Liquid biopsies in pediatric oncology : towards clinical implementation

Abstract

During the last few years, liquid biopsies based on cell-free DNA (cfDNA) have become approved for clinical use in several areas of adult oncology. Due to the minimally invasive nature of this procedure, liquid biopsies could serve many purposes in childhood oncology as well. For instance, the method can provide highly specific biomarkers for molecular diagnostics, enable identification therapeutic targets and, more frequently, be used to assess therapeutic response or disease recurrence. This information could aid clinical decision-making in tailoring treatment and follow-up for each individual patient. However, developing customized assays for children's cancers is demanding for several reasons. First, childhood cancer is rare compared to adult cancers, making it difficult to get the numbers needed for statistically powered studies. Second, children’s cancers generally have few and often private mutations, and display a different mutational landscape than adult counterparts, meaning that generic assays are very challenging to develop. Third, small children mean small liquid biopsy volumes and high risk for subsampling errors, demanding ultrasensitive techniques for disease detection. Due to these circumstances, liquid biopsy assays for children with cancer require some innovative strategies. In this thesis, we describe a few liquid biopsy pilot studies focusing on the two main groups of childhood cancers; acute lymphoblastic leukemia and central nervous system (CNS) tumors.

In paper I, we use whole genome sequencing to identify structural variants (SVs) in leukemic cells and use the resulting unique sequences as targets for patient-specific droplet digital PCR (ddPCR) assays. Based on analysis of samples from six children, we show that it is technically feasible to use this approach to accurately quantify measurable residual disease in bone marrow as well as in cfDNA from plasma. The molecular assays also enabled detection of potential low-grade CNSinvolvement in half of patients at diagnosis. In paper II, we use a similar approach analyzing samples from 12 children with medulloblastoma, this time combining single nucleotide variants and SVs in our assays. We show that post-operatively, all tumors that grew in contact with cerebrospinal fluid (CSF) prior to resection, and where imaging supported or could not rule out residual tumor after surgery, molecular signs of the disease were seen in liquid biopsies on at least one occasion within three weeks of surgery. Furthermore, most plasma samples at diagnosis also bore molecular signs of the tumors. In paper III, we analyze cfDNA in CSF from a child with an inoperable brainstem tumor for BRAF V600E/K/R mutations. After confirmation of a BRAF mutation, molecular targeted therapy was initiated with dramatic clinical response during the first nine months of treatment. In paper IV, we used multiomics data to seek a molecular diagnosis for a child born with a large CNS tumor. The data revealed a rare type of glioma, most likely caused by a novel fusion gene; SNRNP70::ALK. Molecular targeted therapy was initiated and four months into treatment, there are clinical and radiological signs of treatment effect. Taken together, the results indicate that our approach to design molecular assays could expand the utility of ddPCR for liquid biopsy analysis. Furthermore, in clinically challenging cases, multiomics can resolve complex diagnoses and liquid biopsies can guide choice of treatment.

These studies are the initial efforts of our research group to promote liquid biopsies as a powerful precision diagnostic tool in pediatric oncology. Apart from the 20 patients described in these studies, another 130 children have been included for research and a liquid biopsy biobank of >1800 samples has been established. We have also set up clinical liquid biopsy analysis for BRAF V600 and H3-3A K27M mutations during this period. This was achieved through a close collaboration between the Departments of Clinical Genetics, Pediatric Oncology, Pediatric Neurosurgery, Clinical Pathology and The Swedish Childhood Tumor Biobank.