Making Reflexive Testing the Standard of Care in Lung Cancer: Robert Kratzke, MD

Delays in genomic testing continue to hamper timely lung cancer treatment, but strategies like reflexive biomarker testing and artificial intelligence (AI)–enabled record integration could help close the gap.

In the second part of this Q&A with Robert Kratzke, MD, professor of medicine and head of medical oncology at the University of Minnesota, he emphasized that policy and payer standards must ensure every non–small cell lung cancer (NSCLC) biopsy undergoes next-generation sequencing and immunohistochemistry to match patients with life-prolonging therapies

In the first part, Kratzke argued that molecular testing should be standardized for all patients with NSCLC, especially in underserved areas, and encouraged managed care organizations to support these efforts. Here, he explains which policies need to be put in place to fulfill this vision.

This transcript was lightly edited for clarity.

Transcript

AJMC: Given the rise of biomarker-driven therapies, what policies or best practices can managed care organizations adopt to streamline prior authorization and reduce delays in treatment initiation?

Kratzke: I'm particularly interested in the genomic biomarkers. In lung cancer, given the large number of markers with specific therapies identified, I do think that we have to all try to make reflexive testing the standard of care. Meaning, if a biopsy sample says lung cancer, the pathologists are more or less required to send that for genomic testing, and that there's not a delay.

Quite often, in my experience, the impediment there is that the pathology department doesn't necessarily benefit from it. It's just an extra task, and we have to streamline it such that it's really just what has to be done. Once again in the breast cancer community, they faced this decades ago on HER2 testing, that it was an extra task.

Finally, my understanding is that the American College of Pathology said, “You have to do this,” and insurers started to say, “You have to do this,” because once again, these therapies worked—they saved lives and they saved money. I think we have to streamline that process from the tissue sample to the diagnostic lab, whether it's genomic or immunohistochemistry, and just make it our standard of care.

AJMC: What infrastructure investments, such as in electronic health records or care coordination, are most critical for health systems to ensure equitable access to precision oncology? Does AI play a role anywhere in your practice?

Kratzke: I think AI could very much have a role. I don't use AI on a routine basis in seeing my patients, but maybe I should. I mentioned at the conference, I was very impressed that when given very precise instructions or questions, ChatGPT gave a very precise answer. I'm not here to advertise for one AI system or another, but I think some education there to providers—knowing that AI is their friend and not their enemy—could really help quite a bit.

Of course, the electronic medical record gets a lot of disdain from providers because they say it's just another hurdle. But I think for something like this, it's just critical for the biomarker results to be integrated into the electronic medical record. And that then is something that [could help] a provider in a very small town, a provider in the inner city, a provider on a Native American reservation, perhaps populations that would not necessarily have access to the most recent treatments. The electronic medical record should, in my opinion, level that playing field dramatically.

Once again, a lot of that education is for the doctors and for the providers. The electronic medical record is not our enemy; AI is not our enemy. If we let it, it would really help us and help our patients dramatically.

AJMC: What policy changes do you think would have the biggest impact in reducing disparities in access to targeted and immuno-oncology therapies for NSCLC?

Kratzke: I do think what I said earlier, in that we should make standard testing for doing next-generation sequencing testing on all NSCLC samples and aggressively standardize immunohistochemistry for targets such as HER2 and now MET, since there are antibody-drug conjugates for those targets.

I'm not really blaming the pathology departments, but either at the level of the certification by the American College of Pathology or at the level of insurance providers, there has to be pressure to make sure that every pathology department—whether it's a small town or the biggest city in the country—is aggressively testing these samples for all the new therapies that we have.

I realize it's not easy, not trivial, keeping up on these things. And once again, many of them are send outs, and I'm confident that none of the hospitals are very enthusiastic about taking time to do send outs, but somebody somewhere has to, in my opinion, make sure that every patient with NSCLC in the US has their sample tested at the DNA level and at the appropriate immunohistochemistry level for targeted changes and mutations.