NEW YORK (GenomeWeb) – Some cancers have become poster children for precision and personalized medicine, some have benefitted less so from advances in genomic testing, and others have been largely cutoff from the sea change of genomic medicine and molecular diagnostics.
Retinoblastoma has in some ways been both, having been one of the first cancers in which driving tumor suppressor gene mutations were discovered and are now tested for, but lacking genomic predictors of disease severity or response to drugs that can help personalize treatment.
But a new discovery published late last week, researchers from Children's Hospital Los Angeles were able to extract and analyze cancer DNA from samples of aqueous humor, a pocket of liquid in the front of the eye.
The findings open up possibilities a new generation of clinical genetic studies and new avenues for prognostic subtyping, personalized treatment, and research into molecularly targeted drugs.
Reporting their findings in JAMA Ophthalmology, the team analyzed six aqueous humor samples from three retinoblastoma eyes — two samples from eyes that had been removed and four taken before patients were treated with an injection of chemotherapy.
Using low-depth whole genome sequencing, the investigators were able to identify chromosomal copy number variations representing changes in the tumor.
Jesse Berry, the study's first author and an ocular oncologist at CHLA, said this week that the impetus for the experiment was much like what has driven genomic discovery in other tumor types.
Just in the last few years, retinoblastoma treatment has expanded from two options — removal of the eye or treatment with systemic chemotherapy — to three, with the addition of injected chemo for cases in which tumor cells have seeded away from the retina into the eye's vitreous humor.
But clinicians lack tools to be able to predict which kids will respond to this injected treatment, and which will still have to have their eye removed, potentially after numerous failed charges.
"Two kids come in and they look exactly alike, and I would not know with more than 50 percent certainty whether these tumors will respond," Berry said.
As liquid biopsy techniques have advanced in blood, researchers have been looking to see if tumor genetic material can be assayed from blood samples in retinoblastoma cases, but without much luck, Berry added. And based on the early findings from work she and her colleagues have conducted, that probably makes sense.
"In our study, even in eyes being treated, the concentration [of tumor DNA] in the aqueous was really low, so if it got into the bloodstream it would be even lower."
Interestingly, Berry and her colleagues initially hoped they might find intact circulating cancer cells in the aqueous humor samples, as well as cell-free DNA fragments.
"Jim Hicks, the senior author does a lot of work with CTCs, and he [said], 'All I need is one cell, find me one cell and we've got it.' But we didn't. There were no cells," Berry said.
The team did detect cell-free DNA alteration, though. More precisely, they were able to compare cell-free DNA copy number profiles with tumor tissue analyses, which showed that the DNA floating in the aqueous humor was likely tumor derived.
In addition, the team picked up one nonsense RB1 mutation in one child's aqueous humor during treatment with chemotherapy and was later able to find the same mutation after the eye was eventually removed.
The data is very preliminary, but it opens up significant new possibilities for genomic study and the personalized treatment of children with these tumors.
Berry said she and her colleagues have already expanded their study to 50 individuals. The first insights they are targeting are mainly prognostic — looking for genomic features that distinguish more aggressive and less aggressive cases, that could be used to guide decisions on whom to treat with chemotherapy and who would be better served by immediate surgery.
But down the line, she said that the genomics of aggressive cases may yield leads for exploring treatment with targeted drugs.
On the genomics side, the challenge will be finding and optimizing the right methodology to analyze the scant circulating tumor DNA in these aqueous humor samples.
In a commentary accompanying the team's study in JAMA, the University of Miami's William Harbour reflected that detecting RB1 mutations from aqueous fluid for routine clinical purposes would be difficult.
"RB1 mutation screening is challenging even under optimal circumstances in which DNA quantity and quality are not limiting. This is a result of the large size of the RB1 gene and the virtually unlimited forms in which mutations can manifest," he wrote.
Berry said she and her team are working on that, and considering the advances being made in blood-based liquid biopsy, it seems like the technology will be there to support them.
Other labs have reached out to the CHLA team to do the sequencing, "but we want to keep the research here for right now," Berry said.