DNA liquid biopsy-based prediction of cancer-associated venous thromboembolism.

Cancer-associated venous thromboembolism (VTE) is a major source of oncologic cost, morbidity and mortality. Identifying high-risk patients for prophylactic anticoagulation is challenging and adds to clinician burden. Circulating tumor DNA (ctDNA) sequencing assays ('liquid biopsies') are widely implemented, but their utility for VTE prognostication is unknown. Here we analyzed three plasma sequencing cohorts: a pan-cancer discovery cohort of 4,141 patients with non-small cell lung cancer (NSCLC) or breast, pancreatic and other cancers; a prospective validation cohort consisting of 1,426 patients with the same cancer types; and an international generalizability cohort of 463 patients with advanced NSCLC. ctDNA detection was associated with VTE independent of clinical and radiographic features. A machine learning model trained on liquid biopsy data outperformed previous risk scores (discovery, validation and generalizability c-indices 0.74, 0.73 and 0.67, respectively, versus 0.57, 0.61 and 0.54 for the Khorana score). In real-world data, anticoagulation was associated with lower VTE rates if ctDNA was detected (n = 2,522, adjusted hazard ratio (HR) = 0.50, 95% confidence interval (CI): 0.30-0.81); ctDNA- patients (n = 1,619) did not benefit from anticoagulation (adjusted HR = 0.89, 95% CI: 0.40-2.0). These results provide preliminary evidence that liquid biopsies may improve VTE risk stratification in addition to clinical parameters. Interventional, randomized prospective studies are needed to confirm the clinical utility of liquid biopsies for guiding anticoagulation in patients with cancer.

Microsatellite Instability, Tumor Mutational Burden, and Response to Immune Checkpoint Blockade in Patients with Prostate Cancer.

PURPOSE: Patients with microsatellite instability high/mismatch repair deficient (MSI-H/dMMR) and high tumor mutational burden (TMB-H) prostate cancers are candidates for pembrolizumab. We define the genomic features, clinical course, and response to immune checkpoint blockade (ICB) in patients with MSI-H/dMMR and TMB-H prostate cancers without MSI (TMB-H/MSS). METHODS: We sequenced 3,244 tumors from 2,257 prostate cancer patients. MSI-H/dMMR prostate cancer was defined as MSIsensor score ≥10 or MSIsensor score ≥3 and <10 with a deleterious MMR alteration. TMB-H was defined as ≥10 mutations/megabase. PSA50 and RECIST responses were assigned. Overall survival (OS) and radiographic progression-free survival (rPFS) were compared using log rank test. RESULTS: 63 (2.8%) men had MSI-H/dMMR and 33 (1.5%) had TMB-H/MSS prostate cancers. Patients with MSI-H/dMMR and TMB-H/MSS tumors more commonly presented with grade group 5 and metastatic disease at diagnosis. MSI-H/dMMR tumors had higher TMB, indel and neoantigen burden compared with TMB-H/MSS. 27 patients with MSI-H/dMMR and 8 patients with TMB-H/MSS tumors received ICB, none of whom harbored POLE mutations. 45% of MSI-H/dMMR patients had a RECIST response and 65% had a PSA50 response. No TMB-H/MSS patient had a RECIST response and 50% had a PSA50 response. rPFS tended to be longer in MSI-H/dMMR patients than in TMB-H/MSS patients who received immunotherapy. Pronounced differences in genomics, TMB or MSIsensor score were not detected between MSI-H/dMMR responders and non-responders. CONCLUSIONS: MSI-H/dMMR prostate cancers have greater TMB, indel and neoantigen burden compared with TMB-H/MSS prostate cancers, and these differences may contribute to more profound and durable responses to ICB.

Precision oncology: current and future platforms for treatment selection.

Genomic profiling of hundreds of cancer-associated genes is now a component of routine cancer care. DNA sequencing can identify mutations, mutational signatures, and structural alterations predictive of therapy response and assess for heritable cancer risk, but it has been less useful for identifying predictive biomarkers of sensitivity to cytotoxic chemotherapies, antibody drug conjugates, and immunotherapies. The clinical adoption of molecular profiling platforms such as RNA sequencing better suited to identifying those patients most likely to respond to immunotherapies and drug combinations will be critical to expanding the benefits of precision oncology. This review discusses the potential advantages of innovative molecular and functional profiling platforms designed to replace or complement targeted DNA sequencing and the major hurdles to their clinical adoption.

Genomic Landscape of Adenocarcinomas Across the Gastroesophageal Junction: Moving on from the Siewert Classification.

OBJECTIVE: To investigate how the Siewert classification of gastroesophageal junction adenocarcinomas correlates with genomic profiles. SUMMARY/BACKGROUND DATA: Current staging and treatment guidelines recommend that tumors with an epicenter less than 2 cm into the gastric cardia be treated as esophageal cancers, while tumors with epicenter greater than 2 cm into the cardia be staged and treated as gastric cancers. To date, however, few studies have compared the genomic profiles of the 3 Siewert classification groups to validate this distinction. METHODS: Using targeted tumor sequencing data on patients with adenocarcinoma of the gastroesophageal junction previously treated with surgery at our institution, we compared genomic features across Siewert classification groups. RESULTS: A total of 350 patients were included: 121 had Siewert type I, 170 type II, and 59 type III. Comparisons by Siewert location revealed that Siewert type I and II were primarily characterized as the chromosomal instability (CIN) molecular subtype and displayed Barrett's metaplasia and p53 and cell cycle pathway dysregulation. Siewert type III tumors, by contrast, were more heterogeneous, including higher proportions of microsatellite instability (MSI) and genomically stable (GS) tumors and more frequently displayed ARID1A and somatic CDH1 alterations, signet ring cell features, and poor differentiation. Overall, Siewert type I and II tumors demonstrated greater genomic overlap with lower esophageal tumors, while Siewert type III tumors shared genomic features with gastric tumors. CONCLUSIONS: Overall, our results support recent updates in treatment and staging guidelines. Ultimately, however, molecular rather than anatomic classification may prove more valuable in determining staging, treatment, and prognosis.

Clinicogenomic predictors of outcomes in patients with hepatocellular carcinoma treated with immunotherapy.

INTRODUCTION: Immune checkpoint inhibitor (ICI) combinations extend overall survival (OS) while anti-PD-1/L1 monotherapy is non-inferior to sorafenib in treatment-naïve, patients with advanced hepatocellular carcinoma (HCC). Clinicogenomic features are posited to influence patient outcomes. METHODS: The primary objective of this retrospective study was to define the clinical, pathologic, and genomic factors associated with outcomes to ICI therapy in patients with HCC. Patients with histologically confirmed advanced HCC treated with ICI at Memorial Sloan Kettering Cancer Center from 2012 to 2022 were included. Association between clinical, pathological, and genomic characteristics were assessed with univariable and multivariable Cox regression model for progression-free survival (PFS) and OS. RESULTS: Two-hundred and forty-two patients were treated with ICI-based therapy. Patients were predominantly male (82%) with virally mediated HCC (53%) and Child Pugh A score (70%). Median follow-up was 28 months (0.5-78.4). Median PFS for those treated in 1st line, 2nd line and ≥ 3rd line was 4.9 (range: 2.9-6.2), 3.1 (2.3-4.0), and 2.5 (2.1-4.0) months, respectively. Median OS for those treated in 1st line, 2nd line, and ≥ 3rd line was 16 (11-22), 7.5 (6.4-11), and 6.4 (4.6-26) months, respectively. Poor liver function and performance status associated with worse PFS and OS, while viral hepatitis C was associated with favorable outcome. Genetic alterations were not associated with outcomes. CONCLUSION: Clinicopathologic factors were the major determinates of outcomes for patients with advanced HCC treated with ICI. Molecular profiling did not aid in stratification of ICI outcomes. Future studies should explore alternative biomarkers such as the level of immune activation or the pretreatment composition of the immune tumor microenvironment.

Early On-Treatment Assessment of T Cells, Cytokines, and Tumor DNA with Adaptively Dosed Nivolumab + Ipilimumab: Final Results from the Phase 2 ADAPT-IT Study.

PURPOSE: The Adaptively Dosed ImmunoTherapy Trial (ADAPT-IT;NCT03122522) investigated adaptive ipilimumab discontinuation in melanoma based on early radiographic assessment. Initial findings indicated similar effectiveness compared with conventional nivolumab-ipilimumab (nivo-ipi). Exploratory biomarker analyses and final clinical results are now reported. PATIENTS AND METHODS: Patients with unresectable melanoma received two doses of nivo-ipi. Radiographic assessment at Week 6 informed continuation of ipilimumab before nivolumab maintenance. The primary endpoint was overall response rate at Week 12. Plasma was assayed for circulating tumor DNA and 10 cytokines using a multiplex immunoassay. Flow cytometry of peripheral blood mononuclear cells was performed with an 11-color panel. RESULTS: Among the treated patients, expansion of proliferating T-cell populations was observed in responders and nonresponders. Baseline IL6 levels were low in patients achieving an objective radiographic response (median 1.30 vs. 2.86 pg/mL; P = 0.025). High baseline IL6 levels were associated with short progression-free survival [PFS; HR = 1.24, 95% confidence interval (CI), 1.01-1.52; P = 0.041]. At Week 6, patients with response had lower average tumor variant allele fractions than nonresponders (median 0.000 vs. 0.019; P = 0.014). Greater increases in average variant allele fractions from baseline to Week 6 correlated with short PFS (HR = 1.11, 95% CI, 1.01-1.21; P = 0.023). Week 12 overall response rate was 47% (95% CI, 35%-59%) with a median follow-up of 34 months among survivors. Median PFS was 21 months (95% CI, 10-not reached); 76% of responses (95% CI, 64%-91%) persisted at 36 months. CONCLUSIONS: Adaptively dosed nivo-ipi responses are durable and resemble historical data for conventional nivo-ipi. Baseline IL6 and circulating tumor DNA changes during treatment warrant further study as biomarkers of nivo-ipi response.

Natural History and Genomic Landscape of Chemotherapy-Resistant Muscle-Invasive Bladder Cancer.

PURPOSE: Patients with residual invasive bladder cancer after neoadjuvant chemotherapy (NAC) and radical cystectomy have a poor prognosis. Data on adjuvant therapy for these patients are conflicting. We sought to evaluate the natural history and genomic landscape of chemotherapy-resistant bladder cancer to inform patient management and clinical trials. METHODS: Data were collected on patients with clinically localized muscle-invasive urothelial bladder cancer treated with NAC and cystectomy at our institution between May 15, 2001, and August 15, 2019, and completed four cycles of gemcitabine and cisplatin NAC, excluding those treated with adjuvant therapies. Survival was estimated using the Kaplan-Meier method, and multivariable Cox proportional hazards models were used to identify predictors of recurrence-free survival (RFS). Genomic alterations were identified in targeted exome sequencing (Memorial Sloan Kettering Integrated Mutation Profiling of Actionable Cancer Targets) data from post-NAC specimens from a subset of patients. RESULTS: Lymphovascular invasion (LVI) was the strongest predictor of RFS (hazard ratio, 2.15 [95% CI, 1.37 to 3.39]) on multivariable analysis. Patients with ypT2N0 disease without LVI had a significantly prolonged RFS compared with those with LVI (70% RFS at 5 years). Lymph node yield did not affect RFS. Among patients with sequencing data (n = 101), chemotherapy-resistant tumors had fewer alterations in DNA damage response genes compared with tumors from a publicly available chemotherapy-naïve cohort (15% v 29%; P = .021). Alterations in CDKN2A/B were associated with shorter RFS. PIK3CA alterations were associated with LVI. Potentially actionable alterations were identified in more than 75% of tumors. CONCLUSION: Although chemotherapy-resistant bladder cancer generally portends a poor prognosis, patients with organ-confined disease without LVI may be candidates for close observation without adjuvant therapy. The genomic landscape of chemotherapy-resistant tumors is similar to chemotherapy-naïve tumors. Therapeutic opportunities exist for targeted therapies as adjuvant treatment in chemotherapy-resistant disease.

Cell-free DNA from nail clippings as source of normal control for genomic studies in hematologic malignancies.

Comprehensive genomic sequencing is becoming a critical component in the assessment of hematologic malignancies, with broad implications for patient management. In this context, unequivocally discriminating somatic from germline events is challenging but greatly facilitated by matched analysis of tumor:normal pairs. In contrast to solid tumors, conventional sources of normal control (peripheral blood, buccal swabs, saliva) could be highly involved by the neoplastic process, rendering them unsuitable. In this work we describe our real-world experience using cell free DNA (cfDNA) isolated from nail clippings as an alternate source of normal control, through the dedicated review of 2,610 tumor:nail pairs comprehensively sequenced by MSK-IMPACT-heme. Overall, we find nail cfDNA is a robust source of germline control for paired genomic studies. In a subset of patients, nail DNA may have tumor DNA contamination, reflecting unique attributes of the hematologic disease and transplant history. Contamination is generally low level, but significantly more common among patients with myeloid neoplasms (20.5%; 304/1482) compared to lymphoid diseases (5.4%; 61/1128) and particularly enriched in myeloproliferative neoplasms with marked myelofibrosis. When identified in patients with lymphoid and plasma-cell neoplasms, mutations commonly reflected a myeloid profile and correlated with a concurrent/evolving clonal myeloid neoplasm. For nails collected after allogeneic stem-cell transplantation, donor DNA was identified in 22% (11/50). In this cohort, an association with recent history of graft-vs-host disease was identified. These findings should be considered as a potential limitation for the use of nail as normal control but could also provide important diagnostic information regarding the disease process.

Deep-Learning Model for Tumor-Type Prediction Using Targeted Clinical Genomic Sequencing Data.

Tumor type guides clinical treatment decisions in cancer, but histology-based diagnosis remains challenging. Genomic alterations are highly diagnostic of tumor type, and tumor-type classifiers trained on genomic features have been explored, but the most accurate methods are not clinically feasible, relying on features derived from whole-genome sequencing (WGS), or predicting across limited cancer types. We use genomic features from a data set of 39,787 solid tumors sequenced using a clinically targeted cancer gene panel to develop Genome-Derived-Diagnosis Ensemble (GDD-ENS): a hyperparameter ensemble for classifying tumor type using deep neural networks. GDD-ENS achieves 93% accuracy for high-confidence predictions across 38 cancer types, rivaling the performance of WGS-based methods. GDD-ENS can also guide diagnoses of rare type and cancers of unknown primary and incorporate patient-specific clinical information for improved predictions. Overall, integrating GDD-ENS into prospective clinical sequencing workflows could provide clinically relevant tumor-type predictions to guide treatment decisions in real time. SIGNIFICANCE: We describe a highly accurate tumor-type prediction model, designed specifically for clinical implementation. Our model relies only on widely used cancer gene panel sequencing data, predicts across 38 distinct cancer types, and supports integration of patient-specific nongenomic information for enhanced decision support in challenging diagnostic situations. See related commentary by Garg, p. 906. This article is featured in Selected Articles from This Issue, p. 897.

Clinical utility of plasma cell-free DNA in pancreatic neuroendocrine neoplasms.

In advanced pancreatic neuroendocrine neoplasms (PanNEN), there are little data detailing the frequency of genetic alterations identified in cell free DNA (cfDNA), plasma-tissue concordance of detected alterations, and clinical utility of cfDNA. Patients with metastatic PanNENs underwent cfDNA collection in routine practice. Next-generation sequencing (NGS) of cfDNA and matched tissue when available was performed. Clinical actionability of variants was annotated by OncoKB. Thirty-two cfDNA samples were analyzed from 25 patients, the majority who had well-differentiated intermediate grade disease (13/25; 52%). Genomic alterations were detected in 68% of patients and in 66% of all cfDNA samples. The most frequently altered genes were DAXX (28%), TSC2 (24%), MEN1 (24%), ARID1B (20%), ARID1A (12%), and ATRX (12%). Twenty-three out of 25 (92%) patients underwent tumor tissue NGS. Tissue-plasma concordance for select genes was as follows:DAXX (95.7%), ARID1A (91.1%), ATRX (87%), TSC2 (82.6%), MEN1 (69.6%). Potentially actionable alterations were identified in cfDNA of 8 patients, including TSC2 (4; level 3b), ATM (1; level 3b), ARID1A (2; level 4), and KRAS (1; level 4). An ETV6:NTRK fusion detected in tumor tissue was treated with larotrectinib; at progression, sequencing of cfDNA identified an NTRK3 G623R alteration as the acquired mechanism of resistance; the patient enrolled in a clinical trial of a second-generation TRK inhibitor with clinical benefit. In metastatic PanNENs, cfDNA-based NGS identified tumor-associated mutations in 66% of plasma samples with a high level of plasma-tissue agreement in PanNEN-associated genes. Clonal evolution, actionable alterations, and resistance mechanisms were detected through circulating cfDNA genotyping.

High-risk and silent clonal hematopoietic genotypes in patients with nonhematologic cancer.

ABSTRACT: Clonal hematopoiesis (CH) identified by somatic gene variants with variant allele fraction (VAF) ≥ 2% is associated with an increased risk of hematologic malignancy. However, CH defined by a broader set of genotypes and lower VAFs is ubiquitous in older individuals. To improve our understanding of the relationship between CH genotype and risk of hematologic malignancy, we analyzed data from 42 714 patients who underwent blood sequencing as a normal comparator for nonhematologic tumor testing using a large cancer-related gene panel. We cataloged hematologic malignancies in this cohort using natural language processing and manual curation of medical records. We found that some CH genotypes including JAK2, RUNX1, and XPO1 variants were associated with high hematologic malignancy risk. Chronic disease was predicted better than acute disease suggesting the influence of length bias. To better understand the implications of hematopoietic clonality independent of mutational function, we evaluated a set of silent synonymous and noncoding mutations. We found that silent CH, particularly when multiple variants were present or VAF was high, was associated with increased risk of hematologic malignancy. We tracked expansion of CH mutations in 26 hematologic malignancies sequenced with the same platform. JAK2 and TP53 VAF consistently expanded at disease onset, whereas DNMT3A and silent CH VAFs mostly decreased. These data inform the clinical and biological interpretation of CH in the context of nonhematologic cancer.