The precision medicine paradigm is a powerful one–for cancer patients, and for the entire oncology ecosystem. Over 50 targeted (‘tailored’ or ‘genome-informed’) oncology drugs have been approved by the FDA for clinical use since the 1990s, and propelled into practice with growing payor coverage of tumor genomic profiling. This wave of targeted therapy approvals began with profoundly impactful medicines like trastuzumab in 1998 for HER2+ breast cancer, imatinib in 2001 for chronic myeloid leukemia, and erlotinib in 2004 for advanced EGFR-mutant lung cancer–and continues through 2023 with a series of recent approvals in just Q1 of this year, in indications ranging from lymphoma to colorectal cancer to subtypes of (mismatch repair deficient, or dMMR) endometrial cancer. The pace of therapeutic innovation has been breathtaking.
The patient <> therapy matching problem in oncology
And yet, patients and their care teams also face several harsh realities. The median response rate across all new oncology therapies approved over the last two decades is only 41%, only about 25% of all cancer patients are eligible for ‘genome-informed’ therapies on the basis of DNA alterations in their tumors, and the percentage of all cancer patients who will actually benefit from an approved genome-informed therapy is likely still under 15%.
In practice, off-label use continues to grow and the indications for existing targeted therapies continue to expand. But for the majority of cancer patients, we have no way of knowing whether they will respond to a given therapy…until after they actually receive the therapy.
Put yourself in the shoes of an advanced stage cancer patient today. You may have heard that many new cancer drugs have been developed–great news, but you have a limited number of shots on goal, and only really care about the drugs that will actually work for you. To learn whether certain drugs might work for you, your oncologist may offer comprehensive genetic profiling (CGP) of your tumor, but there’s a >75% chance that nothing actionable for you will be found on that testing. Let’s say you are one of the 3 in 4 patients in whom DNA sequencing does not point to a specific therapy choice. You are left wondering: doesn’t my cancer have an Achilles heel, too? Might one of those newly approved targeted cancer medicines be functional against my tumor? How could we find out–quickly, and accurately–so that I don’t have to waste time trying therapies that don’t work?
This is a conundrum for the broader oncology community as well. For example, medical oncologists are often picking between two or more treatment options which may have similar supporting evidence at population-scale, but may have wildly different efficacy between individual patients. Payors, or insurance companies, are often bearing the cost for therapies to be tried empirically for several months, only to later learn that the patient may have received no benefit at all–or worse, experienced (costly) toxicities. And for biopharma companies on the hunt for novel cancer medicines, stratifying patients by their likelihood of response is a clinical development game changer, shrinking both time and cost to bring new therapies to patients.
In short, from every perspective, we still have a vexing patient <> therapy matching problem in oncology. And this ‘problem’ is only getting worse as the targeted therapy arsenal continues to grow. What if it could be turned into a massive opportunity instead?
Building the data platform for a function-forward future
If translational oncology research has taught us anything over the past several decades, it is that cancer is creative. Cancer cells can find ways to co-opt normal function and exploit pathways in so many diverse ways for their gain. And as we have learned, these mechanisms are not always encoded in DNA changes–they may be manifested in epigenetic changes, RNA-level changes, splicing changes, protein conformational changes, and the list goes on. As such, while it is disappointing that tumor DNA sequencing only offers actionable clarity to 1 in 4 patients, it is not fundamentally surprising.
What if we could probe the actual gene function (vs. gene sequence alone) of different targets in a patient’s tumor cells? What if we could find the functional Achilles heel(s) of each patient’s own cancer, before prescribing therapy? While many teams have built various ex vivo (outside the body) diagnostics assays in which tumor cells are exposed to a panel of drugs, none are used in routine practice today, and few have offered mechanistic insight into the precise functional pathways on which a patient’s tumor may be dependent.
Enter Function Oncology. The team is building the first-ever, real-world functional oncology data platform, a continuously growing and evolving dataset of primary patient cancer cells that have been broadly profiled to identify patient-specific functional dependencies: genes, proteins, and pathways to which patients’ individual tumors are addicted. If we could know this information about every patient’s cancer, at the right time in their clinical journey, the emergent opportunities are endless. The ‘problem’ of too many available targeted therapies suddenly becomes an opportunity: functional profiling could drive personalized therapy selection in the clinic, intelligent re-purposing and off-label utilization of many already approved medicines, and smarter pre-clinical and clinical development of novel oncology drugs across our industry.
CRISPR strikes again
So how does Function Oncology learn the functional dependencies of a patient’s tumor? To start, they are leveraging CRISPR tools in a very creative way. CRISPR technology has taken the bio world and even popular science by storm over the last decade, spawning a wave of exciting programmable medicine companies, the creation of numerous biotech platform companies, and even a Nobel Prize.
But as many in both academia and biopharma know, some of the most transformative impacts of CRISPR capabilities have actually been felt in pre-clinical research: the ability to perturb (knock out, dampen, or even activate) a single gene’s function in cells with a known genetic background is a game-changer, because it can closely mimic (“phenocopy”) what a therapeutic medicine might aim to do. CRISPR functional genomic screens–perturbations across a large panel of genes to identify novel gene targets or modifiers–are now a cutting-edge capability at many modern biotech companies, but Function Oncology is taking this several steps further: into primary patient cancer cells, for real-world diagnostic and clinical use. As the Function Oncology founders explained to me, sometimes CRISPR is the therapy–but here, perhaps CRISPR can help us choose the right therapy for the right patients. We found this vision–a fresh take on a ‘biologically engineered future’–inspiring.
The FxOnc team: twins, technologists, tirelessly patient-focused
A former Flatiron / Foundation Medicine colleague, Alex Parker, first introduced us to the Function Oncology (FxOnc) founders. He mentioned the team with a suspenseful “…I have a feeling you’ll just really enjoy chatting with them.” As our smiles below suggest (a memory from the day we agreed to work together!), this proved to be so true.
Srinath and Hari, both MD/PhD physician scientists, are lifelong collaborators of sorts–twin brothers!–and all three founders Srinath, Hari, and Christian had been working together for many years as investigators and leaders at Novartis Research Foundation’s Genomics Institute (GNF) in San Diego prior to starting Function Oncology. They bring a unique combination of familiarity and expertise in clinical medicine, basic science, and therapeutic research & development. Like many of our best startup teams, they are scrappy but also rigorous, ambitious, and tirelessly patient-focused.
Here’s to massively multiplying the number of cancer patients to whom our therapies can offer hope. We are honored to partner with the Function Oncology team in their mission to advance a new personalized oncology paradigm.
***
