P02.11.B An hypothesis generating study of MRI-Derived Radiomics on tumor and microenvironment tissue heterogeneity to guide post-operative management of glioblastoma: toward personalized radiation treatment volume delineation

Background

The glioblastoma’s bad prognosis is primarily due to intra-tumor heterogeneity, demonstrated from several studies that collected molecular biology, cytogenetic data and more recently radiomic features for a better prognostic stratification.The GLIFA project (GLIoblastoma Feature Analysis) is a multicentric project planned to investigated the role of radiomic analysis in GBM management, to verify if radiomic features in the tissue around the resection cavity which may guide the radiation target volume delineation.

Material and Methods

We retrospectively analyze from three centers radiomic features extracted from 90 patients with total or near total resection, who completed the standard adjuvant treatment and for whom we had post-operative images available for features extraction.The Manual segmentation was performed on post gadolinium T1w MRI sequence by 2 radiation oncologist reviewed by a neuroradiologist, both with at least 10 years of experience. The Region of interest (ROI) considered for the analysis were: the surgical cavity +/- post-surgical residual mass (CTV_cavity); the CTV a margin of 1.5 cm added to CTV_cavity and the volume resulting from subtracting the CTV_cavity from the CTV was defined as CTV_Ring. Radiomic analysis and modelling were conducted in RStudio. Z-score normalization was applied to each radiomic feature. A radiomic model was generated using the 226 features extracted from the Ring to perform a binary classification and predict the PFS at 6 months (statistical, morphological and textural features). A 3-fold cross-validation repeated five times was implemented for internal validation of the model.

Results

Two-hundred and seventy ROIs were contoured. The proposed radiomic model was given by the best fitting logistic regression model, and included the following 3 features: F_cm_merged.contrast, F_cm_merged.info.corr.2, F_rlm_merged.rlnu. A good agreement between model predicted probabilities and observed outcome probabilities was obtained (p-value of 0.49 by Hosmer and Lemeshow statistical test). The ROC curve of the model reported an AUC of 0.78 (95% CI: 0.68 - 0.88).

Conclusion

This is the first hypothesis-generating study who applies a radiomic analysis focusing on healthy tissue ring around the surgical cavity on post-operative MRI. This study provides a preliminary model for a decision support tool for a customization of the radiation target volume in GBM patients in order to achieve a margin reduction strategy.

Poziotinib for EGFR exon 20-mutant NSCLC: Clinical efficacy, resistance mechanisms, and impact of insertion location on drug sensitivity

We report a phase II study of 50 advanced non-small cell lung cancer (NSCLC) patients with point mutations or insertions in EGFR exon 20 treated with poziotinib (NCT03066206). The study achieved its primary endpoint, with confirmed objective response rates (ORRs) of 32% and 31% by investigator and blinded independent review, respectively, with a median progression-free survival of 5.5 months. Using preclinical studies, in silico modeling, and molecular dynamics simulations, we found that poziotinib sensitivity was highly dependent on the insertion location, with near-loop insertions (amino acids A767 to P772) being more sensitive than far-loop insertions, an observation confirmed clinically with ORRs of 46% and 0% observed in near versus far-loop, respectively (p = 0.0015). Putative mechanisms of acquired resistance included EGFR T790M, MET amplifications, and epithelial-to-mesenchymal transition (EMT). Our data demonstrate that poziotinib is active in EGFR exon 20-mutant NSCLC, although this activity is influenced by insertion location.

Abstract 3473: Use of DNA methylation from tumor and plasma to identify four major small cell lung cancer subtypes with distinct biology and therapeutic vulnerabilities

Small cell lung cancer (SCLC) is a highly aggressive cancer with limited treatment options and generally poor prognosis. Treatment of SCLC has not considerably changed over the last decades with therapeutic options focusing on unselected populations. Although in the past SCLC was thought to be a relatively homogenous malignancy, recent reports from our group and others identified four major distinct subgroups of SCLC, each with different therapeutic vulnerabilities. Three of the subtypes are defined by the expression of a specific transcription factor, ASCL1 (SCLC-A), NEUROD1 (SCLC-N) and POU2F3 (SCLC-P) while the fourth subtype is defined by an inflamed phenotype (SCLC-I). While our initial subtyping of SCLC is based on a gene expression signature comprised of ~1300 genes, which makes routine implementation challenging, we hypothesized that DNA methylation as a proxy to gene expression might be a more suitable approach for biomarker development in SCLC. We assembled a cohort of 105 SCLC formalin-fixed paraffin embedded (FFPE) samples (82/105 Stage > IIIb) and performed matched RNA-Sequencing (RNAseq) and methylation profiling using reduced-representation bisulfite sequencing (RRBS). To validate our findings and expand our analysis across different sample types, we profiled a panel of 59 fully characterized SCLC cell lines as well as 68 patient-derived xenograft models. We found that methylation levels differ markedly between the four subtypes, with the SCLC-N presenting with a hypermethylated phenotype and the SCLC-P with a hypomethylated phenotype across the genome, highlighting the profound differences in the underlying epigenetic regulation among the SCLC subtypes and supporting DNA methylation analysis as a potential readout for identifying SCLC subtypes. Furthermore, in order to subtype the clinical SCLC samples, we developed a predictive model using an extreme gradient boost model using RNA expression and DNA methylation, respectively, to allow the classification with 94.5% accuracy in the tissue testing cohort. Using a cohort of matched plasma samples, we further demonstrated that the DNA methylation differences were indeed preserved in cell-free DNA (cfDNA) allowing subtype classification with an accuracy of 87.5%. These data indicate that DNA methylation can be used for reliable subtyping of SCLC in tissue and in liquid biopsy samples. In summary, using a large cohort of predominantly extensive stage SCLC clinical samples, we were able to identify profound differences in DNA methylation that can be exploited as a novel biomarker for the classification of SCLC into four distinct subtypes with both tissue biopsy and non-invasive using plasma. Considering the previously shown therapeutic vulnerabilities of the four subtypes, these findings will enable the rapid initiation of personalized clinical trials in SCLC.

Citation Format: Simon Heeke, Carl M. Gay, Marcos R. Estecio, Allison Stewart, Hai Tran, Bingnan Zhang, Ximing Tang, Gabriela Raso, Kyle Concannon, Luana Guimaraes De Sousa, Whitney E. Lewis, Monique Nilsson, Yuanxin Xi, Lixia Diao, Qi Wang, Jianjun Zhang, Jing Wang, Ignacio I. Wistuba, Lauren A. Byers, John V. Heymach. Use of DNA methylation from tumor and plasma to identify four major small cell lung cancer subtypes with distinct biology and therapeutic vulnerabilities [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 3473.

Abstract 1827: CD70 is a novel therapeutic target for EGFR mutant NSCLC with acquired, EMT-associated EGFR TKI resistance

Approximately 15% of all patients with non-small cell lung cancer (NSCLC) and nearly 35% of Asian patients with NSCLC harbor activating mutations within the epidermal growth factor receptor (EGFR). Although these patients are initially highly sensitive to first or second generation EGFR tyrosine kinase inhibitors (TKIs) including erlotinib or third-generation inhibitors including osimertinib, EGFR TKI-refractory disease inevitably emerges. While therapeutic strategies to target resistant disease that emerges though secondary EGFR mutations or MET amplification have been developed, there remains a void of therapeutic options for patients where resistance occurs through EGFR-independent mechanism such as epithelial to mesenchymal transition (EMT) or transformation to small cell lung cancer (SCLC). To identify cell surface proteins that could be targeted by antibody-based or adoptive cell therapy approaches we interrogated RNAseq data from EGFR mutant NSCLC cell lines (HCC827 and HCC4006) and their associated EGFR TKI resistant variants previously shown to have developed resistance through EMT and filtered gene expression data to include only genes which transcribed proteins localized to the cell surface. We identified CD70 as a being highly upregulated in EGFR TKI resistant cells (p = 7.2e-42). Given that CD70 expression is highly restricted and only transiently expressed on immune cells, CD70 was selected as a top candidate cell surface protein for targeting studies. Western blotting and flow cytometry analysis confirmed CD70 protein levels to be highly upregulated in EGFR TKI resistant cells that had undergone EMT but not in cells harboring secondary EGFR mutations or MET amplifications. We also observed CD70 upregulation in osimertinib-treated drug tolerant persister cells, indicating that CD70 upregulation is an early event in the evolution of TKI resistance. Moreover, patient-derived models of acquired EGFR TKI resistance also exhibited CD70 positivity. Our data also indicated that in EGFR mutant NSCLC cells, CD70 could be upregulated through decreased CD70 promoter methylation as well as by the EMT regulators, transforming growth factor-β (TGF-β) and ZEB1, both of which were upregulated in TKI resistant cells. In EGFR TKI resistant cells, CD70 knockdown impaired cell viability and invasiveness, and stimulation of CD70 using the exogenous binding partner CD27 resulted in activation of AKT and MAPK, pathways known to be re-activated with acquired TKI resistance. CD70-targeting approaches including anti-CD70 antibody drug conjugates (ADCs) and CD70-targeting CAR T cell and CAR NK cells showed promising in vitro and in vivo activity against CD70 positive tumor cells and in osimertinib drug-tolerant persister cells. These results identify CD70 as a novel therapeutic target for EGFR mutant tumors with acquired EGFR TKI resistance that merits further investigation in the clinic.

Abstract 3260: Enhanced lineage plasticity in RTK-independent TKI-resistant EGFR-mutant NSCLC

Background: Resistance to targeted tyrosine kinase inhibitors (TKI) inevitably develops in metastatic EGFR-mutant non-small cell lung cancer (NSCLC). Resistance mechanisms are diverse, and mechanisms beyond receptor tyrosine kinase (RTK) pathway mutations are poorly understood. We hypothesized that the use of osimertinib as first-line therapy is increasing the prevalence of RTK-independent resistance mechanisms, and that RTK-independent resistant tumors undergo enhanced tumor cell lineage plasticity as an escape mechanism to EGFR TKI therapy.

Methods: We identified patients who developed osimertinib resistance (OR) after first line (1L, n=54) and second line (2L, n=42) treatments and determined the resistance mechanisms based on clinical sequencing and histopathology. We also performed single-cell RNA-seq of 24 samples from 13 patients with EGFRm NSCLC at TKI treatment-naïve (TN, n=2), residual disease (RD, n=4), and progression disease (PD, n=7) stages.

Results: Compared to 2L OR tumors, 1L OR tumors had increased RTK-independent mechanisms of resistance (76% vs. 46%, p=0.002), including 8% with small cell transformation (n=4), 2% with squamous transformation (n=1) and 66% with unknown mechanisms (n=34). To understand inter- and intra-tumor heterogeneity, we analyzed transcriptomic profiles of 76,266 single cells. Lung developmental lineages were assigned to 10,250 EpCAM+ cells, including 4,735 cells classified as malignant cells by inferCNV and RTK signaling analysis. In the two EGFRm TN tumors, the malignant cells demonstrated bronchoalveolar lineage and moderate EGFR expression. In the TKI resistant cases (PD, n=7), both RTK-dependent and RTK-independent resistance were observed. The RTK-dependent tumors (EGFR T790M n=1; ERBB2 amplification n=1) demonstrated preserved bronchoalveolar lineage identity. In the RTK-independent resistant tumors (n=5), one had complete lineage switch from epithelial to small cell neuroendocrine and very low expression level of EGFR. The remaining 4 PD tumors displayed varying expression of epithelial-to-mesenchymal transformation (EMT) features. One tumor had sarcomatoid histology and a high proportion of cells having positive VIM expression (84%) and 92% of cells having complete loss of NAPSA expression; 3 tumors had partial EMT demonstrated by heterogeneous proportion of cells having VIM expression (18-56%) and loss of NAPSA (26-67%). Interestingly, some of the cells with EMT and partial-EMT had moderate levels of EGFR expression, similar to the levels in the TN tumors.

Conclusion: With osimertinib use at 1L, the incidence of RTK-independent resistance has increased to become the dominant mechanism, whereas RTK-dependent resistance has decreased. Increased lineage plasticity (small cell neuroendocrine, squamous and EMT) potentially serves as an RTK-independent TKI-resistance mechanism in EGFRm NSCLC.

Citation Format: Xiuning Le, Ruiping Wang, Natalie Vokes, Yasir Elamin, Neda Kalhor, Daniel McGrail, Yuanxin Xi, Santiago Treviño III, Lingzhi Hong, Robyn Du, George Blumenschein, Carl Gay, Marcelo Negrao, Mehmet Altan, Hai Tran, Limei Hu, Jing Wang, Simon Heeke, Monique Nilsson, Jacqulyne Robichaux, Minghao Dang, Guangchun Han, Lauren Byers, Anne Tsao, Boris Sepesi, Chantale Bernatchez, Jianjun Zhang, Linghua Wang, John Heymach. Enhanced lineage plasticity in RTK-independent TKI-resistant EGFR-mutant NSCLC [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 3260.

Abstract 673: The EMT activator ZEB1 initiates polarized secretion of pro-tumorigenic effector proteins to drive lung adenocarcinoma progression

There are few effective therapeutic options for metastatic lung adenocarcinomas (LUADs) that lack actionable mutations in receptor tyrosine kinases, and there is an urgent need to develop a better understanding of the biological basis of LUAD growth and metastasis. Our group has developed mouse models of LUAD metastasis and found that LUAD metastasis is epithelial-to-mesenchymal transition (EMT)-dependent. EMT causes cancer cells to switch their axis of polarity from apical-basal to front-rear, which orients organelles and actinbased cytoskeletal structures in ways that facilitate purposeful cancer cell motility and metastasis. The EMT-activating transcription factor ZEB1 silences microRNAs that target key inhibitors of the polarity axis switch. These data support the general belief that EMT-dependent metastasis is a cell-autonomous process. However, our new findings here show that ZEB1 coordinates the components of secretory trafficking machinery, Rab6A and Rab8A, via silencing miR-148a that target these effectors, to drive polarized secretory trafficking toward the leading edge of LUAD cells, to enhance surface exposure of MMP14 so that triggers ECM degradation, accelerates focal adhesion turnover and promotes cancer cell migration. Moreover, the ZEB1-driven secretory trafficking increases the secretions of cytokines and immunomodulation factors, to generate an immunosuppressive tumor microenvironment, and to promote LUAD metastasis. These findings advance a paradigm in which EMT drives LUAD metastasis through a cell non-autonomous mechanism. The novelty rests in our results that demonstrate a transcriptional governance of polarized vesicular transport, providing the potential to target ZEB1-driven secretory pathways for the purpose of blocking metastasis in LUAD.

Citation Format: Guan-Yu Xiao, Xiaochao Tan, Leticia Rodriguez, Xin Liu, Jiang Yu, Mayra Vasquez, Hai Tran, William Russell, Don Gibbons, Jonathan Kurie. The EMT activator ZEB1 initiates polarized secretion of pro-tumorigenic effector proteins to drive lung adenocarcinoma progression [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 673.

0348 Sleep Staging Using End-to-End Deep Learning Model Based on Nocturnal Sound for Smartphones

Introduction

Convenient sleep tracking with mobile devices such as smartphones is desirable for people who want to easily objectify their sleep. The objective of this study was to introduce a deep learning model for sound-based sleep staging using audio data recorded with smartphones during sleep.

Methods

Two different audio datasets were used. One (N = 1,154) was extracted from polysomnography (PSG) data and the other (N = 327) was recorded using a smartphone during PSG from independent subjects. The performance of sound-based sleep staging would always depend on the quality of the audio. In practical conditions (non-contact and smartphone microphones), breathing and body movement sounds during night are so weak that the energy of such signals is sometimes smaller than that of ambient noise. The audio was converted into Mel spectrogram to detect latent temporal frequency patterns of breathing and body movement sound from ambient noise. The proposed neural network model consisted of two sub-models. The first sub-model extracted features from each 30-second epoch Mel spectrogram and the second one classified sleep stages through inter-epoch analysis of extracted features.

Results

Our model achieved 70 % epoch-by-epoch agreement for 4-class (wake, light, deep, rapid eye movement) stage classification and robust performance across various signal-to-noise conditions. More precisely, the model was correct in 77% of wake, 73% of light, 46% of deep, and 66% of REM. The model performance was not considerably affected by existence of sleep apnea but degradation observed with severe periodic limb movement. External validation with smartphone dataset also showed 68 % epoch-by-epoch agreement. Compared with some commercially available sleep trackers such as Fitbit Alta HR (0.6325 in mean per-class sensitivity) and SleepScore Max (0.565 in mean per-class sensitivity), our model showed superior performance in both PSG audio (0.655 in mean per-class sensitivity) and smartphone audio (0.6525 in mean per-class sensitivity).

Conclusion

To the best of our knowledge, this is the first end (Mel spectrogram-based feature extraction)-to-end (sleep staging) deep learning model that can work with audio data in practical conditions. Our proposed deep learning model of sound-based sleep staging has potential to be integrated in smartphone application for reliable at-home sleep tracking.

Support (If Any)

 

The Impact of COVID-19 Vaccinations on Emergency Department Presentations

Objective

The aim of the present study was to describe the burden of patients presenting to the ED with symptoms occurring after receiving a COVID‐19 vaccination.

Methods

This was a retrospective cohort study performed over a 4‐month period across two EDs. Participants were eligible for inclusion if it was documented in the ED triage record that their ED attendance was associated with the receipt of a COVID‐19 vaccination. Data regarding the type of vaccine (Comirnaty or ChAdOx1) were subsequently extracted from their electronic medical record. Primary outcome was ED length of stay (LOS) and secondary outcomes included requests for imaging and ED disposition destination.

Results

During the study period of 22 February 2021 to 21 June 2021, 632 patients were identified for inclusion in the present study, of which 543 (85.9%) had received the ChAdOx1 vaccination. The highest proportion of COVID‐19 vaccine‐related attendances occurred in June 2021 and accounted for 21 (8%) of 262 total daily ED attendances. Patients who had an ED presentation related to ChAdOx1 had a longer median ED LOS (253 vs 180 min, P < 0.001) compared to Comirnaty and a higher proportion had haematology tests and imaging requested in the ED. Most patients (n = 588, 88.8%) were discharged home from the ED.

Conclusion

There was a notable proportion of ED attendances related to recent COVID‐19 vaccination administration, many of which were associated with lengthy ED stays and had multiple investigations. In the majority of cases, the patients were able to be discharged home from the ED.

Chest X-ray Severity Score as a Putative Predictor of Clinical Outcome in Hospitalized Patients: An Experience From a Vietnamese COVID-19 Field Hospital

Background

Through the coronavirus disease 2019 (COVID-19) pandemic, portable radiography was particularly useful for assessing and monitoring the COVID-19 disease in Vietnamese field hospitals. It provides a convenient and precise picture of the progression of the disease. The purpose of this study was to evaluate the predictive value of chest radiograph reporting systems (Brixia and total severity score (TSS)) and the National Early Warning Score (NEWS) clinical score in a group of hospitalized patients with COVID-19.

Methods

This retrospective cohort study used routinely collected clinical data from polymerase chain reaction (PCR)-positive severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) patients admitted to Field Hospital District 8, Ho Chi Minh City, Vietnam, from August 2021 to September 2021. The initial chest radiographs were scored based on the TSS and Brixia scoring systems to quantify the extent of lung involvement. After the chest radiograph score was reported, two residents calculated the rate of all-cause in-hospital mortality with the consultation of expert radiologists. In this study, NEWS2 scores on hospital admission were calculated. The gradient boosting machines (GBMs) and Shapley additive exPlanations (SHAP) were applied to access the important variable and improve the accuracy of mortality prediction. The adjusted odds ratio for predictor was presented by univariate analysis and multivariate analysis.

Results

The chest X-rays (CXRs) at the admission of 273 patients (mean age 59 years +/-16, 42.1% were male) were scored. In the univariate analysis, age, vaccination status, previous disease, NEWS2, a saturation of peripheral oxygen (Sp02), the Brixia and TSS scores were significant predictors of mortality (p-value < 0.05). In multivariate analysis, there were statistically significant differences in mortality between age, Sp02, Brixia score, and patients with previous diseases were independent predictors of mortality and hospitalization.

A gradient boosting machine was performed in the train data set, which showed that the best hyperparameters for predicting the mortality of patients are the Brixia score (exclude TSS score). In the top five predictors, an increase in Brixia, age, and BMI increased the logarithmic number of probability clarifying as death status. Although the TSS and Brixia scores evaluated chest imaging, the TSS score was not essential as the Brixia score (rank 6/11). It was clear that the BMI and NEWS2 score was positively correlated with the Brixia score, and age did not affect this correlation. Meanwhile, we did not find any trend between the TSS score versus BMI and NEWS2.

Conclusion

When integrated with the BMI and NEWS2 clinical classification systems, the severity score of COVID-19 chest radiographs, particularly the Brixia score, was an excellent predictor of all-cause in-hospital mortality.

Continuous Thoracic Paravertebral Block to Treat Electrical Storm

Continuous ganglion block is increasingly being used to help manage ventricular tachyarrhythmias. We present the cases of 2 patients in whom we used continuous left thoracic paravertebral block to achieve sympathetic denervation and improvement in drug-refractory ventricular tachyarrhythmias. Whether as destination therapy or bridging therapy, we conclude that the block is safe, improves patients' comfort, and is superior in several ways to stellate ganglion block and other single-injection techniques.

Synthesis of Conjugated Molecules Based on Dithienopyrrole Derivatives and Pyrene as Chemosensor for Mesotrione Detection

The conjugated molecules based on pyrene and dithienopyrrole derivatives including 4-(2-ethylhexyl)-2-(pyren-1-yl)-4H-dithieno[3,2-b:2’,3’-d]pyrrole (EP4HP) and 4-(2-ethylhexyl)- 2,6-di(pyren-1-yl)-4H-dithieno[3,2-b:2’,3’-d]pyrrole (EDP4HP) have been successfully synthesized via C–H direct arylation reaction. The structures of these conjugated molecules were determined via nuclear magnetic resonance and Fourier-transform infrared spectroscopy (FTIR). The optical properties of conjugated molecules were evaluated via ultraviolet visible (UV-Vis) and fluorescence spectroscopies. The conjugated molecules exhibited the efficient fluorescence quenching toward herbicide mesotrione as nitroaromatic pesticides that could be promising candidates as the chemosensor for tracing of nitroaromatic pesticides.

Plasma Based Protein Signatures Associated with Small Cell Lung Cancer

Small-cell-lung cancer (SCLC) is associated with overexpression of oncogenes including Myc family genes and YAP1 and inactivation of tumor suppressor genes. We performed in-depth proteomic profiling of plasmas collected from 15 individuals with newly diagnosed early stage SCLC and from 15 individuals before the diagnosis of SCLC and compared findings with plasma proteomic profiles of 30 matched controls to determine the occurrence of signatures that reflect disease pathogenesis. A total of 272 proteins were elevated (area under the receiver operating characteristic curve (AUC) ≥ 0.60) among newly diagnosed cases compared to matched controls of which 31 proteins were also elevated (AUC ≥ 0.60) in case plasmas collected within one year prior to diagnosis. Ingenuity Pathway analyses of SCLC-associated proteins revealed enrichment of signatures of oncogenic MYC and YAP1. Intersection of proteins elevated in case plasmas with proteomic profiles of conditioned medium from 17 SCLC cell lines yielded 52 overlapping proteins characterized by YAP1-associated signatures of cytoskeletal re-arrangement and epithelial-to-mesenchymal transition. Among samples collected more than one year prior to diagnosis there was a predominance of inflammatory markers. Our integrated analyses identified novel circulating protein features in early stage SCLC associated with oncogenic drivers.

Estrogen Promotes Resistance to Bevacizumab in Murine Models of NSCLC

INTRODUCTION

Subgroup analyses from clinical studies have suggested that among patients with metastatic NSCLC receiving chemotherapy, females may derive less benefit from the addition of the vascular endothelial growth factor (VEGF) monoclonal antibody bevacizumab (BV) than males. This has raised the question of whether estrogen may affect the response to antiangiogenic therapy.

METHODS

To address this, we investigated the effects of estrogen on tumor growth, angiogenesis, and the response to BV in human xenograft models of NSCLC.

RESULTS

We observed that estrogen induced marked resistance to BV, which was accompanied by a 2.3-fold increase in tumor vascular pericyte coverage (p = 0.01) and an up-regulation of proangiogenic factors, VEGF and platelet-derived growth factor-BB. We also investigated the role of infiltrating myeloid cells, a population that has been associated with resistance to anti-VEGF therapies. We observed that estrogen induced a greater than twofold increase (p = 0.001) in the recruitment of tumor-infiltrating myeloid cells and concomitant increases in the myeloid recruitment factors, G-CSF and CXCL1. Blockade of the estrogen receptor pathway using fulvestrant resensitized tumors to VEGF targeting as evidenced by reduced tumor vasculature and an increase in overall survival in our NSCLC xenograft models.

CONCLUSIONS

Collectively, these data provide evidence that estrogen may promote resistance to VEGF-targeted therapies, potentially by enhancing pericyte coverage and myeloid recruitment, and suggest that estrogen receptor blockade merits further investigation as an approach to enhance the effects of antiangiogenic therapy.

Abstract 1670: Circulating biomarkers are associated with recurrence following complete resection of non-small cell lung cancer

Background: We have previously demonstrated that the presence of a proinflammatory peripheral cytokine milieu correlates with high levels of circulating neutrophils at the time of surgery and reduced overall survival and recurrence-free survival. In this study, we hypothesized that functional immune features or states in circulation may indicate early tumor recurrence when assessed longitudinally. We also investigated whether potential deficiencies in peripheral immune functionality at the time of lung cancer resection could identify correlates with subsequent outcome.

Methods: We performed flow cytometry and luminex profiling of blood samples collected from patients with stage I-IIIA resected NSCLC (n=150) and enrolled on the Immunogenomic profiling of NSCLC (ICON) prospective protocol. Patient characteristics include 75 adenocarcinomas, 30 squamous and 12 mixed or other histologies. Only patients who underwent primary cancer resection without neoadjuvant therapy were included. At a median follow up of 18.2 months, 37 patients had disease recurrence. Blood was collected at the time of primary lung cancer resection, at 4 weeks, and 4 months thereafter with PBMCs utilized for flow cytometry and plasma for cytokine assessment. Changes in cytokines were assessed by normalizing to baseline levels.

Results: Larger tumors as well as advanced clinical and pathological stages were associated with higher frequencies of proliferating Ki67+CD4+ and Ki67+CD8+ T cells in circulation at time of resection, suggesting an activated circulating immune response. We identified novel strong correlations in the plasma between soluble BTLA and Tim3 (r= 0.87, p=1.74e-140), PD1 and CD80 (r=0.72, p=6.41e-74) and moderate correlations between soluble PD1 and PDL1 (r=0.39, p=8.41e-18). Efforts are ongoing to determine the association of specific circulating immune states with the presence of these soluble receptors. CD8+Tim3+ T cells (FC=5.2%, p=0.028) and CTLA4+NK cells (FC=47.7%, p=0.002) as well as CTLA4+Tregs (FC=25.2%, p=0.05) were found to be increased in circulation at pre-recurrence time points (either 4 weeks or 4 months) relative to resection. This suggests the emergence of a suppressive cell type as well as induction of specific checkpoint receptors on effector cells that correlate with tumor recurrence. Finally, we identified a cytokine signature associated with recurrence by testing three sets within the ICON cohort with a training set AUC = 0.76 and the test set AUC=0.72, which was validated in a third set of patients yielding an AUC = 0.76.

Conclusions: We identified circulating immune features associated with initial tumor size and overall stage as well as unique associations among soluble proteins. Increased presence of potentially inhibited or suppressed CD8+ T cells and NK cells as well as Tregs are suggestive of mechanisms of immune suppression relative to tumor recurrence.

Citation Format: Younghee Lee, Daniel McGrail, Hai Tran, Mayra E. Vasquez, Carlos Ramos, Alexandre Reuben, Ara A. Vaporciyan, Annikka Weissferdt, Chantale Bernatchez, Tina Cascone, Ignacio I. Wistuba, Jianjun Zhang, John Heymach, Marcelo V. Negrao, Don L. Gibbons, Boris Sepesi, Cara L. Haymaker. Circulating biomarkers are associated with recurrence following complete resection of non-small cell lung cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 1670.

A YAP/FOXM1 axis mediates EMT-associated EGFR inhibitor resistance and increased expression of spindle assembly checkpoint components

Acquired resistance to tyrosine kinase inhibitors (TKIs) of epidermal growth factor receptor (EGFR) remains a clinical challenge. Especially challenging are cases in which resistance emerges through EGFR-independent mechanisms, such as through pathways that promote epithelial-to-mesenchymal transition (EMT). Through an integrated transcriptomic, proteomic, and drug screening approach, we identified activation of the yes-associated protein (YAP) and forkhead box protein M1 (FOXM1) axis as a driver of EMT-associated EGFR TKI resistance. EGFR inhibitor resistance was associated with broad multidrug resistance that extended across multiple chemotherapeutic and targeted agents, consistent with the difficulty of effectively treating resistant disease. EGFR TKI-resistant cells displayed increased abundance of spindle assembly checkpoint (SAC) proteins, including polo-like kinase 1 (PLK1), Aurora kinases, survivin, and kinesin spindle protein (KSP). Moreover, EGFR TKI-resistant cells exhibited vulnerability to SAC inhibitors. Increased activation of the YAP/FOXM1 axis mediated an increase in the abundance of SAC components in resistant cells. The clinical relevance of these finding was indicated by evaluation of specimens from patients with EGFR mutant lung cancer, which showed that high FOXM1 expression correlated with expression of genes encoding SAC proteins and was associated with a worse clinical outcome. These data revealed the YAP/FOXM1 axis as a central regulator of EMT-associated EGFR TKI resistance and that this pathway, along with SAC components, are therapeutic vulnerabilities for targeting this multidrug-resistant phenotype.