AUTHORS
Felicia Andresen, MD, Ben Goldberg2, Kasiani C. Myers I, MD3, Ashley Gutierrez4, Diana Schwarz5, Christopher R. Reilly, MD6, Taizo Nakano Jr, MD7, Marcin W Wlodarski, MD, PhD8, Karyn Brundige, MSN9, Sara Loveless, RN, BSN, BMTCN10, Leah Cheng, MA, CCRP11, Edie Weller12, R. Coleman Lindsley, MD, PhD13, Akiko Shimamura, MD PhD14
AFFILIATIONS
1 Dana-Farber/Boston Children's Cancer and Blood Disorders Center, Boston, Massachusetts, United States
2 Boston Children's Hospital, Boston, Massachusetts, United States
3 Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio, United States
4 Dana-Farber/Boston Children's Cancer and Blood Disorders Center, Boston, Massachusetts, United States
5 Dana-Farber/Boston Children's Cancer and Blood Disorders Center, Boston, Massachusetts, United States
6 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
7 Children's Hospital Colorado, Aurora, Colorado, United States
8 St. Jude Children's Research Hospital, Memphis, Tennessee, United States
9 Dana-Farber/Boston Children's Cancer and Blood Disorders Center, Boston, Massachusetts, United States
10 Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio, United States
11 Dana-Farber/Boston Children's Cancer and Blood Disorders Center, Boston, Massachusetts, United States
12 Boston Children’s Hospital, Boston, Massachusetts, United States
13 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
14 Dana-Farber/Boston Children's Cancer and Blood Disorders Center, Boston, Massachusetts, United States
ABSTRACT
Background: Early diagnosis of hereditary hematologic malignancy predisposition (HHMp) provides an opportunity for surveillance and preventive intervention before progression to malignancy. Expert consensus panels recommend regular bone marrow examinations to guide transplant timing, but patient’s willingness to undergo repeated marrow sampling is often limited. Molecular testing of blood offers a minimally invasive alternative approach to clone detection. Effective surveillance requires sensitive detection of small clones and assessment of clone trajectories over time. Whether blood testing meets these requirements in HHMp has not been determined. Here, we paired blood and marrow sequencing in Shwachman-Diamond syndrome (SDS), a disorder predisposing to TP53-mutated myeloid malignancy, to define the performance of blood-based clone detection against the marrow reference standard.
Methods: All prospectively collected paired marrow and blood samples from patients with SBDS-associated SDS enrolled in the North American SDS Registry were studied, excluding samples obtained at the time of myeloid malignancy or after HSCT. We performed error-corrected next-generation sequencing (NGS) on 337 blood and marrow sample pairs (674 samples total) from 119 SDS patients. Sensitivity and specificity of blood to detect marrow clones were estimated at the variant, gene, and sample level, defining a pair as concordant if both samples detected: (1) the same variant, (2) any variant in the same gene, or (3) at least one somatic variant, irrespective of gene or variant. We evaluated detection across variant allele fraction (VAF) thresholds and, in serial pairs, compared clone trajectories between compartments. The association of myeloid to lymphoid ratio (M:L) or absolute neutrophil count with blood-marrow concordance was evaluated using a beta regression model.
Results: TP53 and EIF6 were the most frequently mutated genes in both marrow and blood. At VAFs ≥0.02, blood NGS for TP53 variants had a sensitivity of 89% and specificity of 99%. At the sample level, blood NGS had a sensitivity of 91% and specificity of 95%. At VAFs ≥0.02, 91% of patients with a positive blood NGS test also had marrow clonal hematopoiesis, and TP53 gene-level positive predictive value was 94%. Conversely, 95% of patients with a negative blood test had negative marrow testing (sample-level negative predictive value (NPV) 95%, TP53 gene-level NPV 98%). Variant-level sensitivity was dependent on VAF thresholds and marrow clone size. Variant-level sensitivity was lower and depended on both the blood VAF threshold and marrow clone size. At thresholds of ≥0.005, ≥0.01, and ≥0.02, variant-level sensitivity of blood was 72%, 77%, 78%, respectively. At blood VAF thresholds of ≥0.02, variant-level sensitivity for clones with marrow VAFs 0.02-<0.05 was 62% and increased to >99% for variants with a marrow VAF ≥0.05. Variant-level specificity was >99% at all thresholds. Neither neutropenia nor M:L ratio in blood affected concordance (p≥0.8). While blood and marrow VAFs were generally correlated (R2=0.91) and similar over time (variance of ΔVAF(marrow-blood VAF) <0.0001), compartmental differences in clone size and trajectories were observed in individual patients. One patient had a progressive rise in TP53 marrow VAF from 0.02 to 0.19 while the concomitant blood VAF remained low at 0.01 to 0.03 on serial testing.
Conclusion: Blood NGS detected marrow clones, including TP53 clones, with high sensitivity and specificity at the sample and gene level, but variant-level sensitivity fell as marrow clone size decreased. In individual patients, blood VAF did not reliably reflect marrow clone size or its change over time. These data support blood NGS as a screening tool for emerging marrow clonal hematopoiesis in HHMp surveillance, tailored to the mechanisms of clonal evolution expected for a given condition. Detection of a clone in blood should prompt a marrow examination to identify the specific variants, determine clone size, assess trajectory, and evaluate concomitant morphologic and chromosomal abnormalities. A negative blood test does not exclude small marrow clones. The same approach may inform blood-based surveillance in other settings of myeloid neoplasm risk, such therapy-related clonal hematopoiesis.
DISCLOSURES
Felicia Andresen, MD: No relevant conflicts of interest to declare.