Primary tumor type prediction based on US nationwide genomic profiling data in 13,522 patients

Timely and accurate primary tumor diagnosis is critical, and misdiagnoses and delays may cause undue health and economic burden. To predict primary tumor types based on genomics data from a de-identified US nationwide clinico-genomic database (CGDB), the XGBoost-based Clinico-Genomic Machine Learning Model (XC-GeM) was developed to predict 13 primary tumor types based on data from 12,060 patients in the CGDB, derived from routine clinical comprehensive genomic profiling (CGP) testing and chart-confirmed electronic health records (EHRs). The SHapley Additive exPlanations method was used to interpret model predictions. XC-GeM reached an outstanding area under the curve (AUC) of 0.965 and Matthew's correlation coefficient (MCC) of 0.742 in the holdout validation dataset. In the independent validation cohort of 955 patients, XC-GeM reached 0.954 AUC and 0.733 MCC and made correct predictions in 77% of non-small cell lung cancer (NSCLC), 86% of colorectal cancer, and 84% of breast cancer patients. Top predictors for the overall model (e.g. tumor mutational burden (TMB), gender, and KRAS alteration), and for specific tumor types (e.g., TMB and EGFR alteration for NSCLC) were supported by published studies. XC-GeM also achieved an excellent AUC of 0.880 and positive MCC of 0.540 in 507 patients with missing primary diagnosis. XC-GeM is the first algorithm to predict primary tumor type using US nationwide data from routine CGP testing and chart-confirmed EHRs, showing promising performance. It may enhance the accuracy and efficiency of cancer diagnoses, enabling more timely treatment choices and potentially leading to better outcomes.

Prior anti-CTLA-4 therapy impacts molecular characteristics associated with anti-PD-1 response in advanced melanoma

Immune checkpoint inhibitors (ICIs), including CTLA-4- and PD-1-blocking antibodies, can have profound effects on tumor immune cell infiltration that have not been consistent in biopsy series reported to date. Here, we analyze seven molecular datasets of samples from patients with advanced melanoma (N = 514) treated with ICI agents to investigate clinical, genomic, and transcriptomic features of anti-PD-1 response in cutaneous melanoma. We find that prior anti-CTLA-4 therapy is associated with differences in genomic, individual gene, and gene signatures in anti-PD-1 responders. Anti-CTLA-4-experienced melanoma tumors that respond to PD-1 blockade exhibit increased tumor mutational burden, inflammatory signatures, and altered cell cycle processes compared with anti-CTLA-4-naive tumors or anti-CTLA-4-experienced, anti-PD-1-nonresponsive melanoma tumors. We report a harmonized, aggregate resource and suggest that prior CTLA-4 blockade therapy is associated with marked differences in the tumor microenvironment that impact the predictive features of PD-1 blockade therapy response.

Abstract 3818: Infrequent chromosomal loss and recurrent gains lead to imbalanced expression of HLA genes in melanoma

Introduction: Loss-of-heterozygosity (LOH) events in chromosome 6p, comprising the human leukocyte antigen (HLA) genes, have been reported in about 10% of cutaneous melanoma (compared to 20-40% of squamous cell carcinomas), while copy number gains in this region have been observed in over 50% of melanoma. Recent studies focused in HLA allelic loss have been restricted to DNA-based approaches, and have not been validated orthogonally at the RNA or protein levels. Here, using clinical melanoma biopsies and patient-derived melanoma cell lines, we show that genetic alterations in HLA genes results in imbalanced allele expression, subsequently skewing antigen presentation by melanoma cells.

Methods: Whole exome and RNA sequencing (WES, RNAseq) analyses were performed on 760 melanoma biopsies and 60 patient-derived melanoma cell lines. Patient-matched normal WES was used to perform HLA haplotyping across Class I and II HLA genes. Tumor WES was analyzed for copy number alterations in chromosome 6, identifying which alleles were lost or gained. Differential expression of HLA alleles was quantified in tumor RNAseq, correlating the allelic imbalance at the DNA and RNA levels. Melanoma cell lines heterozygous for HLA-A*02, A*03, and A*24 were analyzed by flow cytometry for surface-level HLA protein expression using allele-specific antibodies to quantify allelic densities and compare the imbalance of HLA-A alleles at the DNA, RNA, and protein levels.

Results: Across 760 melanoma biopsies, copy number alterations in chromosome 6p were identified in 76% of tumors; 12% had LOH, and 54% had copy number gains that resulted in imbalanced copies of alleles. In paired tumor WES and RNAseq (N=682), genetic imbalance was correlated with imbalanced expression of HLA alleles in the classical Class I HLA genes (Spearman rho=0.64-0.7; p=2.2e-16); this association was strengthened in tumors with high tumor cellularity, and was not associated with the total expression of the HLA genes.These patterns were explored in a 60 patient-derived melanoma cell lines with matched tumor WES and RNAseq, confirming that alleles gained at the genetic level were also expressed at higher levels than alleles that were not gained or lost. In 10 cell lines heterozygous for either HLA-A*02, A*03, or A*24, allelic imbalance at the DNA and RNA level resulted in correlative imbalanced surface presentation of alleles at the protein level.

Conclusions: Evaluation of paired tumor WES and RNAseq revealed orthogonal validation of HLA allelic imbalance, and analysis in cell lines suggested that these patterns were likely tumor intrinsic. Experimental validation of these findings at the protein level suggests that antigen presentation density can be modulated by chromosomal gains, and not just allelic loss, in HLA genes. This knowledge is important for the design of cancer vaccines or T cell therapies targeting neoantigens presented by HLA class I complexes.

Citation Format: Katie M. Campbell, Justin Saco, Egmidio Medina, Meelad Amouzgar, Shannon M. Pfeiffer, Cynthia R. Gonzalez, Gabriela Steiner, Ameya Champhekar, Cristina Puig Saus, Jesse Zaretsky, Gabriel Abril Rodriguez, Agustin Vega-Crespo, Ignacio Baselga Carretero, Mito Tariveranmoshabad, Anusha Kalbasi, Christine Spencer, Zachary L. Skidmore, Malachi Griffith, Obi L. Griffith, Daniel K. Wells, Antoni Ribas. Infrequent chromosomal loss and recurrent gains lead to imbalanced expression of HLA genes in melanoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 3818.

Circulating KRAS variant-specific shedding and association with survival in patients with metastatic pancreatic ductal adenocarcinoma (mPDAC) receiving chemoimmunotherapy

2548

Background: Circulating tumor DNA (ctDNA) is increasingly used as a prognostic marker with high ctDNA shedding associated with poor survival. Gene-, but not variant-specific, differences in ctDNA shedding have been reported. Tumor burden, mitotic rate, and cell death rate have been proposed as contributors to ctDNA shedding. Here we investigate associations of ctDNA shedding for the two most common mPDAC KRAS variants, G12D and G12V, with tumor burden, mitotic score, and overall survival (OS). Methods: Pretreatment (baseline) ctDNA was analyzed by droplet digital PCR for 86 (including 44 G12D, 30 G12V) patients with mPDAC receiving front-line chemoimmunotherapy in a randomized open-label Phase II study (NCT03214250). Baseline tumor burden in total, within the pancreas, and distally, was assessed by sum of RECIST target lesion diameters. Tumor tissue variant allele fraction (tVAF) and mitotic score (geometric mean expression of 65 mitosis-associated genes) were calculated from DNA and RNA sequencing. Results: ctKRAS shedding (dichotomized at median mutant copies/mL) was associated with OS for G12D bearing tumors (p = 0.03) but not G12V (p = 0.17, log-rank test). To identify variant-specific features of shedding, we examined the Spearman correlation for total tumor burden and ctKRAS shedding; G12D but not G12V shedding was correlated with tumor burden (p = 0.01 and p = 0.22 respectively). However, the higher tVAF in G12V compared to G12D tumors (p = 0.048, Mann-Whitney test) could result from differences in purity, ploidy, and KRAS copy number. Thus, we used tVAF as a scalar to calculate an adjusted tumor burden which was significantly correlated with both G12D and G12V ctDNA shedding (p = 0.004 and 0.02, respectively). When a patient’s distal vs. pancreatic lesions were analyzed separately, pancreatic tumor burden was not correlated with G12D or G12V shedding (p = 0.10 and 0.33, respectively) but distal burden was correlated with both (p = 0.001 and 0.02, respectively). While there was no difference by KRAS variant for the correlation between adjusted tumor burden and shedding, these results do suggest that, in patients with mPDAC, distal rather than primary tumor burden may drive ctDNA shedding. Finally, tumor mitotic rate was combined with adjusted total tumor burden in a linear regression model; both were significant for predicting G12D shedding (p = 0.007 and p < 0.0001, respectively) but not for G12V (p = 0.045 and p = 0.16, respectively). Conclusions: These data suggest that ctDNA shedding and survival associations may be KRAS variant-specific in mPDAC. Tumor mitotic score and location of tumors may explain some variant-specific differences in shedding. As clinical ctDNA tests become more widely used, further investigation of variant-specific shedding in KRAS and other genes may be key for proper interpretation of ctDNA tests.

Distinct biosignatures associate with survival after chemoimmunotherapy in a randomized, three-arm phase II study in patients with metastatic pancreatic cancer

4010

Background: Preclinical and small clinical studies of chemoimmunotherapy for metastatic pancreatic ductal adenocarcinoma (mPDAC) point to a yet unrealized potential of clinically significant immune activation. In our phase II study of the CD40 agonist antibody sotigalimab (sotiga) and/or nivolumab (nivo) with gemcitabine and nab-paclitaxel (chemo), we observed promising improvements in overall survival (OS) in 105 patients with newly diagnosed mPDAC (NCT03214250); the primary endpoint of 1-year OS rate was 57.7% (p = 0.006) in the nivo/chemo arm, 48.1% (p = 0.062) in the sotiga/chemo arm and 41.3% (p = 0.233) in the nivo/sotiga/chemo arm (O’Hara, ASCO 2021) as compared to a historical control of 35%. Here, we report results of multi-omic translational analyses designed to identify signatures predictive of OS benefit. Methods: Longitudinal blood and tumor tissue samples were collected for immune and tumor biomarker analysis. Tumor samples underwent RNA sequencing and multiplex immunofluorescence (mIF). Peripheral blood was analyzed by mass cytometry time of flight (CyTOF), high parameter flow cytometry, and proteomics. Machine learning (ML) algorithms were applied to the data to identify biosignatures related to OS in each arm. Results: Comprehensive multi-omic, multi-parameter immune and tumor biomarker analyses identified distinct pretreatment immune signatures predictive of longer OS specific to nivo/chemo or sotiga/chemo (Table, representative examples). Because patients in each arm received chemotherapy, these and other arm-unique biomarkers suggest a relationship to the immunotherapy rather than chemotherapy in this randomized study. There was evidence of immune exhaustion in the sotiga/nivo/chemo arm that may explain the lack of survival benefit. Conclusions: From in-depth translational and ML analyses of randomized phase II trial of first-line chemoimmunotherapy in mPDAC patients, we identified novel biomarkers that associated with OS distinctly in each arm. Clinical trials in first-line mPDAC exploiting these previously unappreciated biomarkers and aiming to enrich patients for response, are warranted to further advance chemoimmunotherapy in this disease. Clinical trial information: NCT03214250. [Table: see text]

Sotigalimab and/or nivolumab with chemotherapy in first-line metastatic pancreatic cancer: clinical and immunologic analyses from the randomized phase 2 PRINCE trial

Chemotherapy combined with immunotherapy has improved the treatment of certain solid tumors, but effective regimens remain elusive for pancreatic ductal adenocarcinoma (PDAC). We conducted a randomized phase 2 trial evaluating the efficacy of nivolumab (nivo; anti-PD-1) and/or sotigalimab (sotiga; CD40 agonistic antibody) with gemcitabine/nab-paclitaxel (chemotherapy) in patients with first-line metastatic PDAC ( NCT03214250 ). In 105 patients analyzed for efficacy, the primary endpoint of 1-year overall survival (OS) was met for nivo/chemo (57.7%, P = 0.006 compared to historical 1-year OS of 35%, n = 34) but was not met for sotiga/chemo (48.1%, P = 0.062, n = 36) or sotiga/nivo/chemo (41.3%, P = 0.223, n = 35). Secondary endpoints were progression-free survival, objective response rate, disease control rate, duration of response and safety. Treatment-related adverse event rates were similar across arms. Multi-omic circulating and tumor biomarker analyses identified distinct immune signatures associated with survival for nivo/chemo and sotiga/chemo. Survival after nivo/chemo correlated with a less suppressive tumor microenvironment and higher numbers of activated, antigen-experienced circulating T cells at baseline. Survival after sotiga/chemo correlated with greater intratumoral CD4 T cell infiltration and circulating differentiated CD4 T cells and antigen-presenting cells. A patient subset benefitting from sotiga/nivo/chemo was not identified. Collectively, these analyses suggest potential treatment-specific correlates of efficacy and may enable biomarker-selected patient populations in subsequent PDAC chemoimmunotherapy trials.

Real-world data prognostic model of overall survival in patients with advanced NSCLC receiving anti-PD-1/PD-L1 immune checkpoint inhibitors as second-line monotherapy

Background and aim

The objective of this retrospective, observational, noninterventional cohort study was to investigate prognostic factors of overall survival (OS) in patients with advanced non‐small cell lung cancer (aNSCLC) and to develop a novel prognostic model.

Methods

A total of 4049 patients with aNSCLC diagnosed between January 2011 and February 2020 who received atezolizumab, nivolumab, or pembrolizumab as second‐line monotherapy were selected from a real‐world deidentified database to build the cohort. Patients could not have received first‐line treatment with clinical study drug(s) nor immune checkpoint inhibitors including anti‐programmed cell death 1 (PD‐1)/programmed death‐ligand 1 (PD‐L1), and anti‐cytotoxic T‐lymphocyte‐associated protein 4 therapies.

Results

Patients had a median age of 69 years; 45% were female, 75% White, 70% had stage IV at initial diagnosis, and 70% had nonsquamous histology. A Cox proportional hazards model with lasso regularization was used to build a prognostic model for OS using 18 baseline demographic and clinical factors based on the real‐world data cohort. The risk‐increasing prognostic factors were abnormally low albumin and chloride levels, Eastern Cooperative Oncology Group performance status score ≥ 2, and abnormally high levels of alkaline phosphatase and white blood cells. The risk‐decreasing prognostic factors were PD‐L1 positivity, longer time from advanced diagnosis to start of first‐line therapy, and higher systolic blood pressure. The performance of the model was validated using data from the OAK trial, and the c‐index for the OAK trial validation cohort was 0.65 and 0.67 for the real‐world data cohort.

Conclusions

Based on baseline demographic and clinical factors from a real‐world setting, this prognostic model was developed to discriminate the risk of death in patients with aNSCLC treated with checkpoint inhibitors as second‐line monotherapy, and it performed well in the real‐world data and clinical trial cohorts.