YAP1 status defines two intrinsic subtypes of LCNEC with distinct molecular features and therapeutic vulnerabilities.

PURPOSE: Large cell neuroendocrine carcinoma (LCNEC) is a high-grade neuroendocrine malignancy that, like small cell lung cancer (SCLC), is associated with an absence of druggable oncogenic drivers and dismal prognosis. In contrast to SCLC, however, there is little evidence to guide optimal treatment strategies which are often adapted from SCLC and non-small cell lung cancer (NSCLC) approaches. EXPERIMENTAL DESIGN: To better define the biology of LCNEC, we analyzed cell line and patient genomic data and performed immunohistochemistry and single-cell (sc)RNAseq of core needle biopsies from LCNEC patients and preclinical models. RESULTS: Here, we demonstrate that the presence or absence of YAP1 distinguishes two subsets of LCNEC. The YAP1-high subset is mesenchymal and inflamed and characterized, alongside TP53 mutations, by co-occurring alterations in CDKN2A/B and SMARCA4. Therapeutically, the YAP1-high subset demonstrates vulnerability to MEK and AXL targeting strategies, including a novel preclinical AXL CAR-T cell. Meanwhile, the YAP1-low subset is epithelial and immune-cold and more commonly features TP53 and RB1 co-mutations, similar to those observed in pure SCLC. Notably, the YAP1-low subset is also characterized by expression of SCLC subtype-defining transcription factors - especially ASCL1 and NEUROD1 - and, as expected given its transcriptional similarities to SCLC, exhibits putative vulnerabilities reminiscent of SCLC, including Delta-like ligand 3 (DLL3) and CD56 targeting, as with novel preclinical DLL3 and CD56 CAR T-cells, and DNA damage repair (DDR) inhibition. CONCLUSION: YAP1 defines distinct subsets of LCNEC with unique biology. These findings highlight the potential for YAP1 to guide personalized treatment strategies for LCNEC.

PIKing up and AKTing on resistance mutations in osimertinib-treated EGFR-mutated NSCLC.

A recent study identified high rates of PI3K-AKT pathway mutations from the FLAURA and AURA3 osimertinib trials and pre-clinically validated that these mutations decreased osimertinib sensitivity in EGFR-mutated NSCLC. The AKT inhibitor capivasertib was found to overcome this resistance, providing important rationale for the development of AKT inhibitors in NSCLC.

Integrative genomic and transcriptomic profiling of pulmonary sarcomatoid carcinoma identifies molecular subtypes associated with distinct immune features and clinical outcomes.

Background: Pulmonary sarcomatoid carcinoma (PSC) is a rare and aggressive subtype of non-small cell lung cancer (NSCLC), characterized by the presence of epithelial and sarcoma-like components. The molecular and immune landscape of PSC has not been well defined. Methods: Multiomics profiling of 21 pairs of PSCs with matched normal lung tissues was performed through targeted high-depth DNA panel, whole-exome, and RNA sequencing. We describe molecular and immune features that define subgroups of PSC with disparate genomic and immunogenic features as well as distinct clinical outcomes. Results: In total, 27 canonical cancer gene mutations were identified, with TP53 the most frequently mutated gene, followed by KRAS. Interestingly, most TP53 and KRAS mutations were earlier genomic events mapped to the trunks of the tumors, suggesting branching evolution in most PSC tumors. We identified two distinct molecular subtypes of PSC, driven primarily by immune infiltration and signaling. The Immune High (IM-H) subtype was associated with superior survival, highlighting the impact of immune infiltration on the biological and clinical features of localized PSCs. Conclusions: We provided detailed insight into the mutational landscape of PSC and identified two molecular subtypes associated with prognosis. IM-H tumors were associated with favorable recurrence-free survival and overall survival, highlighting the importance of tumor immune infiltration in the biological and clinical features of PSCs.

CoCo-ST: Comparing and Contrasting Spatial Transcriptomics data sets using graph contrastive learning.

Traditional feature dimension reduction methods have been widely used to uncover biological patterns or structures within individual spatial transcriptomics data. However, these methods are designed to yield feature representations that emphasize patterns or structures with dominant high variance, such as the normal tissue spatial pattern in a precancer setting. Consequently, they may inadvertently overlook patterns of interest that are potentially masked by these high-variance structures. Herein we present our graph contrastive feature representation method called CoCo-ST (Comparing and Contrasting Spatial Transcriptomics) to overcome this limitation. By incorporating a background data set representing normal tissue, this approach enhances the identification of interesting patterns in a target data set representing precancerous tissue. Simultaneously, it mitigates the influence of dominant common patterns shared by the background and target data sets. This enables discerning biologically relevant features crucial for capturing tissue-specific patterns, a capability we showcased through the analysis of serial mouse precancerous lung tissue samples.

The evolution of lung adenocarcinoma precursors is associated with chromosomal instability and transition from innate to adaptive immune response/evasion.

Studying lung adenocarcinoma (LUAD) early carcinogenesis is challenging, primarily due to the lack of LUAD precursors specimens. We amassed multi-omics data from 213 LUAD and LUAD precursors to identify molecular features underlying LUAD precancer evolution. We observed progressively increasing mutations, chromosomal aberrations, whole genome doubling and genomic instability from precancer to invasive LUAD, indicating aggravating chromosomal instability (CIN). Telomere shortening, a crucial genomic alteration linked to CIN, emerged at precancer stage. Moreover, later-stage lesions demonstrated increasing cancer stemness and decreasing alveolar identity, suggesting epithelial de-differentiation during early LUAD carcinogenesis. The innate immune cells progressively diminished from precancer to invasive LUAD, concomitant with a gradual recruitment of adaptive immune cells (except CD8+ and gamma-delta T cells that decreased in later stages) and upregulation of numerous immune checkpoints, suggesting LUAD precancer evolution is associated with a shift from innate to adaptive immune response and immune evasion mediated by various mechanisms.

Joint multi-omics discriminant analysis with consistent representation learning using PANDA.

Integrative multi-omics analysis provides deeper insight and enables better and more realistic modeling of the underlying biology and causes of diseases than does single omics analysis. Although several integrative multi-omics analysis methods have been proposed and demonstrated promising results in integrating distinct omics datasets, inconsistent distribution of the different omics data, which is caused by technology variations, poses a challenge for paired integrative multi-omics methods. In addition, the existing discriminant analysis-based integrative methods do not effectively exploit correlation and consistent discriminant structures, necessitating a compromise between correlation and discrimination in using these methods. Herein we present PAN-omics Discriminant Analysis (PANDA), a joint discriminant analysis method that seeks omics-specific discriminant common spaces by jointly learning consistent discriminant latent representations for each omics. PANDA jointly maximizes between-class and minimizes within-class omics variations in a common space and simultaneously models the relationships among omics at the consistency representation and cross-omics correlation levels, overcoming the need for compromise between discrimination and correlation as with the existing integrative multi-omics methods. Because of the consistency representation learning incorporated into the objective function of PANDA, this method seeks a common discriminant space to minimize the differences in distributions among omics, can lead to a more robust latent representations than other methods, and is against the inconsistency of the different omics. We compared PANDA to 10 other state-of-the-art multi-omics data integration methods using both simulated and real-world multi-omics datasets and found that PANDA consistently outperformed them while providing meaningful discriminant latent representations. PANDA is implemented using both R and MATLAB, with codes available at https://github.com/WuLabMDA/PANDA.

Enhancing NSCLC recurrence prediction with PET/CT habitat imaging, ctDNA, and integrative radiogenomics-blood insights.

While we recognize the prognostic importance of clinicopathological measures and circulating tumor DNA (ctDNA), the independent contribution of quantitative image markers to prognosis in non-small cell lung cancer (NSCLC) remains underexplored. In our multi-institutional study of 394 NSCLC patients, we utilize pre-treatment computed tomography (CT) and 18F-fluorodeoxyglucose positron emission tomography (FDG-PET) to establish a habitat imaging framework for assessing regional heterogeneity within individual tumors. This framework identifies three PET/CT subtypes, which maintain prognostic value after adjusting for clinicopathologic risk factors including tumor volume. Additionally, these subtypes complement ctDNA in predicting disease recurrence. Radiogenomics analysis unveil the molecular underpinnings of these imaging subtypes, highlighting downregulation in interferon alpha and gamma pathways in the high-risk subtype. In summary, our study demonstrates that these habitat imaging subtypes effectively stratify NSCLC patients based on their risk levels for disease recurrence after initial curative surgery or radiotherapy, providing valuable insights for personalized treatment approaches.

Clinical Benefit from Docetaxel +/- Ramucirumab Is Not Associated with Mutation Status in Metastatic Non-Small-Cell Lung Cancer Patients Who Progressed on Platinum Doublets and Immunotherapy.

Docetaxel +/- ramucirumab remains the standard-of-care therapy for patients with metastatic non-small-cell lung cancer (NSCLC) after progression on platinum doublets and immune checkpoint inhibitors (ICIs). The aim of our study was to investigate whether the cancer gene mutation status was associated with clinical benefits from docetaxel +/- ramucirumab. We also investigated whether platinum/taxane-based regimens offered a better clinical benefit in this patient population. A total of 454 patients were analyzed (docetaxel +/- ramucirumab n=381; platinum/taxane-based regimens n=73). Progression-free survival (PFS) and overall survival (OS) were compared among different subpopulations with different cancer gene mutations and between patients who received docetaxel +/- ramucirumab versus platinum/taxane-based regimens. Among patients who received docetaxel +/- ramucirumab, the top mutated cancer genes included TP53 (n=167), KRAS (n=127), EGFR (n=65), STK11 (n=32), ERBB2 (HER2) (n=26), etc. None of these cancer gene mutations or PD-L1 expression was associated with PFS or OS. Platinum/taxane-based regimens were associated with a significantly longer mQS (13.00 m, 95% Cl: 11.20-14.80 m versus 8.40 m, 95% Cl: 7.12-9.68 m, LogRank P=0.019) than docetaxel +/- ramcirumab. Key prognostic factors including age, histology, and performance status were not different between these two groups. In conclusion, in patients with metastatic NSCLC who have progressed on platinum doublets and ICIs, the clinical benefit from docetaxel +/- ramucirumab is not associated with the cancer gene mutation status. Platinum/taxane-based regimens may offer a superior clinical benefit over docetaxel +/- ramucirumab in this patient population.

Synthetic PET from CT improves diagnosis and prognosis for lung cancer: Proof of concept.

[18F]Fluorodeoxyglucose positron emission tomography (FDG-PET) and computed tomography (CT) are indispensable components in modern medicine. Although PET can provide additional diagnostic value, it is costly and not universally accessible, particularly in low-income countries. To bridge this gap, we have developed a conditional generative adversarial network pipeline that can produce FDG-PET from diagnostic CT scans based on multi-center multi-modal lung cancer datasets (n = 1,478). Synthetic PET images are validated across imaging, biological, and clinical aspects. Radiologists confirm comparable imaging quality and tumor contrast between synthetic and actual PET scans. Radiogenomics analysis further proves that the dysregulated cancer hallmark pathways of synthetic PET are consistent with actual PET. We also demonstrate the clinical values of synthetic PET in improving lung cancer diagnosis, staging, risk prediction, and prognosis. Taken together, this proof-of-concept study testifies to the feasibility of applying deep learning to obtain high-fidelity PET translated from CT.

Brief Report: Comprehensive Clinicogenomic Profiling of Small Cell Transformation From EGFR-Mutant NSCLC Informs Potential Therapeutic Targets.

Introduction: NSCLC transformation to SCLC has been best characterized with EGFR-mutant NSCLC, with emerging case reports seen in ALK, RET, and KRAS-altered NSCLC. Previous reports revealed transformed SCLC from EGFR-mutant NSCLC portends very poor prognosis and lack effective treatment. Genomic analyses revealed TP53 and RB1 loss of function increase the risk of SCLC transformation. Little has been reported on the detailed clinicogenomic characteristics and potential therapeutic targets for this patient population. Methods: In this study, we conducted a single-center retrospective analysis of clinical and genomic characteristics of patients with EGFR-mutant NSCLC transformed to SCLC. Demographic data, treatment course, and clinical molecular testing reports were extracted from electronic medical records. Kaplan-Meier analyses were used to estimate survival outcomes. Next generation sequencing-based assays was used to identify EGFR and co-occurring genetic alterations in tissue or plasma before and after SCLC transformation. Single-cell RNA sequencing (scRNA-seq) was performed on a patient-derived-xenograft model generated from a patient with EGFR-NSCLC transformed SCLC tumor. Results: A total of 34 patients were identified in our study. Median age at initial diagnosis was 58, and median time to SCLC transformation was 24.2 months. 68% were female and 82% were never smokers. 79% of patients were diagnosed as stage IV disease, and over half had brain metastases at baseline. Median overall survival of the entire cohort was 38.3 months from initial diagnoses and 12.4 months from time of SCLC transformation. Most patients harbored EGFR exon19 deletions as opposed to exon21 L858R alteration. Continuing EGFR tyrosine kinase inhibitor post-transformation did not improve overall survival compared with those patients where tyrosine kinase inhibitor was stopped in our cohort. In the 20 paired pretransformed and post-transformed patient samples, statistically significant enrichment was seen with PIK3CA alterations (p = 0.04) post-transformation. Profiling of longitudinal liquid biopsy samples suggest emergence of SCLC genetic alterations before biopsy-proven SCLC, as shown by increasing variant allele frequency of TP53, RB1, PIK3CA alterations. ScRNA-seq revealed potential therapeutic targets including DLL3, CD276 (B7-H3) and PTK7 were widely expressed in transformed SCLC. Conclusions: SCLC transformation is a potential treatment resistance mechanism in driver-mutant NSCLC. In our cohort of 34 EGFR-mutant NSCLC, poor prognosis was observed after SCLC transformation. Clinicogenomic analyses of paired and longitudinal samples identified genomic alterations emerging post-transformation and scRNA-seq reveal potential therapeutic targets in this population. Further studies are needed to rigorously validate biomarkers and therapeutic targets for this patient population.

Tumor- and circulating-free DNA methylation identifies clinically relevant small cell lung cancer subtypes.

Small cell lung cancer (SCLC) is an aggressive malignancy composed of distinct transcriptional subtypes, but implementing subtyping in the clinic has remained challenging, particularly due to limited tissue availability. Given the known epigenetic regulation of critical SCLC transcriptional programs, we hypothesized that subtype-specific patterns of DNA methylation could be detected in tumor or blood from SCLC patients. Using genomic-wide reduced-representation bisulfite sequencing (RRBS) in two cohorts totaling 179 SCLC patients and using machine learning approaches, we report a highly accurate DNA methylation-based classifier (SCLC-DMC) that can distinguish SCLC subtypes. We further adjust the classifier for circulating-free DNA (cfDNA) to subtype SCLC from plasma. Using the cfDNA classifier (cfDMC), we demonstrate that SCLC phenotypes can evolve during disease progression, highlighting the need for longitudinal tracking of SCLC during clinical treatment. These data establish that tumor and cfDNA methylation can be used to identify SCLC subtypes and might guide precision SCLC therapy.

The role of APOBEC3B in lung tumor evolution and targeted cancer therapy resistance.

In this study, the impact of the apolipoprotein B mRNA-editing catalytic subunit-like (APOBEC) enzyme APOBEC3B (A3B) on epidermal growth factor receptor (EGFR)-driven lung cancer was assessed. A3B expression in EGFR mutant (EGFRmut) non-small-cell lung cancer (NSCLC) mouse models constrained tumorigenesis, while A3B expression in tumors treated with EGFR-targeted cancer therapy was associated with treatment resistance. Analyses of human NSCLC models treated with EGFR-targeted therapy showed upregulation of A3B and revealed therapy-induced activation of nuclear factor kappa B (NF-κB) as an inducer of A3B expression. Significantly reduced viability was observed with A3B deficiency, and A3B was required for the enrichment of APOBEC mutation signatures, in targeted therapy-treated human NSCLC preclinical models. Upregulation of A3B was confirmed in patients with NSCLC treated with EGFR-targeted therapy. This study uncovers the multifaceted roles of A3B in NSCLC and identifies A3B as a potential target for more durable responses to targeted cancer therapy.

Distinct patterns of auto-reactive antibodies associated with organ-specific immune-related adverse events.

The roles of preexisting auto-reactive antibodies in immune-related adverse events (irAEs) associated with immune checkpoint inhibitor therapy are not well defined. Here, we analyzed plasma samples longitudinally collected at predefined time points and at the time of irAEs from 58 patients with immunotherapy naïve metastatic non-small cell lung cancer treated on clinical protocol with ipilimumab and nivolumab. We used a proteomic microarray system capable of assaying antibody reactivity for IgG and IgM fractions against 120 antigens for systemically evaluating the correlations between auto-reactive antibodies and certain organ-specific irAEs. We found that distinct patterns of auto-reactive antibodies at baseline were associated with the subsequent development of organ-specific irAEs. Notably, ACHRG IgM was associated with pneumonitis, anti-cytokeratin 19 IgM with dermatitis, and anti-thyroglobulin IgG with hepatitis. These antibodies merit further investigation as potential biomarkers for identifying high-risk populations for irAEs and/or monitoring irAEs during immunotherapy treatment. Trial registration: ClinicalTrials.gov identifier: NCT03391869.

ATM Mutations Associate with Distinct Co-Mutational Patterns and Therapeutic Vulnerabilities in NSCLC.

PURPOSE: Ataxia-telangiectasia mutated (ATM) is the most frequently mutated DNA damage repair gene in non-small cell lung cancer (NSCLC). However, the molecular correlates of ATM mutations and their clinical implications have not been fully elucidated. EXPERIMENTAL DESIGN: Clinicopathologic and genomic data from 26,587 patients with NSCLC from MD Anderson, public databases, and a de-identified nationwide (US-based) NSCLC clinicogenomic database (CGDB) were used to assess the co-mutation landscape, protein expression, and mutational processes in ATM-mutant tumors. We used the CGDB to evaluate ATM-associated outcomes in patients treated with immune checkpoint inhibitors (ICI) with or without chemotherapy, and assessed the effect of ATM loss on STING signaling and chemotherapy sensitivity in preclinical models. RESULTS: Nonsynonymous mutations in ATM were observed in 11.2% of samples (2,980/26,587) and were significantly associated with mutations in KRAS, but mutually exclusive with EGFR (q < 0.1). KRAS mutational status constrained the ATM co-mutation landscape, with strong mutual exclusivity with TP53 and KEAP1 within KRAS-mutated samples. Those ATM mutations that co-occurred with TP53 were more likely to be missense mutations and associate with high mutational burden, suggestive of non-functional passenger mutations. In the CGDB cohort, dysfunctional ATM mutations associated with improved OS only in patients treated with ICI-chemotherapy, and not ICI alone. In vitro analyses demonstrated enhanced upregulation of STING signaling in ATM knockout cells with the addition of chemotherapy. CONCLUSIONS: ATM mutations define a distinct subset of NSCLC associated with KRAS mutations, increased TMB, decreased TP53 and EGFR co-occurrence, and potential increased sensitivity to ICIs in the context of DNA-damaging chemotherapy.

SwarmDeepSurv: swarm intelligence advances deep survival network for prognostic radiomics signatures in four solid cancers.

Survival models exist to study relationships between biomarkers and treatment effects. Deep learning-powered survival models supersede the classical Cox proportional hazards (CoxPH) model, but substantial performance drops were observed on high-dimensional features because of irrelevant/redundant information. To fill this gap, we proposed SwarmDeepSurv by integrating swarm intelligence algorithms with the deep survival model. Furthermore, four objective functions were designed to optimize prognostic prediction while regularizing selected feature numbers. When testing on multicenter sets (n = 1,058) of four different cancer types, SwarmDeepSurv was less prone to overfitting and achieved optimal patient risk stratification compared with popular survival modeling algorithms. Strikingly, SwarmDeepSurv selected different features compared with classical feature selection algorithms, including the least absolute shrinkage and selection operator (LASSO), with nearly no feature overlapping across these models. Taken together, SwarmDeepSurv offers an alternative approach to model relationships between radiomics features and survival endpoints, which can further extend to study other input data types including genomics.

Brief Report: Clinical Response, Toxicity, and Resistance Mechanisms to Osimertinib Plus MET Inhibitors in Patients With EGFR-Mutant MET-Amplified NSCLC.

Introduction: MET amplification is a known resistance mechanism to EGFR tyrosine kinase inhibitor (TKI) treatment in EGFR-mutant NSCLC. Dual EGFR-MET inhibition has been reported with success in overcoming such resistance and inducing clinical benefit. Resistance mechanisms to dual EGFR-MET inhibition require further investigation and characterization. Methods: Patients with NSCLC with both MET amplification and EGFR mutation who have received crizotinib, capmatinib, savolitinib, or tepotinib plus osimertinib (OSI) after progression on OSI at MD Anderson Cancer Center were included in this study. Molecular profiling was completed by means of fluorescence in situ hybridization (FISH) and next-generation sequencing (NGS). Radiological response was assessed on the basis of Response Evaluation Criteria in Solid Tumors version 1.1. Results: From March 2016 to March 2022, 23 treatments with dual MET inhibitor and osi were identified with a total of 20 patients included. Three patients received capmatinib plus OSI after progression on crizotinib plus OSI. Median age was 64 (38-89) years old and 75% were female. MET amplification was detected by FISH in 14 patients in the tissue, NGS in 10 patients, and circulating tumor DNA in three patients. Median MET gene copy number was 13.6 (6.4-20). Overall response rate was 34.8% (eight of 23). In assessable patients, tumor shrinkage was observed in 82.4% (14 of 17). Median time on treatment was 27 months. Two of three patients responded to capmatinib plus OSI after progression on crizotinib plus OSI. Dual EGFR-MET inhibition was overall well tolerated. Two patients on crizotinib plus OSI and one pt on capmatinib plus OSI discontinued therapy due to pneumonitis. One pt discontinued crizotinib plus OSI due to gastrointestinal toxicity. Six patients were still on double TKI treatment. At disease progression to dual EGFR-MET inhibition, FISH and NGS on tumor and plasma were completed in six patients. Notable resistance mechanisms observed include acquired MET D1246H (n = 1), acquired EGFR C797S (n = 2), FGFR2 fusion (n = 1, concurrent with C797S), and EGFR G796S (n = 1, concurrent with C797S). Four patients lost MET amplification. Conclusions: Dual EGFR and MET inhibition yielded high clinical response rate after progression on OSI. Resistance mechanisms to EGFR-MET double TKI inhibition include MET secondary mutation, EGFR secondary mutation, or loss of MET amplification.

Clinical and Molecular Features of Long-term Response to Immune Checkpoint Inhibitors in Patients with Advanced Non-Small Cell Lung Cancer.

PURPOSE: We sought to identify features of patients with advanced non-small cell lung cancer (NSCLC) who achieve long-term response (LTR) to immune checkpoint inhibitors (ICI), and how these might differ from features predictive of short-term response (STR). EXPERIMENTAL DESIGN: We performed a multicenter retrospective analysis of patients with advanced NSCLC treated with ICIs between 2011 and 2022. LTR and STR were defined as response ≥ 24 months and response < 12 months, respectively. Tumor programmed death ligand 1 (PD-L1) expression, tumor mutational burden (TMB), next-generation sequencing (NGS), and whole-exome sequencing (WES) data were analyzed to identify characteristics enriched in patients achieving LTR compared with STR and non-LTR. RESULTS: Among 3,118 patients, 8% achieved LTR and 7% achieved STR, with 5-year overall survival (OS) of 81% and 18% among LTR and STR patients, respectively. High TMB (≥50th percentile) enriched for LTR compared with STR (P = 0.001) and non-LTR (P < 0.001). Whereas PD-L1 ≥ 50% enriched for LTR compared with non-LTR (P < 0.001), PD-L1 ≥ 50% did not enrich for LTR compared with STR (P = 0.181). Nonsquamous histology (P = 0.040) and increasing depth of response [median best overall response (BOR) -65% vs. -46%, P < 0.001] also associated with LTR compared with STR; no individual genomic alterations were uniquely enriched among LTR patients. CONCLUSIONS: Among patients with advanced NSCLC treated with ICIs, distinct features including high TMB, nonsquamous histology, and depth of radiographic improvement distinguish patients poised to achieve LTR compared with initial response followed by progression, whereas high PD-L1 does not.

MCT4-dependent lactate secretion suppresses antitumor immunity in LKB1-deficient lung adenocarcinoma.

Inactivating STK11/LKB1 mutations are genomic drivers of primary resistance to immunotherapy in KRAS-mutated lung adenocarcinoma (LUAD), although the underlying mechanisms remain unelucidated. We find that LKB1 loss results in enhanced lactate production and secretion via the MCT4 transporter. Single-cell RNA profiling of murine models indicates that LKB1-deficient tumors have increased M2 macrophage polarization and hypofunctional T cells, effects that could be recapitulated by the addition of exogenous lactate and abrogated by MCT4 knockdown or therapeutic blockade of the lactate receptor GPR81 expressed on immune cells. Furthermore, MCT4 knockout reverses the resistance to PD-1 blockade induced by LKB1 loss in syngeneic murine models. Finally, tumors from STK11/LKB1 mutant LUAD patients demonstrate a similar phenotype of enhanced M2-macrophages polarization and hypofunctional T cells. These data provide evidence that lactate suppresses antitumor immunity and therapeutic targeting of this pathway is a promising strategy to reversing immunotherapy resistance in STK11/LKB1 mutant LUAD.

Predicting benefit from immune checkpoint inhibitors in patients with non-small-cell lung cancer by CT-based ensemble deep learning: a retrospective study.

BACKGROUND: Only around 20-30% of patients with non-small-cell lung cancer (NCSLC) have durable benefit from immune-checkpoint inhibitors. Although tissue-based biomarkers (eg, PD-L1) are limited by suboptimal performance, tissue availability, and tumour heterogeneity, radiographic images might holistically capture the underlying cancer biology. We aimed to investigate the application of deep learning on chest CT scans to derive an imaging signature of response to immune checkpoint inhibitors and evaluate its added value in the clinical context. METHODS: In this retrospective modelling study, 976 patients with metastatic, EGFR/ALK negative NSCLC treated with immune checkpoint inhibitors at MD Anderson and Stanford were enrolled from Jan 1, 2014, to Feb 29, 2020. We built and tested an ensemble deep learning model on pretreatment CTs (Deep-CT) to predict overall survival and progression-free survival after treatment with immune checkpoint inhibitors. We also evaluated the added predictive value of the Deep-CT model in the context of existing clinicopathological and radiological metrics. FINDINGS: Our Deep-CT model demonstrated robust stratification of patient survival of the MD Anderson testing set, which was validated in the external Stanford set. The performance of the Deep-CT model remained significant on subgroup analyses stratified by PD-L1, histology, age, sex, and race. In univariate analysis, Deep-CT outperformed the conventional risk factors, including histology, smoking status, and PD-L1 expression, and remained an independent predictor after multivariate adjustment. Integrating the Deep-CT model with conventional risk factors demonstrated significantly improved prediction performance, with overall survival C-index increases from 0·70 (clinical model) to 0·75 (composite model) during testing. On the other hand, the deep learning risk scores correlated with some radiomics features, but radiomics alone could not reach the performance level of deep learning, indicating that the deep learning model effectively captured additional imaging patterns beyond known radiomics features. INTERPRETATION: This proof-of-concept study shows that automated profiling of radiographic scans through deep learning can provide orthogonal information independent of existing clinicopathological biomarkers, bringing the goal of precision immunotherapy for patients with NSCLC closer. FUNDING: National Institutes of Health, Mark Foundation Damon Runyon Foundation Physician Scientist Award, MD Anderson Strategic Initiative Development Program, MD Anderson Lung Moon Shot Program, Andrea Mugnaini, and Edward L C Smith.

Machine Learning Models for the Identification of Prognostic and Predictive Cancer Biomarkers: A Systematic Review.

The identification of biomarkers plays a crucial role in personalized medicine, both in the clinical and research settings. However, the contrast between predictive and prognostic biomarkers can be challenging due to the overlap between the two. A prognostic biomarker predicts the future outcome of cancer, regardless of treatment, and a predictive biomarker predicts the effectiveness of a therapeutic intervention. Misclassifying a prognostic biomarker as predictive (or vice versa) can have serious financial and personal consequences for patients. To address this issue, various statistical and machine learning approaches have been developed. The aim of this study is to present an in-depth analysis of recent advancements, trends, challenges, and future prospects in biomarker identification. A systematic search was conducted using PubMed to identify relevant studies published between 2017 and 2023. The selected studies were analyzed to better understand the concept of biomarker identification, evaluate machine learning methods, assess the level of research activity, and highlight the application of these methods in cancer research and treatment. Furthermore, existing obstacles and concerns are discussed to identify prospective research areas. We believe that this review will serve as a valuable resource for researchers, providing insights into the methods and approaches used in biomarker discovery and identifying future research opportunities.