Expanded genetic testing of GIST patients identifies high proportion of non-syndromic patients with germline alterations

Authors

Mandelker D

Marra A

Mehta N

Selenica P

Yelskaya Z

Yang C

Somar J

Mehine M

Misyura M

Basturk O

Latham A

Carlo M

Walsh M

Stadler ZK

Offit K

Bandlamudi C

Hameed M

Chi P

Reis-Filho JS

Ceyhan-Birsoy O

Doi

10.1038/s41698-022-00342-z

PMID: 36593350 · DOI: 10.1038/s41698-022-00342-z · Journal: npj Precision Oncology (2023)

TL;DR

This study analyzed matched tumor-germline sequencing (MSK-IMPACT) from 103 GIST patients and an expanded cohort of 499 tumor-normal pairs at Memorial Sloan Kettering. It found that 69% of KIT/PDGFRA-wildtype GIST patients harbored germline pathogenic/likely pathogenic variants in GIST-associated genes (SDHA, SDHB, SDHC, NF1, KIT), with 63% of those patients lacking any syndromic features. The authors recommend germline testing for all KIT/PDGFRA-wildtype GIST patients regardless of syndromic presentation.

Cohort & data

  • 103 GIST patients who consented to germline analysis at MSKCC (April 2015 to June 2021), sequenced with MSK-IMPACT (341, 410, or 505 gene panels).
  • Extended validation cohort of 499 de-identified GIST tumor-normal pairs from the same institution.
  • Cancer type: GIST.
  • Dataset: gist_msk_2022 (publicly available on cBioPortal).

Key findings

  • 23% (24/103) of unselected GIST patients had a germline P/LP variant in a GIST-associated gene.
  • 69% (24/35) of KIT/PDGFRA-wildtype GIST patients harbored germline P/LP variants: 46% in SDH-complex genes, 20% in NF1, 3% in KIT.
  • 96% (24/25) of patients with somatic driver-negative tumors had germline P/LP variants.
  • 63% (15/24) of patients with germline GIST-associated variants had no personal or family history of syndromic features.
  • Patients with germline variants had younger median age-of-onset (39.5 vs 52 years, p=0.01), more metastatic disease (50% vs 20%, p=0.01), and more multifocal lesions (50% vs 18%, p=0.01).
  • In 499-patient tumor-only analysis, P/LP variants in BRCA1/BRCA2 and SDHB were almost exclusively germline in origin; KIT variants (430/431) were primarily somatic.
  • ESMO-recommended VAF thresholds (>30% for SNVs, >20% for indels) correctly distinguished all germline from somatic P/LP variants.

Genes & alterations

  • KIT — somatic activating mutations in 56% of GISTs; one germline P/LP variant (p.Lys509Ile) identified.
  • PDGFRA — somatic mutations in 10% of GISTs.
  • SDHA — germline LOF in 7/35 KIT/PDGFRA-wildtype patients (20%).
  • SDHB — germline LOF in 7/35 KIT/PDGFRA-wildtype patients (20%); 10/11 variants in tumor-only sequencing were germline.
  • SDHC — germline LOF in 2/35 KIT/PDGFRA-wildtype patients; SDHC promoter methylation tested and ruled out in driver-negative tumors.
  • NF1 — germline P/LP in 7/35 KIT/PDGFRA-wildtype patients (20%); 86% had NF1 clinical features.
  • BRAF — somatic activating mutations in 2 KIT/PDGFRA-wildtype tumors.
  • BRCA1 / BRCA2 — germline P/LP (incidental findings in KIT/PDGFRA-mutant GISTs); 4/4 BRCA2 variants in tumor-only were germline.
  • TP53 — one germline P/LP variant (incidental finding).
  • MLH1 — one germline P/LP variant; tumor was microsatellite-stable (likely incidental).

Clinical implications

  • Germline testing should be offered to all patients with KIT/PDGFRA-wildtype GISTs regardless of syndromic features.
  • SDH-deficient and NF1-related GISTs respond poorly to imatinib; distinguishing germline drivers guides therapy selection.
  • VAF thresholds from tumor-only sequencing can identify patients who should be referred for confirmatory germline testing (BRCA1/2, SDHB variants almost always germline).
  • Germline findings have cascade implications for at-risk family members (paragangliomas/pheochromocytomas for SDH carriers; nerve sheath tumors, breast cancer, gliomas for NF1 carriers).

Limitations & open questions

  • Clinical syndromic assessment was only available for the 103 consented patients, not the full 499-patient cohort.
  • Structural variants and low-level mosaicism may be missed by the MSK-IMPACT assay.
  • No standard clinical guidelines yet exist for universal germline testing in GIST; additional large-scale studies needed.
  • The role of incidental germline findings (BRCA2, TP53, MLH1) in GIST development remains unclear.

Citations from this paper used in the wiki

  • “69% (24/35) of individuals with a KIT/PDGFRA-wildtype GIST harbored a germline P/LP variant in a GIST-associated gene.”
  • “63% (15/24) of patients with germline GIST-associated variants did not have personal or family history of syndromic features.”
  • “P/LP variants identified in tumor-only sequencing of GISTs were almost exclusively germline in certain genes, such as BRCA1, BRCA2 (4/4 variants), and SDHB (10/11 variants).”
  • “Patients with germline GIST-associated variants had a younger median age-of-onset (39.5 vs 52 years, p=0.01), and were more likely to have metastatic disease (50% vs 20%; p=0.01) and multifocal lesions (50% vs 18%, p=0.01).”

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