Sourcing of patient samples proved a main area of interest for researchers gathering at the Clinical Proteomic Technologies for Cancer initiative's pre-application meeting this month in Bethesda, Md.
The several dozen scientists attending raised a variety of questions, including what tumor types would be investigated, the requirements for bringing in outside samples, and the formats of the samples to be studied.
Most pointedly, George Mason University researcher Emanuel Petricoin expressed concerns about the project's plan to use samples from different patient cohorts for its discovery and verification stages.
The second issuance of the National Cancer Institute's CPTC comprises two parts: a discovery stage wherein the Proteome Characterization Centers established by the program will look for potential biomarkers in ovarian, kidney, breast, and glioblastoma tumors; and a verification stage wherein the researchers will investigate smaller biomarker sets composed of the most promising candidates identified in discovery (PM 7/9/2010).
The NCI-funded Cancer Genome Atlas is providing tumor tissue for the discovery stage, along with genome analysis for the tissue, including sequencing, copy number variation, DNA methylation, gene expression, miRNA expression, SNP analysis, and somatic mutations. TCGA doesn't have blood from these patients for use in the verification stage of the CPTC initiative, however, which, Petricoin observed, means the researchers will be using samples from different cohorts in the two stages of the project.
"It's unclear to me [how] the linkage of the tissue analysis [in the discovery stage] with the blood serum analysis in the verification stage … is going to be maintained," he told ProteoMonitor in a conversation after the CPTC meeting.
In an e-mail to ProteoMonitor, CPTC director Henry Rodriguez said that the agency plans to match "patients providing samples for the verification stage to the tumors used in the discovery stage," citing as an example that "samples used for verification from ovarian cancer patients would involve patients with serous cystadenocarcinoma of the ovary, and exclude mucinous, clear cell, and endometriod ovarian cancers, as the discovery-stage samples involve patients with ovarian serous cystadenocarcinoma."
Petricoin said that such histological distinctions weren't his primary concern, however. He was most interested, he said, in whether the tumor and blood samples would be taken from patients with similarly sized tumors.
"I'm not talking about histology." he said. "I'm talking about size of tumor. Because nothing predicts outcome better than the size of the tumor. I'm talking about overt clinical measurements – stage, size of tumor, previous treatment – all of those things."
Petricoin noted at the meeting and in his conversation with ProteoMonitor that the initiative has "a hardwire bias into starting with tumors that have to be large from the outset," due to the fact that the project needs enough tumor tissue to distribute among the six to eight Proteome Characterization Centers that it will fund.
"I understand why [the CPTC is] doing it that way — to be able to distribute enough of it around for people to analyze – but it's a big issue," he said. "The serum and plasma that you'll be getting in the verification stage, is that going to be taken from patients with large tumors?"
Leigh Anderson, founder and CEO of the Plasma Proteome Institute, told ProteoMonitor that questions about using discovery and verification samples from different patient cohorts were "legitimate," but that "there is ample opportunity to select [the plasma and serum samples] to represent exactly what one would like to go after."
"It would be ideal, of course, to have the serum specimens from the same people as you had tissue tumor from, but I don't really know of any collections that actually have that on a large enough scale to do this," he added.
Once researchers reach the verification stage, there will be "much more precision in terms of what the diagnostic question is," Anderson said. "Are you trying to do early detection, or stage tumors, or determine who's going to benefit from therapy? Those are all individual questions you would approach by designing a carefully selected set of verification plasmas. And since that's basically to be defined, I think there's plenty of opportunity to do that."
Anderson noted that, because the discovery samples available from TCGA are "fixed at this point, the scientific questions you can ask, the diagnostic questions, are somewhat circumscribed," but using those samples will allow researchers to "provide a lot of proteomic characterization for samples that already have genomic characterization."
"TCGA results provide a unique, fruitful platform for guiding searches in the complex space of the proteome," Rodriguez said. "Complementing TCGA's genomics pipeline with CPTC's proteomics pipeline could produce a unique continuum that will allow the cancer research community to begin connecting cancer genotype to phenotype."
Petricoin, however, suggested that, while such research could be useful, TCGA's sample sets weren't necessarily ideal from a clinical proteomics perspective.
"They're trying to leverage resources that have been funded through our taxes. And I understand that," he said. "We have genomic sequencing done on [TCGA's tissue], so wouldn't it be interesting to see what the proteome of this is and be able to walk back and forth? That's all interesting research, but this is a clinical proteomics initiative. I think it requires a different layer of discussion around the tissue input."
He also questioned the decision to focus on plasma in the verification stage, saying that applications like "personalized therapy and aggressive versus indolent disease prognosis" would "require a tissue sample as an input" in a clinical setting.
"The focus is on blood as clinical applications where tissue is not readily accessible are practical if a blood biomarker is discovered," Rodriguez said. However, he noted that "tissue protein biomarkers may be relevant" in clinical applications like "prognosis and chemotherapeutic choice."
As a result, he said that the Proteome Characterization Centers are "encouraged to use tissues as appropriate to test proof of concept for a given verification assay," adding that samples for the verification stage have not yet been identified and that "use of tissue samples for a full-scale verification study is a possibility."
During the meeting Rodriguez noted that the initiative was open to obtaining verification samples from a range of sources, including participating researchers and various governmental bodies.