The Evolving Landscape of Molecular Alterations in NSCLC: Optimizing Testing and Treatment to Improve Outcomes - Episode 12

Testing Strategies for Actionable Mutations in Non–Small Cell Lung Cancer

Expert oncologist John Heymach, MD, PhD, provides comprehensive insight on testing strategies for actionable mutations in NSCLC.


John V. Heymach, MD, PhD: Hello and welcome to this OncLive® program titled “The Evolving Landscape of Molecular Alterations in NSCLC: Optimizing Testing and Treatment to Improve Outcomes.” I’m John Heymach, the chair of thoracic head and neck medical oncology at [The University of Texas] MD Anderson Cancer Center [in Houston, Texas]. I’m going to give a number of updates on the testing and treatment of molecular alterations in advanced non–small cell lung cancer. We’ll discuss data in the context of the landscape and its impact on clinical practice. Let’s get started with our first topic.

The landscape of actionable alterations in non–small cell lung cancer has changed dramatically in the past few years. Going back to 2003 and 2004, that’s when EGFR mutations were first discovered. After that, ALK fusions [were discovered]. Since then we’ve had a steady stream of new actionable alterations. For a patient with non–small lung cancer who’s diagnosed today, we have to profile 8 different oncogenes at least. In the near future it’s going to be 9: EGFR, ALK, ROS1, BRAF, KRAS—especially KRAS G12C alterations—MET, RET,NTRK, and we think HER2 [human epidermal growth factor receptor 2] in the near future. That’s going to be a relevant target in terms of HER2 mutations as well.

In addition to all those targetable alterations, we typically want to do PD-L1 testing. But from my perspective, the targetable alterations are much more important than PD-L1 testing, select our therapy. Because whether or not we have PD-L1 data, we still have select immunotherapy data. But having the driver mutation data is the most important point. We always make a point to ask the pathologist to do the genomic testing first and then do the PD-L1 on the leftover tissue.

The way we test for these different alterations has changed in recent years. When we only had to check for 1or 2 alterations like EGFRorALK, we could use a PCR [polymerase chain reaction]–based assay for EGFR, which is well suited to look at 1 or a couple of alterations and can be done inexpensively. It’s very sensitive when it’s being done from tissue, and we would use FISH [fluorescence in situ hybridization] assays for the fusions or break apart FISH assays more specifically for ALK, RET, and so forth. Each FISH assay would require its own tissue sections.

Now that we have so many things to check for it, the recommendation is to use a broad-based NGS [next-generation sequencing] panel to look for all of these at once. In fact, if you use a broad-based NGS panel, commonly you would check 50 or 60 genes at 1 time, including all 8 of the actionable alterations, which you have to check for right now. This is much more efficient than doing 1 assay and then another and another, where you commonly would run out of tissue after doing a couple. A number of NGS assays are commonly available. All of them can commonly do the job. Sometimes, putting together a DNA- and an RNA-based assay is a more efficient way to detect fusions because fusions like ALKor RET fusions are sometimes a little trickier to detect using purely DNA-based approaches.

What type of sample should be used for testing? If we’re talking about tissue sequencing, diagnostic assays are fine-needle aspirates, and we always ask our radiologists to try to get core biopsies instead of fine-needle biopsies to make sure they get enough tissue. If there’s a surgical specimen available, that’s always preferable. That should provide more than enough tissue. What about if there’s a malignant pleural effusion and you’ve done a thoracentesis? Theoretically, you can get the sequence from that, but it’s not routinely done. That can be used for diagnosing non–small cell [lung cancer], but it’s not the routine way, so we would assay it.

The final thing I’ll point out is that you can do a diagnosis of lung cancer from the biopsy of a bone lesion, but it’s hard to do NGS from a bone lesion. You have to do special preparation to do it, and most labs aren’t prepared to do it. When I can, I always try a biopsy of a tissue lesion. I told the interventional radiologist, pulmonologist, or whoever is doing the assay to get core tissue and to get multiple passes if possible.

The last thing is, if the assay is being done by bronchoscopy, and they’re going to be biopsying a lymph node, you often don’t get a lot of tissue because a lymph node may often have a little cancer at a given time. We try to avoid using lymph node biopsies for next-generation sequencing. Sometimes you get it, sometimes you don’t. In my experience, the yield is relatively low. By contrast, if there’s a lung or liver lesion, those usually get plenty of tissue. Or if you’ve got a soft-tissue lesion or a subcutaneous nodule, those are appropriate. Remember, biopsies from bone require special preparation and aren’t preferred for NGS assays.

If we look at the NCCN [National Comprehensive Cancer Network] Guidelines about testing, a few years ago, it only recommended testing EGFRand ALKin nonsquamous patients. Now it’s much broader. It recommends testing in all patients with non–small cell lung cancer using a broad molecular NGS panel. That panel should include EGFR, ALK, ROS1, BRAF, KRAS, MET—specifically MET exon 14 skipping mutations, not MET amplification; MET amplification is commonly provided, but that isn’t what we base our therapies on right now—RET fusions but not RET mutations, and NTRK fusions. In the next year HER2 is going to become a routinely included assay, and recommendations now include that in all patients. Sometimes for patients with squamous histology, people choose not to test. My recommendation is to test them. Studies from our group and others have shown that a sizable portion of those patients, more than 10%, have an actionable alteration. Broad NGS testing includes all 8 of those genes at a minimum likely that HER2 will be required in the near future as well.

Transcript edited for clarity.