NEW YORK – Researchers at Children's Hospital Los Angeles (CHLA) have developed a liquid biopsy assay that uses low-pass whole-genome sequencing assay to help diagnose or provide prognostic information for pediatric solid tumors.
The researchers detailed their development and validation of the assay in a paper published this month in NPJ Precision Oncology. They tested the assay on 143 samples from 73 patients with a newly diagnosed or recurrent malignant bone or soft-tissue sarcomas or germ cell, hepatic, thyroid, or renal tumors, along with 19 non-cancer control patients.
Jaclyn Biegel, chief of genomic medicine at the Center for Personalized Medicine at Children's Hospital Los Angeles and senior author of the study, said that the most recent work was spurred by a CHLA colleague's prior development of an ultra-low-pass sequencing assay for retinoblastoma patients using aqueous humor samples. Biegel's team wanted to develop a test that could be used across pediatric tumors and use different sources for liquid biopsy samples.
The resulting test was developed to screen for copy number alterations and gene fusions in cell-free DNA (cfDNA) found in aqueous humor, cerebral spinal fluid, and blood samples. CHLA launched the test late last year, Biegel said, and has clinically tested about 10 patients with it so far.
"We've had phenomenal success so far," she said. "The numbers are small, but it has been really helpful in identifying patients where the diagnosis isn't clear, such as distinguishing hepatoblastoma from embryonal sarcoma of the liver or identifying poor prognostic features in patients with retinoblastoma. This assay has complemented our standard diagnostic testing that we do in the lab."
In the paper, the researchers first evaluated whether the low-pass WGS assay could detect circulating tumor DNA (ctDNA) and copy number alterations. It detected ctDNA in 70 percent of diagnostic samples, or samples from patients collected at initial diagnosis prior to chemotherapy or definitive surgery, and in 43 percent of relapse samples, or samples from patients who enrolled after relapse. Of the diagnostic samples, 67 percent of patients with localized disease and 80 percent of patients with metastatic disease had copy number alterations detected by the low-pass WGS test.
Then, they evaluated whether the test could detect copy number alterations associated with specific tumor types. Among 17 osteosarcoma patients, the assay could detect genomic instability, a characteristic of the disease, in 53 percent of patients. Of 13 Ewing sarcoma patients, copy number alterations were detected in nine patients, with seven patients exhibiting gain of chromosome 8, an abnormality frequently observed in this disease.
In other tumor types, copy number alterations were detectable in 75 percent of newly diagnosed renal tumor patients and in all five relapsed renal tumor patients. All three patients with newly diagnosed germ cell tumor and one of two patients with germ cell tumor at relapse had detectable alterations.
Three of four patients with hepatic tumors had detectable copy number alterations at diagnosis. The assay also detected alterations in one patient with a peripheral malignant nerve sheath tumor, one patient with alveolar rhabdomyosarcoma, and one patient with synovial cell sarcoma. Neither of the two patients with thyroid tumors had detectable alterations.
Of the 59 patients who had detectable alterations, 34 patients (69 percent) had abnormal copy number alteration profiles by the low-pass WGS assay compared to 49 abnormal profiles identified in tumor tissue analysis.
Finally, the researchers examined whether the low-pass WGS assay could detect pathogenic variants and translocations. There was next-generation sequencing data from CHLA's in-house OncoKids assay available for 53 patients in the study, 26 of whom harbored mutations in clinically significant cancer genes. By comparison, the low-pass WGS assay was able to identify pathogenic variants in 14 patients, about half as many as the NGS assay. The mutations identified by low-pass WGS included alterations in TP53, RB1, NF1, KIT, and KRAS. The researchers noted that more studies are needed to determine if the low-pass WGS assay is clinically feasible to detect these mutations, or whether it requires a companion targeted capture NGS-based assay.
The first version of the test, which can detect copy number alterations to help diagnosis, launched last year. But the researchers are developing a second version that can screen for mutations and gene fusions, which is important in pediatric cancer, Biegel said.
"We see the proof-of-concept for version two in this manuscript where we are able to retrospectively look at that whole-genome sequencing data and find mutations where we knew where they were located," she said. "Going forward [with version two] will be more challenging because we don't have as much sequence data and we don't know what mutations we're looking for, so to distinguish the background noise from the driver mutations is going to be more challenging."
To aid identification of driver mutations and gene fusions, the researchers have also developed more targeted panels based on the WGS data. In the paper, they detail two targeted panels to detect EWSR1 and FOXO1 fusions that can help diagnose Ewing sarcoma and alveolar rhabdomyosarcoma, respectively. They found that the WGS assay could detect EWSR1 fusions in 10 of 12 Ewing sarcoma patients who had these mutations previously confirmed by OncoKids NGS. Biegel noted they will continue to develop these targeted panels for other tumor types.
Previous studies have explored the benefits of genomic profiling for pediatric cancers. A 2022 study found about one-third of adult and pediatric sarcoma patients harbored actionable mutations and about 10 percent of patients were diagnosed with the wrong sarcoma subtype after analyzing NGS results. A number of studies have also explored the frequency of potentially actionable mutations in pediatric cancers. One study found 86 percent of pediatric cancer patients with extracranial solid tumors harbored genomic alterations that could potentially impact their care.
Midhat Farooqi, director of medical oncology at the Center for Pediatric Genomic Medicine at Children's Mercy Kansas City, who was not involved with the study, noted that the CHLA assay's ability to detect copy number alterations and gene fusions could aid in diagnosis and prognosis for some pediatric tumor types.
However, Farooqi said the CHLA assay would likely need to be supplemented by other diagnostics. "To their credit, the authors of the manuscript note … that three of the fusion-positive patients were not detected by their low-pass whole-genome sequencing liquid biopsy assay and so comprehensive characterization, via addition of a higher coverage targeted panel, was recommended," he said.
Farooqi also said the assay has a different design than many commercially available liquid biopsy assays. He noted several important differences: The CHLA assay has lower coverage and sensitivity due to the low-pass design, it sequences the whole genome, and it is focused on detecting copy number alterations.
"Most commercially available assays are the opposite: They target certain regions of the genome, have higher coverage, and are able to detect sequence alterations, instead of copy number alterations, though certain liquid biopsy assays can do both, plus [detect] fusions," he added.
To optimize their assay for pediatric patients, the CHLA researchers designed the low-pass WGS assay to work with smaller samples. Biegel noted that compared to a typical adult blood draw of 15 ml to 20 ml, blood draws from pediatric patients are often around 5 ml. While there are many commercially available liquid biopsy assays, they can be challenging to use for pediatric patients because of these small sample sizes or because they do not screen for copy number alterations, Biegel said.
The researchers also wanted to develop a test for newly diagnosed patients because many liquid biopsy pediatric assays focus on relapsed or metastatic patients where ctDNA is more easily detected, Biegel said.
"We really wanted to see if we could use this upfront for patients where surgery might not be an option or where there's no primary site identified," Biegel said. "There's an example in the paper of a patient who had a suspected germ cell tumor, based on the fact that they had high alpha-fetoprotein levels, but there was no primary site that they could biopsy. We were very quickly able to identify copy number alterations that were characteristic for a germ cell tumor in the peripheral blood."
The researchers expect to publish more data validating the clinical utility of the assay, which will help support potential insurance reimbursement. The next study will use cerebral spinal fluid as a liquid biopsy for patients with brain tumors, followed by a study exploring serial testing with the assay to monitor treatment response and screen for early signs of relapse, Biegel said.
"In this paper, we looked retrospectively at samples collected over a year [for one patient], and we were able to detect ctDNA reflective of recurrence at as early as three months after therapy, whereas the patient didn't show evidence for recurrence until a year after treatment," Biegel said. "The additional studies are really going to be [longer term] and more targeted sequencing approaches to identify additional sequence variants and gene fusions."