Skip to main content
Premium Trial:

Request an Annual Quote

Q&A: William Nelson on Johns Hopkins' Personalized Rx Focused 'Learning Healthcare' System


Johns Hopkins University last month received a $30 million gift from the Commonwealth Foundation for Cancer Research to establish the Center for Personalized Cancer Medicine at the university's Kimmel Cancer Center.

With the new funds, JHU researchers will conduct genomic and epigenomic research in various cancer indications with the aim of applying the knowledge gleaned from these investigations to treat patients in the university's healthcare system.

JHU launched the center in July, bringing together university experts from various disciplines, such as oncology, public health, and bioinformatics.

The university is already a leader in genomics and epigenomics research and has outlicensed several technologies to commercial entities in this regard. One method developed at the institution, called personalized analysis of rearranged ends, or PARE, is already being used to gauge deleterious mutations in cancer patients and determine the best treatment course for them.

The goal of the new center, according to William Nelson, director of the Johns Hopkins Kimmel Cancer Center, is to build a "learning healthcare system." JHU plans not only to use genomic strategies in its healthcare environment, but then track whether these interventions are improving patient outcomes and are cost effective compared to the standard of care.

Nelson recently spoke to PGx Reporter about JHU's individualized medicine initiative. Below is an edited transcript of the interview.

Why did Johns Hopkins choose to launch a personalized medicine center at this point?

Formally, at this point, it's an initiative. The entire university, including the health system and all the other schools, was looking to engage in [projects] in which all the assets of the university could be brought together to solve important problems. Among those ideas was a notion of personalized medicine. We have over the past couple of years been meeting as a university and assessing as a committee what we do well and the unmet needs in the world. Now, what we have come up with is an initiative … on individualized health. By that general description you can take the notion that genomics, epigenomics, and those types of data are increasingly going to be available with improvements in technology … in the context of healthcare delivery over the next decade.

They aren't the only massive sources [of data] out there. You can think about imaging data and think about how [patient healthcare data] is becoming more organized into electronic health records. In personalized cancer medicine, you now have agents that work in a very specifically defined group of people. If you develop and launch such a drug then you have to design a healthcare system that would deliver to those people that kind of treatment. Then, giving the right drug at the right time to the right person, you would have better outcomes as a healthcare system, and in the end you'd reduce cost because you'd reduce ineffective care.

Broadly, the individualized healthcare initiative encompasses genomics and epigenomics, but also other aspects of health information and how that information is used.

What resources do you need to start this initiative? Would you need to hire more staff? What changes would you need to implement to get this initiative off the ground?

If the goal heading forward is to create a learning health system, I think one of the things that needs to be recognized is that there are a number of centers and institutes that are focusing on gathering and curating specific types of information. Places like the Broad Institute, the Sanger Center, and the like are looking at genomics very carefully. But most of these places aren't doing it in a way that is integrated with its health service delivery in a streamlined way. We began to recognize that as the Johns Hopkins health system has grown to have six hospitals and four sets of primary care, ranging from insurance products that we own and manage, to provider networks, … we have the kind of health system in which you can deliver individualized health at a population scale.

For example, one of the groups that Hopkins manages, called Priority Partners, is effectively a Medicaid HMO. The folks in that group have been able to identify, using more conventional [methods,] who in that group, which is primarily an urban, poor population with low health literacy, are at exceptionally high risk to progressed end-stage renal disease over a period of time. [Doctors] intervened aggressively in that group, which some people are calling "high touch" medical care, where you make sure they have their medications to treat blood pressure and blood sugar, and you make sure they're educated on how to take them and what adverse reactions they can expect. By diverting those resources very intensively in that group most at risk to progress to end-stage renal disease, you avoid a substantial amount of it and thus need to pay for less dialysis. So, in the end, it's cost effective and you have fewer people progressing to end-stage renal disease. And that doesn't use genomics at all.

Clearly, in cancer medicine it's a disease of genes, so genomics looms larger. But I think you're going to see this in other dimensions of healthcare to great effects.

Can you provide some details on the projects that the personalized medicine initiative will focus on?

We're imagining that these types of projects will be very large in scope in order to transform a healthcare delivery system that uses these principles and reaps the benefits. [As far as the] philanthropic support for these efforts, we engaged some people who are very excited about some of the particular pieces of the projects. The Commonwealth Foundation with Bill and Mary Goodwin has been specifically interested in cancer. Some newer ideas will organically grow as we get more funding support, but they will be in the [individualized medicine] theme.

Some of the identified projects focus on highly divergent phenotypes for sets of genomes and epigenomes. For instance, in prostate cancer there are men who will respond to hormonal therapy and live 15 years without the disease getting worse and then there are men who don't respond at all. In looking at those extremes it gets around some of the … problems you have trying to estimate differences in genotypes and gene differences associated with a slightly increased risk of a bad or good outcome. Some of the projects we're exploring have that architecture — looking for extremes in response to treatment and extremes in the behavior of the diseases.

Another set of projects will try to look at exome sequences, or they may be collected as whole genomes, and variances in proteins that are expressed. The thought is that universe, might be a good substrate for peptide antigens that are presented by functioning cells in the immune system. So they may be a good universe of cancer-specific antigens. They may be different for each person, so it's personalizing immunotherapy. … There are now these immune checkpoint [inhibitors] that are appearing and gaining clinical momentum. One can imagine that personalized vaccines along with one of these immune checkpoint inhibitors might be something we see in the future. So there is an idea around that.

Leukemia is another area. Washington University has been aggressive in looking at this disease using whole-genome sequencing of expressed RNAs and the like to look at acute myelogenous leukemia. This is a disease where we're already making genome-based decisions regarding treatment and clinical care with cytogenetics, in a low-resolution manner. There are a number of known coding genes that are associated with risk. These are already sort of conventionally done, but the Washington University group is looking at taking people at intermediate risk and nudging their classification toward lower or higher risk. The general idea of risk in that scenario is that if you're at favorable risk … then treatment with chemotherapy and targeted agents alone is likely to gain a good outcome. If you have high risk, then once you are in remission you're more likely to progress on to need bone marrow transplantation. So, researchers are figuring out whether that can be shaped more.

Here, we build a lot of tools for epigenome profiling. We're a major center for epigenomics research. [Researchers at JHU] are trying to take epigenetic drugs into myelodysplasia, leukemia, and other diseases. With that we have people now who we know respond or don't respond to certain drugs and genomic or epigenomic tools of response can make it clearer who will benefit and who won't. That is something else we will be studying through this initiative.

In terms of the technology providers that you'll be using in this research, will you be partnering with platform providers or will you be developing tests yourselves?

All of the above. We are working to create a laboratory-medicine based approach to genome sequencing and epigenome analysis, through which CLIA-type data can be available. I think some of the technology platforms available today are getting ready for that kind of use, and other centers are employing these tools. We don't yet have a partner with any particular vendor of sequencing machines, although I can imagine we might, if they find what we're doing interesting.

The other interesting thing is that there have been spin-out companies related to identifying and turning epigenome features into tests. A lot of the intellectual property that has been generated here has been licensed through to [epigenomics test developer] MDxHealth. Many discoveries could work their way through that kind of system, where we generate intellectual property that others may find useful.

There is [also] a spin-out company from our cancer center … that is trying to develop a capability for the PARE, or personalized analysis of rearranged ends [assay].

So this is going to be a dynamo of intellectual property that will be generated that might be commercializeable. If you look broadly at the overall individualized healthcare initiative, the various decision-support and analysis tools … will [also] likely be a great substrate for commercialization. The major vendor that the health system is going to be using for the base architecture for the electronic healthcare [records] … will be Epic Systems. Any tools that will be built [on top of that] will be scalable and deployable, perhaps through commercial opportunities.

You're starting with cancer. What are some of the other areas you're eyeing in the long term?

If you're asking how big could this get, we'll need to recruit people in information sciences, systems engineering, and in computer science, to gain expertise in using and archiving these vast data sets. In that sense, as we move forward, building the entire healthcare infrastructure … we'll need to make those investments. Initially, we're starting with demonstration projects before a humungous building [is] built. Initially, we want to begin to generate a series of stories where [individualized medicine] approaches have been successfully introduced, and we can see the benefit and cost savings with these interventions. When we see that these approaches can work and are scalable, then we will be able to track more philanthropic investments for other aspects of medical care.

Other universities have launched classes to explore the political, social, and ethical aspects of genomic medicine. Will there be any of this done at JHU as part of the individualized medicine initiative?

This will be a trans-university effort that will not only involve genetics, genomics, and biology, but, for example, also the school of engineering where our computer science department is housed and they're experts in systems engineering. We also operate the Berman Institute of Bioethics. [They can explore issues] like [whether] we are having a fundamentally different interaction with the people we have in our healthcare system, [establishing] different expectations as to what is research, and what constitutes risk and benefit in a learning healthcare system.

By the time all is said and done this will be a very large educational effort. The good thing about being at a university is that education won't be the neglected thing, but a critical element.

The Scan

US Booster Eligibility Decision

The US CDC director recommends that people at high risk of developing COVID-19 due to their jobs also be eligible for COVID-19 boosters, in addition to those 65 years old and older or with underlying medical conditions.

Arizona Bill Before Judge

The Arizona Daily Star reports that a judge weighing whether a new Arizona law restricting abortion due to genetic conditions is a ban or a restriction.

Additional Genes

Wales is rolling out new genetic testing service for cancer patients, according to BBC News.

Science Papers Examine State of Human Genomic Research, Single-Cell Protein Quantification

In Science this week: a number of editorials and policy reports discuss advances in human genomic research, and more.