Implementation of next generation sequencing into pediatric hematology-oncology practice: moving beyond actionable alterations

Authors

Jennifer A. Oberg

Julia L. Glade Bender

Maria Luisa Sulis

Danielle Pendrick

Anthony N. Sireci

Susan J. Hsiao

Andrew T. Turk

Filemon S. Dela Cruz

Hanina Hibshoosh

Helen Remotti

Rebecca J. Zylber

Jiuhong Pang

Daniel Diolaiti

Carrie Koval

Stuart J. Andrews

James H. Garvin

Darrell J. Yamashiro

Wendy K. Chung

Stephen G. Emerson

Peter L. Nagy

Mahesh M. Mansukhani

Andrew L. Kung

Doi

10.1186/s13073-016-0389-6

PMID: 28007021 · DOI: 10.1186/s13073-016-0389-6 · Journal: Genome Medicine (2016)

TL;DR

Oberg et al. report the first 101 high-risk pediatric hematology-oncology patients enrolled in the Precision in Pediatric Sequencing (PIPseq) program at Columbia University Medical Center. Using CLIA-certified cancer whole-exome sequencing of paired tumor/normal tissue plus tumor RNA-seq (with a 467-gene FFPE panel as backup), they show that potentially actionable somatic alterations occur in 38% of patients but only 16% of those receive matched targeted therapy. By taking a broader view of “clinical impact” — diagnostic, prognostic, and pharmacogenomic value beyond drug-target identification — sequencing meaningfully informed clinical management in 66% of all 101 cases and in 75% (45/60) of cases that received the full cWES + RNA-seq combination. Germline cancer-predisposition variants were identified in 14% of patients, and ACMG secondary findings in six. Data were deposited in the PIPseq cBioPortal study PMID:28007021.

Cohort & data

Key findings

  • Genomic aberrations were reported in 92/101 patients (91%). A total of 180 somatic mutations and 20 fusions were reported; solid tumors averaged 2.91 aberrations per sample (median 2, range 1–6) and hematologic samples averaged 5.2 (median 4, range 1–12) PMID:28007021.
  • Mean mutational load was 216.9 variants per patient (SD 829.3, median 69), higher in solid tumors than hematologic conditions; pediatric loads were lower than published adult-cancer benchmarks PMID:28007021.
  • Most frequently mutated genes: TP53 in solid tumors (n=9, 9%); RAS-pathway genes in hematologic samples (NRAS n=5 / 5%; KRAS n=3 / 3%) PMID:28007021.
  • Potentially actionable alterations identified in 38/101 patients (38%) — 21/65 solid (32%) and 17/36 hematologic (47%). Only 6/38 (16%) actually received matched targeted therapy PMID:28007021.
  • Non-targetable somatic findings were clinically informative for diagnostic, prognostic, or pharmacogenomic purposes in 38 additional patients (38%), giving 66/101 patients (66%) clinically impactful results overall (Fig. 4) PMID:28007021.
  • In the 60 patients with full cWES + RNA-seq, 72 clinically impactful results were found in 45 cases (75%); cWES contributed 85% of findings (tumor/normal WES 45%, n=32; RNA-seq 40%, n=29; CNV 7%, n=5; RNA-seq + CNV combined 8%, n=6) (Fig. 5) PMID:28007021.
  • RNA-seq supplied unique clinical impact in 23/33 patients (70%) through fusion detection, expression-based subtyping, and identification of BCR-ABL1-like signaling patterns PMID:28007021.
  • Germline alterations were sequenced in 90 patients; clinically impactful germline findings in 18/90 (20%), with 14% carrying alterations in cancer-predisposition genes — slightly higher than published rates of 8.5–10% PMID:28007021.
  • ACMG secondary findings returned to six patients, including germline BRCA1 (in two patients), TP53, TNNT2 (dilated cardiomyopathy), RYR1 (malignant hyperthermia), and VHL PMID:28007021.
  • Cost and reimbursement. Estimated per-case cost: $4,459 for cWES + $1,764 for RNA-seq (variable + fixed costs, excluding administrative overhead). Reimbursement averaged $2,747 from commercial payers, $2,918 from managed government plans, and $0 from straight government plans PMID:28007021.
  • Patient consent uptake was high. Of 67 patients consented under the clinical cWES protocol, only 4 (6%) opted out of learning ACMG secondary findings; 21 (31%) opted out of having those findings appear in the EMR; only 1 (2%) opted out of future contact PMID:28007021.

Genes & alterations

Targetable somatic alterations (Table 2)

  • KIT — Activating Asn655Lys mutation in a 7-year-old boy with AML; palliative imatinib gave near-complete clearing of peripheral blood blasts with sustained 9-month response. A second AML patient harbored c.2446G>C (D816H) co-occurring with TET2 and FLT3 variants PMID:28007021.
  • BCR–ABL1 fusion — Diagnostic of CML (PIPID 14-85546); also detected as a BCR-ABL1-like RNA-seq signature in a 9-year-old girl with relapsed/refractory B-cell BLL, where confirmatory RT-PCR identified a NUP214–ABL1 fusion; addition of dasatinib to third-line induction produced deep remission enabling curative bone-marrow transplant PMID:28007021.
  • IDH1 — R132C in AML; IDH-inhibitor target. Same variant separately diagnostic of Maffucci syndrome PMID:28007021.
  • NRAS — Q61H, Q61R, Q61K, G13A, G13D activating mutations across multiple AML, ALL, NBL, and RMS cases — MEK-inhibitor targets PMID:28007021.
  • KRAS — Q61H and G12C in ALL and NBL; MEK-inhibitor targets PMID:28007021.
  • JAK3 — A573V activating mutation in AML; JAK-inhibitor target. JAK1 K1026E in T-BLL (T-ALL) PMID:28007021.
  • KMT2A–AFF1 (MLL-AFF1) fusion in AML — DOT1L-inhibitor target PMID:28007021.
  • NUP98–NSD1 fusion in AML — DOT1L-inhibitor target PMID:28007021.
  • FOXP1–ABL1 fusion in BLL — TKI target (treated) PMID:28007021.
  • PTPN11 — G503V mutation in AML (MEK-inhibitor target); separately a PTPN11 D61Y co-mutation with SETBP1 D868N (reported as “SETB1” in the paper) reclassified one 4-year-old’s diagnosis from de novo AML to JMML (which evolved into AML) PMID:28007021.
  • KMT2C (MLL3) — E704X nonsense in AML; BET-inhibitor target PMID:28007021.
  • SMARCA4 — R1192C in hepatic rhabdoid tumor (MRTL); EZH2-inhibitor target. SMARCB1 homozygous deletion of chr22q11.23 with biallelic loss of expression in another hepatic rhabdoid tumor — also EZH2-inhibitor target PMID:28007021.
  • SMARCC2–PDGFRB fusion in BLL — TKI target PMID:28007021.
  • TSC1 — c.2503-1G>C splice-site mutation in OS; mTOR-inhibitor target PMID:28007021.
  • PIK3CA — N345K in nephroblastomatosis (related to WT); PI3K/AKT/mTOR-inhibitor target. The same variant also supported the diagnosis of nephroblastomatosis PMID:28007021.
  • TMEM106B–BRAF fusion in PXA; MEK-inhibitor target PMID:28007021.
  • FGFR4 — G528C, V550L, R650L hotspot mutations in two patients with RMS; FGFR4-inhibitor targets PMID:28007021.
  • CDK2 — Gain at 12q14.1 involving CDK2 in glioblastoma multiforme; CDK4/6-inhibitor target PMID:28007021.
  • CDK4, MDM2 — co-overexpression in NBL; enrolled on the NEPENTHE trial (NCT02780128) PMID:28007021.
  • PTCH1, SUFU, ZIC3 — overexpression in MBL; SMO-inhibitor target PMID:28007021.
  • H3-3A (H3F3A) — K28M (i.e., K27M) in glioma and glioblastoma multiforme; HDAC-inhibitor target. Also paired with FGFR1 N577K in one glioma patient PMID:28007021.
  • MYC, MCL1, CCNE1 — overexpression in OS; BET- and CDK4/6-inhibitor targets in the same patient PMID:28007021.
  • CUL4A — overexpression in OS; NAE (NEDD8-activating enzyme) inhibitor target PMID:28007021.
  • RAD51C — F8L variant in OS; PARP-inhibitor target PMID:28007021.
  • PDGFRA, KDR (VEGFR2) — overexpression in OS; multi-targeted kinase inhibitor (MTKI) target (treated) PMID:28007021.
  • EML4–NTRK3 fusion in congenital fibrosarcoma (IFS) — supported diagnostic reclassification from undifferentiated sarcoma to infantile fibrosarcoma; ALK-inhibitor target PMID:28007021.
  • VCAN–IL23R fusion in inflammatory myofibroblastic tumor (IMT) — JAK-inhibitor target (treated) PMID:28007021.
  • ALK — Q1146K in adrenocortical carcinoma (ACC); ALK-inhibitor target (treated) PMID:28007021.
  • STAT5B — I704L mutation in T-BLL (paired with KRAS V14I and JAK1 K1026E); BCL2/BCL-XL-inhibitor target. Same variant separately diagnostic of gamma-delta T-cell lymphoma PMID:28007021.

Diagnostic / prognostic / pharmacogenomic findings (Table 3)

Clinically impactful germline findings (Table 4)

  • KMT2D (MLL2) — homozygous frameshift M3881Cfs*9 establishing Kabuki syndrome in a 2-month-old hospitalized for fulminant hemophagocytic syndrome; familial HLH was ruled out and stem-cell transplant was averted PMID:28007021.
  • C1QA — homozygous Q208* nonsense, diagnostic of C1Q deficiency in a 2-year-old with HLH PMID:28007021.
  • GATA2 — heterozygous frameshift (de novo) in AML with myelodysplastic syndrome — transplant-recommendation modifier PMID:28007021.
  • PMS2 — homozygous S459X diagnostic of constitutional mismatch-repair deficiency syndrome in a patient with T-cell lymphoblastic lymphoma and consanguineous parentage PMID:28007021.
  • XIAP — heterozygous R443P missense diagnostic of X-linked lymphoproliferative syndrome 2 (XLP2) in a 6-year-old girl with HLH and recurrent EBV infections PMID:28007021.
  • RUNX1 — heterozygous splice-site c.806-2A>G in AML, diagnostic of familial platelet disorder; HLA-matched sibling and father carried the same variant, prompting selection of an unrelated donor PMID:28007021.
  • APC — R1114* in hepatoblastoma diagnostic of familial adenomatous polyposis (FAP); E1554fs in poorly differentiated carcinoma diagnostic of newly appreciated Gardner syndrome PMID:28007021.
  • RB1 — splice-site c.1216-3A>G in OS, increased risk for second cancers PMID:28007021.
  • ATM — R189K + K2756* missense + nonsense in MBL, increased risk for other cancers PMID:28007021.
  • UGT1A1 — homozygous *28 (TA)7TAA allele in two patients with PB and hepatocellular carcinoma — pharmacogenomic flag for irinotecan/SN-38 toxicity PMID:28007021.
  • DICER1 — frameshift L1603Pfs in pineoblastoma — health maintenance for Sertoli-Leydig tumor risk PMID:28007021.
  • TP53 — S215N missense in AML — explained lack of response to conventional therapy and increased other-cancer risk PMID:28007021.
  • BRCA1 — frameshift in nested stromal epithelial liver tumor and a separate frameshift in EPM — breast-cancer health-maintenance flag (also returned as ACMG secondary findings) PMID:28007021.
  • RYR1 — V2280I missense in NBL — malignant-hyperthermia anesthesia precaution (ACMG secondary) PMID:28007021.
  • TNNT2 — R141Q heterozygous in OS — dilated cardiomyopathy (ACMG secondary) PMID:28007021.
  • VHL — I180T in EPM — Von Hippel-Lindau syndrome (ACMG secondary) PMID:28007021.
  • VOUS returned to four families under stringent criteria: ITK V175V in Hodgkin lymphoma + EBV, SDHC G75D in BLL, DICER1 D609Y in anaplastic large cell lymphoma (ALCL), and APC V1822D in Ewing sarcoma PMID:28007021.

Clinical implications

  • A more inclusive definition of “clinical utility” reframes the value of pediatric NGS. Limiting benefit to actionable-target identification understates the impact: 66% of all 101 cases and 75% of cases with cWES + RNA-seq received clinically meaningful information, including diagnostic clarification, prognostic risk-stratification, and pharmacogenomic guidance PMID:28007021.
  • Tumor RNA-seq is essential for comprehensive characterization, not optional. RNA-seq independently produced ~40% of clinically impactful findings in the cWES cohort, including translocation detection (e.g., PAX7–FOXO1, C11orf95–RELA), expression-pattern subtyping (BCR-ABL1-like signature in BLL, Ewing prognostic signature, medulloblastoma subgrouping, ATRT classification), and overexpression-based therapeutic targeting (CDK4/6 inhibitors for CDK4/MDM2 co-overexpression) PMID:28007021.
  • Germline analysis should be routine in pediatric oncology. Cancer-predisposition variants in 14% of patients (slightly above the 8.5–10% literature baseline) carry implications for the patient (e.g., transplant-donor selection in RUNX1 familial platelet disorder; avoiding transplant in Kabuki syndrome misdiagnosed as familial HLH) and for at-risk family members PMID:28007021.
  • Matched targeted therapy remains a bottleneck despite actionable findings. Only 16% of patients with potentially actionable alterations received matched therapy, due to lack of pediatric-approved drugs, limited single-agent efficacy of MEK inhibitors in RAS-mutant tumors, insurance non-coverage of off-label use, and limited compassionate-use access PMID:28007021.
  • Patients and families overwhelmingly choose to receive secondary findings. 94% of clinically consented patients opted in to learn ACMG secondary findings, undercutting the paternalistic argument for restricting return of results PMID:28007021.
  • Cost-benefit framing should weigh test-replacement value. Comprehensive NGS may replace multiple sequential conventional assays (karyotype, FISH, single-gene panels), particularly important when initial diagnostic biopsies are minimally invasive and tissue is limited PMID:28007021.
  • cBioPortal data deposit. VCFs, translocations, and gene expression are accessible through the PIPseq cBioPortal study under CUMC IRB approval PMID:28007021.

Limitations & open questions

  • Single-institution experience. All 101 cases come from Columbia University Medical Center; generalizability across pediatric oncology programs with different referral patterns or molecular tumor-board composition is not established.
  • Selection bias. The cohort represents ~32% of the institution’s pediatric oncology practice, restricted to <50% 5-year survival, outlier phenotype, rare cancers without standard-of-care, suspected predisposition, or relapsed disease — explicitly enriched for high-risk patients, possibly inflating both predisposition-allele frequency and actionable-alteration yield relative to unselected pediatric oncology populations PMID:28007021.
  • Retrospective review of clinical impact. “Clinically impactful” is a broad investigator-defined category; the authors do not report patient-outcome differences attributable to genomic-guided versus standard management.
  • Matched-therapy translation gap. Of 38 patients with actionable alterations, only 6 received matched targeted therapy, leaving open the question of whether the genomic findings would change outcome if barriers (drug access, insurance coverage, pediatric labeling) were removed PMID:28007021.
  • Variant-calling thresholds. The pipeline’s tolerance (5% allelic fraction in normal, 10% in tumor) was set to maximize sensitivity for tumor-board review; specificity tradeoffs against orthogonal validation in a multi-center setting are not characterized.
  • Reimbursement uncertainty. 80% of patients with reimbursement decisions received only partial reimbursement, and government plans paid $0 — sustainability of the program outside grant-funded contexts remains an open question.
  • Reporting policy heterogeneity. Carrier status, adult-onset variants, and most VOUS were not returned; this conservative policy may underestimate the patient-relevant yield of WES that other programs (with different return policies) would report.
  • No comparison against standard-of-care molecular testing. Whether comprehensive cWES + RNA-seq replaces (vs. supplements) karyotyping, FISH, or targeted panels for specific diagnoses is asserted but not formally benchmarked.

Citations from this paper used in the wiki

  • “NGS was performed on tumor and/or normal tissue from 101 high-risk pediatric patients. Potentially actionable alterations were identified in 38% of patients, of which only 16% subsequently received matched therapy” (Abstract).
  • “RNA-seq was clinically impactful in 37/65 patients (57%) providing diagnostic and/or prognostic information for 17 patients (26%) and identified therapeutic targets in 15 patients (23%)” (Abstract).
  • “Known or likely pathogenic germline alterations were discovered in 18/90 patients (20%) with 14% having germline alternations in cancer predisposition genes. American College of Medical Genetics (ACMG) secondary findings were identified in six patients” (Abstract).
  • “66% of cases tested through our program had clinically impactful findings and samples interrogated with both WES and RNA-seq resulted in data that impacted clinical decisions in 75% of cases” (Abstract).
  • “The mean mutational load across patients was 216.9 variants (SD=829.3, median =69), with a higher mean mutational load in solid tumors compared to hematologic malignancies” (p. 6).
  • “The most commonly mutated gene was TP53 (n=9, 9%) in solid tumor samples and RAS pathway constituents (NRAS: n=5, 5%; KRAS: n=3, 3%) in hematologic samples” (p. 6).
  • “the identification of a cKIT (p.Asn655Lys) mutation in a 7-year-old boy with acute myeloid leukemia (AML), who was subsequently treated with palliative imatinib and achieved a near-complete clearing of peripheral blood leukemic blasts with a sustained response for 9 months” (p. 7).
  • “Subsequent analysis identified a NUP214-ABL1 fusion by real-time polymerase chain reaction (RT-PCR) and the addition of dasatinib to the third-line induction regimen resulted in a deep remission allowing for a curative bone marrow transplant” (p. 7).
  • “Datasets are available through the cBioPortal for Cancer Genomics (http://cbioportal.org)” (p. 5).
  • “germline WES identified a pathogenic homozygous mutation in MLL2 (p.M3881Cfs*9) establishing the diagnosis of Kabuki syndrome … familial HLH was ruled out … and subsequently plans for a bone marrow transplant were averted” (p. 10).
  • “The estimated cost of WES (tumor/normal) was $4459 and the cost of RNA-seq was $1764” (p. 6).
  • “we identified germline alterations that predispose to cancer in 14% of our patients. This is slightly higher than other studies that have demonstrated approximately 8.5–10% frequency of germline risk alleles in pediatric oncology patients” (p. 11).

This page was processed by crosslinker on 2026-05-14.