Q&A

The core challenge of rare cancer

Dr. Razelle Kurzrock: The biggest challenge in rare cancers is making an accurate diagnosis. Pathological diagnosis alone can often be difficult, and even when it’s established, it doesn’t always reveal what’s truly driving the disease. Fundamentally, I believe genomics is the diagnosis, understanding the genomic alterations that define a tumor is key. My own belief is that utilizing genomic profiling for these cancers is perhaps the single most important thing we can do to ensure patients receive both an accurate diagnosis and the most appropriate treatment plan. Ideally, this should be done reflexively on path tissue.

Redirecting treatment with genomics

Dr. Razelle Kurzrock: Absolutely. We’ve seen this repeatedly. We find a pathognomonic alteration, often a fusion, and it redirects our thinking about what the cancer is. But more importantly, it redirects our treatment.

The impact of tumor-agnostic approvals

Dr. Razelle Kurzrock: There are now 9 solid tumor-agnostic approvals by the FDA, and one of the reasons the FDA moved towards this approach is to address the unmet needs of patients with rare cancers. Most are for genomic aberrations. If you don’t know your patient has one of those, you cannot give them an FDA-approved, on-label therapy that is often tremendously active. For rare cancers, where there’s so little data, I think it’s even more critical.

The role of industry partnerships

Dr. Razelle Kurzrock: There is almost unimaginable value in this. This is true for all cancers, but especially true for rare cancers, where historically, there has been data scarcity. The value is at every level, from new discoveries to understanding therapy. Imagine that you have a database that can give you answers to your questions with real-world data. This is a whole new world. The advantage of not having to wait years for patients who have urgency, it’s something I call real-time discovery . We can make discoveries in time for the patient in the clinic.

AI and data complexity

Dr. Razelle Kurzrock: This is a really, really important issue. The average metastatic tumor has 5 genomic alterations. There are 700 cancer-causing genes. If you do the math, there’s over a trillion patterns. Nobody, at least nobody that I know, can do a trillion patterns in their head. I think, ultimately, there is no other solution for optimizing therapy than an AI-driven solution.

A case study: Circumventing the “undruggable”

Dr. Razelle Kurzrock: Yes, this is a patient with a high-grade myxoid liposarcoma, a rare sarcoma. The patient has a FUS-DIT3 fusion, which is an undruggable fusion. We’ve found that if you have an undruggable fusion and you try to drug the other secondary alterations, you don’t get very far.

Dr. Ezra Cohen: So if the primary driver is undruggable, how do you find a way around that?

Dr. Razelle Kurzrock: This is what’s so important. This patient’s molecular profile also revealed a somatic BRCA alteration. This is critical because it may provide a pathway to circumvent the undruggable fusion issue. It points to homologous repair deficiency (HRD), and we can address that with a synthetic lethality approach by giving PARP inhibitors or platinums. The genomic profile provided a treatment pathway that would not have been obvious otherwise.

Future directions: Multi-omics and trial design

Dr. Razelle Kurzrock: It’s really going to be about multi-omics. We’ve seen solutions for undruggable targets by looking at how the RNA is affected. Beyond that, it’s looking at the proteome, looking at epigenetic effects, and looking at the immune environment in its entirety.

Dr. Ezra Cohen: And do you see AI platforms helping design new types of trials for these ultra-rare cancers?

Dr. Razelle Kurzrock: Absolutely. But going even beyond designing prospective trials, I think we’re coming very close to the era where we may be able to get the information that we need from real-world data, in some cases without doing the prospective clinical trial. Even if we could skip that step some of the time, it would be a huge benefit for patients who need answers and treatment options urgently, rather than waiting years for those insights to emerge from clinical trials.