8011 Background: Outcomes for pts with ES-SCLC remains suboptimal despite the addition of ICIs to platinum-based chemotherapy. Somatostatin receptor 2 (SSTR2) is expressed in the majority of SCLC lesions. SSTR2 targeted radiopharmaceutical therapy (RPT) has proven safe and effective when used in other neuroendocrine tumors. The multicentre, open-label, phase I LuSato-1 (ACTRN12623000185662) study investigated the use of 177 Lu-satoreotide Tetraxetan ( 177 Lu-SSO110) with 68 Ga-Satoreotide Trizoxetan ( 68 Ga-SSO120) companion imaging in pts with ES-SCLC, who are on 1L maintenance ICIs. Methods: Adult pts with ES-SCLC received 1L induction therapy with carboplatin, etoposide and atezolizumab for 4 cycles. Following induction, eligible pts without PD were enrolled and received up to 4 doses of 177 Lu-SSO110 in 6 to 9 week intervals, with up to 3 additional treatments per investigator choice. Activities were escalated from 3.7GBq to 5.2GBq, following a BOIN design with a target DLT rate of 0.30. The primary objective was to investigate the safety and tolerability of 177 Lu-SSO110 with 68 Ga-SSO120 companion imaging in pts on maintenance ICI, with secondary objective of anti-tumor activity of the combination. Concordance between 68 Ga-SSO120 PET/CT and contrast-enhanced CT was an exploratory endpoint. Results: Of the 35 screened pts, 20 were deemed eligible (median age 65.5 years (range 47-83), 55% female, 25% brain mets). With a median 11.1 month follow up (data cut off: 08/12/25), the TEAE rate associated with 177 Lu-SSO110 was 0% at 3.7GBq (dose level 1), 71.4% at 4.5GBq (dose level 2), and 83.3% at 5.2GBq (dose level 3). The grade 3/4 TEAE rate associated with 177 Lu-SSO110 was 0% (dose level 1), 57.1% (dose level 2), and 41.7% (dose level 3). Grade 2 to 4 thrombocytopenia rate related to 177 Lu-SSO110 was 45%. Grade 3/4 irAEs related to ICI was 20%. There was 10% pneumonitis rate attributed to ICI. The treatment interruption rate was 50% and treatment discontinuation rate was 10% (2 attributed to 177 Lu-SSO110 associated thrombocytopenia). From the start of C1 induction chemoICI, the ORR was 85%, including 1 CR. From the start of maintenance ICI, median PFS was 3.7 months and median OS was 8 months. Conclusions: Addition of SSTR2 targeted RPT to maintenance ICI was well tolerated and demonstrated clinically meaningful anti-tumor activity. This supports further investigation of SSTR2-targeted RPT in combination with ICI in patients with ES-SCLC. Clinical trial information: ACTRN12623000185662. ORR, PFS and OS. Efficacy ES-SCLC patients N=20 Best response rates (from C1 induction) CR 1 (5%) PR 16 (80%) SD 3 (15%) ORR (from C1 induction) 17 (85%) Median PFS (months) (from C1 maintenance) 3.7 months Median OS (months) (from C1 maintenance) 8 months

8006 Background: Brain metastases (BM) are common in patients (pts) with SCLC and are associated with poor outcomes. Tarlatamab, a bispecific T-cell engager (BiTE) immunotherapy, demonstrated superior overall survival versus CTx in pts with SCLC following progression on or after platinum-based CTx in the DeLLphi-304 study, including pts with history of BM (prior or current) at baseline (OS HR 0.45 [95% CI: 0.31–0.65]). Here we compare the intracranial efficacy of tarlatamab vs CTx. Methods: Pts were randomized 1:1 to receive tarlatamab or CTx (topotecan, lurbinectedin or amrubicin) as 2L treatment for SCLC. Pts with stable asymptomatic brain metastases were eligible; prior CNS treatment was required until protocol amendment 3. Baseline brain imaging by contrast enhanced MRI was mandatory for all pts at screening and repeated at all subsequent imaging assessments for pts with a history of BM at baseline. A post hoc analysis on intracranial efficacy was performed by BICR per mRANO-BM. Given that most pts had prior CNS treatment, specified outcomes were CR, non-CR/non-PD, and PD. Results: BM at baseline were present in 98/254 pts (39%) in the tarlatamab arm and 99/255 (39%) in the CTx arm, of whom 75/98 (77%) and 69/99 (70%) had prior CNS treatment, respectively. A CNS full analysis set (FAS) was specified to include pts who had both a baseline scan and ≥ 1 postbaseline scan (tarlatamab-67; CTx-56 pts). In pts in FAS, treatment with tarlatamab resulted in longer CNS PFS than CTx (median: 6.5 mos vs 4.2 mos; HR, 0.40 [95% CI: 0.24–0.66]; Table). CNS tumor shrinkage of ≥30% was observed in 56% of pts with tarlatamab vs 38% with CTx. CNS complete response was observed in 15% of pts with tarlatamab vs 5% with CTx, with longer CNS duration of complete response (DOCR) (not estimable [NE] vs 3.6 mo) and longer CNS duration of disease control (DODC) (8.2 vs 5.2 mo) for pts in the tarlatamab arm. Pts with BM at baseline had longer OS with tarlatamab vs CTx (median OS: 13.9 vs 6.8 mos; HR, 0.51 [95% CI: 0.34–0.74]). In pts with BM at baseline, treatment-emergent adverse events (TEAEs) of any grade (gr)/gr 3/gr 4/gr 5 occurred in 99%/38%/9%/7% for tarlatamab vs 100%/38%/40%/10% for CTx. In pts treated with tarlatamab, the incidence of CRS and ICANS was 54% and 9% in pts with BM at baseline vs 58% and 4% in pts without BM at baseline, respectively. Conclusions: Tarlatamab demonstrated increased intracranial efficacy with longer CNS PFS and OS vs CTx in pts with stable, treated and untreated asymptomatic BM. These results affirm tarlatamab as the 2L standard of care for SCLC, even in pts with BM. Clinical trial information: NCT05740566 . Tarlatamabn = 67 CTxn = 56 CNS PFS, mos (95% CI) 6.5 (4.3, 13.7) 4.2 (2.9, 5.5) CNS Complete Response, n (%) 10 (15%) 3 (5%) CNS DOCR, mos (95% CI) NE (2.9, NE) 3.6 (3.1, NE) CNS DODC, mos (95% CI) 8.2 (6.3, NE) 5.2 (4.2, 6.2) CNS tumor shrinkage of ≥30%, % (n/N) a 56% (9/16) 38% (5/13) a Assessed in pts with ≥1 lesion that was ≥ 10 mm.

INTRODUCTION:Novel agents that improve outcomes in NSCLC are needed. Co-inhibition of T-cell immunoreceptor with immunoglobulin and immunoreceptor tyrosine-based inhibitor motif domains and programmed cell death protein 1 (PD-1) may enhance anti-PD-1 activity. AdvanTIG-302 is a phase 3 trial of ociperlimab plus tislelizumab (A), pembrolizumab (B), and tislelizumab (C) in first-line stage III/IV NSCLC with programmed cell death-ligand 1 expression of 50% or more (NCT04746924). METHODS:Eligible patients were randomized (5:5:2) to A, B, or C. The primary objective was overall survival (OS) in A versus B; key secondary analyses were progression-free survival (PFS) and overall response rate (ORR). RESULTS:As of May 30, 2025, the prespecified interim analysis (N = 662) was terminated due to futility; efficacy analyses are descriptive only. Median OS (95% confidence interval [CI]) was 31.9 (25.7-not estimable [NE]), 29.4 (25.8-35.0), and 27.7 (20.0-NE) months for arms A, B, and C, respectively; the stratified hazard ratio for A versus B was 0.97 (95% CI: 0.76-1.23). For A, B, and C, median PFS (95% CI) was 14.3 (11.5-16.0), 10.5 (8.4-12.6), and 16.6 (8.9-26.3) months and ORR (95% CI) was 61.0% (55.1-66.7), 48.8% (42.9-54.7), and 55.7% (44.7-66.3), respectively. In arms A, B, and C, treatment-related adverse events occurred in 84.3%, 79.4%, and 79.3% of patients, respectively. CONCLUSION:Ociperlimab plus tislelizumab had no improvement in OS compared with pembrolizumab; median OS was comparable across arms. Data for B versus C should be interpreted cautiously given the descriptive nature of this comparison. The safety profiles of all treatment arms were well tolerated, with no new safety signals.

BACKGROUND:Extensive-stage small cell lung cancer (ES-SCLC) is associated with a high symptom burden and impaired health-related quality of life (HRQoL). This prespecified analysis from the phase 3 DeLLphi-304 trial evaluated patient-reported outcomes (PROs) for tarlatamab versus standard-of-care (SoC) chemotherapy following first-line platinum-based therapy. METHODS:DeLLphi-304 is a multicenter, open-label, randomized phase 3 study in adults with ES-SCLC. PROs were assessed using validated instruments, including the EORTC QLQ-C30, EORTC QLQ-LC13, FACT-G GP5, BPI-SF, and the EQ-5D-5L visual analogue scale. Change from baseline, response rates, and time to deterioration in these PROs were analyzed. RESULTS:PRO data from all 509 patients enrolled were evaluated. Compliance with QLQ-C30 and QLQ-LC13 assessments remained above 69% through 19 weeks. A higher proportion of patients receiving tarlatamab achieved symptom or functional improvement at 19 weeks compared with SoC in chest pain (19% vs 10%), cough (35% vs 26%), dyspnea (22% vs 7%), physical functioning (13% vs 8%), and global health status (23% vs 15%), respectively. Tarlatamab also delayed deterioration in symptoms, physical functioning, and pain at worst relative to SoC. FACT-G GP5 results indicated that patients receiving tarlatamab were less bothered by treatment side effects over time. CONCLUSIONS:In addition to its previously reported antitumor activity, tarlatamab demonstrated clinically meaningful improvements in symptoms and HRQoL compared with SoC. These findings support a favorable benefit-risk profile of tarlatamab in patients previously treated for ES-SCLC.

BACKGROUND:Targeted therapies have improved outcomes in ROS1-rearranged NSCLC "ROS1", but real-world evidence on evolving treatment sequencing and survival to understanding this rare cancer remains sparse. METHODS:ROS1 cases identified from the Australian multicentre AURORA cohort between 2012 and 2025 were analysed. Demographic, diagnostic, and treatment data were extracted, with systemic therapies mapped to report sequencing and discontinuation. Progression-free survival (PFS) and overall survival (OS) were estimated using Kaplan-Meier methods. Multivariable Cox models assessed clinical correlates of outcome. RESULTS:Among 5,189 NSCLC cases within AURORA, 115 (2.2 %) were ROS1 positive. Median age was 58 years; 64 % female; 85 % had de novo advanced disease, 14 % baseline brain metastases. Amongst n = 16 with early-stage, 59 % recurred. In total n = 104/115 (90 %) were treated for advanced disease, median 2 lines; range 1-8, 32 % did not receive a second line. First-line ROS1-inhibitor therapy (ROS1i) was given to 74 % (n = 77), including early-generation (n = 63; crizotinib/entrectinib) and later-generation ROS1i's (n = 14; repotrectinib/zidesamtinib/lorlatinib). Across all lines, 96 % received a ROS1i including 63 % with later-generation, and 53 % participated in a clinical trial. Median PFS with first line early-generation ROS1i was 17 months(mo), 48mo for first line later-generation ROS1i (p = 0.071). Median OS was 56mo overall, 80mo with first line later-generation ROS1i. Median OS was 26mo in those with brain metastases at diagnosis versus 58mo without (p = 0.010) and 41mo in those with PD-L1 ≥ 50 % versus 62mo 0-49 % (p = 0.017). CONCLUSION:This multicentre real-world cohort describes longitudinal ROS1 management with evolving treatments. Favourable survival likely reflects reflex molecular testing, access to ROS1i, and high clinical trial enrolment.

1661 Background: Integration of geriatric assessment into cancer care improves outcomes for older adults but remains limited by resource intensity and reliance on physician time, hence restricting scalability. We evaluated the implementation of a co-created multidisciplinary (MDT) nurse-led geriatric oncology (GO) model developed to address these barriers. Methods: Implementation evaluation was conducted across two tertiary cancer centers. Older adults (≥65 years) with solid cancers underwent nurse-led structured geriatric assessment and MDT review involving medical oncology and geriatric medicine prior to oncology consultation. MDT recommendations addressed identified vulnerabilities, with referral for geriatrician-led CGA reserved for complex cases. Outcomes included G8 screening, referral for CGA, SACT delivery, treatment modification due to frailty, treatment discontinuation, delays, dose modification due to toxicity, and health service utilization (ED, SURC and unplanned hospital admissions within 12 weeks). Demographic and clinical characteristics were summarized using descriptive statistics. Associations with health service use were determined using logistic regression analyses. Results: Of 431 eligible pts, 204 (47.3%) participated in the GO assessment. Median age was 76 years. Most pts had lung cancer (73.1%), followed by GU (14.8%) and upper GI cancers (8.6%). Disease stage did not differ between groups; GO pts were more likely to have ECOG PS 1–2 and 81.4% scored ≤14 on the G8 screening tool. Half (50.7%) were referred for CGA and 67.1% of this cohort received SACT while 32.4% of the cohort not referred for a CGA received SACT. Upfront treatment modification due to frailty was numerically higher in GO pts compared to non-participants (21.1% vs 13.1%, p=0.076). GO participation was associated with higher receipt of SACT (70.6% vs 59.7%, p=0.019), increased use of monotherapy (22.5% vs 13.7%, p=0.016), and lower rates of BSC alone (41.0% vs 59.0%, p=0.033). There were no differences between groups in treatment discontinuation (p=0.64), delays (p=0.074), or dose modifications due to toxicity (p=0.42). GO pts were more likely to be admitted to hospital within 12 weeks of initial oncology consultation. Conclusions: GO assessment supported detailed evaluation of older pts with borderline ECOG PS and increased access to SACT, particularly monotherapy, without increasing treatment-related toxicity. Higher hospitalization rates likely reflect appropriate escalation of care following identification of frailty-related needs. GO vs no GO. GO (n=204) No GO (n=227) p-value Age median (range) 76.0 (72.0, 81.0) 76.0 (71.0, 81.0) 0.77 Monotherapy 46 (22.5%) 31 (13.7%) 0.016 Doublet, or more therapy 98 (48%) 101 (44.5%) 0.46 ED presentation 93 (45.6%) 86 (37.9%) 0.11 SURC use 43 (21.1%) 49 (21.6%) 0.90 Inpatient admission 74 (36.3%) 62 (27.3%) 0.046

BACKGROUND:The inflammasome is a critical immunological sensor comprised of NLRP-3, ASC, and CASPASE-1. Mutations in NLRP-3 are prevalent in inflammatory diseases. However, the role of NLRP-3 in cancer is controversial. This study investigates whether NLRP-3 expression is associated with clinical outcomes in patients with solid cancers. METHODS:PubMed (MEDLINE), Embase, Cochrane, and Google Scholar were searched for articles reporting NLRP-3 expression and disease outcome data in cancer patients. RevMan Review Manager was used to calculate pooled hazard ratios and Mantel-Haenszel pooled odds ratios. RNA sequencing datasets from the TCGA Pan-Cancer (PANCAN) were used for external validation. RESULTS:Patients with higher NLRP-3 expression showed a significant association with larger tumor size, advanced tumor grade, TNM stage, and presence of metastasis. High NLRP-3 expression has a significant association with poor OS (HR:2.12, 95% CI = 1.49-3.03), p < 0.0001) and DFS (HR:1.86, 95% CI = 1.30- 2.65, p = 0.0007). Subgroup analysis showed that higher NLRP-3 expression is associated with worse OS in head and neck cancer (HR: 2.77, 95% CI = 1.88-4.09, p < 0.00001), colorectal cancers (HR:2.14, 95% CI= 1.59- 2.87, p < 0.00001), and pancreatic cancer patients (HR: 3.19, 95% CI = 1.73-5.91, p = 0.0002). CONCLUSION:High NLRP-3 expression is associated with advanced disease and poor outcomes in many solid tumours.

Supplementary Table 4: Systemic therapy received by each patient prior to recruitment into the trial.

Supplementary Figure 5. Functional markers expressed in CD8+ T cells in gPFS patients

Supplementary Table 5. CAVATAK samples: time between biopsies and biopsy sites.

INTRODUCTION:ALK-rearranged NSCLC is a distinct molecular subtype that disproportionally affects younger never-smoking females, which overlaps with the reproductive age group. Alectinib is a second-generation ALK inhibitor and is generally considered a safer option during pregnancy compared with lorlatinib. Here, we describe a case of a patient who became pregnant while on alectinib complicated by rapid disease progression, with a focus on maternal-fetal outcomes and placental health. METHODS:A brief report was conducted detailing clinical management by a multidisciplinary team, pharmacokinetic analysis of alectinib levels, and post-delivery placental examination, including histopathology and mitochondrial respirometry function assessment as an indicator of placental health. RESULTS:A 29-year-old woman with metastatic ALK-rearranged NSCLC became pregnant while on alectinib therapy. At 24 weeks of gestation, she experienced rapid systemic disease progression manifesting as marantic endocarditis with multi-territory small cerebral infarcts, thrombocytopenia, an elevated D-dimer level, and development of two large middle cerebral ischemic strokes requiring endovascular clot retrieval. Carboplatin and paclitaxel chemotherapy was added to alectinib. The patient responded well to the treatment and successfully delivered an infant through elective caesarean section at 34+5 weeks. The patient was switched to lorlatinib plus pemetrexed postpartum. Although she achieved a complete metabolic response with lorlatinib plus pemetrexed, disease progression occurred after several months, and she died 28 months after initial diagnosis. Placental analysis revealed features indicative of fetal growth restriction. Mitochondrial function appeared within normal limits, despite changes in placental morphology. The infant has shown normal development at 9 months of age. CONCLUSIONS:This is the first reported case of concurrent alectinib and cytotoxic chemotherapy during pregnancy, successfully used in the setting of progressive ALK-rearranged NSCLC. The placental findings align with observations in pregnancies complicated by fetal growth restriction. This case highlights the potential feasibility and safety of intensification of systemic therapy during pregnancy and underscores the importance of individualized multidisciplinary care.

Basic and translational research in lung cancer is a rapidly evolving field with a transformational impact on early detection, diagnosis, therapeutic development, and personalization of care. Recent advances have greatly increased our understanding of the molecular genomics, proteomics, pathogenesis, and cellular biology of this deadly malignancy. The International Association for the Study of Lung Cancer (IASLC) recently formed a Basic and Translational Science (BaTS) Committee to further enhance the scientific leadership of IASLC in thoracic cancer research. This review by members of the committee highlights the breadth of current research in NSCLC, with a focus on molecular risk factors and processes in tumorigenesis, heterogeneity, phenotypic plasticity, metabolic reprogramming, immunobiology, the immune microenvironment, and microbiome. This review also identifies future research areas that may lead to further improvement in survival outcomes and curative therapies especially for patients with advanced NSCLC.

PURPOSE:Acquired or de novo resistance to immune checkpoint inhibitors occurs in the majority of advanced non-small cell lung cancers. There is an unmet need to improve outcomes for patients with this condition. Oncolytic viruses represent an attractive treatment approach because of their dual activity in inducing tumor cell lysis directly and potentially augmenting antitumor immunity. In this study, we present the safety, efficacy, and translational findings from a phase I/II single-arm trial utilizing CVA21, an oncolytic coxsackievirus, in combination with pembrolizumab in patients with advanced pretreated non-small cell lung cancers. PATIENTS AND METHODS:We performed paired pre- and posttreatment biopsies in 10 patients who received intravenous CVA21 and pembrolizumab, eight of whom had prior treatment with immune checkpoint inhibitor therapy. Whole-genome sequencing and spatial proteomics were performed to comprehensively characterize the response to CVA21. RESULTS:Combination CVA21/pembrolizumab (anti-PD-1) therapy was well tolerated with no serious treatment-related adverse events. Partial responses were seen in two patients with prior acquired anti-PD-1 resistance and disease stabilization in six patients, giving a clinical benefit rate of 80%. High baseline tumor mutational burden and PD-L1 expression were observed in patients with better response to treatment. Interestingly, an increase in antigen presentation and CD8+ T-cell infiltration was observed on-treatment compared with baseline in patients with better progression-free survival. CONCLUSIONS:This study demonstrates the potential of CVA21 to modulate the immunogenicity of tumor cells and remodel the tumor microenvironment, providing insights for patient selection for trials involving novel immunotherapeutic approaches.

Supplementary Figure 2. Tumor cell characteristics at baseline and on treatment in gPFS and loPFS patients

Supplementary Figure 1. Survival analysis.

Most mesothelioma patients either have primary refractory disease or develop acquired resistance to immune checkpoint inhibitors (ICIs). A more sophisticated understanding of the tissue microenvironment (TME) and intercellular communication will enable opportunities for overcoming ICI resistance. Here, we utilize the Visium 10x spatial transcriptomics (ST) technology to address this question. We conducted ST on a cohort of 12 mesothelioma patients (28 samples) treated with ICIs (Table 1). Nine patients had a response to ICIs with a median response duration of 9 months and three had primary ICI resistant disease. We show the transition from the immune infiltrated tumor in a pre-treatment sample transition to immune exclusion in the post-treatment (ICI-resistant) sample. In addition, we were able to map the immune cell sub-types in relation to cancer cells. In a responder, we identified CXCL12-CXCR4 mediated communication at the mesothelial-immune interface. We mapped a 48-gene prognostic signature to generate spatial heat maps showing specific regions within tissue with elevated poor prognosis signatures in treatment naïve samples that expanded in matched relapsed samples. We also identified remarkable heterogeneity showing pre- and post-ICI treatment samples to be transcriptionally diverse. In another patient, we assessed four samples from the primary tumor prior to treatment, two from PET-positive and two from PET-negative areas. Moreover, we complemented spatial transcriptomics with multi-region whole genome sequencing. This revealed that these sites are genomically stable but transcriptionally distinct. Further deconvolutional analysis to approximate cell composition, cell co-localization, and cell evolution/plasticity measures are underway. Together, this deep spatial investigation provides a key understanding of oncogenic communication in TME, before and after ICI therapy. Table 1. Patient No Sex Histological Subtype Treatment Pre-Treatment samples Post Treatment samples Response Response duration 1 M Sarcomatoid PD-1 5 - Yes 6 2 M Epithelioid PD-1 3 - No - 3 M Biphasic PD-1/CTLA-4 4 - Yes 87 4 F Biphasic PD-1/CTLA-4 1 - No - 5 M Epithelioid PD-1/CTLA-4 2 1 Yes 12 6 F Sarcomatoid PD-1 1 - Yes 52 7 M Epithelioid PD-1 1 1 Yes 17 8 M Sarcomatoid PD-1/CTLA-4 2 - Yes 75 9 M Epithelioid PD-1/CTLA-4 1 1 No - 10 M Biphasic PD-1 1 1 Yes 18 11 M Sarcomatoid PD-1/CTLA-4 1 1 Yes 22 12 M Epithelioid PD-1/CTLA-4 - 1 Yes 16 Muhammad Alamgeer, Daniel Gough, Vinvent Xue, Beena Kumar, Adrian Pick, Fathima Sharafath Yoosuf, Surein Arulananda, Jhoseph Marquez, Vinod Ganju. Illuminating mesothelioma immune microenvironment in response to immune checkpoint inhibitors using spatial transcriptomics [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 3977.

Background: KRAS G12D and G12C mutations have distinct biological traits influencing treatment response. This study examines real-world demographics, clinical characteristics, and first-line treatment outcomes in metastatic non-small-cell lung cancer (NSCLC) patients with these mutations. Methods: This retrospective, multi-institution observational study used data from the AURORA database. Patients aged 18 years or older, diagnosed with metastatic KRAS G12D or G12C NSCLC between January 1, 2010, and April 30, 2024, were included. Descriptive statistics compared patient characteristics, and time-to-event outcomes were assessed using Cox proportional hazards regression. Results: A total of 298 (216 KRAS G12C and 82 KRAS G12D) patients were included. The KRAS G12D group had a higher proportion of never smokers (15 % vs. 1 %, p < 0.01) and PD-L1 < 1 % (36 % vs. 21 %, p = 0.06). No significant differences were observed in overall survival (OS) (HR 1.09, 95 % CI 0.80-1.48, p = 0.60) or real-world progression-free survival (rwPFS) (HR 1.21, 95 % CI 0.92-1.59, p = 0.18) between mutation groups. In KRAS G12C, monotherapy immunotherapy (HR 0.61, 95 % CI 0.39-0.97, p = 0.04) and chemo-immunotherapy (HR 0.59, 95 % CI 0.37-0.94, p = 0.03) improved OS compared to chemotherapy. For KRAS G12D, neither immunotherapy (HR 0.74, 95 % CI 0.29-1.89, p = 0.53) nor chemo-immunotherapy (HR 0.73, 95 % CI 0.34-1.57, p = 0.42) improved OS compared to chemotherapy alone. Conclusion: KRAS G12C and G12D mutations demonstrate distinct clinical characteristics and treatment responses, with poorer immunotherapy outcomes in KRAS G12D patients. Prospective studies are needed to validate these findings.