First FDA Recognition for Tumor Marker Database 

Helen Leask

November 19, 2021

In the 20 years since the first targeted cancer therapy was approved, more than 100 such products are now on the market or in development. These targeted drugs offer a different approach to the carpet bombing of chemotherapy — they are precise missiles, homing in on specific targets in the tumor's genetics.

But which targets? In which tumors?

Oncologists are increasingly concerned that the fire-hose of biomarker information now coming at them causes confusion, and worry that patients may be missing out on effective therapies.

A new database aims to change all that.

Last month, the US Food and Drug Administration (FDA) granted its first recognition of a precision-oncology knowledge base linking genetic biomarkers to specific cancer types — the Memorial Sloan Kettering Cancer Center's Oncology Knowledge Base (OncoKB).

"OncoKB's origins are rooted in the recognition that it was asking too much of clinical oncologists to memorize the rapidly growing number of tumor-type- specific cancer mutations that can act as predictive biomarkers to targeted therapies," said Debyani Chakravarty, PhD, Lead Scientist of OncoKB at the MSKCC's Center for Molecular Oncology.

"A clinical-decision support system to provide this information in the point-of-care setting was urgently needed," she added.

OncoKB made its public debut on May 16, 2017, with the publication of a paper by Chakravarty and colleagues in JCO Precision Oncology. At the time, the database boasted hand-curated levels of evidence for more than 3000 alterations in 418 cancer-associated genes and 19 cancer types. The roster is now 5685 alterations in 682 genes in 127 cancer types.

In October, OncoKB became the first tumor-mutation database to be included in the FDA's database recognition program.

Physicians or researchers can assess the evidence linking combinations of cancers, genetic biomarkers, and drugs through a simple interface.

"It's hard to figure out what is an actionable mutation and what is not without this kind of data," said Timothy Rebbeck, PhD, professor of cancer prevention at the Dana-Farber Cancer Institute, Boston, Massachusetts, who was  approached for comment.

He said the FDA nod for OncoKB makes sense: "[It's] relevant for the FDA to be involved because of the treatment implications of these mutations."

Many cancer genomic databases exist. The list includes MSKCC's own CBioPortal, as well as MyCancerGenome, the Precision Medicine Knowledge Base, Cancer Genome Interpreter, Cancer Driver Log, Tumor Portal, and others.

Chakravarty acknowledges that OncoKB is not alone. "But that being said, it is now the first somatic human cancer variant database to be recognized by the FDA," she added.

At the heart of OncoKB is a rolling curation process involving more than 50 clinicians and geneticists at the cancer center. "We have found a way to codify the scientific and medical expertise of MSK clinical oncologists", said Chakravarty.

The team asks itself three deceptively simple questions: what the gene is, whether or not the mutation in that gene is oncogenic and, when found in a specific tumor type, the associated therapeutic implications.

The supporting data are then assigned an OncoKB therapeutic level of evidence on a scale that runs from "standard care" to "hypothetical."

"What surprised me most was, even just getting those three pieces of information and ensuring their accuracy was quite a challenge," Chakravarty said.

The new FDA recognition is deemed to be 'partial' because it does not cover all of OncoKB: the agency has not reviewed all the linkages between drugs and biomarkers on the site. However, the developers of OncoKB created — and submitted for recognition — an FDA-specific zone that lists oncology biomarkers that already have a regulatory green light.

The FDA-recognized zone of OncoKB provides similar data to the rest of the site, but it is categorized using the FDA's own level-of-evidence system.

The FDA assigns one of three levels of evidence to genetic variants in regulatory submissions that it receives. FDA Level 1 is reserved for approved companion diagnostics to targeted therapies; FDA Level 2 is assigned to "cancer mutations with evidence of clinical significance"; and FDA Level 3 indicates "cancer mutations with potential clinical significance".

The FDA-recognized zone of OncoKB currently lists 46 actionable genes in 40 cancer types with FDA Level-2 evidence and 38 actionable genes in 33 cancer types with FDA Level-3 evidence.

The FDA first released guidance on how to achieve endorsement of genetic-variant databases in April 2018. Bearing in mind that cancer genomics is an enormous, energetic field, Medscape asked the FDA why it took 3½ years to recognize a tumor-mutation database.

FDA press officer Jim McKinney placed the onus on database developers to apply for official status: "The FDA is unable to direct outside entities to apply with the FDA for database recognition," McKinney said. As for OncoKB, the review timeline was "impacted by the COVID-19 pandemic, which affected agency resources."

McKinney was not able to share how many other oncology genetic databases the FDA is currently considering for recognition. However, he did confirm that the partial FDA recognition of OncoKB means that "test developers can use these data to support the clinical validity of tumor-profiling tests in premarket submissions."

Chakravarty noted that commercial licensing requests, which historically have trickled in at a rate of one per week, "spiked" in the days following FDA recognition.

Chakravarty and Rebbeck have disclosed no relevant financial relationships.

For more from Medscape Oncology, join us on Twitter and Facebook

Comments

3090D553-9492-4563-8681-AD288FA52ACE
Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.

processing....