Molecular profiling of 888 pediatric tumors informs future precision trials and data-sharing initiatives in pediatric cancer

Despite advancements in pediatric oncology, cancer remains a leading cause of death among children in the United States, and many survivors experience long-term treatment sequelae. Progress has been uneven, with some diagnoses showing minimal improvement in outcomes. Precision oncology, which involves characterizing the genome and employing molecularly targeted therapies, offers a path towards improved outcomes and reduced toxicity. However, the rarity of many pediatric cancers hinders the generation of large, meaningful genomic datasets to guide precision oncology. This study leverages data from the Profile Cancer Research Study at Dana-Farber/Boston Children's Cancer and Blood Disorders Center, which has performed clinical-grade targeted next-generation sequencing (NGS) on tumor samples since 2013. The study aimed to analyze the genetic features of common and rare pediatric cancers, facilitating clinical trial design, real-world precision oncology practice, and contributions to data-sharing initiatives like the National Cancer Institute's Childhood Cancer Data Initiative (CCDI). The study analyzed data from a cohort of 888 pediatric patients with solid tumors, aiming to comprehensively characterize the genomic landscape of these tumors and match genomic alterations to existing clinical trial protocols. This work provides crucial insights for the design of future pediatric oncology clinical trials and promotes data sharing to improve the treatment of rare pediatric cancers.

The literature extensively documents the improved survival rates in pediatric oncology over the past 50 years, largely attributed to intensified chemotherapy, risk stratification, and multi-modal treatments. However, the continued high mortality rates from cancer and the long-term effects on survivors remain major concerns. The application of precision oncology, tailored to specific genomic characteristics, holds significant promise. Existing studies, such as TARGET and others have provided valuable insights into the genomic landscape of common pediatric cancers. Nevertheless, these studies often underrepresent the large number of rare pediatric cancers, hindering the development of targeted therapies for these diseases. This study addresses this gap by focusing on a diverse cohort of patients encompassing a broad range of both common and rare pediatric solid tumors.

Between September 2013 and March 2019, 1120 pediatric patients with intracranial or extracranial solid tumors consented to the Profile Cancer Research Study. Targeted NGS of tumor samples was conducted using the OncoPanel assay at the Center for Advanced Molecular Diagnostics (CAMD). Successful sequencing was achieved for 848 (76%) of the participants. The final analytic cohort included 888 patients. Diagnoses were classified according to the ICD-O-3.2 standard by a multidisciplinary team. The OncoPanel assay detected single-nucleotide variants (SNVs), insertions and deletions, copy number alterations (CNAs), and structural variants (SVs). Variant allele fraction (VAF) ≥5% was used as a cutoff, with lower VAF variants included if deemed oncogenic by molecular pathologists. Data from ClinVar and gnomAD databases were used to filter variants. Tumor mutation burden was calculated. Genomic alterations were matched to the actionable mutation lists (AMLs) of three precision oncology trials: NCI-COG Pediatric MATCH, NCI-MATCH, and ASCO TAPUR. Matches were defined as precise matches on the AML list or the same gene and variant type. Medical records were reviewed to determine if patients received molecularly targeted therapy corresponding to the matched variants. Clinical data were collected from electronic medical records (EMRs), including demographics, histology, treatment regimens, and imaging data. The cohort's generalizability was assessed by comparing diagnostic proportions to the National Cancer Institute's National Childhood Cancer Registry (NCCR). Statistical methods were not used to predetermine sample size, and no data were excluded from the analyses. Data were submitted to the NCI's Childhood Cancer Data Initiative (CCDI) and dbGaP.

The study included 888 pediatric patients with 95 distinct diagnoses. 55% of patients had one of ten common pediatric cancers, while 45% had one of 85 rare diagnoses. The cohort's extracranial solid tumor diagnoses proportionally resembled those in the NCCR. Expected genomic events were observed in specific diagnoses (e.g., TP53 rearrangements in osteosarcoma). Activating PIK3CA alterations (18 cases, 2% of the cohort) and ARID1A inactivating mutations (10 cases, 1%) were distributed across multiple cancer diagnoses. In rare tumors not included in previous pan-cancer analyses, recurrently altered genes included CTNNB1, DICER1, and NF1 (extracranial) and CTNNB1, NF2, and KIT (CNS). 33% (289/888) of patients had at least one variant matching a precision oncology trial protocol, and 14% (41/289) received matched molecularly targeted therapy. The study also highlighted the feasibility of using the ICD-O-3.2 standard for classifying pediatric cancer diagnoses.

This study provides a comprehensive genomic landscape of pediatric solid tumors, including both common and rare cancers. The high proportion of patients with rare cancers highlights the need for large-scale data sharing initiatives to fully understand the genomic features and identify therapeutic targets for these diseases. The findings demonstrate the potential of integrating genomic data from routine clinical sequencing into precision oncology clinical trials. The significant number of patients with alterations matching existing trial protocols underscores the potential for expanding access to targeted therapies. The study also emphasizes the importance of standardized diagnostic classification, using systems like ICD-O-3.2, to facilitate data sharing and research. The data sharing efforts through CCDI and dbGaP will enable further research and collaborative efforts to improve outcomes in pediatric oncology.

This large-scale molecular profiling study of pediatric solid tumors revealed a substantial number of patients with targetable genomic alterations, emphasizing the potential for precision oncology. The study's comprehensive assessment of both common and rare cancers underscores the critical need for collaborative data-sharing initiatives to advance our understanding of and treatments for these cancers. Future research should focus on validating the findings in larger, independent cohorts and on developing novel targeted therapies based on the identified genomic alterations. Standardizing clinical data reporting will further enhance the utility of these data.

The study is limited by its retrospective design and single-institution setting, which may affect the generalizability of the findings. The use of different versions of the OncoPanel assay across the study period might introduce some technical variability. The reliance on EMR data for clinical information might introduce inconsistencies or missing data. The relatively low number of patients with matched variants receiving molecularly targeted therapies should be considered when interpreting the impact of this approach.