Prevalence of Cancer Predisposition Variants in a Pediatric Pineoblastoma Cohort

 5
Prevalence of Cancer Predisposition Variants in a Pediatric Pineoblastoma Cohort

Annual Clinical Genetics Meeting, March 2026

Fang Fang, PhD, LGG Fellow, Nationwide Children's Hospital, Fang Fang, PhD, Kareesma Parbhoo, BSc, Vijayakumar Jayaraman, MS, Don Corsmeier, MSc, Douglas Depoorter, MS, Jessica L. Howard, MB(ASCP), Liana Hernandez, MS, Amy Everest, PhD, Swetha Ramadesikan, PhD, Benjamin J. Kelly, MS, Sean Caruthers, MS, Alejandro O. Bravo, Ph.D, Shalini Reshmi, MS, PhD, Karen Tsuchiya, MD, Daniel R Boué R. Boué, MD, Kathleen M. Schieffer, PhD, FACMG, Marco L. Leung, PhD, FACMG, Mariam T. Mathew, PhD, FACMG, Yassmine Akkari, PhD, FACMG, Elaine R. Mardis, PhD, Catherine E. Cottrell, PhD, Ying-Chen Claire Hou, PhD, FACMG

Introduction: Pineoblastoma (PB) is a poorly differentiated embryonal neoplasm that originates in the pineal region and accounts for less than 1% of all pediatric brain tumors. PB is a highly aggressive tumor and is often associated with a poor outcome with a median age at diagnosis of six years. To date, only a few cohorts of PB patients have undergone molecular profiling by genome-wide DNA methylation analysis, and the prognostic value of molecular subgrouping has been shown to be significant for clinical management of PB patients. Given the rarity of PB, the specific cancer predisposition genes and the prevalence and spectrum of cancer predisposition variants in PB remain largely unclear. Germline variants in DICER1 and RB1 are known to be associated with specific molecular subgroups of PB. Recent data also suggest that germline variants in DROSHA, a ribonuclease enzyme involved in miRNA biogenesis, may be linked to an increased risk of developing PB. Here, we present findings on germline cancer predisposition in a cohort of pediatric patients with PB from the National Cancer Institute’s Molecular Characterization Initiative.

Methods: Comprehensive genomic profiling, including paired exome sequencing of a comparator germline sample and disease-involved tissue, was performed for 28 patients (median age: 7.3 y [range: 2.8-20.6]) with PB enrolled between Mar. 2022 to Jan. 2025.  The capture panel targets coding genes in concert with a genomic “backbone” tiled every 0.3Mb, enabling the detection of single-nucleotide variants (SNV), copy number variants (CNV), and loss of heterozygosity (LOH). DNA methylation profiling and solid tumor fusion analysis were also performed.

Results: In this cohort, 53.6% (n=15) of individuals harbored germline loss-of-function (LOF) variants in genes involved in the miRNA processing pathway, including SNVs in 73.3% (n=11), and CNV in 26.7% (n=4). The affected genes were DICER1 (n=8), DROSHA (n=4), and DGCR8 (n=3). Among individuals with DICER1 germline variants, somatic second-hit variants included whole chromosome 14 loss or copy neutral LOH (n=6) and DICER1 nonsense variants (n=2), consistent with Knudson’s two-hit hypothesis. Similarly, for the four individuals with DROSHA germline variants, somatic second-hit LOF variants were also identified in all cases, suggesting that DROSHA may function as a tumor suppressor gene. In this cohort, we observed three individuals with 22q11.2 deletion syndrome, lacking one copy of DGCR8. Somatic second-hit LOF DGCR8 variants were detected in the corresponding tumor specimens, predicted to result in biallelic inactivation of DGCR8, supporting its potential role as a tumor suppressor. DNA methylation profiling of the corresponding tumors of individuals with germline variants showed epigenetic signature of pineoblastoma, miRNA pathway altered, group 1 or 2. While gene rearrangement events were not identified in any of these tumors, somatic copy number alterations were commonly observed involving gains of chromosomes 7, 12, 14, and 17, as well as loss or copy-neutral LOH of chromosome 14, representing a somatic second-hit of DICER1 variants.

Conclusion: In summary, 53.6 % of individuals in this cohort harbored germline LOF variants in genes involved in miRNA processing pathway. Studies from various cohorts have shown that approximately 10% of pediatric cancer patients have an underlying cancer predisposition syndrome. Our findings suggest that germline defects in miRNA biogenesis play a clinically significant role in PB pathogenesis and support that DROSHA and DGCR8 may function as tumor suppressor genes. Additionally, our results broaden the spectrum of malignancies associated with 22q11.2 deletion syndrome. The notable prevalence of germline predisposition among those with pineoblastoma highlights the need for systematic germline testing and the development of long-term surveillance.