Paper Details

ABSTRACT

Background: Accurate detection of clinically relevant genomic alterations is essential for molecularly informed diagnosis, classification, risk stratification, and treatment decisions in patients with hematologic malignancies. However, current standard-of-care testing relies on multiple complementary assays, including cytogenetics, fluorescence in situ hybridization (FISH), and targeted next-generation sequencing (NGS), resulting in fragmented workflows, longer turnaround times, and increased complexity for physicians. Whole-genome sequencing (WGS) has the potential to consolidate these approaches into a single comprehensive assay but remains largely unavailable outside specialized academic centers. To address the need for broader access to this unified profiling approach, we developed and validated Caris ChromoSeq™, a WGS-based assay for hematologic malignancies. Caris ChromoSeq comprehensively detects genomic abnormalities including International Consensus Classification (ICC) and World Health Organization (WHO) disease-defining single-nucleotide variants (SNVs), insertions/deletions (INDELs), copy number alterations (CNAs), and structural variants (SVs) in acute myeloid leukemia (AML), myelodysplastic syndromes (MDS), myeloproliferative neoplasms (MPNs), and B-cell acute lymphoblastic leukemia (B-ALL).

Methods: Assay performance was evaluated in 391 blood or bone marrow samples from Washington University (WashU) and Tampa General Hospital, seven HapMap reference standards, and a reference cell line. Positive percent agreement (PPA) and positive predictive value (PPV) for clinically relevant variant classes were determined by comparison with validated orthogonal assays, including WashU ChromoSeq® (n = 167) and targeted NGS, cytogenetics, and FISH methods (n = 224). Libraries were prepared using a custom workflow based on KAPA HyperPrep and EvoPrep chemistries, and WGS was performed on NovaSeq X instruments. Variant calling was based on well-characterized tools such as SAMtools, Pindel, and SnpEff. CNAs, tumor purity, and ploidy were inferred using PyreCN (a Python implementation of PureCN), and SVs were identified using the break-end caller GRIDSS2.

Results: Analytical validation established a minimum input requirement of 250 ng genomic DNA with optimal input of 500 ng. In this range, 200x target sequencing depth was achieved with low variability and high coverage uniformity. The limit of detection was identified as 5% variant allele frequency (VAF) for SNVs/INDELs and 20% tumor fraction for CNAs/SVs; these thresholds were applied for all reported performance metrics. Caris ChromoSeq demonstrated 95.4% overall accuracy of variant detection in a well-characterized reference cell line and detection of all known variants in HapMap reference standards. Across pathogenic/likely-pathogenic DNA variants, PPA and PPV were 99.1% (1099/1109) and 98.9% (1099/1111), respectively, compared with WashU ChromoSeq and 95.4% (597/626) and 97.2% (597/614), respectively, compared with targeted NGS. Using WashU ChromoSeq as the comparator for clinically relevant CNAs, PPA and PPV were 98.5% (384/390) and 80.7% (384/476), respectively, for amplifications and 97.1% (402/414) and 83.2% (402/483), respectively, for deletions. For clinically relevant SVs including fusions, PPA and PPV were both 98.5% (66/67) relative to WashU ChromoSeq. SV accuracy (PPA) relative to targeted NGS, karyotype, or FISH was 98.1% (53/54). Precision across reagent lots and operators exceeded 98% for SNVs/INDELs and reached 100% for CNAs/fusions.

Conclusions: The validation data presented here demonstrates that Caris ChromoSeq can detect a broad range of clinically relevant alterations in AML, MDS, MPN, and B-ALL, displaying high concordance with clinically validated assays. One limitation of this validation is the availability of less sensitive comparators, resulting in reduced PPV for some CNA analyses. By consolidating variant detection traditionally performed using fragmented profiling methods, Caris ChromoSeq has the potential to streamline diagnostic testing for hematologic malignancies, expand access to comprehensive genomic profiling beyond specialized academic centers, and support timely, guideline-informed clinical decision-making.