Recurrent somatic mutations as predictors of immunotherapy response

Reannotation of cancer mutations based on expressed RNA transcripts reveals functional non-coding mutations in melanoma

The role of synonymous mutations in cancer pathogenesis is currently underexplored. We developed a method to detect significant clusters of synonymous and missense mutations in public cancer genomics data. In melanoma, we show that 22% (11/50) of these mutation clusters are misannotated as coding mutations because the reference transcripts used for their annotation are not expressed. Instead, these mutations are actually non-coding. This, for instance, applies to the mutation clusters targeting known cancer genes kinetochore localized astrin (SPAG5) binding protein (KNSTRN) and BCL2-like 12 (BCL2L12), each affecting 4%-5% of melanoma tumors. For the latter, we show that these mutations are functional non-coding mutations that target the shared promoter region of interferon regulatory factor 3 (IRF3) and BCL2L12. This results in downregulation of IRF3, BCL2L12, and tumor protein p53 (TP53) expression in a CRISPR-Cas9 primary melanocyte model and in melanoma tumors. In individuals with melanoma, these mutations were also associated with a worse response to immunotherapy. Finally, we propose a simple automated method to more accurately annotate cancer mutations based on expressed transcripts. This work shows the importance of integrating DNA- and RNA-sequencing data to properly annotate mutations and identifies a number of previously overlooked and wrongly annotated functional non-coding mutations in melanoma.

Clinical development of immuno-oncology therapeutics

Immuno-oncology (IO) is one of the fastest growing therapeutic areas within oncology. IO agents work indirectly via the host's adaptive and innate immune system to recognize and eradicate tumor cells. Despite checkpoint inhibitors being only introduced to the market since 2011, they have become the second most approved product category. Current Food and Drug Administration (FDA)-approved classes of IO agents include: immune checkpoint inhibitors (ICIs), chimeric antigen receptor T-cell therapy (CAR-T), bi-specific T-cell engager (BiTE) antibody therapy, T-cell receptor (TCR) engineered T cell therapy, tumor-infiltrating lymphocyte (TIL) therapy, cytokine therapy, cancer vaccine therapy, and oncolytic virus therapy. Cancer immunotherapy has made progress in multiple cancer types including melanoma, non-small cell lung cancer (NSCLC), renal cell carcinoma (RCC), and urothelial carcinoma; however, several cancers remain refractory to immunotherapy. Future directions of IO include exploration in the neoadjuvant/perioperative setting, combination strategies, and optimizing patient selection through improved biomarkers.

Comprehensive assessment of computational methods for cancer immunoediting

Cancer immunoediting reflects the role of the immune system in eliminating tumor cells and shaping tumor immunogenicity, which leaves marks in the genome. In this study, we systematically evaluate four methods for quantifying immunoediting. In colorectal cancer samples from The Cancer Genome Atlas, we found that these methods identified 78.41%, 46.17%, 36.61%, and 4.92% of immunoedited samples, respectively, covering 92.90% of all colorectal cancer samples. Comparison of 10 patient-derived xenografts (PDXs) with their original tumors showed that different methods identified reduced immune selection in PDXs ranging from 44.44% to 60.0%. The proportion of such PDX-tumor pairs increases to 77.78% when considering the union of results from multiple methods, indicating the complementarity of these methods. We find that observed-to-expected ratios highly rely on neoantigen selection criteria and reference datasets. In contrast, HLA-binding mutation ratio, immune dN/dS, and enrichment score of cancer cell fraction were less affected by these factors. Our findings suggest integration of multiple methods may benefit future immunoediting analyses.

Bioinformatics analysis reveals novel tumor antigens and immune subtypes of skin cutaneous melanoma contributing to mRNA vaccine development

Introduction

Skin cutaneous melanoma (SKCM) is a common malignant skin cancer with high mortality and recurrence rates. Although the mRNA vaccine is a promising strategy for cancer treatment, its application against SKCM remains confusing. In this study, we employed computational bioinformatics analysis to explore SKCM-associated antigens for an mRNA vaccine and suitable populations for vaccination.

Methods

Gene expression and clinical data were retrieved from GEO and TCGA. The differential expression levels and prognostic index of selected antigens were computed via GEPIA2,while genetic alterations were analyzed using cBioPortal. TIMER was utilized to assess the correlation between antigen-presenting cell infiltration and antigen. Consensus clustering identified immune subtypes, and immune characteristics were evaluated across subtypes. Weighted gene co-expression network analysis was performed to identify modules of immune-related genes.

Results

We discovered five tumor antigens (P2RY6, PLA2G2D, RBM47, SEL1L3, and SPIB) that are significantly increased and mutated, which correlate with the survival of patients and the presence of immune cells that present these antigens. Our analysis revealed two distinct immune subtypes among the SKCM samples. Immune subtype 1 was associated with poorer clinical outcomes and exhibited low levels of immune activity, characterized by fewer mutations and lower immune cell infiltration. In contrast, immune subtype 2 showed higher immune activity and better patient outcomes. Subsequently, the immune landscape of SKCM exhibited immune heterogeneity among patients, and a key gene module that is enriched in immune-related pathways was identified.

Conclusions

Our findings suggest that the identified tumor antigens could serve as valuable targets for developing mRNA vaccines against SKCM, particularly for patients in immune subtype 1. This research provides valuable insights into personalized immunotherapy approaches for this challenging cancer and highlights the advantages of bioinformatics in identifying immune targets and optimizing treatment approaches.

Paired CRISPR screens to map gene regulation in cis and trans

Recent massively-parallel approaches to decipher gene regulatory circuits have focused on the discovery of either cis -regulatory elements (CREs) or trans -acting factors. Here, we develop a scalable approach that pairs cis - and trans -regulatory CRISPR screens to systematically dissect how the key immune checkpoint PD-L1 is regulated. In human pancreatic ductal adenocarcinoma (PDAC) cells, we tile the PD-L1 locus using ∼25,000 CRISPR perturbations in constitutive and IFNγ-stimulated conditions. We discover 67 enhancer- or repressor-like CREs and show that distal CREs tend to contact the promoter of PD-L1 and related genes. Next, we measure how loss of all ∼2,000 transcription factors (TFs) in the human genome impacts PD-L1 expression and, using this, we link specific TFs to individual CREs and reveal novel PD-L1 regulatory circuits. For one of these regulatory circuits, we confirm the binding of predicted trans -factors (SRF and BPTF) using CUT&RUN and show that loss of either the CRE or TFs potentiates the anti-cancer activity of primary T cells engineered with a chimeric antigen receptor. Finally, we show that expression of these TFs correlates with PD-L1 expression in vivo in primary PDAC tumors and that somatic mutations in TFs can alter response and overall survival in immune checkpoint blockade-treated patients. Taken together, our approach establishes a generalizable toolkit for decoding the regulatory landscape of any gene or locus in the human genome, yielding insights into gene regulation and clinical impact.

Tumour mutational burden: clinical utility, challenges and emerging improvements

Tumour mutational burden (TMB), defined as the total number of somatic non-synonymous mutations present within the cancer genome, varies across and within cancer types. A first wave of retrospective and prospective research identified TMB as a predictive biomarker of response to immune-checkpoint inhibitors and culminated in the disease-agnostic approval of pembrolizumab for patients with TMB-high tumours based on data from the Keynote-158 trial. Although the applicability of outcomes from this trial to all cancer types and the optimal thresholds for TMB are yet to be ascertained, research into TMB is advancing along three principal avenues: enhancement of TMB assessments through rigorous quality control measures within the laboratory process, including the mitigation of confounding factors such as limited panel scope and low tumour purity; refinement of the traditional TMB framework through the incorporation of innovative concepts such as clonal, persistent or HLA-corrected TMB, tumour neoantigen load and mutational signatures; and integration of TMB with established and emerging biomarkers such as PD-L1 expression, microsatellite instability, immune gene expression profiles and the tumour immune contexture. Given its pivotal functions in both the pathogenesis of cancer and the ability of the immune system to recognize tumours, a profound comprehension of the foundational principles and the continued evolution of TMB are of paramount relevance for the field of oncology.

Personalized Immunotherapy Achieves Complete Response in Metastatic Adenoid Cystic Carcinoma Despite Lack of Conventional Biomarkers

There is currently no effective treatment strategy for recurrent/metastatic adenoid cystic carcinoma (R/M ACC). Furthermore, recent single-agent and combination immunotherapy trials have failed in unselected ACC cohorts, unlike non-ACC salivary gland cancers. Genomic profiling revealed no actionable targets but NOTCH1 and KDM6A frameshift and CTCF splice site mutations (no MYB/L fusion) with a low tumor mutational burden (TMB), microsatellite stable (MSS) and negative programmed death ligand 1 (PD-L1) were observed. We recommended an anti-programmed cell death protein 1 (anti-PD-1) plus anti-Cytotoxic T-lymphocyte-associated protein 4 (anti-CTLA-4) combination based on TMB 2-fold greater-than-median TMB in ACC, tumor harboring multiple immunogenic frameshift or splice site mutations, and PD-L1 negativity. Accordingly, we achieved a complete response in a radiotherapy (RT) and chemotherapy (CT)-refractory patient with locally recurrent lacrimal gland (LG) ACC and lung metastasis following personalized immunotherapy in combination with integrative therapeutics. Therefore, it is crucial to assess not only conventional immune biomarkers but also patient-specific parameters, especially in "immune-cold" cancer types.

The long-term effectiveness and mechanism of oncolytic virotherapy combined with anti-PD-L1 antibody in colorectal cancer patient

Colorectal cancer (CRC) is known to be resistant to immunotherapy. In our phase-I clinical trial, one patient achieved a 313-day prolonged response during the combined treatment of oncolytic virotherapy and immunotherapy. To gain a deeper understanding of the potential molecular mechanisms, we performed a comprehensive multi-omics analysis on this patient and three non-responders. Our investigation unveiled that, initially, the tumor microenvironment (TME) of this responder presented minimal infiltration of T cells and natural killer cells, along with a relatively higher presence of macrophages compared to non-responders. Remarkably, during treatment, there was a progressive increase in CD4+ T cells, CD8+ T cells, and B cells in the responder's tumor tissue. This was accompanied by a significant upregulation of transcription factors associated with T-cell activation and cytotoxicity, including GATA3, EOMES, and RUNX3. Furthermore, dynamic monitoring of peripheral blood samples from the responder revealed a rapid decrease in circulating tumor DNA (ctDNA), suggesting its potential as an early blood biomarker of treatment efficacy. Collectively, our findings demonstrate the effectiveness of combined oncolytic virotherapy and immunotherapy in certain CRC patients and provide molecular evidence that virotherapy can potentially transform a "cold" TME into a "hot" one, thereby improving sensitivity to immunotherapy.

Pan-cancer analysis of T-cell proliferation regulatory genes as potential immunotherapeutic targets

T cells are the key to killing tumor cells. However, the exact mechanism of their role in cancer is not fully understood. Therefore, a comprehensive understanding of the role of T-cell proliferation regulatory genes in tumors is needed. In our study, we investigated the expression levels of genes controlling T-cell proliferation, their impact on prognosis, and their genetic variations. Additionally, we explored their associations with TMB, MSI, ESTIMATEScore, ImmuneScore, StromalScore, and immune cell infiltration. We examined the role of these genes in cancer-related pathways using GSEA. Furthermore, we calculated their activity levels across various types of cancer. Drug analysis was also conducted targeting these genes. Single-cell analysis, LASSO Cox model construction, and prognosis analysis were performed. We observed distinct expression patterns of T-cell proliferation regulatory genes across different malignant tumors. Their abnormal expression may be caused by CNA and DNA methylation. In certain cancers, they also showed complex associations with TMB and MSI. Moreover, in many tumors, they exhibited significant positive correlations with ESTIMATEScores, ImmuneScore, and StromalScore. Additionally, in most tumors, their GSVA scores were significantly positively correlated with various T-cell subtypes. GSEA analysis revealed their involvement in multiple immune pathways. Furthermore, we found that model scores were associated with patient prognosis and related to tumor malignancy progression. T-cell proliferation regulatory genes are closely associated with the tumor immune microenvironment (TIM), especially T cells. Targeting them may be an essential approach for cancer immunotherapy.

A lipid/PLGA nanocomplex to reshape tumor immune microenvironment for colon cancer therapy

Immune checkpoint blockade therapy provides a new strategy for tumor treatment; however, the insufficient infiltration of cytotoxic T cells and immunosuppression in tumor microenvironment lead to unsatisfied effects. Herein, we reported a lipid/PLGA nanocomplex (RDCM) co-loaded with the photosensitizer Ce6 and the indoleamine 2,3-dioxygenase (IDO) inhibitor 1MT to improve immunotherapy of colon cancer. Arginine-glycine-aspartic acid (RGD) as the targeting moiety was conjugated on 1,2-distearoyl-snglycero-3-phosphoethanolamine lipid via polyethylene glycol (PEG), and programmed cell death-ligand 1 (PD-L1) peptide inhibitor DPPA (sequence: CPLGVRGK-GGG-d(NYSKPTDRQYHF)) was immobilized on the terminal group of PEG via matrix metalloproteinase 2 sensitive peptide linker. The Ce6 and 1MT were encapsulated in PLGA nanoparticles. The drug loaded nanoparticles were composited with RGD and DPPA modified lipid and lecithin to form lipid/PLGA nanocomplexes. When the nanocomplexes were delivered to tumor, DPPA was released by the cleavage of a matrix metalloproteinase 2-sensitive peptide linker for PD-L1 binding. RGD facilitated the cellular internalization of nanocomplexes via avβ3 integrin. Strong immunogenic cell death was induced by 1O2 generated from Ce6 irradiation under 660 nm laser. 1MT inhibited the activity of IDO and reduced the inhibition of cytotoxic T cells caused by kynurenine accumulation in the tumor microenvironment. The RDCM facilitated the maturation of dendritic cells, inhibited the activity of IDO, and markedly recruited the proportion of tumor-infiltrating cytotoxic T cells in CT26 tumor-bearing mice, triggering a robust immunological memory effect, thus effectively preventing tumor metastasis. The results indicated that the RDCM with dual IDO and PD-L1 inhibition effects is a promising platform for targeted photoimmunotherapy of colon cancer.

Genetic insights into carbohydrate sulfotransferase 8 and its impact on the immunotherapy efficacy of cancer

Immune checkpoint blockade (ICB) is a promising therapy for solid tumors, but its effectiveness depends on biomarkers that are not precise. Here, we utilized genome-wide association study to investigate the association between genetic variants and tumor mutation burden to interpret ICB response. We identified 16 variants (p < 5 × 10-8) probed to 17 genes on 9 chromosomes. Subsequent analysis of one of the most significant loci in 19q13.11 suggested that the rs111308825 locus at the enhancer is causal, as its A allele impairs KLF2 binding, leading to lower carbohydrate sulfotransferase 8 (CHST8) expression. Breast cancer cells expressing CHST8 suppress T cell activation, and Chst8 loss attenuates tumor growth in a syngeneic mouse model. Further investigation revealed that programmed death-ligand 1 (PD-L1) and its homologs could be sulfated by CHST8, resulting in M2-like macrophage enrichment in the tumor microenvironment. Finally, we confirmed that low-CHST8 tumors have better ICB response, supporting the genetic effect and clinical value of rs111308825 for ICB efficacy prediction.

Biology-aware mutation-based deep learning for outcome prediction of cancer immunotherapy with immune checkpoint inhibitors

The response rate of cancer immune checkpoint inhibitors (ICI) varies among patients, making it challenging to pre-determine whether a particular patient will respond to immunotherapy. While gene mutation is critical to the treatment outcome, a framework capable of explicitly incorporating biology knowledge has yet to be established. Here we aim to propose and validate a mutation-based deep learning model for survival analysis on 1571 patients treated with ICI. Our model achieves an average concordance index of 0.59 ± 0.13 across nine types of cancer, compared to the gold standard Cox-PH model (0.52 ± 0.10). The "black box" nature of deep learning is a major concern in healthcare field. This model's interpretability, which results from incorporating the gene pathways and protein interaction (i.e., biology-aware) rather than relying on a 'black box' approach, helps patient stratification and provides insight into novel gene biomarkers, advancing our understanding of ICI treatment.

Biomarker-Driven Personalization of Neoadjuvant Immunotherapy in Melanoma

The introduction of immunotherapy has ushered in a new era of anticancer therapy for many cancer types including melanoma. Given the increasing development of novel compounds and combinations and the investigation in earlier disease stages, the need grows for biomarker-based treatment personalization. Stage III melanoma is one of the front-runners in the neoadjuvant immunotherapy field, facilitating quick biomarker identification by its immunogenic capacity, homogeneous patient population, and reliable efficacy readout. In this review, we discuss potential biomarkers for response prediction to neoadjuvant immunotherapy, and how the neoadjuvant melanoma platform could pave the way for biomarker identification in other tumor types.

SIGNIFICANCE

In accordance with the increasing rate of therapy development, the need for biomarker-driven personalized treatments grows. The current landscape of neoadjuvant treatment and biomarker development in stage III melanoma can function as a poster child for these personalized treatments in other tumors, assisting in the development of new biomarker-based neoadjuvant trials. This will contribute to personalized benefit-risk predictions to identify the most beneficial treatment for each patient.

Immune checkpoint therapy for solid tumours: clinical dilemmas and future trends

Immune-checkpoint inhibitors (ICBs), in addition to targeting CTLA-4, PD-1, and PD-L1, novel targeting LAG-3 drugs have also been approved in clinical application. With the widespread use of the drug, we must deeply analyze the dilemma of the agents and seek a breakthrough in the treatment prospect. Over the past decades, these agents have demonstrated dramatic efficacy, especially in patients with melanoma and non-small cell lung cancer (NSCLC). Nonetheless, in the field of a broad concept of solid tumours, non-specific indications, inseparable immune response and side effects, unconfirmed progressive disease, and complex regulatory networks of immune resistance are four barriers that limit its widespread application. Fortunately, the successful clinical trials of novel ICB agents and combination therapies, the advent of the era of oncolytic virus gene editing, and the breakthrough of the technical barriers of mRNA vaccines and nano-delivery systems have made remarkable breakthroughs currently. In this review, we enumerate the mechanisms of each immune checkpoint targets, associations between ICB with tumour mutation burden, key immune regulatory or resistance signalling pathways, the specific clinical evidence of the efficacy of classical targets and new targets among different tumour types and put forward dialectical thoughts on drug safety. Finally, we discuss the importance of accurate triage of ICB based on recent advances in predictive biomarkers and diagnostic testing techniques.

A Comprehensive Benchmark of Transcriptomic Biomarkers for Immune Checkpoint Blockades

Immune checkpoint blockades (ICBs) have revolutionized cancer therapy by inducing durable clinical responses, but only a small percentage of patients can benefit from ICB treatments. Many studies have established various biomarkers to predict ICB responses. However, different biomarkers were found with diverse performances in practice, and a timely and unbiased assessment has yet to be conducted due to the complexity of ICB-related studies and trials. In this study, we manually curated 29 published datasets with matched transcriptome and clinical data from more than 1400 patients, and uniformly preprocessed these datasets for further analyses. In addition, we collected 39 sets of transcriptomic biomarkers, and based on the nature of the corresponding computational methods, we categorized them into the gene-set-like group (with the self-contained design and the competitive design, respectively) and the deconvolution-like group. Next, we investigated the correlations and patterns of these biomarkers and utilized a standardized workflow to systematically evaluate their performance in predicting ICB responses and survival statuses across different datasets, cancer types, antibodies, biopsy times, and combinatory treatments. In our benchmark, most biomarkers showed poor performance in terms of stability and robustness across different datasets. Two scores (TIDE and CYT) had a competitive performance for ICB response prediction, and two others (PASS-ON and EIGS_ssGSEA) showed the best association with clinical outcome. Finally, we developed ICB-Portal to host the datasets, biomarkers, and benchmark results and to implement the computational methods for researchers to test their custom biomarkers. Our work provided valuable resources and a one-stop solution to facilitate ICB-related research.

Predicting response and toxicity to immune checkpoint inhibitors in lung cancer using antibodies to frameshift neoantigens

PURPOSE

To evaluate a new class of blood-based biomarkers, anti-frameshift peptide antibodies, for predicting both tumor responses and adverse immune events to immune checkpoint inhibitor (ICI) therapies in advanced lung cancer patients.

EXPERIMENTAL DESIGN

Serum samples were obtained from 74 lung cancer patients prior to palliative PD-(L)1 therapies with subsequently recorded tumor responses and immune adverse events (irAEs). Pretreatment samples were assayed on microarrays of frameshift peptides (FSPs), representing ~ 375,000 variant peptides that tumor cells can be informatically predicted to produce from translated mRNA processing errors. Serum-antibodies specifically recognizing these ligands were measured. Binding activities preferentially associated with best-response and adverse-event outcomes were determined. These antibody bound FSPs were used in iterative resampling analyses to develop predictive models of tumor response and immune toxicity.

RESULTS

Lung cancer serum samples were classified based on predictive models of ICI treatment outcomes. Disease progression was predicted pretreatment with ~ 98% accuracy in the full cohort of all response categories, though ~ 30% of the samples were indeterminate. This model was built with a heterogeneous sample cohort from patients that (i) would show either clear response or stable outcomes, (ii) would be administered either single or combination therapies and (iii) were diagnosed with different lung cancer subtypes. Removing the stable disease, combination therapy or SCLC groups from model building increased the proportion of samples classified while performance remained high. Informatic analyses showed that several of the FSPs in the all-response model mapped to translations of variant mRNAs from the same genes. In the predictive model for treatment toxicities, binding to irAE-associated FSPs provided 90% accuracy pretreatment, with no indeterminates. Several of the classifying FSPs displayed sequence similarity to self-proteins.

CONCLUSIONS

Anti-FSP antibodies may serve as biomarkers for predicting ICI outcomes when tested against ligands corresponding to mRNA-error derived FSPs. Model performances suggest this approach might provide a single test to predict treatment response to ICI and identify patients at high risk for immunotherapy toxicities.

Screening and Verification of Key Ubiquitination Genes Related to Immune Infiltration in Stage III/IV Hepatocellular Carcinoma

Introduction

Immune checkpoint therapy (ICIs) effectively improves the prognosis of advanced (stage III/IV) hepatocellular carcinoma (HCC) patients. However, its objective response rate (ORR) is below 20%, significantly limiting ICI use in advanced HCC patients. The level of tumour immune infiltration influences ICI response rate. Recent studies have found ubiquitinase to be an important factor that regulates tumour immune infiltration. Therefore, the aim of this study is to explore the key ubiquitination genes that regulate immune infiltration in advanced HCC and further validate them.

Methods

A biotechnological process was performed as a means of classifying 90 advanced HCC patients into three immune subtypes and identifying associations with immune infiltration in the co-expressed modules. Ubiquitination-related genes were then screened with WGCNA. Gene enrichment analysis was performed for the target module and 30 hub genes were screened out by protein-protein interaction network (PPI). ssGSEA, single-gene sequencing and the MCP counter were used for exploring immune infiltration. TIDE score was applied for predicting drug efficacy and GSEA was used for exploring potential pathways. Finally, GRB2 expression in HCC tissue was validated by in vitro experiments.

Results

GRB2 expression was found to have a significant correlation with the pathological stage and prognosis of HCC patients and a positive correlation with immune infiltration and tumour mutation burden (TMB). In addition, significant correlations with the efficacy of ICIs, sorafenib and transarterial chemoembolization (TACE) were identified. GRB2 was found to be most significantly associated with the JAK-STAT signalling pathway and cytosolic DNA sensing pathway. Finally, it was found that GRB2 expression is closely related to the prognosis, tumour size and TMN stage.

Conclusion

A significant association was observed between the ubiquitinated gene GRB2 and the prognosis and immune infiltration of advanced HCC patients and it may potentially be used for predicting therapy efficacy in advanced HCC patients in the future.

Exome Sequencing of Paired Colorectal Carcinomas and Synchronous Liver Metastases for Prognosis and Therapy Prediction

PURPOSE

Analysis of somatic variant profiles in retrospectively collected pairs of primary tumors and synchronous liver metastases from surgically treated patients with colorectal carcinomas. Mutational profiles were compared between groups of patients stratified by response to chemotherapy and survival.

PATIENTS AND METHODS

The study used whole-exome sequencing of tumor sample pairs from 20 patients diagnosed and treated at a single center. The Cancer Genome Atlas COAD-READ data set (n = 380) was used for validation in silico, where possible.

RESULTS

The most frequently altered oncodrivers were APC (55% in primaries and 60% in metastases), TP53 (50/45), TRIP11 (30/5), FAT4 (20/25), and KRAS (15/25). Harboring variants with a high or moderate predicted functional effect in KRAS in primary tumors was significantly associated with poor relapse-free survival in both our sample set and the validation data set. We found a number of additional prognostic associations, including mutational load, alterations in individual genes, oncodriver pathways, and single base substitution (SBS) signatures in primary tissues, which were not confirmed by validation. Altered ATM, DNAH11, and MUC5AC, or a higher share of SBS24 signature in metastases seemed to represent poor prognostic factors, but because of a lack of suitable validation data sets, these results must be treated with extreme caution. No gene or profile was significantly associated with response to chemotherapy.

CONCLUSION

Taken together, we report subtle differences in exome mutational profiles between paired primary tumors and synchronous liver metastases and a distinct prognostic relevance of KRAS in primary tumors. Although the general scarcity of primary tumor-synchronous metastasis sample pairs with high-quality clinical data makes robust validation difficult, this study provides potentially valuable data for utilization in precision oncology and could serve as a springboard for larger studies.

Biological knowledge graph-guided investigation of immune therapy response in cancer with graph neural network

The determination of transcriptome profiles that mediate immune therapy in cancer remains a major clinical and biological challenge. Despite responses induced by immune-check points inhibitors (ICIs) in diverse tumor types and all the big breakthroughs in cancer immunotherapy, most patients with solid tumors do not respond to ICI therapies. It still remains a big challenge to predict the ICI treatment response. Here, we propose a framework with multiple prior knowledge networks guided for immune checkpoints inhibitors prediction-DeepOmix-ICI (or ICInet for short). ICInet can predict the immune therapy response by leveraging geometric deep learning and prior biological knowledge graphs of gene-gene interactions. Here, we demonstrate more than 600 ICI-treated patients with ICI response data and gene expression profile to apply on ICInet. ICInet was used for ICI therapy responses prediciton across different cancer types-melanoma, gastric cancer and bladder cancer, which includes 7 cohorts from different data sources. ICInet is able to robustly generalize into multiple cancer types. Moreover, the performance of ICInet in those cancer types can outperform other ICI biomarkers in the clinic. Our model [area under the curve (AUC = 0.85)] generally outperformed other measures, including tumor mutational burden (AUC = 0.62) and programmed cell death ligand-1 score (AUC = 0.74). Therefore, our study presents a prior-knowledge guided deep learning method to effectively select immunotherapy-response-associated biomarkers, thereby improving the prediction of immunotherapy response for precision oncology.

Feasibility of Longitudinal ctDNA Assessment in Patients with Uterine and Extra-Uterine Leiomyosarcoma

Background: Leiomyosarcomas (LMS) are aggressive malignancies with a propensity for early relapse. Current surveillance modalities include physical exam and imaging; however, radiological response to therapy may only manifest after 4-6 cycles of treatment. Herein, we evaluated the feasibility of longitudinal circulating tumor DNA (ctDNA) assessment in LMS patients to identify disease progression. Methods: We performed a retrospective review of patients with LMS who underwent treatment at Stanford Cancer Center between September 2019 and May 2022. ctDNA detection was performed using a personalized, tumor-informed ctDNA assay. Genomic analysis was conducted to characterize tumor mutation burden (TMB) and known driver mutations. Results: A total of 148 plasma samples were obtained from 34 patients with uterine (N = 21) and extrauterine (N = 13) LMS (median follow-up: 67.2 (19-346.3) weeks] and analyzed for ctDNA presence. Nineteen patients had metastatic disease. The most frequently mutated driver genes across sub-cohorts were TP53, RB1, and PTEN. Patients were stratified into four sub-cohorts (A-D) based on ctDNA kinetics. ctDNA levels tracked longitudinally with progression of disease and response to therapy. Conclusion: Our results indicate that while undetectable ctDNA may suggest a lower likelihood of relapse, ctDNA positivity may indicate progressive disease, enabling closer monitoring of patients for early clinical intervention.

Genomic and Transcriptomic Predictors of Response to Immune Checkpoint Inhibitors in Melanoma Patients: A Machine Learning Approach

Immune checkpoint inhibitors (ICIs) became one of the most revolutionary cancer treatments, especially in melanoma. While they have been proven to prolong survival with lesser side effects compared to chemotherapy, the accurate prediction of response remains to be an unmet gap. Thus, we aim to identify accurate clinical and transcriptomic biomarkers for ICI response in melanoma. We also provide mechanistic insight into how high-performing markers impose their effect on the tumor microenvironment (TME). Clinical and transcriptomic data were retrieved from melanoma studies administering ICIs from cBioportal and GEO databases. Four machine learning models were developed using random-forest classification (RFC) entailing clinical and genomic features (RFC7), differentially expressed genes (DEGs, RFC-Seq), survival-related DEGs (RFC-Surv) and a combination model. The xCELL algorithm was used to investigate the TME. A total of 212 ICI-treated melanoma patients were identified. All models achieved a high area under the curve (AUC) and bootstrap estimate (RFC7: 0.71, 0.74; RFC-Seq: 0.87, 0.75; RFC-Surv: 0.76, 0.76, respectively). Tumor mutation burden, GSTA3, and VNN2 were the highest contributing features. Tumor infiltration analyses revealed a direct correlation between upregulated genes and CD8+, CD4+ T cells, and B cells and inversely correlated with myeloid-derived suppressor cells. Our findings confirmed the accuracy of several genomic, clinical, and transcriptomic-based RFC models, that could further support the use of TMB in predicting response to ICIs. Novel genes (GSTA3 and VNN2) were identified through RFC-seq and RFC-surv models that could serve as genomic biomarkers after robust validation.