Hope S. Rugo, MD: There are some institutions, 1 in California, that are actually looking at the NGS [next-generation sequencing] data and perhaps sending 5 different NGS tests, proteomics, and so forth. They then design these personalized combination therapies based on preclinical data or data from other diseases. I've always felt a little bit concerned about that approach. Obviously, each patient is an individual, and we want to do the most we can, but then I worry a little bit about toxicity. I don't know what the payers are thinking about. Mostly this particular institution can get the drugs covered, actually.

John Fox, MD, MHA: I would say first of all that for using NGS testing alone to inform a treatment decision outside of trial, I don't know of any payers that would cover that. My best example of that was in melanoma when patients had a BRAF V600E mutation. You used targeted therapy, or a BRAF V600E mutation with a MEK mutation, and you could use combination therapy in those patients. But if you use that same drug in patients who have a BRAF V600E mutation in colorectal cancer, the outcomes are actually worse.

Simply extrapolating that patients are going to benefit based on a molecular mutation is not evidence-based, and again, most payers would not cover that knowingly. That doesn't mean you can't get the drug on a compassionate use basis, but payers should be denying that because there is no evidence to support it.

Hope S. Rugo, MD: That’s interesting. I think the management, utilization, and costs, and all these issues play a huge role. Sometimes there are studies where there are few patients who have been treated who have breast cancer, and yet the drug gets approved all across the board, and presumably the payers will cover those drugs for patients.

John Fox, MD, MHA: Again, if it's in the NCCN [National Comprehensive Cancer Network] guidelines, it's going to get covered. Interestingly though, oftentimes, especially outside of the cancer setting, the FDA approved label is broad, and so we look at the clinical trial population that was used for the FDA approval to identify the patient subsets that are appropriate.

With atezolizumab, if you don't have a positive PD-L1 titer, we're not going to approve that drug. That's in the label, but there are some providers who would want to use that off-label. And providers still want to use nivolumab in first-line non–small cell lung cancer even though their clinical trials failed. That doesn't mean that providers can't be compassionate, but they also have to be evidence-based.

Hope S. Rugo, MD: Yes, indeed. It's a hard thing when you're doing all these peer-to-peer comparisons and need to provide all the data and the patient is there going, “I want that drug.” Anyway, we're always dealing with that. Priyanka, how do you deal with those issues, and do you do NGS in all of your patients too?

Priyanka Sharma, MD: Like everybody else, at time of first-line therapy for most patients, I'm doing it. There's clear utilization in those with hormone-positive disease but triple negative. There are those rare activating HER2 mutations or MSI [microsatellite instability]-high that you find that you can match them to drugs that are available. Pembrolizumab is now approved for MSI-high across all solid tumors. I think that trial only had 2 patients with breast cancer, so it goes back to getting these drugs approved. We don't necessarily know how well they're going to work in our specific tumor types. Beyond that it is to look for trials for PI3 kinase inhibitors or Akt inhibitors. Recently we've had some data on high TMB [tumor mutational burden] and efficacy of pembrolizumab monotherapy, although that’s not FDA approved yet. That trial did show us some reasonable efficacy, so I would consider getting pembrolizumab off-label in that kind of population with high TMB.

John Fox, MD, MHA: That's a great example because I think there may be anecdotal evidence to support use in patients who are TMB high. As payers, we wouldn't cover that because it's not an FDA approved indication. It's not in the NCCN guidelines. But interestingly, that doesn't prevent you as a prescriber from getting that drug in a compassionate use basis for a patient paying out of pocket. So what do you do in a circumstance where something is used off-label without evidence and yet the patient responds?

I think we've dealt with that several times actually as a payer because patients do get NGS testing or TMB testing, and they do respond. In that circumstance, we've made the decision as a payer that if a patient is deemed to be a responder, either a partial or complete response, then we are going to pay for the therapy the third month on. You can't say it’s experimental or investigational if the patient is considered a responder, in an N-of-1 experiment.

Hope S. Rugo, MD: That's a really important distinction that you just made because I think perhaps all of us have run into the situation where we had a patient who was responding say to single-agent pembrolizumab given on a compassionate use, and the insurer maybe a year or 2 into it says, “We’re not paying for the infusions anymore.” We have successfully argued that that doesn't seem to be reasonable in that situation.

John Fox, MD, MHA: Our primary reason for denying that coverage is that it’s considered investigational, experimental or unproven, and by the time you have a patient who is a responder using RECIST [Response Evaluation Criteria in Solid Tumors] criteria, then it's very hard to argue that in this N-of-1 experiment that they're not a responder.

Hope S. Rugo, MD: Yes, absolutely. I think we've heard from everybody. Claudine, do you have any final comments about that?

Claudine J. Isaacs, MD: No, I think that we've addressed most of them. I think one of the other things to bring up that others have alluded to is that we sometimes find mutations in genes in the HRD [homologous recombination deficiency] pathway, and certainly the role of PARP inhibitors is being looked at beyond BRCA1 and BRCA2 and looking at somatic mutations as well. There are some clues that come up that I think could be very useful in everyday life, although truly more so in directing patients toward clinical trials.

Hope S. Rugo, MD: Absolutely, and I think what Priyanka brought up about these HER2 somatic mutations also are really important since it seems to be more in our ER [estrogen receptor]-positive group where we haven't been focusing as much on NGS, I think, except for PIK3CA mutations, but now we're going to pick it up.

Transcript Edited for Clarity