EGFR and KRAS testing

[levels x3]

---

Members don't see this ad.

What are your labs doing today when it comes to EGFR testing in lung adenocarcinoma and KRAS in colorectal carcinoma? I'm particularly (but not exclusively) interested in your practice if you, like me, are at a community hospital that does not have a molecular lab. Do you reflexively send these tests out on all cases, colon and/or lung; or do you wait for a request to come down?

I ask this because our oncologists are being detailed by many upstart "reference" labs offering expanded molecular panels (ResponseDX is the latest). Often these panels will have something of significance, like EGFR, and then a number of other investigational markers that will not yet affect management. Then they want to know why we're not sending these out on every case.

As a follow-up, it's easy to meet minimum specimen requirements on a colon resection with a giant tumor to sample, but essentially all of our lung specimens are core biopsies...very small core biopsies. What's your experience with sending such small specimens for molecular testing, after you've already cut levels for an H&E and perhaps 2-4 stains?

 

[yaah]

We wait for the lung cases until it is requested. Usually colon tumors too. It depends on the relationship you have with your clinicians. At our place we meet with them and have these discussions. Our oncologists do not want that info on every case do we don't send it. And we also discussed what tests are appropriate and which provide value, and those are the ones we use.

 

[levels x3]

Which did you decide were of value? EGFR and KRAS are obvious candidates. BRAF could be also. What about RRM1 and others? And for EGFR, do you do mutation only, or do you look for expression as well?

 

[deleted314957]

In colon cancer I usually inform the clinicians of the availability of these assays and when they are appropriate. When we do KRAS we do a sequence of testing going from KRAS, and if neg for mutation to NRAS, and if neg for mutation to BRAF. If all are neg for mutation then that excludes 99%+ of mutations that would negate the appropriateness of Anti-EGFR therapy. We have similar algorithms for NSCLC. Although I am a solo path at a community hospital we have been absorbed(= bought) by Ameripath
( since 1996 ) and subsequently by Quest ( since 2007 ) and it is super easy for me to get these assays when needed. It is also nice that the really esoteric folks at Nichols and Specialty Labs ( who are also part of oue evil dominion ) are VERY research focused and often lead in the literature the best sequence, etc of the algorithmic testing of these tumors.

 

[yaah]

Which did you decide were of value? EGFR and KRAS are obvious candidates. BRAF could be also. What about RRM1 and others? And for EGFR, do you do mutation only, or do you look for expression as well?

Current literature suggests that mutation analysis for EGFR is probably better than expression (FISH or CISH or Impox). Older data says the expression was useful but newer data is more convincing that it isn't. And EGFR mutation data is only useful if KRAS is wild type. So the algorithm is either 1) do KRAS and if negative do EGFR or 2) Do EGFR and if positive do KRAS. Some add BRAF but others don't since it's not that common. Some institutions also add ALK now. I am not aware of RRM1 being helpful for anything except "predicting prognosis" which is unlikely to be really useful, our clinicians don't care about it.

 

[levels x3]

Thanks to both of you. What you are doing is in line with our thinking as well.

Yaah, with regard to your algorithms, everything I have read recently suggests that EGFR and KRAS mutations are mutually exclusive. If you are going to do it stepwise rather than at the same time, why not do EGFR first and stop if positive; do KRAS only if EGFR is negative. I realize that practically it would be easier to do them together if it's a send-out. But for discussion's sake, am I missing something? Have you read convincing reports of, or encountered cases with mutations of both EGFR and KRAS?

 

[lipomas]

If you are going to do it stepwise rather than at the same time, why not do EGFR first and stop if positive; do KRAS only if EGFR is negative.

You have it backwards. KRAS only matters in lung cancer if EGFR is positive, because a KRAS mutation invalidates the EGFR mutation. At least, I think so. I suppose some clinicians might still want to know KRAS status even if EGFR was negative.

 

[levels x3]

Hmmm...it seems I may need to find some new sources. I'm not questioning at all doing these tests together for efficiency's sake. I'm just trying to get a better understanding of the biology.

I'll have to read some more, but my recent reading states that these mutations essentially never occur together. This is likely because they are found in different tumor types, e.g. mutant EGFR more commonly in adenocarcinomas, particularly those with BAC features. But, I'm sure my understanding is incomplete.

I appreciate the comments.

 

[GeoLeoX]

While the biology is similar - EGFR activation leading to downstream KRAS activation and subsequent increased gene expression, the testing algorithms are different. The mutations are almost always mutually exclusive (of course there have been rare exceptions) with (IN GENERAL) EGFR mutations associated with increased susceptibility to anti-EGFR therapy and KRAS mutations associated with decreased susceptibility. As an exception EGFR T790M is associated with loss of response to anti-EGFR, but that usually develops under pressure of therapy.

In colon cancer most skip EGFR testing and go straight to KRAS testing for the reasons above. In lung patients with wild-type EGFR status treated with first line anti-EGFR therapy (compared to conventional chemo) do worse than those with mutant EGFR irrespective of the KRAS status. I don't think the same applies to colon cancer (though I am not sure).

It seems like the tide is turning against expression testing for EGFR and for sequence testing in lung (not so much for colon) for good reason, but there are mutation-specific antibodies developed that may be useful (especially since sequencing is time- and cost-intensive). However, whenever you test for specific mutations you run the risk of missing others.

For lung I would test for both EGFR and KRAS (if requested or if I were at some fancy-pants academic institution). There are clinical trials for KRAS mutant-specific therapy, so it is useful information even if there have been some recent trials that show that KRAS mutations may not be predictive of response to anti-EGFR therapy in non-small cell lung cancer (different than colon cancer). If both are negative one could argue for ALK testing (which as I understand is exclusive of EGFR and KRAS mutations and for which there is at least one clinical trial for ALK-inhibitors in lung cancer). The rest of the mutations that ResponseDX (grrr..) tests for lung (ERCC1, RRM1, TS) are probably low-yield. Did you know ResponseDX Lung tests both EGFR expression level and sequence? What a waste.

For colon I like the algorithm that MikeSheree lists above.

For lung check out Curr Oncol Rep (2010) 12:335-348.

Then there's MSI testing....

 

[yaah]

There are still clinicians who want BOTH mutation status and FISH for EGFR. Overkill. But lots of pathologists are willing to play along with that because FISH can have a reasonable professional component to the billing.

And yes - algorithms are different. MSKCC, for example, has the algorithm of:

1) EGFR mutation
If negative: Stop testing
If positive: Test KRAS
This is because KRAS mutation will render the EGFR activating mutation redundant - if you block EGFR with TKIs then the KRAS will still drive proliferation. Whereas if KRAS is wildtype blocking EGFR will be successful.

However, the lab where we send our tests has an algorithm where they test KRAS first. If KRAS is negative, then they test EGFR.

We usually just order both KRAS and EGFR mutation analysis because the clinicians still want to know KRAS even if EGFR is negative. But yes, adding FISH or CISH or Impox is unnecessary and only serves to drive up medical costs.

 

[zao275]

Regarding the question of which cases of colon adenoca to test for KRAS (and/or EGFR), at my program, most of the cases being tested were older resection specimens that now presented with metastatic disease. The reason for this is that anti-EGFR therapeutic agents (like cetuximab and panitumumab) are indicated for metastatic colorectal adenocarcinoma (http://www.cancer.gov/cancertopics/druginfo/cetuximab). Of course, that doesn't mean that they are always used in that manner. It seems reasonable to me to perform these tests on patients with metastatic disease rather than all new diagnoses of CRC.

We just did KRAS and rarely did EGFR, although I never really understood why EGFR was not tested. We performed our KRAS in-house, and used pyro sequencing as the method (just FYI).