Objectives 1. Describe the clinical characteristics of patients presenting with an overlap of immune-related myositis and myasthenia gravis (MG) secondary to immune checkpoint inhibitors (ICIs). 2. Evaluate the management strategies utilized to care for these patients. 3. Identify clinical and treatment-related factors associated with patient outcomes. Methods Clinical data was abstracted from medical records, including demographics, oncologic history, pre-existing autoimmune conditions, manifestations of disease, laboratory findings, treatment, and outcomes. Results 18 patients were identified with overlapping myositis and MG. 50% of MG diagnoses were supported by positive electromyography (EMG) or MRI, and 50% were diagnosed clinically by ocular or bulbar symptoms, and/or fatigable muscle weakness. Myositis was diagnosed by CK elevation, MRI or EMG, and physical findings of sustained proximal muscle weakness. All patients received initial high-dose glucocorticoids, with improvement in 12 of 18 patients. In addition to glucocorticoids, 7 received intravenous immunoglobulins, 1 plasma exchange, 8 acetylcholinesterase inhibitor therapy (pyridostigmine), and 3 disease-modifying antirheumatic drug (DMARD) therapy. DMARD therapies included mycophenolate (symptom improvement in 2 of 3 patients), infliximab (improvement in 1 of 3 patients), and rituximab (deterioration in 1 patient). 7 patients received pyridostigmine, of which 6 had symptom improvement. None required ventilatory support. 3 of 18 patients died from MG/myositis syndrome. Mortality was associated with Grade 4 symptom severity, bulbar involvement, and profound muscle weakness. The presence of MG− or myositis associated antibodies did not predict mortality. Conclusion This multi-center study characterizes a spectrum of clinical presentations among patients with overlapping immune-related myositis and MG. Mortality was associated with Grade 4 symptom manifestations, bulbar symptoms, and need for ventilatory support, highlighting the critical role of early recognition and aggressive management in severe cases.[1] All patients received high-dose glucocorticoids, with some receiving additional immunosuppressive therapies and supportive care. Addition of acetylcholinesterase inhibitor therapy improved overlap symptoms more than DMARD therapy alone. Among those with EMG-confirmed MG, the majority improved with pyridostigmine, further suggesting true neuromuscular junction pathology and arguing against the interpretation that MG diagnoses merely represent severe myositis, which reinforces the concept of distinct overlap syndrome rather than a single disease spectrum. As in previous studies, clinical outcomes were influenced by disease severity and organ involvement rather than serologic markers, as antibody positivity did not correlate with mortality.[2] These findings support the need for multidisciplinary care, early diagnostic evaluation, and individualized treatment strategies to optimize outcomes in patients with overlap syndromes. References [1.] Haugh A. Expert Opin Drug Saf 2020;19:479-88. [2.] Alghabban A. JTO Clin Res Resp 2024;6:100772.

Background: Immune mediated biliary injury (ICI-cholangiopathy) is a rare toxicity in patients treated with immune checkpoint inhibitors (ICI) and clinical characteristics, treatment responses and optimal management remain poorly characterised. This study aimed to characterize clinical profiles and treatment response in ICI-cholangiopathy. Methods: Patients treated with anti-PD-1, PDL-1 or CTLA-4 therapy between January 2010 and December 2025 were included in this retrospective study. Patients with ICI-cholangiopathy were identified from patients with significant elevation in cholestatic liver enzymes (R value <2) within 12 months of ICI exposure and clinical records were reviewed. Results: Of 5103 patients treated with ICI during the study period, 14 patients developed ICI-cholangiopathy (0.3%). Median age was 71 years and most (64%) were male. Most common primary malignancy was lung cancer (57%) and most were treated with pembrolizumab (71%). Most patients (8, 57%) had abnormal biliary imaging, with the majority reporting intra- and extra-hepatic biliary wall thickening and dilatation. All patients were treated with immunosuppression. ALP normalised over 3-6 months in 10 patients (71%; termed acute ICI-c) and improved but remained elevated < 2.5 x ULN in 4 (29%, chronic ICI-c). There were no significant differences in age, sex, primary cancer type, presence of liver metastases, duration of ICI therapy, or extent of ALP or ALT elevation at ICI-c diagnosis between patients with acute or chronic ICI-c. Patients with acute ICI-c improved faster than patients with chronic ICI-c (median time to ALP < 2.5 x ULN 16.5 vs 270 days, p = 0.007). Percent ALP reduction at 7 days distinguished patients with acute and chronic ICI-c (area under the receiver operating characteristic curve (AUROC) 0.92 (95% CI 0.63 - 0.99, p < 0.0001)). An optimal cut-off of <=20% ALP reduction at treatment day 7 was identified. Conclusion: ICI-cholangiopathy, defined using biochemical criteria, is an uncommon complication of ICI treatment. Most patients (71%) respond to immunosuppression and normalise ALP over time. ALP reduction <=20% at treatment day 7 may identify patients who will respond poorly to treatment, allowing management to be tailored accordingly, and this approach should be evaluated in future prospective studies.

S1. The associations of ACO2 expression with NSCLC patient outcomes, and with cell line proliferation in vitro and in vivo.

Kaplan–Meier estimates showing OS in all trial patients with a valid baseline (detectable or non-detectable) plasma ctDNA result in A, patients from the AURA3 trial (n = 291) and B, patients from the FLAURA trial (n = 499). Censored data are indicated by tick marks. Abbreviations: CI, confidence interval; EGFRm, epidermal growth factor receptor mutation (ex19del or L858R); HR, hazard ratio; mOS, median overall survival; NC, not calculable.

Abstract Background Since the Omicron variant emerged, uptake of booster doses of COVID-19 vaccines has gradually decreased both in the general population and among individuals at increased risk of severe COVID-19, such as patients with lung cancer. COVID-19 vaccine effectiveness (VE) has not been estimated in patients with lung cancer in the Omicron era. Methods In this test-negative design study using linked population-based cancer registry, health administrative, vaccination, and public health surveillance databases, we included all patients with active lung cancer or mesothelioma living in Ontario, Canada, who were tested for SARS-CoV-2 by RT-PCR from January 2, 2022, to August 31, 2023. We estimated VE against COVID-19-related severe outcomes (hospitalization or death) 7-179 days and ≥180 days following vaccination. Results During the study period, 13,622 patients with active lung cancer or mesothelioma underwent SARS-CoV-2 testing, with 1,371 (10.1%) positive. After exclusions, we analyzed 6,037 testing episodes, including 1,354 test-positive and 4,683 randomly selected test-negative episodes. Overall, COVID-19-associated hospitalization and mortality rates were 7.3% and 4.0%, respectively. Across both groups, 126/439 hospitalized patients (28.7%) and 116/5,598 non-hospitalized patients died (2.1%) (p=0.001). After multivariable adjustment, VE against severe COVID-19 was 56% (95%CI, 29%, 72%) 7-179 after vaccination and 10% (-45%, 44%) ≥180 days after vaccination. Conclusion Overall, COVID-19 VE against severe outcomes appears to be considerably lower for patients with active lung cancer or mesothelioma than the general population and decreases substantially beyond 180 days after vaccination. Our finding supports administration of COVID-19 booster doses in patients with lung cancer and mesothelioma every 6 months. Disclosures All Authors: No reported disclosures

BACKGROUND:Manual extraction of real-world clinical data for research can be time-consuming and prone to error. We assessed the feasibility of using natural language processing (NLP), an AI technique, to automate data extraction for patients with advanced lung cancer (aLC). We assessed the external validity of our NLP-extracted data by comparing our findings to those reported in the literature. METHODS:Patients diagnosed with stage IIIB or IV lung cancer between January 2015 to December 2017 at Princess Margaret Cancer Centre who received at least one dose of systemic therapy were included. Their electronic health records were provided to Pentavere's NLP platform, DARWENTM, in March 2019. Descriptive statistics summarized baseline patient and cancer characteristics, molecular biomarkers, and first-line systemic therapies. Cox multivariate models were used to evaluate prognostic factors for advanced non-small cell lung cancer (NSCLC) and small-cell lung cancer (SCLC) cohort. RESULT:NLP extracted clinical information (n = 333 patients) in a total of 8 hours, with only a few missing data for smoking status (n = 2), and Eastern Cooperative Oncology Group (ECOG) status (n = 5). Baseline patient and cancer characteristics summarized from NLP-extracted data were comparable to those in previous studies and population reports. For NSCLC patients, being male (HR 1.44, 95 % CI [1.04, 2.00]), having worse ECOG (1.48 [1.22, 1.81]), and having liver (2.24 [1.45, 3.46]), bone (2.09 [1.48, 2.96]), or lung metastases (2.54 [1.05, 2.26]) were associated with worse survival outcomes. For SCLC patients, having older age (HR 1.70 per 10 years, 95 % CI [1.10, 2.63]) and liver metastases (3.81 [1.61, 9.01]) were associated with worse survival outcomes. CONCLUSION:Our study demonstrated that automated data extraction using NLP is feasible and time efficient. Additionally, the NLP-extracted data can be used to identify valid and useful clinical endpoints for research. NLP holds significant potential to accelerate the extraction of real-world data for future observational studies.

Clinical information and proteome quantification

Univariable and multivariable analyses for OS from the AURA3 trial (ctDNA evaluable population, n = 291) and FLAURA trial (ctDNA evaluable population, n = 499).

BACKGROUND:Anaplastic lymphoma kinase inhibitors (ALKi) are a mainstay of therapy for patients with advanced non-small cell lung cancers (NSCLC). ALKi are associated with increased serum creatinine, which may represent reduced renal tubular creatinine secretion and/or true acute kidney injury (AKI). METHODS:We performed a retrospective study of patients who received ALKi for NSCLC (2013-2022). The primary outcomes were incidence of AKI within 90 days of ALKi start and chronic kidney disease (CKD) within 1-year. AKI and CKD were defined via KDIGO criteria, using the CKD-EPI equation with creatinine (and cystatin C in a subcohort). We performed logistic regression for AKI risk factors and Kaplan-Meier analysis for overall survival (OS). Spline curves were generated for eGFR means over time. RESULTS:Among 114 NSCLC patients, median age 61 years; 53% were female; 191 ALKi treatments were initiated: Alectinib (n = 91), Crizotinib (n = 33), Ceritinib (n = 27), Brigatinib (n = 21) and Lorlatinib (n = 19). There were 20 (10%) AKI events within 90 days after ALKi initiation, with 4 treatment changes attributed to AKI. Twenty-eight (14%) patients developed CKD, with 10 treatment changes; none requiring dialysis. In multivariate analysis, hypertension and male sex were associated with AKI. OS did not differ by AKI status. CONCLUSION:AKI/CKD events were frequent post-ALKi initiation (using creatinine-based eGFR), with a minority resulting in treatment change. Mean eGFR declined in the 90-days post-ALKi start. Most patients had mild CKD, with eGFR recovering postdrug cessation. AKI did not impact OS. Our findings suggest that most patients may continue ALKi therapy despite creatinine-based eGFR changes.

Kaplan–Meier estimates showing investigator-assessed PFS in FLAURA trial patients by clearance or non-clearance of plasma EGFRm status at Weeks 3 or 6 in patients who had baseline detectable plasma EGFRm. For comparison, patients with baseline non-detectable plasma EGFRm are included. A, Osimertinib arm by Week 3 plasma EGFRm status (n = 238). B, Comparator EGFR-TKI arm by Week 3 plasma EGFRm status (n = 243). C, Osimertinib arm by Week 6 plasma EGFRm status (n = 240). D, Comparator EGFR-TKI arm by Week 6 plasma EGFRm status (n = 235). Censored data are indicated by tick marks. Abbreviations: CI, confidence interval; EGFRm, epidermal growth factor receptor mutation (ex19del or L858R); NC, not calculable; mPFS, median PFS; TKI, tyrosine kinase inhibitor.

Representativeness of study participants

AURA3 flow diagram for ctDNA analysis (n = 291). Clearance refers to non-detectable plasma EGFRm, where EGFRm was detectable at baseline. *There were 9 patients in the osimertinib arm and 1 patient in the platinum-pemetrexed arm without a valid ddPCR result at Weeks 3 and 6; these samples were excluded from the ctDNA clearance analysis. Abbreviations: ctDNA, circulating tumor DNA; EGFRm, epidermal growth factor receptor mutation (ex19del or L858R).

Correlation between EGFRm allelic fraction at baseline by ddPCR (Biodesix) and NGS (Guardant Health) analyses in A, patients from the AURA3 trial (n = 202) and B, patients from the FLAURA trial (n = 347). Abbreviations: ddPCR, droplet digital PCR; EGFRm, epidermal growth factor receptor mutation (ex19del or L858R); NGS, next-generation sequencing.

Differentially expressed proteins and pathways analyses

Statistical association of clinical and pathologic characteristics with engraftment

Background: TP53 mutations (TP53-MUT) are common in NSCLC and have been reported as predictive of response and prognostic of poor outcome in EGFR-mutant NSCLC. The impact of TP53-MUT in NSCLCs with rarer driver mutations and approved targeted treatments is unclear. Methods: Records of 436 patients were reviewed and associations between TP53 status, demographics, and outcomes (overall response [ORR], survival [OS] and progression-free survival [PFS], and incidence of brain metastases [BM]), were investigated. Results: TP53-MUTs were found in 184/436 (42.4 %) with a significant demographic difference noted in stage distribution (p = 0.044). Median (m)OS was significantly shorter in TP53-MUT at 23.3 m (95 %CI 19.6-30.7) v 66.4 m (CI 55.0-not reached [NR]) (stage adjusted harzard ratio [aHR] 2.62, CI 1.98-3.46, p < 0.001). mRFS in early stage trended towards worse outcomes in TP53-MUT (p = 0.142). mPFS on first line treatment was shorter in TP53-MUT 5.0 m (CI 3.4-7.6) v 10.0 m (CI 6.1-13.5) (aHR 1.64, CI 1.23-2.19, p < 0.001). ORR to first instance of targeted treatment was 62 % v 70 % (TP53-MUT v WT) with more progressive disease as best response 24 % v 11 % (p = 0.0177). TP53-MT had a non-significant trend towards more BM at any point (46 % v 34 % p = 0.063) and a higher cumulative incidence of BM in stage I-III patients (Gray's test p < 0.001). When comparing disruptive versus non-disruptive TP53 mutations, there were no significant differences in demographics or survival outcomes. Conclusions: Co-occurring TP53 variants with rare driver mutations are predictive of a poor response to targeted treatments and associated with shorter OS and PFS in NSCLC.

S2. The associations of ACO2 expression and iron on cell proliferation and expression of proteins involved in iron regulation and ferroptosis.

8550 Background: Immune checkpoint inhibitor treatment failure constitutes a significant clinical challenge in non-small cell lung cancer (NSCLC). Molecular residual disease (MRD) detection in NSCLC may allow earlier detection of disease recurrence/progression and enable early treatment intensification or clinical trial enrolment. We have developed a novel tissue-agnostic genome-wide methylation enrichment platform based on cell free methylated DNA immunoprecipitation and high throughput sequencing (cfMeDIP-seq). Here, we present data on its application as an MRD assay to predict early recurrence or progression in patients (pts) with NSCLC receiving immunotherapy. Methods: The study population consists of pts with stage III/IV NSCLC at the Princess Margaret Cancer Centre, treated with definitive chemoradiation followed by consolidative durvalumab (stage III) or with PD-1 inhibitors +/- chemotherapy (stage IV). Pts underwent serial blood collection prior to initiation of treatment, 2-4 weeks after treatment initiation and approximately 6-8 weeks thereafter until progression. 5-10 ng of cfDNA was isolated from plasma. A classifier was trained on an independent set of lung and non-cancer samples to quantify relative circulating tumor DNA (ctDNA) content. The analysis considered multiple timepoints. Results were considered "positive” if there was a detected result at any follow-up timepoint. Results were considered "negative” if all follow-up timepoints were reported as not detected. Progression-free survival (PFS) was compared between groups using a log-rank test. Hazard ratio (HR) was estimated using Cox proportional hazards model. Results: A total of 187 samples from 63 unique pts (44% stage III and 56% stage IV) were analyzed and correlated with PFS. Pts with a positive MRD test showed significantly worse PFS than those who tested negative (HR 4.8; 95% CI, 2.1-10.8, P<0.0001), sensitivity 80%, specificity 91%. The lead time between MRD positivity and progression was up to 12.6 months, with a mean of 5.1 months. Secondary analysis of pts with stage III NSCLC revealed significantly worse PFS in MRD-positive pts compared to MRD-negative pts (HR 8; 95% CI, 1.4-46.7, P=0.007). Conclusions: MRD detection using genome-wide methylome enrichment correlates strongly with PFS in pts with advanced NSCLC receiving immunotherapy. This tissue agnostic assay shows promise for early identification of treatment failure, enabling timely selection of patients for treatment intensification or clinical trials.

Background:The role of prophylactic cranial irradiation (PCI) is not well-defined in extensive-stage SCLC (ES-SCLC), with conflicting results from randomized trials and a lack of relevant data for patients who received consolidative thoracic radiotherapy (CTRT). We sought to evaluate the impact of PCI on the outcomes of ES-SCLC patients who were all treated with CTRT. Methods:A retrospective analysis of ES-SCLC patients without brain metastases who were all treated with CTRT between 2013-2021 at our institution was conducted. Overall survival (OS) and incidence of brain failure (BFR) were estimated using Kaplan-Meier estimation and cumulative incidence function. Multivariable Cox or Fine-Gray's proportional hazard regression analysis (MVA) were performed to determine association between PCI and OS. Results:47 patients met inclusion criteria and were theoretically eligible for PCI, 27 (57.4 %) received PCI and CTRT while 20 (42.6 %) received CTRT alone. Baseline characteristics were similar except for age, where patients receiving PCI were younger (median age 62) compared to patients who did not receive PCI (median age 72). Median OS with PCI was 19.2 months, compared to 10.8 months without PCI (P = 0.0334). This improved OS remained apparent in patients who received post-chemotherapy MRI restaging (P = 0.0245). BFR was reduced with PCI (HR = 0.22 [0.09-0.52], P = 0.0004). On MVA, PCI was significantly and independently associated with improved OS (HR = 0.39 [0.19-0.80], P = 0.01) and reduced BFR (HR = 0.20 [0.09-0.44], P = < 0.001). Conclusion:This real-world study found PCI was independently associated with improved OS and reduced BFR in ES-SCLC patients treated with CTRT compared to patients treated with CTRT not receiving PCI, including after post-chemotherapy brain MRI. The role of PCI with CTRT should be evaluated in prospective studies.

FLAURA flow diagram for ctDNA analysis (n = 499). Clearance refers to non-detectable plasma EGFRm, where EGFRm was detectable at baseline. *There were 3 patients in the osimertinib arm and 7 patients in the comparator EGFR-TKI arm without a valid ddPCR result at Weeks 3 and 6; these samples were excluded from the ctDNA clearance analysis. Abbreviations: ctDNA, circulating tumor DNA; EGFRm, epidermal growth factor receptor mutation (ex19del or L858R); TKI, tyrosine kinase inhibitor.