PierianDx Launches the Clinical Genomicist Workstation for SaaS NGS Analysis

By Allison Proffitt 
 
July 23, 2014 | Washington University is throwing its hat in the clinical sequencing ring with a new spin out: PierianDx. The company launched about two months ago, and announced last week that Moffitt Cancer Center in Tampa, Fla., and Fletcher Allen Health Care, affiliated with the University of Vermont College of Medicine, have signed on as customers of the software-as-a-service tool provided through a HIPAA-compliant cloud. 
 
PierianDx’s flagship product, the Clinical Genomicist Workstation (CGW), is older than the company by about two years. Rakesh Nagarajan, now PierianDx’s Chief Biomedical Informatics Officer, was first tapped to build the informatics infrastructure for Wash U’s newly-established Genomics and Pathology Services (GPS) group in February 2011. 
 
NagarajanAt the time, “there were few to no tools that supported an end-to-end workflow for clinical next-generation sequencing-based assays,” Nagarajan tells Clinical Informatics News. GeneInsight, the tool developed by Partners HealthCare in Boston, solved part of Nagarajan’s problem, but not all of it. 
 
“GeneInsight was the singular product that was actually available to do some work,” Nagarajan says, but he wanted a pipeline that did more. “At that time—and even now—[GeneInsight] focused on writing the report once variants are identified, and really had no support for all of the up-front work: accessioning, sequencing, tracking, analytics to identify variants, annotate them. It started at the interpretation of variants and [helped] to write a report. It also didn’t push [reports to] the medical record.” 
 
Nagarajan wanted to build a broader tool, one that managed every step of the process. The team launched the Clinical Genomicist Workstation (CGW) in November 2011. 
 
Democratizing Clinical Genomics 
 
Over the past two years CGW has been used within Washington University and its affiliated hospitals: Barnes-Jewish Hospital at Washington University Medical Center and the Alvin J. Siteman Cancer Center. Since its launch, the CGW has undergone seven new releases. “We’re road-tested!” Nagarajan says. “The product has been used for well over two years now within Washington University and the GPS in real-day use to sign out thousands of cases of cancer, as well as cardiac disorders and other clinical trials, and things of that nature.” 
 
Now, through PierianDx, the CGW is the now being offered to laboratories. 
 
“We would like to democratize clinical genomics nationally and globally,” Nagarajan says. “We believe that clinical genomics will disseminate very broadly in the healthcare industry, and that… virtually every disease and every encounter will require this type of testing in the future.” 
 
At Moffitt, one of the first two announced customers, Mike Gruidl, Lead Assay Development Scientist in the Clinical Test Development Group, is in the process of validating two Illumina MiSeq sequencing instruments and two genomic testing panels as Moffitt ramps up its in-house sequencing operations.
 
Gruidl says a colleague had recommended he look at the work done at Washington University’s GPS unit while searching for the best software platform to understand and use next-gen sequencing data. The CGW immediately caught his attention. 

Gruidl shared his find with the working group. “What enticed us most was that they had a complete package, which would communicate with our [Cerner] hospital LIMS on both ends—both on ordering and reporting back the results.” 
 
That was exactly Nagarajan’s goal. A treating physician decides to order an NGS assay for a patient, and CGW manages the rest of the process. “That assay or that test is then accessioned, typically in order entry systems or laboratory information systems which then message the order electronically to CGW. The nucleic acid then needs to be isolated, library prepped, and sequenced. Sequence data are then generated,… analyzed to identify variants in that patient, annotate those variants, interpret those variants, [and] write a report,” Nagarajan says. Next, the molecular pathologist or a medical geneticist signs off on the report—a role PierianDx terms clinical genomicists, thus the name of the tool—and CGW then pushes the approved report into medical record. 
 
The end-to-end workflow management is a huge selling point for the CGW, Nagarajan believes, but there are analysis differentiators as well.
 
For identifying variants, CGW relies on pipelines clinically-validated at WashU to identify variants specifically and sensitively across a wide range of assays and disease conditions. CGW is sequencing technology-agnostic and experimental technology-agnostic, Nagarajan says, but “pipelines are unique for somatic cancer as well as constitutional disease, whether amplicon-based or targeted capture sequencing… to identify at a variety of variant types: single nucleotide variants, indels, and structural variants.”
 
From there, variants are automatically annotated with data from public databases and whatever commercial databases clients have licenses for to, for example, associate variants with known polymorphisms or to identify clinically-significant variants. 
 
“Variant classification and interpretation is built on a rules and decision-support engine that organizes variants and infers interpretations appropriately based on the context of the panel and disease that the laboratory is doing,” Nagarajan says.  
 
Developing the Knowledgebase 
 
The final step, before the clinical genomicists step in, relies on PierianDx’s curated and published knowledgebase. CGW’s initial knowledgebase focus was on somatic cancers, Nagarajan says, and has since evolved to support constitutional diseases, such as cardiac disorders. It’s an area of focus that he feels sets PierianDx apart. 
 
“We’ve done curation with dedicated curators of the published literature…wherein manuscripts are cuarted for their clinical relevance and are available just in time as part of the interpretative and sign-out process. That is an extremely active, ongoing process. We have dedicated curators who do this on a daily basis,” Nagarajan says. 
 
That knowledgebase was extremely attractive for Moffitt, says Gruidl. “That’s the beauty of the PierianDx platform. They’ve created a database for what they call the tertiary analysis. What that means is that they assigned a value, 1-5, for the clinical significance of different variants based upon science and the literature. Those classifications can be changed in house here [by] our oncologists and pathologists… [but] they’ve already done the hard work in generating a system, and having the system be able to identify and categorize these variants.” 
 
The final step before CGW pushes the report into the medical record is always review and interpretation by the clinical genomicist. “This system is not intended, nor should it ever be construed in its current state, to provide the final report that can directly go into the medical record,” Nagarajan stresses. “The system really greatly accelerates and facilitates… all of the components that I just told you, but professional review of all of those results as well as interpretations need to be finalized… in order for that report to have any weight in the medical record.” 
 
At Moffitt, Gruidl expects the MiSeq instrument validation to be completed by the end of September. Then Moffitt physicians will be able to order tests, sequencing will be done on campus, analysis done through the CGW and reviewed by Moffitt pathologists, and results returned to the EHR. Gruidl says now the process takes about a week in the lab. Once the workflow is fully incorporated—and pathologists are trained—he guesses it may take a week to ten days, though he’s not yet certain. 
 
In addition to the CGW SaaS tool, PierianDx also offers some consulting and customization services. For Moffitt, PierianDx is optimizing a dedicated Illumina TruSight Tumor Panel pipeline that Gruidl expects may speed the process. PierianDx also plans to offer sequencing for groups who don’t have their own instruments. 
 
Moffitt is using CGW with specific testing panels, but Nagarajan says CGW is capable of whole exome or whole genome workflows. “CGW has this concept of a test. That test can have any number of genes that are sequenced. It can have a subset of those, or all of those that are then reported.” 
 
Because the whole workflow is cloud-based, researchers can choose to substitute their own analytics or enter the analytic and interpretive process at any point. “CGW’s pluggable in that way,” Nagarajan says. “Whether that’s a white box solution that we can deploy for them in our environment, they can still use their own analytics. Or a black box solution, a third-party solution, where they’d like to do analytics on their own in order to generate VCF and possibly BAM files as well, then those can be input into the CGW as well. You can enter the workflow in a variety of ways.”
 
The “cloud” in this instance is a hosted environment at Washington University, but Nagarajan is open to other HIPAA-compliant vendors in the future. But having HIPAA and HITECH compliance and business associate agreements (BAAs) for partners is a “fundamental requirement”, he says. 
 
“We are providing a service to a clinical entity. In fact, we ourselves are business associates to those organizations, and therefore we’re required to provide that level of security in the service that we provide. Therefore, we’ll require that of our vendors as well.” 
 
Gruidl says Moffitt chose the cloud-based system instead of an in-house deployment. “Our informatics and hardware folks looked at all of that, and the clinical side was happy with the security of that system for our patients and providers.” 
 
The cloud platform enables another of PierianDx’s guiding principles: “We believe sharing is a good thing,” Nagarajan says. “Sites and clients are part of the CGW and Pierian partner network, [and] share a variety of information, whether that be assay or panel design details, [or] the interpretations that they each individually write at their sites.” He expects that sharing de-identified patient data—both genomic and clinical outcomes information—will be helpful in improving the quality of interpretation of future patient cases.