Pan-cancer Cohort Research Articles

Germline and somatic MUTYH mutations: Clinicopathologic and mutational analysis in a pan-cancer cohort.

10590 Background: MUTYH is a tumor suppressor gene involved in the base-excision DNA repair pathway. Biallelic pathogenic or likely pathogenic germline variants (gPV) in MUTYH are associated with MUTYH-associated polyposis (MAP), inclusive of colorectal carcinoma (CRC) and additional cancer risks. While monoallelic MUTYH gPV carriers do not have the same cancer risk as those with biallelic MAP, further work is needed to fully characterize secondary somatic events, including allelic imbalances. Methods: Using the MSK-IMPACT targeted NGS to detect both MUTYH gPVs and MUTYH somatic variants in patients (pts) with a solid tumor diagnosis, we analyzed clinicopathologic characteristics at cancer diagnosis, ancestry label as defined per cBioPortal, genomic biomarkers, as well as allele-specific copy number estimates using FACETS annotation. Results: Out of 38,907 samples with germline testing available, 587 pts with a solid tumor and presence of MUTYH gPV were identified. Of these, 15 pts had ≥2 MUTYH gPVs, including 4 pts with homozygous MUTYHgPVs (G396D, T179C, p.G480* and c.389-1G>C) and 11 pts with compound heterozygous MUTYHgPVs. In pts of European ancestry (n=403), the MUTYH G396D and T179C variants were predominant, present in 56% and 19% of cases, respectively. All patients with the R241W variant were listed as being of South Asian ancestry (n=6). Concurrent gPV in other genes were observed in 12% (n=69), with high/moderate-penetrance gPVs in 4% (n=24) including BRCA2 (n=6), BRCA1 (n=5), ATM (n=4) and PALB2 (n=3). Among all pts with MUTYH gPVs, 254/587 pts had concurrent tumor samples with FACETS results. Among pts with monoallelic MUTYH gPVs and FACETS annotation available, 24% of evaluable tumors exhibited biallelic MUTYH inactivation due to loss of heterozygosity (LOH). CRC (29% vs 13%; p=0.004) and cholangiocarcinoma (11% vs. 2%; p=0.04) were enriched in the biallelic MUTYH inactivation cohort vs. the cohort of pts with monoallelic MUTYH gPVs. Co-alterations in KRAS G12C and APC were enriched in the biallelic as compared to the monoallelic cohort (9% vs. 3%, p=0.04 and 27% vs. 12%, p=0.005, respectively). In addition, we identified 244 pts for whom FACETS annotation was available without MUTYH gPVs, but with somatic variants at MUTYH. Of these 244 pts, 117 pts had biallelic MUTYH inactivation either from homozygous deletions (n=87), somatic hit + LOH (n=27) or ≥2 somatic hits (n=3). The predominant cancer types with biallelic somatic MUTYH inactivation were breast cancer (16%), non-small cell lung cancer (14%) and CRC (11%). Conclusions: This study represents the largest cohort of pts with MUTYH gPVs and FACETS analysis on concurrent tumor samples, confirming a high prevalence of LOH in pts with monoallelic MUTYH gPVs and enrichment in KRAS G12Cin pts with MUTYH gPVs and biallelic MUTYH inactivation.

Use of CRTAM expression as a predictive biomarker for immune checkpoint blockade in a pan-cancer cohort.

Use of CRTAM expression as a predictive biomarker for immune checkpoint blockade in a pan-cancer cohort.

Integrative analyses of multi-omics data constructing tumor microenvironment and immune-related molecular prognosis model in human colorectal cancer

The increasing prevalence and incidence of colorectal cancer (CRC), particularly in young adults, underscore the imperative to comprehend its fundamental mechanisms, discover novel diagnostic and prognostic markers, and enhance therapeutic strategies. Here, we integrated multi-omics data, including gene expression, somatic mutation data and DNA methylation data, to unravel the intricacies of tumor microenvironment (TME) in CRC and search for novel prognostic markers. By calculating the immune score for each patient from the expression profile, we delineated the differential immune cell fraction, constructed an immune-related multi-omics atlas, and identified molecular characteristics. The entire colorectal dataset (n = 343) was randomly divided into training (n = 249) and testing datasets (n = 94). We screened 144 immune-related genes, 6 mutant genes, and 38 methylation probes associated with overall survival (OS). These makers were then incorporated into a 10-gene prognostic model using Lasso and Cox regression in the training dataset, and the model's performance was evaluated in an independent validation dataset. The model exhibited satisfactory results (average concordance index [C-index] = 0.77), with the average 1-year, 3-year, and 5-year AUCs being 0.79, 0.76, and 0.76 in the training dataset and 0.74, 0.80, and 0.90 in the testing dataset. Furthermore, the prognostic model demonstrated applicability in guiding chemotherapy for CRC patients and exhibited a degree of pan-cancer utility in risk stratification. In conclusion, our integrated analysis of multi-omics data revealed immune-related genetic and epigenetic characteristics of the TME. We propose an integrative prognostic model that can stratify risk and guide chemotherapy for CRC patients. The generalizability of the model in risk stratification across different cancer types was validated in Pan-Cancer cohort.

Synergistic inhibition of CXCL8 following MRTX1133 treatment with ONC212 or 5-FU in KRAS G12D and G12V mutant CRC and PDAC cell lines.

e15196 Background: CXCL8, serves as a crucial multifunctional cytokine that plays a significant role in regulating tumor growth, invasion, and migration. Although in some cancers the critical role of CXCL8 in patient prognosis and disease progression is stablished but its role in CRC & PDAC patients is under investigation. High levels of CXCL8 in tumor microenvironment (TME) and association with cancer stem cell survival, immune surveillance escape & metastasis make it an interesting therapeutic target. Both CRC & PDAC are among the deadliest cancers with dismal prognosis. The need for novel therapeutics with a focus on TME is emerging. MRTX1133 was identified as an inhibitor of KRASG12D. ONC212 is a fluorinated imipridone with strong anti-cancer activity in nM range and preclinical efficacy against pancreatic and other malignancies. Here we report the synergistic inhibition of CXCL8 in both KRAS G12D & KRAS G12V cell lines with use of KRAS G12D inhibitor in combination with 5-FU or ONC212. Methods: Different CRC & PDAC cells were treated with MRTX1133 with 5-FU or ONC212 cytokine profiling at 48h time point was done. Human protein atlas was used for protein and RNA data &TCGA data for survival was used. Results: We previously have shown the synergistic inhibition of pERK and cytokine alteration between MRTX1133 KRAS G12D inhibitor with 5FU or ONC212 in both cell lines with KRAS G12D & KRAS G12V mutation (Tajiknia V et al. AACR 2023 annual meeting). In LS513 KRAS G12D CRC cell line, treatment of ONC212, MRTX1133 and combination of both showed a decrease of 52.36%,58.83% and 71.65% respectively in CXCL8 levels compared to control at 48h time point. In HPAF-II KRAS G12V PDAC cell line, treatment of ONC212, MRTX1133 and combination of both resulted in 38.69%, 64.44%, 57% inhibition of CXCL8 levels respectively compared to control. In SW480 KRAS G12V CRC cell line, treatment of ONC212, MRTX1133 and combination of both demonstrated 25.7%,47.3% and 66.5% inhibition in CXCL8 level respectively compared to control at 48h time point. Single treatment of MRTX1133 caused a decline in CXCL8/IL-8 levels by 30% while single treatment of 5-FU showed a 36% decrease, the combination of both resulted in more than 55% decrease compared to control in LS513 KRAS G12D cell line. In SW480 KRAS G12V cell line, combination of 5-FU & MRTX1133 showed 30% inhibition while in Capan-2 PDAC KRAS G12V cells the same combination showed a 48.2% reduction in CXCL8 compared to control at 48h time point. TCGA data for RNA shows low specificity of CXCL8 for cancer type but it is significantly high in CRC and PDAC. Protein concentration in the Pan-cancer cohort shows significant high level of CXCL8 in CRC. Conclusions: Considering the important role of KRAS mutations and CXCL8 in pathogenesis of CRC & PDAC, the effective inhibition of CXCL8 with the use of targeted therapy in combination could enhance treatment response and patient survival.

Effect of common medications on immune checkpoint inhibitor efficacy.

e14710 Background: Previous studies have suggested that certain common medications could impact immune checkpoint inhibitor (ICI) efficacy and survival. These studies lacked non-ICI control groups, raising the possibility that survival differences could be explained by confounding. Methods: We developed a pan-cancer cohort of 27,998 patients who received ICIs in the national Veterans Health Administration (VHA) from 2010 to 2023. ICI patients were matched to 66,488 historical control patients treated with chemotherapy or targeted therapy before initial ICI regulatory approval in each cancer. We ascertained baseline use of drugs from 15 common medication classes within 1 year before therapy start and performed propensity-weighted Cox regression analyses to assess the impact of each medication class on overall survival (OS) and time to next treatment (TTNT). Interaction tests measured whether the impact of each medication class was significantly different in ICI versus historical control patients. Results: In the ICI cohort, the most common cancer types were NSCLC (46%) and melanoma (10%). After adjustment for multiple comparisons, only statins (adjusted hazard ratio [aHR] 0.93), PPIs (aHR 1.06), and loop diuretics (aHR 1.19) were associated with overall survival differences among ICI patients; nearly identical associations were observed in historical control patients (interaction p-values > 0.4 for all; Table). Similar results were seen for TTNT and in sensitivity analyses limiting to more recent medication users. Conclusions: Common medications do not appear to impact ICI efficacy. Previously described survival effects are likely the result of confounding. [Table: see text]

Prognostic and predictive value of ultrasensitive ctDNA monitoring in a metastatic pan-cancer cohort treated with immune checkpoint inhibitors in the context of phase 1 clinical trials.

2510 Background: Determining which patients will achieve clinical benefit from immune checkpoint inhibition (ICI) therapy remains an open question. Liquid biopsy tests to assess baseline and early dynamics of circulating tumor DNA (ctDNA) may allow clinicians to track response more granularly and predict ICI response or resistance prior to imaging. However, lack of sensitivity may hinder accurate detection of low ctDNA levels in tumors with low shedding rates or during substantial drops and clearance in response. Methods: This study represents a cohort of 175 patients with refractory metastatic tumors from 18 different cancer types, who received 1-3 successive lines of ICI treatment in the context of phase-1 clinical trials. Thus far, 609 plasma samples from 90 stage-IV cancer patients receiving immune checkpoint inhibitor (ICI) treatment have been processed. ctDNA was profiled using the NeXT Personal assay, a liquid biopsy platform that leverages whole-genome sequencing of tumor and normal samples to create custom, patient-specific panels that include up to 1,800 somatic variants. This enables the detection of molecular residual disease (MRD) down to a detection threshold of 1 part per million (PPM) of ctDNA. Results: NeXT Personal assay detected positive levels of ctDNA in 99% (81/82) of plasma baseline samples, with a wide dynamic range, extending from 4.2 PPM (~0.00042% TF) to ~640,000 PPM (~64% TF) (median limit of detection = 2 PPM, 0.0002% TF). Lower ctDNA values at baseline were correlated with increased duration of PFS (log-rank p=0.017, HR=0.57, 95% CI 0.36-0.91) and extended OS (log-rank p=0.002, HR=0.46, 95% CI 0.28-0.75). Early reduction in ctDNA level from baseline to treatment cycle 3 was associated with significant increases in PFS (log-rank p=0.001, HR=0.36, 95% CI 0.19-0.67) and OS (log-rank p=0.015, HR=0.44, 95% CI 0.22-0.87), representing a >3-fold increase in both median PFS and OS. ctDNA clearance resulted in significant improvement in both PFS and OS (PFS: log-rank p=0.002, HR=0.24, 95% CI 0.09-0.62; OS: log-rank p=0.01, HR=0.29, 95% CI 0.1-0.8). All plasma timepoints from periods of durable complete response (CR) assessed via RECIST 1.1 were ctDNA-negative, with a molecular clearance lead time of 277 days over radiographically confirmed CR. Conclusions: We demonstrate that early ctDNA dynamics are predictive of long-term ICI response in the advanced, pan-cancer ECT setting. Even in this refractory advanced cancer cohort, an ultra-sensitive assay was required for accurate MRD status determination, with low ctDNA detections down to 4.2 PPM that might otherwise have been missed. Taken together, these findings suggest a high sensitivity ctDNA assay is crucial for accurate monitoring of ICI response, providing information for more accurate patient management.

Pervasive loss of antigen presentation in solid tumors driven by loss of HLA class I expression.

e14654 Background: Human Leukocyte Antigens (HLA) play a key role in inducing responses to immune checkpoint inhibitor (ICI) therapy. HLA class I (HLA-I) expression is essential for antigen presentation in cancer cells and its loss leads to immune evasion. Expression of HLA-II shows presence of adaptive immune response via antigen presenting cells. To investigate the complex interplay of HLA class I/II in solid tumors, we study the landscape and association of several HLA phenotypes with biomarkers of tumor response and resistance to ICIs in a large pan-cancer cohort. Methods: A real-world cohort of 24,186 solid tumors was assessed by an immune expression profiling assay of 395 genes that were background subtracted, normalized, and ranked yielding a percentile rank value [0-100]. HLA Class I and class II expressions were estimated as average expression of HLA-A, HLA-B, HLA-C genes, and HLA HLA-DP, HLA-DQ, HLA-DR genes, respectively. Both HLA classes were labelled as high (≥66), moderate (≥33 & <66) or low (<33). We studied the distribution of HLA I/II phenotype defined as Ihi/IIhi comprised hot tumors with no loss of HLA-I and Ilo/IIhi defines immune evasion via loss of HLA-I in hot tumors. Ihi/IIlo phenotype comprises less inflamed tumors with high HLA-I expression and Ilo/IIlo comprises of cold tumors. We also performed differential expression of ICI targets including PD-1, PD-L1, LAG-3, TIM3 and expression-based signatures of inflammation and cell proliferation. Results: Loss of HLA-I expression was observed in 44.1% (n= 10669) of all solid tumors with most frequent loss observed in prostate (55%, n=450) cancer. Similarly, loss of HLA-II expression was observed in 34.7% (n=8396) of solid tumors with highest loss observed in neuroendocrine (59%; n=345) cancers. 49% of all tumors presented with four HLA-phenotypes, where 17.9% (n=4332) were characterized by Ilo/IIlo phenotype. Ihi/IIlo was present in 13% (n=3127) of all tumors. Both Ihi/IIhi and Ilo/IIhi phenotypes were significantly (p<0.001) associated with increased tumor inflammation. Other ICI targets such as PD-L1, TIM3, CTLA4 and LAG3 were significantly (p<0.001) overexpressed in this immune evasion phenotype compared to other phenotypes. Conversely, Ilo/IIlo phenotype was significantly associated with cold tumors (p<0.001), hinting at lack of adaptive immune response in these tumors. Conclusions: This study offers insight into pervasive loss of antigen presentation via HLA expression. Concurrent loss of HLA Class I expression with increased class II expression and additional ICI target co-expression is a potential immune escape mechanism in immunologically active solid tumors. Distinct and significant association of these MHC phenotypes with tumor inflammation, checkpoint markers and PD-L1 highlight the potential use of comprehensive immune profiling for further studying combination strategies to correct soft loss of HLA-I and ICI therapies.

Improved detection of homologous recombination deficiency using WES sequencing: Insights into a Chinese pan-cancer cohort.

e17503 Background: Homologous recombination (HR) deficiency has emerged as a novel biomarker for the clinical use of PARP inhibitors in advanced ovarian cancer, and may also be of value in guiding the clinical use of PARP inhibitors and platinum-based drugs in breast cancer, prostate cancer, and other tumors. This significant finding assists in precision medication and prognosis judgments for patients. Methods: This research incorporated a Chinese pan-cancer cohort comprising 1033 patients diagnosed across thirteen cancer types. All matched tumor tissues and peripheral blood samples were processed for the whole exome sequencing. The HRD scoring was conducted with ScarHRD R package with defined parameters for the heterozygosity (LOH), telomeric allelic imbalance (TAI), and large-scale state transitions (LST) status. HR deficiency were defined as an HRD-score ≥50 or BRCA1/2 deficient, while HR proficiency were defined as an HRD-score <50 and BRCA1/2 intact. Results: Ovarian cancer demonstrates the highest HRD prevalence at 51.85%, surpassing the BRCA1/2 mutation rate of 20.37%. Breast cancer follows with an HRD rate of 45.45%, yet no BRCA1/2 mutations are detected. Other cancers like pancreatic, gastric, and sarcomas also present notable HRD rates (29.41%, 36.84%, and 29.79%, respectively) without corresponding BRCA1/2 mutations. In contrast, cancers such as urological, lung, intestinal, and neurological tumors exhibit a mismatch between HRD rates and BRCA1/2 mutation presence, few patients with notable BRCA1/2 mutations did not have consisted HR-deficient rate, and vice versa. Conclusions: Analysis HRD-score can help to screen a larger population of PARP inhibitor beneficiaries and provide hope to more patients. In the future we will do more dimensional data analysis statistics on this cohort. [Table: see text]

Relationship between dynamic changes in circulating tumor fraction and real-world imaging with real-world survival in patients with solid tumors treated with immunotherapy.

3046 Background: There is an unmet need for a sensitive biomarker that can determine which patients will confer long term benefit from immune checkpoint inhibitors (ICI). While radiologic imaging is the current standard of care for assessing ICI response, imaging can be difficult to interpret and is typically only assessed at 3 month intervals. Dynamic changes in circulating tumor fraction (TF) is a potential biomarker for monitoring ICI response. Here we evaluate the clinical benefit of a molecular biomarker, xM for treatment response monitoring (TRM), in addition to imaging, in a real-world (rw), pan-cancer cohort treated with ICIs. Methods: Using the Tempus liquid biopsy assay, ctDNA TF was computed by applying an ensemble algorithm incorporating copy number variant data and pathogenic variant allele frequencies. Deidentified patient records from the Tempus multimodal database were analyzed if patients had a baseline xF test ≤ 40 days prior to the start of ICI (alone or in combination with chemotherapy [CT], median = 13 days) and an on-treatment xF test 15-180 days post-ICI (median=93 days) and clinical response evaluated via imaging 15-126 days post ICI (median=77 days). Molecular responders (MR) were defined as patients with a ≥ 50% decrease in TF between tests. Rw-imaging, as documented by a provider in their clinical notes, was categorized as complete response (CR), partial response (PR), stable disease (SD), or progressive disease (PD), Rw overall survival (rwOS) was defined as time from the imaging date closest to on-treatment xF test to death. A likelihood ratio test at a two-sided 5% significance level was performed to assess if a Cox proportional hazards model incorporating both xM Monitor and rw-imaging as covariates (full model) had predictive power over a reduced model with only rw-imaging as a covariate. Results: The evaluable cohort (N=51) – 53% female, median age 64 yrs – had >10 advanced solid tumors (NSCLC 29% and breast cancer 18%). The majority of patients (69%) received ICI + CT, while 31% received ICI alone. 24 patients achieved MR (79% on ICI monotherapy, 21% on ICI + CT). Concordance of response markers is shown in the table. In the full model, xM was a significant predictor of rwOS (HR=0.33 [0.13, 0.84], p=0.02). The added information provided from xM was significant and superior compared to the reduced model of only rw-imaging (p=0.02). For patients with nMR and CR/PR/SD predicted median survival was 10 months whereas for patients with MR and CR/PR/SD predicted median survival was 16 months. Conclusions: MRs achieved significantly longer rwOS than nMRs and provides significant power in predicting rwOS beyond rw-imaging response. xM used for TRM can help identify nMRS who can discontinue therapy and be spared ICI toxicity. [Table: see text]

Therapeutic implications of phosphoproteomics in molecular cancer diagnostics.

3068 Background: Precision oncology approaches employing genomics-guided targeted therapies for individual patients have provided significant survival benefits in several cancer types. However, varying response rates in solid malignancies, many patients without actionable genomic lesions, and increasing evidence that non-genomic mechanisms may play an important role in tumors indicate that genomics alone is often insufficient to inform and guide the clinical care of patients. Since most targeted anti-cancer drugs inhibit the activity of protein kinases, measuring the tumor phosphoproteome as a direct readout of oncogenic pathway activity is poised to enhance molecular stratification. Methods: We established a clinical (phospho)proteomic workflow for integration into the molecular tumor board (MTB) workflow of the Germany-wide INFORM (for children with relapsed cancers) and MASTER (for young adults with refractory cancers and patients with rare tumors) registry studies and profiled > 1000 tumor tissue specimen from patients enrolled in the DKFZ/NCT/DKTK MASTER study or the INFORM registry trial. To assess the aberrant activity of druggable signaling pathways, we developed a new tumor pathway activity (TUPAC) scoring methodology using expression data of > 8,000 proteins and > 20,000 phospho-sites per patient in a heterogeneous pan-cancer cohort. TUPAC scores integrate protein expression and phosphopeptide abundance data at several levels to detect aberrant signaling pathway activities of several interdependent oncogenic signaling pathways within one tumor specimen. Results: We show for the first time that comprehensive (phospho)proteome profiling is feasible and informative in a real-world prospective precision oncology setting. Discussion of (phospho)proteomic data of > 500 prospective patients in weekly MTB meetings revealed that adding a (phospho)proteomic layer can supply critical information for personalized therapies that is not discernible from genomic and transcriptomic data. In an independent value evaluation (329 target recommendations in 104 patients), the phosphoproteome layer was decisive in 22% of all recommended targets. Based on anecdotal cases with available clinical follow-up, we provide evidence that (phospho)proteome profiling carries important diagnostic value by detecting actionable tumor-driving kinase signaling in patients without actionable genomic lesions or by functionalizing genomic variants of unknown significance. Conclusions: Measuring the tumor phosphoproteome as a direct readout of oncogenic pathway activity is feasible and adds complementary, therapy-relevant information in a real-world prospective precision oncology setting.

Use of a tissue-free epigenomic circulating tumor DNA (ctDNA) assay for quantification of tumor fraction (TF) and association with outcomes from RADIOHEAD real-world advanced pan-cancer cohort.

Use of a tissue-free epigenomic circulating tumor DNA (ctDNA) assay for quantification of tumor fraction (TF) and association with outcomes from RADIOHEAD real-world advanced pan-cancer cohort.

Integrated analysis of single-cell and bulk RNA sequencing data reveals a myeloid cell-related regulon predicting neoadjuvant immunotherapy response across cancers

BackgroundImmunotherapy has brought about a paradigm shift in the treatment of cancer. However, the majority of patients exhibit resistance or become refractory to immunotherapy, and the underlying mechanisms remain to be explored.MethodsSing-cell RNA sequencing (scRNA‑seq) datasets derived from 1 pretreatment and 1 posttreatment achieving pathological complete response (pCR) patient with lung adenocarcinoma (LUAD) who received neoadjuvant immunotherapy were collected, and pySCENIC was used to find the gene regulatory network (GRN) between cell types and immune checkpoint inhibitor (ICI) response. A regulon predicting ICI response was identified and validated using large‑scale pan-cancer data, including a colorectal cancer scRNA‑seq dataset, a breast cancer scRNA‑seq dataset, The Cancer Genome Atlas (TCGA) pan-cancer cohort, and 5 ICI transcriptomic cohorts. Symphony reference mapping was performed to construct the myeloid cell map.ResultsThirteen major cluster cell types were identified by comparing pretreatment and posttreatment patients, and the fraction of myeloid cells was higher in the posttreatment group (19.0% vs. 11.8%). A PPARG regulon (containing 23 target genes) was associated with ICI response, and its function was validated by a colorectal cancer scRNA‑seq dataset, a breast cancer scRNA‑seq dataset, TCGA pan-cancer cohort, and 5 ICI transcriptomic cohorts. Additionally, a myeloid cell map was developed, and cluster I, II, and III myeloid cells with high expression of PPARG were identified. Moreover, we constructed a website called PPARG (https://pparg.online/PPARG/ or http://43.134.20.130:3838/PPARG/), which provides a powerful discovery tool and resource value for researchers.ConclusionsThe PPARG regulon is a predictor of ICI response. The myeloid cell map enables the identification of PPARG subclusters in public scRNA-seq datasets and provides a powerful discovery tool and resource value.

Establishment of homotrimer collagen type I signature and its association with clinical manifestation and tertiary lymphoid structures formation in liver cancer

Backgroundcollagen type I is a fundamental composition of extracellular matrix. Typically it exists in the form of a heterotrimer, consisting of two α1 chains encoded by COL1A1 and one α2 chain encoded by COL1A2. However, in cancer a homotrimeric form of collagen type I comprises three α1 chains encoded by COL1A1 was founded. There is still a lack of transcriptional and histologic methods for detecting homotrimeric collagen type I. Furthermore, a comprehensive analysis of the pan-cancer distribution pattern and clinical relevance of homotrimeric collagen type I is conspicuously absent. MethodUsing transcriptional and immunoflourance method, we established homocol signature, which is able to transcriptionally and histologically detect homotrimeric collagen type I. We investigated the diagnostic and prognostic potential of homocol as a novel cancer biomarker in a pan-cancer cohort. Furthermore, we assessed its association with clinical manifestations in a liver cancer cohort undergoing treatment at our institute. ResultHomotrimer Collagen Type I is predominantly expressed by cancer cells and is linked to several critical cancer hallmarks, particularly inflammatory response and proliferation. Survival analyses have indicated that a high Homocol expression is correlated with poor outcomes in most types of cancer studied. In terms of cancer detection, Homocol demonstrated strong performance in Receiver Operating Characteristic (ROC) analysis, with an Area Under Curve (AUC) of 0.83 for pan-cancer detection and between 0.72 and 0.99 for individual cancers.In cohorts undergoing PD1 treatment, we noted a higher presence of Homocol in the response group. In a Hepatocellular Carcinoma (HCC) clinical set, high Homocol expression was associated with an increased formation of intra-tumor tertiary lymphoid structures (TLS), larger tumor sizes, more advanced Barcelona Clinic Liver Cancer (BCLC) stages, higher microvascular invasion (MVI) grades, absence of a capsule, and an enriched para-tumor collagen presence. Conclusionour research has led to the development of a novel gene signature that facilitates the detection of Homotrimer Collagen Type I. This may greatly assist efforts in cancer detection, prognosis, treatment response prediction, and further research into Homotrimer Collagen Type I.

PCDH11X mutation as a potential biomarker for immune checkpoint therapies in lung adenocarcinoma.

Immune checkpoint inhibitors (ICIs) have achieved impressive success in lung adenocarcinoma (LUAD). However, the response to ICIs varies among patients, and predictive biomarkers are urgently needed. PCDH11X is frequently mutated in LUAD, while its role in ICI treatment is unclear. In this study, we curated genomic and clinical data of 151 LUAD patients receiving ICIs from three independent cohorts. Relations between PCDH11X and treatment outcomes of ICIs were examined. A melanoma cohort collected from five published studies, a pan-cancer cohort, and non-ICI-treated TCGA-LUAD cohort were also examined to investigate whether PCDH11X mutation is a specific predictive biomarker for LUAD ICI treatment. Among the three ICI-treated LUAD cohorts, PCDH11X mutation (PCDH11X-MUT) was associated with better clinical response compared to wild-type PCDH11X (PCDH11X-WT). While in ICI-treated melanoma cohort, the pan-cancer cohort excluding LUAD, and the non-ICI-treated TCGA-LUAD cohort, no significant differences in overall survival (OS) were observed between the PCDH11X-MUT and PCDH11X-WT groups. PCDH11X mutation was associated with increased PD-L1 expression, tumor mutation burden (TMB), neoantigen load, DNA damage repair (DDR) mutations, and hot tumor microenvironment in TCGA-LUAD cohort. Our findings suggested that the PCDH11X mutation might serve as a specific biomarker to predict the efficacy of ICIs for LUAD patients. Considering the relatively small sample size of ICI-treated cohorts, future research with larger cohorts and prospective clinical trials will be essential for validating and further exploring the role of PCDH11X mutation in the context of immunotherapy outcomes in LUAD. KEY MESSAGES: PCDH11X mutation is associated with better clinical response compared to wild type PCDH11X in three ICIs-treated LUAD cohorts. In ICIs-treated melanoma cohort, the pan-cancer cohort excluding LUAD, and non-ICIs-treated TCGA-LUAD cohortsPCDH11X mutation is not associated with better clinical response, suggesting PCDH11X mutation might be a specific biomarker to predict the efficacy of ICIs treatment for LUAD patients. PCDH11X mutation is associated with increased PD-L1 expression, tumor mutation burden, and neoantigen load in TCGA-LUAD cohort. PCDH11X mutation is associated with hot tumor microenvironment in TCGA-LUAD cohort.

Mapping the pancancer metastasis tumor microbiome

The landscape of the intratumoral microbiome in tumor metastases is largely unchartered. In this issue of Cell, Voest etal. profiled the tumor metastasis-associated microbiome in a pancancer cohort of 4,160 biopsies from 26 cancer types. This dataset offers a useful resource for understanding the role of the microbiome in metastatic cancers.

Inference of Developmental Hierarchy and Functional States of Exhausted T Cells from Epigenetic Profiles with Deep Learning.

Exhausted T cells are a key component of immune cells that play a crucial role in the immune response against cancer and influence the efficacy of immunotherapy. Accurate assessment and measurement of T-cell exhaustion (TEX) are critical for understanding the heterogeneity of TEX in the tumor microenvironment (TME) and tailoring individualized immunotherapeutic strategies. In this study, we introduced DeepEpiTEX, a novel computational framework based on deep neural networks, for inferring the developmental hierarchy and functional states of exhausted T cells in the TME from epigenetic profiles. DeepEpiTEX was trained using various modalities of epigenetic data, including DNA methylation data, microRNA expression data, and long non-coding RNA expression data from 30 bulk solid cancer types in the TCGA pan-cancer cohort, and identified five optimal TEX subsets with significant survival differences across the majority of cancer types. The performance of DeepEpiTEX was further evaluated and validated in external multi-center and multi-type cancer cohorts, consistently demonstrating its generalizability and applicability in different experimental settings. In addition, we discovered the potential relationship between TEX subsets identified by DeepEpiTEX and the response to immune checkpoint blockade therapy, indicating that individuals with immune-favorable TEX subsets may experience the greatest benefits. In conclusion, our study sheds light on the role of epigenetic regulation in TEX and provides a powerful and promising tool for categorizing TEX in different disease settings.

Abstract 2328: Next-generation sequencing-based regression algorithm to determine homologous recombination deficiency scores in a pan-cancer cohort

Abstract Background: Homologous recombination deficiency (HRD) inhibits double strand breaks from being repaired in DNA, leading to cancer cells failing to recover themselves and cell death. Previous studies found that HRD-positive ovarian cancer patients had more significant clinical benefits from poly ADP-ribose polymerase inhibitors (PARPi) treatment. For accurate detection of HRD, next-generation sequencing (whole exome, whole genome) was used to identify large-scale genomic aberrations including telomeric allelic imbalance (HRD-TAI score), loss of heterozygosity profiles (HRD-LOH score), and large-scale state transitions (HRD-LST score). So far, many studies (HRDetect, SigMA, CHORD and shallowHRD) have been conducted to accurately determine HRD in a pan-cancer cohort. As the study progressed, it has been revealed that PARPi is effective when applied to patients with HRD in not only ovarian cancer but also breast cancer, prostate cancer, and pancreatic cancer. Purpose: To develop an algorithm that accurately predict HRD score in patients with ovarian cancer, breast cancer, prostate cancer, and pancreatic cancer utilizing a machine learning algorithm. Methods: We used whole-genome sequencing (WGS) data of 710 samples from 309 patients, whole-exome sequencing (WES) data of 4,650 samples from 2,193 patients from the pan-cancer cohort of the TCGA (TCGA-OV, TCGA-BRCA, TCGA-PRAD, and TCGA-PAAD). The HRD-TAI/HRD-LST/HRD-LOH scores were calculated through structural variation analyses. After data cleaning processes, machine-learning models were trained and tested on 228 out of 2,502 samples, and validation was performed on 1,248 out of 2,502 samples from the TCGA dataset. Results: To assess the performance of the machine-learning regressor, the concordance between predictions and annotations was quantified by calculating the R squared (𝑅;;;^2). As a result of training using machine learning algorithm, we achieved a high 𝑅;;;^2 (0.904) with a RMSE (root mean squared error) score (7.649) for a pan-cancer cohort. Conclusions: Our regressor was robust and accurate when applied to 4 cancer types. Using our systematic pan-cancer analysis, we found novel insights into the mechanisms of HRD across cancer types with potential contribution to clinical practice. Citation Format: Boram Choi, Youn Jin Choi. Next-generation sequencing-based regression algorithm to determine homologous recombination deficiency scores in a pan-cancer cohort [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 2328.

Abstract 5145: Association of a ctDNA biomarker of treatment response with clinical outcomes in a real-world pan-cancer cohort treated with tyrosine kinase inhibitors

Abstract Clinical evidence suggests that early changes in circulating tumor DNA tumor fraction (ctDNA TF) is predictive of patient response to immune checkpoint inhibitors, but the extent to which this approach can be used for patients on Tyrosine Kinase Inhibitors (TKIs) is not known. Here, we use xF Monitor to estimate ctDNA TF in patients that received TKIs and show that changes in circulating TF correlate with outcomes in a real-world pan-cancer cohort. Tempus xF Monitor is a ctDNA assay that measures quantitative molecular changes in ctDNA TF by utilizing diverse genomic events, dynamically weighting somatic variant allele frequencies and copy number variants, while using germline information to inform these estimates and providing single nucleotide variant (SNV) results on 105 genes. Molecular responders (MR) were defined as patients with ≥50% reduction in ctDNA TF between baseline and on-treatment time points, consistent with our previously established threshold. Deidentified patient records from the Tempus multimodal database were analyzed if patients had an xF test ≤ 15 weeks prior to the start of their first TKI regimen and an xF test 15 -180 days post-TKI initiation. Patients received TKI agents as classified in the Tempus Medical Ontology, which assigns a unique ID to key clinical concepts, validated and harmonized across industry standard terminologies. The TKI list was further validated using the NCI Metathesaurus (https://ncim.nci.nih.gov/ncimbrowser/) TKI concept (NCI Code C1967). Clinical endpoints were defined from TKI start to the first progression event or death (rwPFS) or death (rwOS), both censored on the last follow-up in event-free patients. Kaplan Meier analysis and Cox proportional hazard models were fitted to compare MR status and survival outcomes on TKI. The evaluable pan-cancer cohort N=69, median age = 61.5 yo, advanced/metastatic disease = 72%, 65% were female, consisted of > 10 cancer diagnoses (35% NSCLC, 26% BC), 41% 1L. In this cohort, 49% of patients received a TKI combination regimen, 21% with CT, 18% with ICI. 68% of patients with NGS testing received matched targeted mutation therapies based on NCCN guidelines. Patients classified as MR based on xF Monitor results had improved survival outcomes with a median rwOS of 951 days (vs 439 days for non-MRs). MR patients had significantly longer rwPFS (HR = 0.383, p = 0.005, CI = 0.12 - 0.75) and rwOS (HR = 0.205, p = 0.004, CI = 0.07 - 0.61) than non-MRs. xF Monitor is a novel tumor naive serial quantitative ctDNA TF algorithm off the Tempus xF assay that has the potential to be used clinically as a monitoring biomarker to stratify patients who are likely to benefit from TKI therapy. xF monitor can be used as a strategy to identify patients at a high risk of progression who could benefit from early switch or intensifying therapy. Results need to be prospectively validated in a larger cohort. Citation Format: Akash Mitra, John Guittar, Regina Schwind, Terri Driessen, Matthew Berginski, Christine Lo, Michelle M. Stein, Jonathan Freaney, Seung Won Hyun, Chithra Sangli, Kate Sasser, Halla Nimeiri, Aadel Chaudhuri. Association of a ctDNA biomarker of treatment response with clinical outcomes in a real-world pan-cancer cohort treated with tyrosine kinase inhibitors [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 5145.

Abstract 6399: Real-world experience of an Asian pan-cancer combined DNA/RNA next generation sequencing (NGS) tissue biopsy assay

Abstract Introduction: Comprehensive molecular profiling with combined DNA-RNA next-generation sequencing (NGS) on tumor tissue is now increasingly being used as standard-of-care for sequence therapy selection to improve cancer survival. Combined DNA-RNA NGS is also considered standard-of-care for central nervous system (CNS) tumor classification. Despite the genetic heterogeneity and diversity of cancers, most commercial combined DNA/RNA NGS assays are validated in Western populations, resulting in knowledge gaps in patient outcomes. We report here for the first time a range of actionable outcomes from a comprehensive DNA/RNA tissue NGS assay designed specifically to probe Asian-prevalent cancers and biomarkers in an international multicenter real-world study in Asia. Methods: 1,002 formalin-fixed paraffin-embedded tissue samples from 995 Asian cancer patients across >70 centers in 7 countries and/or territories underwent real-world DNA/RNA NGS testing in a CAP-accredited CLIA-certified laboratory. Genomic alterations including SNVs, INDELS, gene fusions, CNVs, TMB, and MSI were analyzed using an Asian-centric pan-cancer hybrid-capture NGS panel that comprehensively profiles 572 cancer-related genes and 91 RNA fusion partners (UNITED). Biomarkers were considered actionable if the therapeutic indication in any solid tumor or hematological cancer was FDA-approved, guidelines-recommended, or supported by well-powered phase III clinical trials. Results: 27 cancer types were represented in 1,002 tissue samples, with lung, breast, colorectal, and prostate cancers comprising 45.5% of clinical volume. 79.5% of samples harbored ≥1 actionable biomarker, with the most frequently altered genes being TP53, KRAS, PIK3CA, EGFR, and NF1. Overall, 42.6% of samples had ≥1 biomarker with an indication in the patient’s cancer type, including 81.5% of lung, 59.0% of breast, 61.8% of colorectal, and 42.1% of prostate cancers. Across the pan-cancer cohort, 13.3% was TMB-high and 1.9% was MSI-high. 26.3% of samples were from cancer types outside the original assay design; the top 3 were pancreatic cancer (7.3%), sarcoma (3.3%), and CNS tumor (3.1%). Cancers of unknown primary (CUP) also constituted a significant proportion (3.8%). In CNS tumors, 38.7% harbored an actionable mutation in IDH1, BRAF, or H3-3A. In CUP, 78.8% harbored an actionable mutation; 50% of samples harbored biomarkers that additionally informed potential tumor origin, including mutations in EGFR, MET, ALK, ROS1, and BRCA1. Conclusion: The use of an Asian-focused pan-cancer DNA/RNA NGS panel resulted in detection of actionable biomarkers in 79.5% of clinical cases. These findings support the utility of DNA/RNA NGS assays in informing standard-of-care therapeutic decisions in Asian cancer patients for maximizing cancer survival. Citation Format: Jason Yongsheng Chan, Jing Yi Lee, Zi Yi Wan, Abner Herbert Lim, Tun Kiat Ko, Cedric Chuan-Young Ng, Donald Poon, Jens Samol, Tsz Him So, Su Pin Choo, Ravindran Kanesvaran, Cheng-Vai Hui, Joseph Siu-Kie Au, Aya El Helali, Timothy Yip, Vikneswari Rajasegaran, Sandy Lim, Ruifen Weng, Puay Hoon Tan, Bin Tean Teh, Min-Han Tan, Jonathan Poh. Real-world experience of an Asian pan-cancer combined DNA/RNA next generation sequencing (NGS) tissue biopsy assay [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6399.

Abstract 2796: Colibactin mutation signatures are associated with younger age of onset in colorectal cancer

Abstract Escherichia coli (E. coli) strains carrying the polyketide synthase (pks) island produce the genotoxin colibactin associated with pathogenesis of colorectal cancer (CRC). Colibactin binds directly to DNA, causing alkylating damage leading to double stranded DNA breaks. Colibactin mutagenesis results in mutations in hexameric A-T rich DNA motifs that can be identified by whole genome sequencing (WGS) as mutation signatures SBS28 and SBS88 (SBS-pks). Since WGS is not routinely used clinically, we developed an approach to specifically identify SBS-pks using a clinical targeted exon capture assay, MSK-IMPACT. In an institutional pan-cancer clinical cohort of 78,905 tumors we identified 15,967 samples with > 20 mutations detected by MSK-IMPACT. We identified 150/1,884 samples with microsatellite stable CRC with >10% mutations attributable to SBS-pks. Examining clinicopathologic and genomic features of SBS-pks+ compared to SBS-pks- CRC tumors, we found that SBS-pks+ signatures are significantly enriched in younger age of onset compared to SBS-pks- CRC patients (median age 55 vs median age 59, p= 0.0028). SBS-pks+ CRC exhibit similar driver mutations and tumor mutational burden (TMB) compared to SBS-pks- CRC but are associated with increased fraction of genome altered (FGA). To mechanistically investigate colibactin-associated damage we developed a novel long-term human colon organoid-microbe co-culture model system that takes advantage of reversal of polarity in suspension culture. WGS of human CRC organoids co-cultured for three months with pks+ and delta-pks E. coli NC101 confirmed induction of pks+ specific SBS-pks signatures, validating the model. Leveraging this model and a biobank of normal colon organoids from donors aged </> 50 years, we are investigating the molecular mechanisms underlying the association of SBS-pks with younger age of onset CRC. Citation Format: Stefanie Gerstberger, Melissa Lumish, Saskia Hartner, Farheen Shah, Seongmin Choi, Anisha Luthra, Qingwen Jiang, Hyung Jun Woo, Ahmed Mahmoud, Henry Walch, Simran Asawa, Mark Donoghue, Andrea Cercek, Rona Yaeger, Andrew McPherson, Francisco Sanchez-Vega, Karuna Ganesh. Colibactin mutation signatures are associated with younger age of onset in colorectal cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 2796.