Sequential genomic analysis using a multisample/multiplatform approach to better define rhabdomyosarcoma progression and relapse

Authors

de Traux de Wardin H

Dermawan JK

Merlin MS

Wexler LH

Orbach D

Vanoli F

Schleiermacher G

Geoerger B

Ballet S

Guillemot D

Frouin E

Cyrille S

Delattre O

Pierron G

Antonescu CR

Doi

10.1038/s41698-023-00445-1

PMID: 37730754 · DOI: 10.1038/s41698-023-00445-1 · Journal: npj Precision Oncology (2023)

TL;DR

This study investigated the genomic evolution from primary to relapsed rhabdomyosarcoma (RMS) in 35 paired primary/relapse samples (18 fusion-positive, 17 fusion-negative) from MSKCC and Institut Curie, using MSK-IMPACT, whole-exome sequencing, and whole-genome sequencing. Fusion-positive RMS (FP-RMS) showed stable genomes at relapse, with secondary alterations in CDKN2A/B, MYCN, and CDK4 present at diagnosis and correlating with worse survival. Fusion-negative RMS (FN-RMS) acquired more alterations at relapse, most commonly SMARCA2 missense mutations. A complementary ctDNA liquid biopsy analysis in 10 patients demonstrated feasibility of detecting PAX3::FOXO1 fusions, SNVs, and CNVs in plasma, with fusion read counts correlating with disease burden.

Cohort & data

  • 35 patients with relapsed/metastatic RMS: 18 fusion-positive (ARMS) and 17 fusion-negative (ERMS), from MSKCC (n=20) and Institut Curie (n=15) PMID:37730754.
  • MSK cohort sequenced on MSK-IMPACT; Curie cohort sequenced by WES or combined targeted DNA (Dragon panel) and WES.
  • Dataset: rms_msk_2023.
  • 10 patients (7 FP-RMS, 3 FN-RMS) had longitudinal ctDNA analysis from 62 plasma samples using a 36-gene custom RMS panel for SNV/fusion detection and shallow WGS for CNV assessment.
  • Archer FusionPlex used for FOXO1 fusion evaluation in the MSK FP-RMS cohort.
  • Median age 14 years (FP-RMS) and 7.8 years (FN-RMS) in the MSK cohort; 5-year overall survival was 33%.

Key findings

  • FP-RMS genomic stability at relapse: FP-RMS had a median of 2 alterations at primary and 3 at relapse (range 0–7). Half of patients acquired additional alterations at relapse (median 1, range 0–5) PMID:37730754.
  • Secondary major alterations in FP-RMS: CDKN2A/CDKN2B alterations (5/18, 28%), MYCN alterations (4/18, 22%), and CDK4 amplifications (3/18, 17%) were the most common recurrent events, mostly present at diagnosis and largely mutually exclusive PMID:37730754.
  • Survival impact in FP-RMS: MYCN alterations (p=0.0012) and CDKN2A deletions (p=0.049) each had statistically significant impact on OS. Combined CDKN2A and MYCN alterations correlated with worse OS (p=0.0017) and PFS (p=0.002) PMID:37730754.
  • Two FP-RMS subgroups: Patients with secondary major alterations at diagnosis acquired zero or very few new alterations at relapse; patients lacking such events developed more alterations at relapse, including de novo AKT1 deletions and CDK4/MYCN alterations PMID:37730754.
  • FN-RMS heterogeneity: FN-RMS acquired an average of 4.3 new alterations per patient at relapse (range 0–17), with 111% overall increase in alterations. Average mutations per sample was higher in FN-RMS vs FP-RMS (5.6 vs 3.2, p=0.002) PMID:37730754.
  • Recurrent FN-RMS alterations: TP53 (7/17, 41%), BCOR (5/17, 30%), PIK3CA pathway genes (HRAS, NRAS, KRAS, PIK3CA; 5/17), MDM2 (4/17), and SMARCA2 (4/17) PMID:37730754.
  • SMARCA2 as relapse-specific event: SMARCA2 missense mutations and frameshift deletions (4/17, 24%) were found exclusively at relapse in FN-RMS, with immunohistochemistry showing loss of SMARCA2 protein expression PMID:37730754.
  • BAP1 deletions (2/17, 12%) also found only at relapse in FN-RMS PMID:37730754.
  • ctDNA detection rates: Combined approach detected alterations in 88% at diagnosis (7/8) and 90% at relapse (9/10). PAX3::FOXO1 fusion detected in 100% at diagnosis (6/6) and 86% at relapse (6/7). Blood plasma showed higher sensitivity than bone marrow (86% vs 61%) PMID:37730754.
  • Fusion reads correlate with disease burden: Higher PAX3::FOXO1 fusion read counts correlated with increased disease burden and relapse in patients with fatal outcomes PMID:37730754.

Genes & alterations

  • PAX3::FOXO1 fusion – pathognomonic driver in FP-RMS (16/18 patients); detected in ctDNA at diagnosis (100%) and relapse (86%).
  • PAX7::FOXO1 fusion – found in 2/18 FP-RMS patients, associated with FOXO1 amplification.
  • CDKN2A/CDKN2B – homozygous deletions and loss-of-function in 28% of FP-RMS; correlated with worse OS (p=0.049) and PFS (p=0.0082).
  • MYCN – amplifications and missense mutations in 22% of FP-RMS; correlated with worse OS (p=0.0012); high MYCN mRNA expression confirmed by RNA-seq.
  • CDK4 – amplifications in 17% of FP-RMS, often co-amplified with GLI1 (2 tumors) or MDM2 (1 tumor) at 12q13-15 locus.
  • AKT1 – deletions acquired at relapse in FP-RMS lacking major secondary events.
  • TP53 – loss-of-function in 41% of FN-RMS patients (33% of samples); previously associated with poor outcome.
  • BCOR – alterations in 30% of FN-RMS patients (24% of samples).
  • SMARCA2 – missense mutations and frameshift deletions in 24% of FN-RMS at relapse only; 4 different point mutations with no redundancy in protein domain.
  • PIK3CA, HRAS, NRAS, KRAS – RAS/PIK3CA pathway alterations in 5/17 FN-RMS.
  • MDM2 – amplifications in 4/17 FN-RMS, detectable in ctDNA.
  • BAP1 – deletions in 2/17 FN-RMS, exclusively at relapse.
  • NF1, FGFR1 – driver events in a subset of FN-RMS.
  • RB1 – SNV detected in ctDNA of FP-RMS.
  • MYOD1 – L122R mutation noted as marker for spindle/sclerosing RMS.
  • IGF2, MED12, NCOR2 – recurrent alterations in FP-RMS.
  • GLI1 – co-amplified with CDK4 in 2 FP-RMS tumors.
  • FBXW7 – SNV detected in ctDNA at relapse in FN-RMS.
  • DMD – 30Kb deletion acquired at relapse in one FN-RMS patient.
  • DNA repair genes (RAD51B, RAD21, RAD51C, ATR, ATM, FANCA, FANCC) – exclusively at relapse in FN-RMS with higher mutational counts.

Clinical implications

  • Secondary alterations in CDKN2A, MYCN, and CDK4 at diagnosis in FP-RMS identify a high-risk subgroup with worse OS and PFS, supporting their use in risk stratification PMID:37730754.
  • Liquid biopsy using a combined targeted DNA panel plus shallow WGS provides reliable detection of PAX3::FOXO1 fusions, SNVs, and CNVs in both FP-RMS and FN-RMS subtypes PMID:37730754.
  • Fusion read number quantification via unique molecular identifiers (UMIs) enables monitoring of disease burden and treatment response over time PMID:37730754.
  • Patients with FP-RMS harboring CDKN2A, CDK4, and MYCN alterations, and FN-RMS with TP53, BCOR, or RAS pathway alterations, may benefit from shorter intervals between cfDNA collections during follow-up PMID:37730754.
  • Rationale provided for implementing ctDNA analysis in the prospective FaR-RMS trial (NCT04625907) PMID:37730754.

Limitations & open questions

  • Small cohort size (n=35 tumor pairs, n=10 ctDNA cases) limits statistical power; findings require validation in larger prospective studies.
  • Retrospective sample collection and lack of centralized blood processing introduced variability; 4 cases excluded due to undetectable ctDNA.
  • The prognostic role of MYCN and CDK4 amplifications in FP-RMS remains controversial, with conflicting findings in prior literature.
  • The functional significance of relapse-specific SMARCA2 alterations in FN-RMS requires further investigation.
  • No correlation found between cfDNA concentration and clinical status, consistent with prior observations that somatic mutation detection is more reliable than cfDNA levels alone.
  • The study did not include digital droplet PCR, which may be more sensitive for fusion detection in low-burden settings.

Citations from this paper used in the wiki

  • “FP-RMS have a stable genome with relapse, with common secondary alterations CDKN2A/B, MYCN, and CDK4 present at diagnosis and impacting survival.” (Abstract)
  • “MYCN alterations (p=0.0012) and CDKN2A deletions (p=0.049) have a statistically significant impact on overall survival (OS) either alone or when combining both altered populations (p=0.0017).” (Results)
  • “SMARCA2 alterations were the most common events acquired in relapsed FN-RMS, with 4 different point mutations found (VAF<50%) with no redundancy in protein domain.” (Results)
  • “ctDNA alterations were evaluated using a targeted 36-gene custom RMS panel at high coverage for single-nucleotide variation and fusion detection, and a shallow whole-genome sequencing for copy number variation.” (Abstract)
  • “All cases had increased fusion reads number at higher disease burden timepoints as well as at relapse in patients following a fatal outcome.” (Results)
  • “Overall the 5-year OS was 33%.” (Results)

This page was processed by crosslinker on 2026-05-04.