The registry from Fribourg shows: The “tail of the curve” is real

Long-term survival in advanced non-small cell lung cancer is now showing up in real-world practice. In this interview, Prof. Dr. med. A.Curioni-Fontecedro explains why protecting that gain depends on the discipline of testing biomarkers upfront, retesting at progression, and refusing to close future treatment doors with unnecessary toxicity.

Today, (chemo-)immunotherapy (IO) has become a standard first-line approach in advanced NSCLC. How has this shift towards chemo-IO combinations influenced the way you manage patients in your daily practice?

A. Curioni-Fontecedro: The advent of chemoimmunotherapy combinations — with the different studies and different combinations — has changed the way we treat patients dramatically, mainly due to the improvement in terms of overall survival and overall responses. In our daily practice patients are usually highly symptomatic when they are first diagnosed with advanced non-small cell lung cancer (NSCLC), and we want to have a reduction of disease burden. Therefore we want to see responses. And these responses have been significantly, and clinically significantly, improved by the combination of chemo- and immunotherapy.

What we need to do as soon as we have a diagnosis of lung cancer, is to evaluate the patient for possible comorbidities or issues that might hamper receiving any treatment. At the same time we need to have reflex testing to understand and to know more about the disease. This reflex testing includes not only immunohistochemistry but also next-generation sequencing for all patients with NSCLC. This allows us to give the patient the best treatment option as soon as possible.

Zooming in on PD-L1-negative advanced NSCLC, what does your clinic’s real-world data show in terms of patient characteristics?

A. Curioni-Fontecedro: As soon as we have a patient diagnosed with NSCLC it is really important to get the biomarker analysis, which includes PD-L1. Let us have a look at what is happening in our daily practice and in our registry data from Fribourg. We see that PD-L1 expression is distributed pretty similarly, with roughly 30% of patients who are PD-L1-negative. What do we know from the literature? PD-L1-expression depends on the different populations of patients. And we also know that it might vary at different time points. I recall a study we performed looking at tumor samples of patients before treatment — in that case, patients before an adjuvant treatment — and then the PD-L1-expression afterwards, when the tumor recurred. There was a difference. That means that we need to eventually retest the patient and not trust or use an old PD-L1 staining, for example, of a tumor that relapsed.

This is what we see in real life, and this is what we see in Fribourg. We have collected roughly 800 cases of NSCLC, with all the analyses — roughly 100 data points for each patient. We want to understand the epidemiology, meaning the features and characteristics of the patients who come to our hospital, because we need to be prepared and give each patient all opportunities. Knowing what the epidemiology looks like helps us a lot to tailor which kinds of clinical trials to open, which options to offer, how many patients we are going to expect, and what the needs of the different populations of patients are.

Along the same lines on PD-L1-expression: We analyze the different metastases only when itʼs necessary — for example, when we do not know whether a lesion is a metastasis or not, or when a patient is suffering from respiratory disease. In those cases we can analyse the different metastases for that individual patient.

What we also saw in our registry — and this is very important — is that we have a large population of patients with more than one type of cancer. Certainly, one of the reasons is that the population is aging. Another one is risk factors. What is probably different in Fribourg is that we have a very large population of smokers, and this certainly impacts the subtype of cancer that we see and the molecular features of these cancers. That is why we decided to develop this registry, which is in the hands of Dr. Kupka, who punctually evaluated all the cases to understand exactly what our population of patients looks like.

And based on those characteristics, what treatment options do you typically consider for PD-L1-negative patients, and what kind of outcomes do you observe in this subgroup?

A. Curioni-Fontecedro: At our institution we follow mostly the ESMO guidelines, where we have a plethora of treatment options in terms of chemo- and chemoimmunotherapy. There is the possibility, of course, to add bevacizumab and so on. Our treatment option in the first-line setting follows the KEYNOTE trials (KEYNOTE-189, KEYNOTE-407), because we have long-term outcome data, which are really important and reflect what we see in our daily practice. Example: when we look at the overall survival curves and we call it the famous “tail of the curve” that we always hoped for as oncologists. This is what we see in daily practice now. In our registry we have roughly 800 patients, and the majority of them has received this combination, because we started the registry 4 to 5 years ago. And we see the tail of the curve even better than in the clinical trials.

We need to keep in mind, when we decide between this plethora of possibilities, what the comorbidities are — for example age, clinical characteristics of the patient, not only the biomarkers, but also how the patient is doing and what other co-medications the patient is taking for the comorbidities. This is very important, because, envisioning that the majority of patients might respond to this treatment, we need to be prepared that — in the maintenance phase, or in case of progression — we need to be ready for a second-line treatment option. That is why we should not preclude the patient from any possible treatment option afterwards by causing too many toxicities, or toxicities that might hamper these possibilities.

We do not preclude the patient from any possible treatment options, and we use mainly a combination with only one immunotherapy, due to the possibility of different toxicities by combining or adding on top, for example, bevacizumab or double immunotherapy. And, having no direct head-to-head comparison, this does not allow us to say: Should we really risk more toxicity for a benefit? We do not know that yet.

How do current clinical guidelines inform about the different options in daily practice for PD-L1-negative patients?

A. Curioni-Fontecedro: We have different guidelines, and they are, of course, very rich in information based on all the different clinical trials. I would say first of all that we need to keep in mind that to date, in 2026, the guidelines for the first-line treatment option of NSCLC are just something to look at, because pa-tients with targetable alterations will not follow the classical algorithm of first-line treatment for non-oncogene-addicted disease. Be very careful that in that algorithm KRAS is also mentioned — and we do not use the KRAS mutation as a targetable oncogenic aberration for treatment decisions in the first-line setting.

Keep this in mind: If I have a patient who comes to my clinic with stage IV NSCLC — let us take the example of adenocarcinoma — with a KRAS mutation, even a classical one such as G12C, currently this does not change the treatment option for the first-line setting. If I had a clinical trial open for this population of patients, yes. That is why we test the patient for all the mutations already at the very beginning, all together.

Do the ESMO, ASCO, or NCCN guidelines help? I think the difficulty is that they list all the studies and all the treatment options. And we, as clinicians, need to be able to define to which patient we want to give a certain kind of treatment regimen, based on comorbidities, other treatment options, the patient’s wishes, and what we have available in the country. When we take this into consideration, we know that the addition of new treatments might add new toxicities. The guidelines often report toxicities as they are reported in the clinical trials but do not guide us in the decision between this plethora of options.

There is something very important with guidelines: We need to always check when they have been updated. We have to take this into consideration, because it might be that a study has been updated in the meantime, or we may have new information about the studies, and this has not yet been included in the guidelines.

We refer in our institution to the ESMO guidelines, but we also have to look at what is approved in Switzerland. So we need to move from a broad guideline to what is available in our country. We are in a very, very good situation because we have access to most of the drugs. I have to say that, when I am visiting professors in other institutions in other countries, this is not always the case — and this is important when we face colleagues from different countries: not to simply say, “Oh, we have every treatment option available”. Look as well at what is available in the country, and then discuss which guidelines can be applied in the specific situation.

What is your personal outlook on desired developments in the PD-L1-negative patient segment? Do you see any promising trials on the horizon or do you have a trial designor treatment pathway in mind that might improve the outcomes in this group?

A. Curioni-Fontecedro: The first important thing would be to understand how PD-L1 expression changes. I always say: Let us retest. Do we really know the biology of this tumor? And the second thing is: Is it relevant for the patient to know that? To date, we do not know. We do not know, because most of the clinical trials that have been performed, especially in the second-line setting, do not use new biopsies or re-biopsies of the tumor. That means that we do not know the actual biology of the cancer when we start the second-line treatment. And this is crucial.

So if I had a wish, it would be to have the possibility of different treatment options within a clinical trial — by understanding as well the biology of the disease and trying to tailor the treatment for the different populations of patients. Of course, this is often not possible and not feasible.

What we are trying to do in research is to understand what the mechanisms of resistance to immunotherapy are — also in the PD-L1-negative population — to then understand which kind of treatment could create a synergy. And then, with these synergies that we can test in the lab, bring this into a clinical trial. We are soon to open in Switzerland the SAKK 18/25 study. This is a new study combining gemcitabine and ivonescimab for the population of patients independent of PD-L1. But I promise that we will look into the biology of this disease, to understand who responds to this treatment and who does not.

Quelle: