Despite advances in genomics and personalized medicine, identifying the best treatment for an individual cancer patient remains challenging. San Francisco-based startup Notable Labs Inc. has developed an approach that aims to improve those odds. Its AI-based platform tests a patient blood sample against hundreds of individual drugs and drug combinations in the lab to help identify the best treatment options for that patient.

In a feasibility study by Stanford Medical Center, Notable was able to provide personalized treatment recommendations based on its work in combination with a genomic analysis by partner Chicago-based Tempus Inc. for 20 refractory myelodysplastic syndrome (MDS) patients within 30 days. The interim data show an average value, both positive and negative, of 84% in predicting whether a specific drug or combination of drugs would benefit an individual patient.

Testing response

"We take peripheral blood or bone marrow from a patient with blood cancer. We obtain a tube of blood where we then remove all the red blood cells but keep all the healthy immune cells as well as the cancer cells for testing in our laboratory. There we take that sample and run hundreds or thousands of different drugs and combinations on the sample to try to discern which drug or combination is most specific for targeting the cancer cells relative to the healthy white blood cells from that patient. So, we do all of that in a short-term test. The turnaround time is typically within about a week," Notable founder and CEO Matt De Silva explained to BioWorld.

Similar approaches have been tried previously, but rapid recent advances in automation, machine learning, miniaturization and stem cell biology have enabled that technique to come together now. And there's a strong need for it, as the pool of genomic-driven oncology drugs, candidates and combos in oncology have exploded in recent years. But the industry hasn't yet gained a lot of traction in reliably predicting responders and nonresponders.

Notable uses flow cytometry to examine samples at a single-cell level, monitoring the activity and response of immune and cancer cells. Previously, it has been difficult to keep the cells alive for analysis in a lab setting, but in the past five years or so the understanding of biology has improved enough to enable that. Given the miniaturization and the cellular level analytics, the process requires only a standard amount of blood in a single standard sample tube.

"We're not growing the cells. We're taking them right out of the body and then, essentially, incubating them with drugs and seeing which drugs have the most activity that we want to measure. So, that happens in a microtiter plate, which you can think of as basically a bunch of small test tubes," said De Silva. "To be able to take one patient blood sample and test it across all of those different drugs at the same time, automation and miniaturization are key for that process."

Finding feasibility

Notable is running a series of feasibility studies to validate the accuracy of its technology at various research institutions, including UCSF, MD Anderson and Seattle Children's Hospital. In addition to work in blood cancers such as MDS, it's also studying its approach in solid tumors.

In the Stanford study, Notable's data were combined with DNA sequencing information from Tempus to provide the basis for a report for each patient to the Stanford MDS tumor board that offered personalized treatment recommendations.

MDS is often highly treatable, but this study focused on refractory patients whose disease had resisted an initial round of treatment. Additional data from the trial and further information on the methodology used will be presented at the European Hematology Association (EHA) annual meeting in Amsterdam in June. In addition, Stanford, Notable and Tempus are preparing a paper on the study that's expected to be published later this year.

"Ex vivo drug sensitivity technology must have a rapid turnaround time, accuracy and efficacy in order to be useful in the clinic," said Peter Greenberg, professor of medicine in hematology and director of the Stanford MDS Center at Stanford University Cancer Center. "Notable Lab's ex vivo drug sensitivity assay screened marrow samples we sent them from patients in our recent biologically focused feasibility trial against a collection of investigational and FDA-approved compounds.

"These patients had higher-risk myelodysplastic syndromes and were refractory to standard therapy. Potentially actionable therapeutic results were returned to us for the patients enrolled in our trial within a clinically actionable time frame," he added. "These data suggest the potential utility of this methodology to aid in decision-making for novel therapeutic drug selection in MDS patients with HMA [hypomethylating agent]-refractory disease."

Currently, Notable is focused on clinical trial use of its technology to map drug responses across a diverse patient population, as well as to predict outcomes for each individual patient. It does not offer a service that is available for use by physicians in standard practice with their patients.

Its technology is currently being offered as a laboratory-developed CLIA test for supporting clinical trials. It is available for research use only. But the company eventually may look to move beyond that to full FDA approval.

In MDS, there are only three drugs approved for treatment: the orally administered immunomodulatory drug Revlimid (lenalidomide, Celgene Corp.) and two parenterally administered nucleoside analogues that are DNA hypomethylating agents, azacitidine and decitabine. Interestingly, the study found that PARP inhibitors have activity in some MDS patients; there are FDA-approved PARP inhibitors, but not specifically in MDS.

Other possibilities include "immuno-oncology drugs as well as targeted kinase and epigenetic therapies that have come through FDA approvals for other types of cancers," said De Silva. "For example, in MDS, we can test a list of drugs that would be acceptable to clinicians if you were able to find activity. We were excited to see activity in a subset of MDS patients with PARP inhibitors, which are approved in other types of cancers like ovarian cancer. It could be used in MDS, or in combination, before initiating a clinical trial to test that hypothesis."

Notable, founded in 2014, has raised more than $20.9 million in funding, including a $14.84 million series A round in September 2017.