CHICAGO (GenomeWeb) – Late last year, NewYork-Presbyterian Hospital launched Weill Cornell Medicine's Exome Cancer Test version 1.0 (EXaCT-1) for clinical cancer care, following three years of development. Meanwhile, Weill Cornell is far along in development of the next generation, EXaCT-2, and is exploring ways to commercialize that test while also looking to bring the original assay to other affiliated hospitals.
EXaCT-1 was the first whole-exome sequencing test for cancer approved in New York State. Weill Cornell researchers subsequently ran a clinical validation study in 1,091 cases, more than half of which were genitourinary cancers, and published their findings in 2016. It has since been introduced to clinical practice, and, significantly, integrated with NYP's Epic Systems electronic health record.
EXaCT-1 examines the genomic mutations of the more than 21,000 genes in each patient's cancer cells with tumor-normal comparison to detect detect point mutations, insertions and deletions, and copy number variations. The test promises relatively quick turnaround, with an average time from receipt to sign-out of 18 days, EXaCT-1 codeveloper Olivier Elemento reported at the annual American Medical Informatics Association's Informatics Summit last month in San Francisco.
Weill Cornell is running whole-exome sequencing on about half of all its metastatic cancer patients, about 500 a year, according to Elemento. "Our goal as soon as possible is to sequence every metastatic cancer patient," likely within the next year or two, said Elemento, director of the Englander Institute for Precision Medicine, associate director of the Institute for Computational Biomedicine, and co-leader of the Genetics, Epigenetics, and Systems Biology Program at Weill Cornell's Meyer Cancer Center.
The 2016 paper, which appeared in NPJ Genomic Medicine, was simply about validation of EXaCT-1. "Since then, what we've done is to implement the app in the clinical setting. Obviously, we are much more advanced and have done a lot more work in terms of integration [with] EHRs and so on," Elemento said.
Clinical reports get pushed automatically into the institution's EHR. Work described in a paper presented at the AMIA meeting took place before the process was fully automated, and all reports were PDF files.
"The EHR [on the clinical side] is more of a vehicle for communication with physicians. We push PDF reports into the EHR. What that means in practice is that physicians can consume information in the EHR," Elemento said.
Data has subsequently been structured, following Health Level Seven International's communication standard, and EHR data integration is fully automated so "you can do queries across genomes of multiple patients and use the combined clinical and genomic database as a way to do queries," Elemento said.
This also helps Weill Cornell find candidates for clinical trials, Elemento explained. "There are additional opportunities given to us as a result of pushing structured data into Epic."
It has taken a while to get to the point where genomic testing reports show up in EHRs like any other test.
"The creation of a report itself happens upstream," Elemento said. After DNA is sequenced and then analyzed, "what comes out of this analysis are mutation calls," he noted. The data has to be annotated to mark the mutations, which Weill Cornell does with its own resource, called the Precision Medicine Knowledge Base.
"The Precision Medicine Knowledge Base … has a lot of interpretations of cancer mutations. Every time we generate a list of mutations from one patient, we look up these interpretations [from] the knowledge base. When we see a match, it means there is an interpretation that can go into the reports that we produce for each individual patient," Elemento explained.
That goes into the template, which Elemento described as a "pre-report," called the NGS Reporter, which lists all the mutations found, with the ones deemed to be most important listed at the top. The NGS Reporter interface allows the pathologist assigned to the case to reclassify mutations based on whether or not they are important to each particular patient's case, he explained.
The forthcoming EXaCT-2 test combines exome sequencing and deep sequencing of hotspots. Weill Cornell has finalized the design of EXaCT-2 and is beginning the validation process to get approval from New York State authorities.
"We've learned a lot in the process of running this for EXaCT-1, so we don't expect this to be as much work as it was for EXaCT-1, but more or less, we have to repeat the work on EXaCT-2," Elemento said. "The good news is, we have an assay now that is quite robust and whose design is really optimized. That should allow us to identify mutations that are in very low abundance … and allow us to cover quite a bit of additional genomic space, for example, some exomes that were not covered in the first EXaCT-1 assay," he continued.
"We have essentially optimized the coverage and the depth of the analysis for EXaCT-2. That test is basically finalized in terms of the design, and we are just about to start the validation process for New York State approval."
As that process plays out, Weill Cornell has already engaged an unspecified partner to license EXaCT-2 for commercial purposes. While Standard Molecular, which helps EXaCT-1 developers connect genomic results and analysis pipeline to the EHR, has been publicly linked to the project, the commercialization partner is a different company, according to Elemento.
Within the NewYork-Presbyterian network, other hospitals have expressed interest in adopting EXaCT-1.
Weill Cornell has two other New York State-approved genomic tests, one for myeloid neoplasms, called My Heme Panel, and a solid-tumor panel of 140 genes, including gene fusions, built on Thermo Fisher Scientific's Oncomine. Both should go live by early summer.
"We are doing quite a bit of work now to be able to integrate with Brooklyn Methodist [Hospital]," Elemento said. "There is a lot of interest in terms of being able to offer EXaCT-1 [and] Oncomine, the myeloid panel, to these other sites, so we are working with them now to enable this genomic testing," but the organization is being deliberate because it is a complicated process. Right now, feasibility pilots are underway.
Among the issues is the fact that Brooklyn Methodist uses a different brand of EHR. EXaCT-1 uses an HL7 interface, so it in theory would work with EHRs not made by Epic, but every installation is different; as the old joke goes, if you've seen one EHR, you've seen one EHR.
At NYP, EXaCT-1 also is integrated with a pathology system, Cerner's CoPath, which involved additional work.
"It will be difficult to have a product that integrates very easily with all of the existing systems because every hospital has a different combination of systems," Elemento said. "There is always going to be some customization that is required."