<p>DNA-based next-generation sequencing (NGS) of tumor tissue has become an important standard in oncology, and many ROS1-positive non–small cell lung cancer (ROS1+ NSCLC) cases are detected in this way. Our experts discuss the optimization of molecular testing with both DNA- and RNA-based NGS for identifying driver mutations in advanced NSCLC, with a focus on ROS1+ patients.</p>

NSCLC implies 1 disease, although we know that NSCLC is actually comprised of different diseases. So, it is imperative that we get the diagnosis correct, and not only the histological diagnosis. Since cancer is a disease of DNA, we also need to interrogate a patient’s DNA to see whether there is any information that may be important for optimizing therapy. It is important that we, as oncologists, deliver the right treatment at the right time to the right patient.

<br>

In my practice, we test tissue for DNA and RNA concurrently. RNA testing increases the yield of some fusions, including ROS1. In my opinion, obtaining circulating tumor DNA (ctDNA) should be complementary to tissue testing. The combination of tissue and plasma optimizes yield and, therefore, the likelihood of finding out whether there is a driver mutation present, which opens up a lot of targeted therapy options. Compared with traditional chemotherapy, targeted therapies are generally more effective, more convenient, and better tolerated in these subsets of patients with NSCLC.

<br>

We reflexively test all appropriate patients, from those with early-stage to stage IV disease, to start the clock sooner so that, by the time the patient sees an oncologist, we know when to expect results. And I think it is imperative that clinicians and pathologists collaborate to determine how and when testing should be done.

<br>

For patients in whom I would need to initiate treatment prior to receiving comprehensive genomic testing results, which is relatively uncommon, I would hold off on using immunotherapy and instead treat with 1 cycle of platinum-based therapy (eg, carboplatin plus pemetrexed or carboplatin, pemetrexed, and bevacizumab). Once I receive the testing results, if a patient is feeling better, I would not necessarily want to give up on an effective treatment that has not been maximized. So, I might consider giving 4 cycles of platinum-based therapy and then transitioning to a targeted therapy. However, if a patient is still symptomatic or is experiencing side effects, I would start treatment with a targeted therapy. We know that the response rates with TKIs in the ROS1+ population are higher than those with chemotherapy.

As the list of things that we look for has expanded, I think that the value of multiplex testing has been proven multiple times over. Once you are looking for multiple analytes, it is more cost-effective to do upfront NGS. As Dr Socinski said, ideally, you are testing both DNA and RNA concurrently. If you see that only DNA-based testing was done for a patient, that might be a “hole” in their panel, particularly if you have not yet found something.

<br>

While ctDNA testing is considered NGS, its sensitivity remains an issue for certain driver mutations, such as ROS1, as it is only DNA based. RNA-based blood testing can be problematic due to rapid degradation. So, it is great if you see a positive result with ctDNA, but if you have a negative result, you have to keep looking.

<br>

Protein-based biomarkers to help inform treatment decisions for ADCs are the next wave of biomarkers. It is important to note that protein-based biomarkers are not as distinct as genetic markers, but they overlap. If you have a patient who is negative for everything else (and probably is positive, too, for an oncogene), I would test MET and/or HER2 immunohistochemistry because we have other therapeutic options.

<br>

Finally, if you have to start treatment prior to knowing the patient’s driver mutation status, I do not like to give just 1 cycle of platinum-based chemotherapy. You do not even know if the chemotherapy is working after 1 cycle. I would, at least, give 2 cycles and then get the scan. If it is not working, then you switch to the TKI. If it is working, you talk about tolerability and other issues. It was actually Dr Wakelee’s work that showed that ROS1+ NSCLC has exaggerated sensitivity to pemetrexed. And so, this is an important line of therapy for some people. I do not think that we should deny them that benefit, and, in that circumstance, the TKI just becomes second-line therapy.

Since you cannot just look at a patient and know what is going on with their tumor, we must look at everything. We will still do fluorescence in situ hybridization testing if we need an immediate answer, but it is not optimal. Where Dr Camidge and I practice, we find ROS1 mutations in a higher percentage of our patients, just by the nature of who comes to see us. If I have a young patient without a history of smoking, I am not going to rest until we have identified the driver mutations.

<br>

ROS1 and other fusions can be missed with standard DNA tissue and/or plasma testing, so RNA testing can be helpful. This is no longer an era in which we can decide what we think is most common and just look for that; we really need to use a comprehensive testing modality.

<br>

If a patient presents with an aggressive tumor and is quite sick, and if you do not yet know what the driver mutations are, we can always start chemotherapy. If there is a high probability of a driver mutation, I will usually not start an ICI at the same time. Then, if a patient is tolerating the treatment relatively well, which many of them do, what if the results come back showing an ROS1 mutation? Many would recommend immediately switching to targeted therapy, but I do not always immediately switch. We know that ROS1+ NSCLC has an exquisite sensitivity to pemetrexed, so I may have a conversation with the patient about this, with a quick trigger to start them on an ROS1-targeted agent if needed.

<br>

The response rates to TKIs are quite high, but they are not 100%, and they do have associated toxicities. Just because it is a pill does not mean that it is necessarily going to be better tolerated. Although we have not done the trials yet, I also wonder about whether adding TKIs to chemotherapy will be a possibility in the future.