Winner Spotlight: Kuveda Provides Personalized Cancer Care with Better Tools for Oncologists
As oncologists take on the arduous task of treating their patients’ cancers, Kuveda Health is giving them better, more sophisticated tools to do their jobs effectively. The startup has spent the past year and a half developing a technology solution that uses more precise genetic mutation patterns to recommend targeted personalized therapies for patients.
The old school of thought was that where a tumor originated was the main factor that determined treatment options. But according to Kuveda’s cofounder Chuck Gershman, his company’s platform builds on new knowledge that downplays this in favor of analyzing the genomic path of the cancer.
Kuveda, which can benefit patients of an array of cancer types, was the health category winner at Challenge Cup San Francisco last week. Gershman will go on to pitch at the Challenge Festival in May. After the regional competition, he broke down what makes Kuveda groundbreaking.
Kuveda is a big data analytics platform for cancer therapeutics. What does that actually mean?
We take in a genomic profile of a cancer patient, as well as an associated tumor sample that is sent to a lab for genetics analysis and protein analysis. Basically, then, we identify the mutation pattern of that specific tumor. We inject that information into our tool and perform an analysis against that mutation information, looking for a drug therapy that we can recommend to oncologists that closely aligns to that molecular profile.
We look into the cellular structure of the mutation pattern and create a model of the cancer tumor. Specifically, we look at the pathways that are adversely affected as well as the signaling pathways of connectivity between the proteins … We then match that to drug therapies that are available, both FDA-approved and non-approved and look at the targeted therapies that we developed to address specific proteins or protein pathways and match the drugs that were intended for these purposes. Drugs that we developed for other proteins or pathways, we eliminate from the analysis, so we’re filtering them out, and come back with a subset of therapies that have a reasonableness model of having some effectiveness for that patient.
Can you put this into context? How has cancer therapeutics been working and what makes this personalization so much better and more advanced than what’s out there?
Generally speaking, about 75 percent of cancer patients are treated through standard of care and become cancer survivors. However, about 25 percent of cancer patients will relapse. They have potentially exhausted standard of care procedures, which can be surgery, chemotherapy, radiation therapy or other targeted drug therapy. And most standard of care is focused on anatomical place of origin. If you have a breast cancer patient you treat them with “breast cancer drugs.”
The problem is it’s reasonably well understood that the anatomical site of origin does not play as significant of a role in successful treatment outcomes as previously believed. The nature of cancer is that there are genetic mutation patterns and these genetic mutation patterns are reflected across different anatomical sites and are similar in nature from a mutation perspective. In other words, it’s not uncommon that a drug prescribed for a breast cancer patient could have value for an ovarian cancer patient. We’re trying to bring down that barrier, number one–having the anatomical site of origin as the primary treatment option. We look at mutation patterns separately. You’ve got 25 percent of the patient pool that potentially relapses and in these cases we’re trying to find targeted therapies that have reasonable probability of being valuable to that patient.
And prior to doing what we’re describing…there’s been a subset [of patients] who would seek personalized data and seek out institutions who would perform this kind of analysis and have this profile done. We’re trying to encapsulate what’s done by these expert tumor teams and put it into a vehicle that allows the majority of oncologists to use it. We’re feeding our tool with similar stimuli to what these expert teams are using and we’re generating very, very quickly the same kinds of recommendations as to what these tumor teams are generating manually. We also give [oncologists] evidence of where the info came from, adverse reactions, failure rates, various tradeoffs they can make on medications that are available.
Who are the specific cancer patients that would benefit from Kuveda’s tools? Since you talk about de-emphasizing the importance of anatomical sites of cancer origins, is it for most cancer subsets?
I’d say that, prior to applying our tool in pilots, we applied our tool to analyze over 50 patients. We went back to patients who had been treated using the techniques we described through the manual tumor team process. We took the data for them and put it into our tool and compared our outcomes to the outcomes by the tumor teams. In every case that we analyzed, our primary finding with how to treat that patient was consistent with the expert tumor teams’ and in almost all cases we had a superset of additional recommendations that the tumor team did not evaluate. And that’s across pancreatic, ovarian, lung, breast, melanoma, etc. All of those were cases of cancer we analyzed before we even made this tool available for pilot to ensure it worked [across a wide range of types].
In terms of how you’ll roll this out, are you selling directly to oncologists or to the hospitals themselves?
Our point of entry is oncologists, yes. Bu in terms of who actually ends up paying for it, it may be oncologists but, more than likely it will be the institution that the oncologist works for–whether that is a clinic or hospital. In terms of the value proposition of the product and where it’s ms appreciated and usability of the tool, that’s all optimized for an oncologist.
You also mentioned during the competition, that there’s a social element for oncologists that enables them to share findings about patient care. How do you facilitate this exchange without compromising patient privacy?
The social aspect that we’re adding in works like this: First off, inside our database all patients are anonymized. Once you’re in our physical database there are no names. However, each oncologist is given log-in privileges for the tool. As part of their log-in privileges, it’s tied to each individual patient that they enter into the system. So the treating oncologist can see the name but only hem and their associated team knows the name of the patient. Everybody else in the system simply sees the raw data.
But we allow an oncologist to search against a patient profile. If a patient’s in the system and they have a particular molecular profile and a particular treatment plan, the physician can search and say, “Show me all patients of similar profile and mutation pattern.” In that manner they can share information with other oncologists about how they’re treating their patients, outcomes they saw, etc. That’s something the oncologists are very interested in. They want to be able to talk to their colleagues about patients most similar. And I use the term “most similar” because, of course, no two patients would be identical.
And we want to go one step further. This is still in development, but we want to make some of this available to patients as well–at least the educational part of what we’re doing–and we’ll work through cancer nonprofits who talk directly with cancer patients. So, at the appropriate time, they can alert their patients to what we’re doing and we can help them in the exchange of options available to them. So they can go to their oncologist and ask for targeted, personalized cancer therapy.
What’s your background and the background of your team that led to Kuveda?
There are three founders of the company. I’m a serial entrepreneur and have been CTO of several companies. My two other partners are big data analytics engineers, but they have spent a lot of me in the genomic space. They both worked on a project called the Cancer Genomics Hub. It is one of the largest repositories of molecular mutation information in the world. They were two of the key contributors developing this repository. It was through this experience that they gained the appreciation for the amount of data, diversity of data and complexity and utility of the data and how it could be applied. They realized that the repository was very, very focused on researchers. In talking to clinicians, we came to the conclusion that the repository was extremely valuable but was very hard to apply on a personal basis and for an individual patient. This was a gap that we could fill.