NEW YORK – Findings from a 10-year multiomics study out of the Translational Genomics Research Institute (TGen) and the Beat Childhood Cancer Research Consortium underscore the importance of profiling pediatric tumors not only at the time of diagnosis, but also at the point of disease relapse.
The findings, published Tuesday in the journal Cancer Research, show that the mutational landscape of pediatric and young adult solid tumors can differ significantly after early lines of treatment, which in turn could have crucial implications for the type of treatments patients receive in the later-line setting. The study, which details genomic and transcriptomic findings from 250 solid tumor samples from around 200 relapsed or refractory patients, are the result of a collaboration between TGen, a nonprofit affiliate of City of Hope, and the Childhood Cancer Research Consortium, a global group of over 40 centers conducting clinical trials for childhood cancers.
"A lot of discovery-oriented genomic profiling of cancer in general has focused on the primary disease setting," said William Hendricks, a TGen researcher and one of the lead authors on the study. "There's still a dearth of information on the genomic underpinnings of relapsed and refractory cancers, and that is particularly true in the pediatric solid tumor space, where these rare tumors have less information on them in general."
At the core of the researchers' sweeping, multi-institutional effort was the goal of identifying key genes, pathways, signatures, and features of relapsed and refractory tumors that could guide future translational efforts and, ultimately, home in on treatment strategies for these difficult-to-treat cancers.
The study included 173 pediatric cancer patients and 29 adolescent and young adult patients — those over age 15 — who were enrolled in one of three multi-institution Beat Childhood Cancer Consortium basket and umbrella trials between 2011 and 2018. These patients had been diagnosed with 46 different kinds of solid tumors, which the researchers grouped into categories including neuroblastomas, sarcomas, central nervous system tumors, and other rare cancer types.
Ashion Analytics, which was TGen's CLIA-certified and CAP-accredited sequencing lab but was purchased by Exact Sciences earlier this year, conducted whole-exome sequencing of patients' tumor tissue and normal blood samples as well as tumor mRNA sequencing.
Using a threshold of at least two somatic mutations per megabase to define tumor mutational burden (TMB)-high cancers — a threshold proposed in previous research based on the fact that mutational frequency is 14 times lower in pediatric tumors than in adult tumors overall — the researchers found 35 percent of relapsed and refractory tumors had this feature, including two "hypermutated" tumors with more than 10 mutations per megabase.
This, the researchers noted, was significantly higher than the 1.3 percent of patients classified as TMB-high in a previously published cohort of pediatric cancers that had been profiled at the point of diagnosis. Both cohorts used the same, two mutation-per-megabase cutoff.
Notably, 43.2 percent of patients' tumors showed mutational signatures associated with prior chemotherapy treatment, and the two hypermutated tumors were enriched for a mutational signature, dubbed SBS11, with a known association with prior treatment with alkylating agents.
"These treatments actually induced a hypermutation phenotype that may have implications for potential immunotherapy response," said TGen's Sara Byron, the study's co-lead author along with Hendricks. Byron highlighted the fact that TMB as well as microsatellite instability status has been shown to correspond with improved responses to immune checkpoint inhibitors in adult tumors.
Indeed, the PD-1 checkpoint inhibitor pembrolizumab (Merck's Keytruda) has tumor-agnostic approval for cancers that are MSI-high or mismatch repair-deficient as well as TMB-high, defined as at least 10 mutations per megabase. Both these tissue-agnostic approvals are for refractory adult and pediatric patients with these biomarkers.
However, whether pediatric cancer patients with treatment-induced hypermutated tumors respond to checkpoint inhibitors remains an open question, according to Byron. "The hypermutation in pediatric cancers have been reported to be linked to germline or inherited mismatch repair deficiencies … but what we're seeing here is different in that it's not germline mismatch repair deficiencies causing the hypermutation. … It's really treatment-induced." The researchers hope their findings will spur further investigations on how well checkpoint inhibitors work in this treatment-induced hypermutation setting specifically in pediatric patients.
Importance of sequential profiling
Beyond the treatment-induced hypermutation findings and potential treatment implications, the researchers found that several actionable alterations emerged in patients' tumors over time. Longitudinal profiling of 20 patients showed a PIK3CA hotspot mutation in a patient with rhabdomyosarcoma as well as an acquired ALK mutation in a neuroblastoma patient.
While Byron, Hendricks, and colleagues acknowledged that these longitudinal findings were a bit limited by the lack of diagnostic biopsies in the relapsed and refractory cohort, they pointed to these findings as an example of the value of rebiopsying tumors at the point of recurrence and noted the need for additional evaluation in a larger dataset.
"There is more work that needs to be done to help us better understand how tumors change over time," said Hendricks, who acknowledged that there can be significant challenges to repeat profiling, in part because doing so requires that patients undergo repeat tumor biopsies, which isn't always possible depending on the cancer type and tissue availability.
"We are increasingly seeing new actionable variants that are coming up at a later time point, which does drive home the importance of being able to measure what's going on in the tumor throughout the treatment course," said Byron. "That additional look later on may identify something else that could be targeted."
More work is also needed on improving pediatric cancer patients' access to tumor profiling in the first place. Although Hendricks noted there has been an uptick in baseline tumor profiling for these patients, it is often only through clinical trials at specialized academic centers that patients undergo repeat profiling.
While other precision oncology trials are analyzing tumor biomarker changes over time, Byron said the in-depth analysis in the TGen study was unique. She referenced, for instance, the ongoing National Cancer Institute's pediatric MATCH trial involving patients with treatment-relapsed tumors. The NCI trial is using a targeted gene panel, whereas TGen's trial used whole-exome tumor/normal sequencing plus tumor RNA sequencing. This allowed for a deeper dive into mutational signatures. "Having that expanded dataset, including the RNA data, was really helpful for fusion detection," she said.
Going forward, Hendricks and Byron hope that others will build on elements of their findings, which will allow the community to zero in on what these tumor characteristics might mean for treating pediatric cancer patients and improve their access to this deep level of sequencing.
"We want to continue to expand understanding of under-researched pediatric cancers [and] the role of new drugs based on a biologic and genomic rationale, whether that's PARP inhibitors or combinations with chemotherapy and standard of care, or immune checkpoint blockade," Hendricks said. "The more we can advance both the logistics of bringing genomics into the clinic for pediatric patients … the better."