Progression-Free Survival in NSCLC: Clinical Utility, Limitations, and Evolving Standards

Rajat Thawani, MD, shares key insights on balancing progression-free and overall survival, the expanding role of biomarker testing, and the promise of multi-omic data integration in reshaping NSCLC treatment.

Please introduce yourself by stating your name, title, and any relevant clinical experience you’d like to share.

Rajat Thawani, MD: I am an assistant professor and a thoracic and phase 1 oncologist at the Knight Cancer Institute in the Oregon Health and Science University.

I treat lung cancer, and I conduct phase 1 clinical trials in my practice. I spend a lot of time thinking about ways to improve the survival and quality of life of patients, as well as thinking about how to make our therapies work better and last longer.

In your view, is progression-free survival (PFS) a sufficient surrogate for overall survival (OS) in the context of modern non-small cell lung cancer (NSCLC) treatment landscapes? Why or why not?

Dr Thawani: That's a great question. In many ways, we do use PFS as a surrogate for survival in non-small cell lung cancer, for example, in patients who have such significant PFS benefit that we almost expect an OS benefit. One example is the CROWN study.

However, in general, the gold standard should always be OS. Even if they're using PFS as a surrogate endpoint and the drugs get a US Food and Drug Administration (FDA) approval, it's still appropriate to discuss the differences between PFS and OS with patients.

It's even more relevant once a patient has progressed. A significant percentage of patients do not see the next line in therapy. PFS still holds a valuable role in treatment decisions there, especially in targeted therapies, where we have to rely on PFS a lot more because patients live longer.

We have seen remarkable improvements in PFS. In a way, I think it's a sufficient surrogate, but it doesn't negate the fact that we need to wait for overall survival data to make better decisions and help patients make better decisions.

When discussing new treatment options with patients, how do you explain the difference between PFS and OS in a way that resonates with their goals and concerns?

Dr Thawani: That's very important because, in my last answer, I said that we discuss PFS and OS. Those terms don’t mean anything to the patients unless you explain it.

PFS is how long until the cancer starts growing again. OS is how long patients live, irrespective of what treatments come after the initial treatment. I often tell patients that PFS would slow down or silence the cancer for a long time, even if it doesn't necessarily mean that they live longer. Cancers might shrink first but may not really improve how long patients live overall.

We discuss that the goal is to select the treatment based on what their wishes are—what they want to do. Do they want to keep the disease controlled? Do they want to minimize side effects or focus on time?

How has the landscape of biomarker testing evolved in NSCLC over the past few years, and what are the most critical markers to test for today?

Dr Thawani: It's evolved a lot. I say that because I was just looking at the history of biomarker testing in non-small cell lung cancer, and I presented it to our fellows yesterday.

In the past, there has been a focus on looking for EGFR and then maybe ALK. But now, the standard testing includes many more biomarkers that we have drugs against. It's almost getting to a point where people really have to count on the reports that they get from vendors who are doing this testing to determine what treatment options patients have.

There are so many drugs that have now been approved. We have EGFR, ALK, ROS1, BRAF, KRAS, MET exon 14, now we have MET IHC, RET, HER2 mutations, HER2 IHC, NTRAK, and PD-L1. There are so many biomarkers that we need to make clinical decisions. The best way to test them is to test all of them.

The most efficient way of testing all of them is to at least do NGS and PD-L1 testing on every patient. NGS would pick up all of the actionable genes and a lot of other genes that probably have trials ongoing. PD-L1 testing is important and required. It's a separate test with IHC to make decisions on immunotherapy with and without chemotherapy. Overall, NGS and PD-L1 are still the most important tests to do upfront.

After frontline progression, there are new approvals for MET IHC. There's a drug—a MET antibody-drug conjugate (ADC)—that requires MET IHC testing. There is also a HER2 ADC that requires HER2 IHC testing. If you have enough tissue, you can do that upfront and make decisions for further lines. You can keep them in your back pocket or, if tissue is an issue—which is the case many times in lung cancer—you can probably wait for progression.

The good thing is that we can do NGS on liquid—specifically on ctDNA using a blood test. It's being used more and more upfront, and it has its pros and cons. But I would say that there are many ways to get NGS, and it's incredibly important in making decisions in lung cancer today.

How do delays in testing (turnaround time, insurance, access) impact clinical decisions and patient outcomes in your experience?

Dr Thawani: It's psychological for the patients and the providers to wait while a patient has stage IV cancer. However, there have been many studies that have shown that we should wait for biomarker testing before we make treatment decisions. It's wise to wait 2 to 3 weeks. Hopefully, that's how long it takes to get NGS testing.

An approach that we have utilized is called concurrent testing, where we send a liquid biopsy at the same time that we send tissue NGS. The liquid biopsy comes back within a week, hopefully, and tissue NGS can take longer. If we have a positive on the liquid biopsy, we can use that to make treatment decisions. If it's negative, we still wait for tissue NGS. That's the psychological aspect of it.

For the medical aspect of it, we have seen that, obviously, waiting to start treatment is not possible for many patients. There are many reasons why we have to wait. For the liquid biopsy and the tissue biopsies that are sent out, vendors are getting better and better about working with insurance companies or absorbing some of the costs of these tests.

There are a lot of consequences for waiting too long or waiting for insurance approvals. Companies are starting to test without having authorization right away. It's a good thing that we're not having to wait as long, and we have tools like concurrent testing to be able to make decisions faster.

What are the most exciting developments you're watching in NSCLC trials right now—and how do you think they'll change the way we approach endpoints and testing in the next few years?

Dr Thawani: In the last couple of years, we've talked a lot about neoadjuvant and perioperative immunotherapy and chemotherapy approvals and trials. It's going to be exciting to see if a surrogate endpoint comes out of it.

Pathologic complete response—if it proves to be a reliable surrogate endpoint for neoadjuvant treatment—is something that will likely shorten trials and improve access to a lot of new treatments in early-stage lung cancer or locally advanced lung cancer.

The other endpoint that I'm looking forward to is seeing how the ctDNA minimal residual disease, which has been used in liquid tumors for a very long time, plays out. CtDNA clearance is expected to predict long-term outcomes. Is minimal residual disease or molecular residual disease going to be an accepted regulatory endpoint? That’s to be determined.

The other space that is exciting are ADCs that have been approved in lung cancer, including trastuzumab deruxtecan (T-DXd) and telisotuzumab vedotin (Teliso-V). We don't have broad approvals, which means that we still have a biomarker that is required for their use in our patients, but there are more therapy options than there were before. It's still exciting, and we'll have to see how the other ADCs that are still in trial play out. There is a lot of new development happening in the targeted therapy space as well, with new small-molecule oral medications being studied for biomarker-driven lung cancer or oncogene-driven lung cancer.

Finally, one of the most exciting things for me is multi-omic profiling. We have a lot of big data in oncology, especially in lung cancer. We sit on a lot of data, such as NGS, clinical outcomes, radiomics, scans, and ctDNA. It'll be exciting to see how all of these can be integrated into one biomarker to help identify high-risk or low-risk patients—those for whom we can stop therapy, de-escalate, or escalate regimens on. It will be exciting to see if that does anything to benefit patients.