy

e21524 Background: In metastatic NSCLC, identification of biomarkers of response to immune checkpoint inhibitors (ICI) is still an unmet need. Liquid biopsy, including the detection of circulating tumor DNA (ctDNA) in patients’ bloodstream, is an emerging tool for detecting and monitoring genetic alterations. Mutations in specific genes as KRAS or STK11, in combination with PDL1, have been suggested as potential biomarkers to select responders to ICI. Methods: Digital NGS of ctDNA was analyzed by Guardant360 in samples from 89 patients with advanced NSCLC treated with ICI between 2017 to 2019. Plasma samples after progression to ICI were obteined from 40 patients and we focus in this subgroup to explore mutations of genes commonly alterated in lung cancer and correlate them with clinical outcomes. Results: Forty patients were included in the analysis with a median age of 64.5 years and 62.5% were diagnosed with adenocarcinoma. Fourteen patients were treated with single agent ICI as first line, 19 patients as second line and 7 in third line. A median of 4 mutations per patient were detected (range 0-13). Mutations (mut) in TP53 were found in 85% of the patients, KRAS in 34%, STK11 in 22% and both genes KRAS and STK11 were alterated in 12.5% of the patients. Globally, median PFS (mPFS) of the patients treated with ICI was 7.5 months. mPFS in patients harboring KRAS mutations was 4.9 m ( vs 10.1 m in patients without KRASmut, logrank test p-value = 0.031) and 4.1 m in those patients with STK11 mutations (vs 9.5 m in patients without STK11mut, logrank test p-value = 0.034). Additionally, mutations in both genes were also associated to statistically significant shorter mPFS 3.9 m vs 9.5 m (logrank test p-value &lt; 0.01). Conclusions: Our results support the idea that tumor alterations identified by liquid biopsy panels can potentially contribute as biomarkers of response to ICI in NSCLC. When analyzing the mutational panorama after ICI treatment, we find that those patients exhibiting KRAS and/or STK11 mutations in their ctDNA have a shorter PFS.

MA03.06 ctDNA Mass-Adjusted bTMB as a Predictive Biomarker in NSCLC Patients Receiving PD-(L)1 Inhibitors

In recent years, immune checkpoint inhibitor (ICI) therapy has revolutionized the treatment of patients with advanced-stage non-small cell lung cancer (NSCLC). The relationship between TP53 mutation and prognosis of non-small cell lung cancer (NSCLC) remains controversial. We aimed to identify advanced-stage NSCLC patients harboring TP53 mutation who would benefit from ICI treatment. Gene mutations and tumor mutational burden (TMB) data of NSCLC patients who received at least one dose of ICI therapy at the Memorial Sloan Kettering Cancer Center between 2013 and 2017 were extracted from the cBioPortal online platform. Gene clustering analyses were performed for patients with short and long overall survival (OS). The top ten significantly different mutated genes were identified. Furthermore, we analyzed the different OS of coexisting TP53 and other significantly different mutated genes to identify NSCLC patients with TP53 mutations who would benefit from immunotherapy. A total of 350 patients were enrolled in the study. Of these a total of 219 (62.6%) patients were found to harbor TP53 mutations, whereas 131 (37.4%) had wild-type TP53. There was no statistically significant difference in OS between TP53 mutated or wild-type NSCLC patients who underwent ICI treatment. However, coexisting TP53 and ZFHX3 mutations were independent prognostic factors. Higher somatic TMB (highest 20% in each histology) and combination of anti-CTLA-4 and anti-PD-1/PD-L1 therapy were also associated with longer OS in multivariate analysis. Coexisting TP53 and ZFHX3 mutations are independent prognostic factors for advanced-stage NSCLC patients undergoing ICI treatment. These findings could help identify patients harboring TP53 mutations that would benefit from ICI treatment.

BackgroundThe aim of the study was to explore associations between elevated red cell distribution width (RDW) and acute kidney injury (AKI) in patients undergoing cardiac surgery (CS-AKI).MethodsPreoperative, intraoperative and postoperative data of 10,274 patients undergoing cardiac surgery, including demographic data, were prospectively collected from January 2009 to December 2014. Propensity score matching was used on the basis of clinical characteristics and preoperative variables. An elevated RDW was defined as the difference between RDW 24 h after cardiac surgery and the latest RDW before cardiac surgery.ResultsA total of 10,274 patients were included in the unmatched cohort, and 3146 patients in the propensity-matched cohort. In the unmatched cohort, the overall CS-AKI incidence was 32.8% (n = 3365) with a hospital mortality of 5.5% (n = 185). In the propensity-matched cohort, the elevated RDW in AKI patients was higher than in patients without AKI (0.3% (0.0%, 0.7%) vs 0.5% (0.1, 1.1%), P < 0.001) and the elevated RDW incidences were 0.4% (0.1%, 0.9%), 0.6% (0.2%, 1.1%) and 1.1% (0.3%, 2.1%) in stage 1, 2 and 3 AKI patients (P < 0.001). Among propensity-matched patients with CS-AKI, the level of elevated RDW in non-survivors was higher than in survivors [1.2% (0.5%, 2.3%) vs 0.5% (0.1%, 1.0%), P < 0.001] and a 0.1% increase in elevated RDW was associated with a 0.24% higher risk of within-hospital mortality in patients with CS-AKI. Estimating the receiver-operating characteristic (ROC) area under the curve (AUC) showed that an elevated RDW had moderate discriminative power for AKI development (AUC = 0.605, 95% CI, 0.586–0.625; P < 0.001) and hospital mortality (AUC = 0.716, 95% CI, 0.640–0.764; P < 0.001) in the propensity-matched cohort.ConclusionsAn elevated RDW might be an independent prognostic factor for the severity and poor prognosis of CS-AKI.

The irruption of the immunotherapy for the treatment of non-small cell lung cancer (NSCLC) based on Immune Checkpoint Inhibitors (ICI) PD-1 and PD-L1 inhibitors is considered a treatment revolution. However, only a small percentage of patients benefit with ICI treatment over the long term, and precise biomarkers that can recognize these individuals before or early during treatment have so far eluded. PD-L1 expression and tumor mutational burden (TMB) are the most well studied biomarkers for predicting response to PD- (L)1 blockade-based ICI prior to treatment. TMB is still being clinically assessed whereas PD-L1 has several drawbacks for prediction of persistent benefit. Our aim is to evaluate the basal state and dynamic changes of cell-free DNA (cfDNA) concentration to predict and monitor response in NSCLC patients starting ICI. A total of 248 cfDNA concentration measurements were performed from 87 NSCLC patients. The quantification was done before the start of the treatment, at the second ICI cycle, after 6 and 12 months in treatment, and at progression if it was within the first 12 months. Quality and quantity of the cfDNA was assessed using Qubit High Sensitivity and Bioanalyzer 2100. We first explored the association with response of the basal cell-free DNA concentration using Mann-Whitney-Wilcoxon test. The longitudinal analysis between different time points was tested with Wilcoxon signed-rank tests. Response was ascertained using RECIST parameters at 3, 6, and 12 months. We evaluated the utility of the cfDNA concentration as a prognostic factor using Mantel-Cox test. The response association results indicate that early and long-term response is associated with lower levels of basal cfDNA (p&amp;lt;0.002, 3 months; p&amp;lt;0.001, 6 months; p&amp;lt;0.05, 12 months). The longitudinal analysis could also detect a decrease in cfDNA from the second cycle to the month 6 of treatment in responders (p&amp;lt;0.05), and an increase in non-responders (p&amp;lt;0.05). To identify a prognostic concentration threshold, statistical stratification of cfDNA levels based on the selected maximum rank method of cfDNA levels was used to group them into low and high concentration. Patients with high basal cfDNA concentrations showed both worse FPS (log-rank 3.4x10-03, Medians: Low 12.57 [8.36 - 18.46] High 5.50 [3.03 - 10.06]) and OS (log-rank 3.2x10-05, Medians: Low 21.26 [15.33 - 33.96] High 5.45 [3.10 - 15.66]). With these results, we provide proof-of-principle that specifically detecting tumor cfDNA by PCR or NGS methods would not be necessary to predict the long-term clinical benefit to ICI in NSCLC patients. Moreover, concentration cut-offs can be defined to categorize the patients in potential responders and non-responders before the start of the treatment. This would constitute a cost-effective and easily implementable prognostic tool that could be particularly useful for avoiding loss of treatment windows of opportunities. Citation Format: Isabel Barragan, Elisabeth Perez-Ruiz, Juan Luis Onieva, Maria Garrido-Ramos, Beatriz Martinez-Galvez, Jaime Dubbelman, Emilio Alba, Juan Zafra, Manuel Cobo, Javier Oliver, Antonio Rueda-Dominguez. Cell-free DNA concentration as a prognostic biomarker in patients with non-small cell lung cancer under immunotherapy treatment [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 1044.

Background: Our previous prospective study demonstrated that non-small cell lung cancer (NSCLC) patients (pts) with pre-sarcopenia or sarcopenia exhibited poor responses to immune checkpoint inhibitors (ICIs). In this study, we evaluated the relationship between immune status and the efficacy of ICIs in NSCLC pts with sarcopenia. Method: Pts with stage IV or recurrent NSCLC receiving ICI as first-line treatment at Niigata University Medical and Dental Hospital were eligible. Sarcopenia and pre-sarcopenia were diagnosed with physical performance and skeletal muscle mass. Before ICI treatments, peripheral mononuclear blood cells (PBMCs) were collected, and immune phenotypic data was obtained by flow cytometry. Result: PBMCs were obtained from 34 pts from December 2021 through September 2024. Six patients and 7 patients were diagnosed with sarcopenia and pre-sarcopenia, respectively. Pts with a higher percentage of Lineage-CD11b+CD14+CD33+DRlow monocytic myeloid-derived suppressor cells (M-MDSCs) or Lineage-CD11b+CD15+CD33+DRlow polymorphonuclear (PMN)-MDSCs significantly shorter OS (median, 12 months (95% CI, 2.9-26.5) vs 33.1 months (95% CI, 11.4-NR); HR 2.7; P=0.035) after ICI treatments. Further, pts with PD-L1+ cells in MDSCs also had poor OS (median, 12 months (95% CI, 2.9-26.5) vs 33.1 months (95% CI, 11.4-NR); HR 2.6; P=0.0387). The higher percentage of M-MDSC (P=0.0534) and the presence of PD-L1+ cells in MDSCs (P=0.0473) tended to be associated with pts with sarcopenia or pre-sarcopenia. Conclusion: MDSCs may play a critical role in immune suppression in NSCLC pts with sarcopenia and could be associated with poor OS following ICI treatment. To improve the survival outcome of pts with sarcopenia, it is essential to establish strategies to overcome MDSC-mediated immune suppression. Citation Format: Ryo Yamazaki, Satoshi Watanabe, Takaaki Masuda, Tomoya Wakabayashi, Yumi Ando, Kunihiro Shono, Naohiro Yanagimura, Masashi Arita, Miyuki Sato, Tomohiro Tanaka, Koichiro Nozaki, Yu Saida, Kenjiro Shima, Yosuke Kimura, Nobumasa Aoki, Yasuyoshi Ohshima, Toshiyuki Koya, Toshiaki Kikuchi. The correlation between immune status and the therapeutic efficacy of immune checkpoint inhibitors in patients with non-small cell lung cancer with sarcopenia [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 7127.

8635 Background: The combination of immune checkpoint inhibitors (ICIs) and chemotherapy has shown high efficacy in non-small cell lung cancer (NSCLC) patients, but further exploration of the predictive biomarkers is needed. KRAS is the most common driver oncogene in patients with NSCLC. However, the influence of different KRAS mutation subtypes on immunotherapy response remains controversial. We aimed to evaluate the value of KRAS mutation subtypes for predicting response to immunotherapy in NSCLC patients and to further analyze the related tumor immune microenvironment (TiME) phenotypes. Methods: In this retrospective study, clinical data were extracted from the medical records of patients with advanced NSCLC treated with ICIs and whose available molecular analysis included KRAS mutation information between January 2017 and December 2022. NSCLC tissue samples were collected for TiME analysis. Differences in immune cell subsets between different groups were analyzed using the Mann‒Whitney U test. Clinical characteristics, ICI treatment efficacy, progression-free survival (PFS), and overall survival (OS) were analyzed using multivariate Cox models. Results: A total of 1913 eligible KRAS-mutated NSCLC patients were screened, and 137 patients who received systemic immunotherapy were enrolled (108 males vs 29 females), with a median age of 62 (range, 39-84) years. Among them, 126 (92.0%) patients had adenocarcinoma, 99 (72.3%) patients were systemic therapy naïve, and 124 (90.5%) patients received combined immunotherapy. Among the patients with KRAS mutations, 48 (35.0%) had G12C, 20 (14.6%) had G12V, 27 (19.7%) had G12D and 42 (30.7%) had other subtypes. PFS was significantly longer in the G12C subgroup than in the other mutation subgroup (median PFS: 11.3 vs 6.8 months, P = 0.001). The median PFS times in the G12C/G12V subgroup and the other groups were 11.1 months and 6.5 months, respectively ( P= 0.0005). Multivariate analysis of the immunotherapy cohort revealed that G12C/G12V was an independent predictor of PFS ( P = 0.003, hazard ratio (HR): 2.021, 95% CI: 1.265-3.227). Further TiME analysis suggested that the infiltration of myeloid-derived suppressor cells (MDSCs, CD33+CD11b+) in the tumor and stromal areas in the G12C/G12V subgroup was significantly lower than that in the other subgroup (tumor area, P= 0.045; stroma area, P= 0.015). However, no significant difference in PD-L1 expression was found between the G12C/G12V group and the other groups. Conclusions: KRAS G12C and G12V were associated with favorable combined ICI treatment efficacy in patients with NSCLC. Patients with G12C or G12V mutations had longer PFS and lower infiltration of MDSCs. KRAS mutation subtypes are potential predictive biomarkers disturbed by MDSCs for combination therapy in NSCLC patients.