06 August 2021
“Smoking is certainly a determinant risk factor for lung cancer, but it’s not the cause”
Interview with Nuno Gil, Director of the Lung Unit
06 August 2021
Interview with Nuno Gil, Director of the Lung Unit
Right now, lung cancer is the most lethal cancer in the world. Eighty percent of its victims are smokers, making smoking a determinant risk factor for this cancer. But according to oncologist Nuno Gil, who heads the Lung Unit at the Champalimaud Clinical Centre, smoking is not the ultimate cause of lung cancer. First, many smokers never develop lung cancer, and second, a growing number of non-smokers have this disease. Nuno Gil would really like to solve this mystery.
When we speak with Nuno Gil, we immediately feel that he cares for people and in particular for his patients. It is not easy to tell someone that they have lung cancer, and a great sense of humanity is needed to help patients deal with the bad news. Lung cancer can be very aggressive and is frequently detected at an advanced stage. Survival statistics are not very encouraging either. Even so, there have been advances, in the last years, in the treatment of certain types of lung cancer – and patient quality of life has improved. Nuno Gil spoke with us about this, and also about a new approach to lung cancer screening being developed by his team that could improve its early diagnosis.
Cases of lung cancer have decreased in the past years, undoubtedly due to the decrease in smoking, a determinant risk factor. But this is not enough for Nuno Gil, who, at 62, thinks that the true cause of lung cancer is yet to be identified.
Actually, cancerous cells confuse Nuno Gil. He would really like “to understand their logic”. They divide, they multiply and then they spread, and the final result is their own death – unlike viruses, which infect an increasing number of hosts and do not disappear. “I find this so odd!”, he exclaims. Nuno Gil views this issue as a great mystery and believes that it will not be possible to get rid of lung cancer as long as it remains. He also gives some insights on how to try to solve this enigma.
It’s the most deadly cancer in the world. Breast cancer, which is more frequent, is less lethal because it can be diagnosed at an early stage. In lung cancer, a relatively common symptom is coughing. And a cough… can be due to an allergy, to side-effects of medication against high blood pressure and to so many other conditions.
Yes. The division of lung cancer into subtypes has evolved historically. Basically, in the mid-twentieth century we divided lung cancer into two main groups: small-cell tumours and tumours – carcinomas – said to be “non-small cell”. Nobody knows why small-cell tumours have small cells. Strange, isn’t it?
In fact, there are several subtypes of non-small cell tumours. In the beginning, the dominant subtype was squamous cell carcinoma, so called because its cells are flattened (whereas respiratory tract cells are cylindrical). The squamous cells are similar to skin cells, so these tumours are also called epidermoid carcinomas. But more recently, adenocarcinomas, which are made of cylindrical cells, became the most common subtype. Nobody knows why.
Not any longer. Since perhaps the first decade of this century, we started to understand that there were a number of adenocarcinoma variants, characterised by specific genetic mutations. We haven’t gone into a classification frenzy: in these adenocarcinoma variants, the nature of the cells is different, the natural history of the disease is different. For example, we are starting to have an idea of what the median survival is for some of them. We estimate that the median survival for adenocarcinoma with a mutation in a gene called EGFR is around 22 to 24 months. But if a patient has a variant with a mutation in another gene, called ALK, the median survival in this case is 81 months. This means there are different diseases within this mix that we initially called non-small cell carcinomas of the lung.
Yes. The difference is mostly that, for known mutations, there has been an increasing number of specific drugs to try to cancel the effects of each mutation. We already have specific treatments for nearly ten of these mutational variants.
If we wanted to be more exact in terms of classification, we should say that there are probably many more variants, because there is no cancer without mutations. But the distinction of the variant is only made when there are specific drugs against the mutation. The current classification depends on the type of specific drugs that are used.
According to data from Globocan, almost 4,800 people died in 2020. The incidence (that is, the number of new patients diagnosed) was 5,400. The mortality rate for lung cancer is very high.
Yes. One of the dogmas of Oncology is that the earlier you detect this disease, the higher the probability of a cure. This is true for all tumours. First problem: how can you detect it earlier? And even before that question, another one: who are the people at greater risk of having the disease? In lung cancer, we even have an advantage, which is the link between this risk and smoking.
I am going to commit a heresy with this statement [he touches his forehead and laughs]: I don’t believe smoking to be the cause of lung cancer. Why do I say this? In argument against what I have just said, we have a statistic: for every hundred diagnoses of lung cancer, 80 involve people who smoke or who have smoked. And when I say this, I’m already contradicting what I have just claimed about a causal link.
But now, let’s turn the argument around and ask: for every hundred people who have smoked their whole life, how many are going to be diagnosed with lung cancer at some point in their life? (A lifelong smoker starts, let’s say, at age 18 and stops at 80, which is the mean life expectancy.) Let’s repeat the question: for every hundred people who have smoked their whole life, do you know how many will be diagnosed with lung cancer? In Portugal, the estimate is 15 percent. And in Sweden, eight percent.
In other words, if the majority of people who are exposed to a supposed cause do not suffer the consequences, then that supposed cause cannot be the cause. Smoking is certainly a determinant risk factor for lung cancer, but it is not the cause. Makes sense, doesn’t it? Those who smoke have a higher propensity to develop lung cancer, but smoking by itself is not sufficient to cause the disease: the remaining 85 percent of smokers will never get lung cancer. And that raises intriguing questions: what is their secret?
It is true that, when we look at lung cancer incidence statistics and compare smokers with non-smokers, we see that the incidence of lung cancer has been decreasing through the years, in tandem with the decline in smoking rates. This has clearly been going on for decades in men and was also true for women until about ten years ago. And of course I believe that the decrease is due to anti-smoking campaigns that make people aware of the risk. Although it is not the cause, smoking is a strong risk factor.
We do. Some ten years ago, we started seeing an interesting phenomenon: the rate of lung cancer was increasing in people who had never smoked, mostly in women. And that’s an issue that has no answer – not only with regard to the cause, but mainly because it means that we are not providing an adequate response in terms of early diagnosis. This is due to the fact that CT scan screening for lung cancer primarily targets smokers, active or passive, depending on the number of cigarettes they smoke or smoked. When it comes to “never-smokers”, smoking as the supposed “cause” doesn’t apply, excluding these people from early diagnosis plans. This is, for me, the most afflicted population, even though it is not the most frequent one.
There is no lung cancer screening in Portugal, period – no matter how strongly clinical studies, from both the US and Europe, have demonstrated that screening decreases mortality. But even if there were, it would never be for never-smokers. They would always be excluded from early detection programmes.
That’s what the numbers suggest. So much so that the initial percentages I gave of 80 percent smokers versus 20 percent non-smokers is decreasing for smokers: it is now more or less 75 percent versus 25 percent. So what is going to happen is that the number of lung cancers we see in never-smokers is going to increase. Of course, we can speculate that, although they are non-smokers, they are still breathing car fumes and pollution in big cities...
Second-hand smoking is very hard to quantify. It not only depends on how much the active smoker is smoking, but also on the environment in which the passive smoker is getting the smoke from the active smoker. It’s one thing if you are in a confined space for days on end and another if you are in an open, large, well-ventilated area; perhaps the associated risk is different. It’s very difficult to quantify that risk and the results from the studies that have looked into this problem vary from 5 percent to more than 20 percent. These are huge margins.
There has been much talk about pollution and several studies have looked into exposure to fumes from cars, mostly from diesel motors. There is also exposure to asbestos, which has been intensively used in construction, shipbuilding, etc. Asbestos is also a risk factor, and so is a natural gas called radon, a radioactive gas that is present mostly in granite rocks – and also, exposure to certain chemicals.
Lung cancer is an environmental problem; it’s not a genetic problem in the sense that it is not hereditary. It’s environmental because, up until now, no mutations have been described, as is the case for breast cancer, for example, that are passed down from generation to generation and are a risk factor for the development of the disease. And today, we are at a stage where we understand most of the genome – or at least a good part of it.
In the literature, only one case has been described so far, in the whole world. It’s the case of a woman from Japan who had adenocarcinoma of the lung with a certain mutation on the EGFR gene, and whose son also developed lung cancer with the same mutation.
But we can also ask, still in relation to smoking as a cause, if countries with a higher percentage of smokers are those with the highest incidence of lung cancer. They aren’t. There is no direct correlation. Do you know which region in Portugal has the highest incidence of lung cancer in proportion to the total number of inhabitants – actually in all types of cancer? The Azores.
The Portuguese Directorate-General of Health hypothesizes that, as there are higher levels of consanguinity in that area – the Azores are a group of islands – lower genetic diversity leads to the emergence of more diseases. And there is also another possible reason in the case of lung cancer: the Azores are the only place in Portugal that currently grows tobacco. There is Azorean tobacco and I think that makes it more accessible and also cheaper. It could be a place where people smoke more, and therefore the risk could be higher. We ourselves tried to perform that epidemiological study, but the extant data by county is not sufficiently precise.
If I wanted to theorise, I would say this: smoking is an ancestral human habit, right? Even for medicinal purposes. Now, from the point of view of evolutionary biology, if a habit is harmful to a species, one of two things can happen: the extinction of the habit or the extinction of the species. But people still smoke today. And not only that: in those National Geographic television programmes, we see isolated populations where everybody smokes since childhood. But I don’t know if there are a lot of lung cancer cases happening there…
The problem may not be the smoking itself, but rather what people are smoking, right? There’s a difference. And taking a step further, what is tobacco? It’s a plant that has to be left to rot. For those who don’t know, the tobacco leaves people smoke are rotten leaves! And rotting is a process that is mediated by microorganisms.
Let me tell you something. Some time ago, me and my colleague Dr. Susana Simões, a pulmonologist at the Lung Unit, went to the Philip Morris research center in Switzerland. We wanted to see what they had to offer for those who wanted to keep enjoying the pleasure of smoking with the least possible risk. For these people, they invented heated tobacco – the so-called IQOS. There are other electronic cigarettes that merely vaporise nicotine, but inside IQOS there is a small roll of tobacco, which is heated when smoked. There is no combustion involved. And they showed – and we should praise them for it – that the amount of nitrosamines (chemical substances that are possibly carcinogenic) and that of carbon monoxide are reduced in IQOS to ten percent of their amount in normal cigarettes. We are talking about two components that are thought to be responsible for many tobacco-related diseases, and not only lung cancer.
But we went there to question them about a different matter: what does tobacco carry with it? They had never considered tobacco as a microorganism transporter. Now: there are microorganisms that resist combustion; they are called extremophiles. They had never thought about that either, that tobacco could be a transporter of disease through microorganisms, and not only through purely chemical compounds. They were very surprised by this hypothesis.
We know there are associations between various tumours and microorganisms. For instance, between cervical cancer and the human papilloma virus (HPV), that one is very well-known. Or between the bacterium Helicobacter pylori and certain stomach tumours and gastric lymphomas. And there are others.
Smoking reduction, which decreased the number of cases of tumours associated with tobacco, is clearly an advance in terms of prevention. But there is still the problem of lung tumours in people who have never smoked and, from that perspective, we have made no advance whatsoever. How can you prevent something if you don’t know what is causing it?
In the case of the lung, studies – European studies, as well as an American study (which was in fact the first) – show that a periodic CT scan of people at risk, who normally have been exposed to tobacco, and who are of a certain age (which varies depending on the study), reduces mortality from lung cancer by around 20 percent (according to the American study) – which is more than the 15 percent reduction of mortality from breast cancer thanks to mammograms. In the European study, the reduction in mortality from lung cancer was even greater, I think between 33 and 44 percent, mainly in women.
There is, however, a problem with screening programmes, which is how to deal with false positives. Even more when the screening is based on images, on a CT scan (mammography has the same problem, though with less false positives than a CT scan). Why is this a problem? Because many of the changes we detect in a CT scan, which are morphologically identical to lung cancer tumour in its initial stages, can in fact be scar tissue or even benign. But once they are detected, they compel doctors to perform a series of invasive examinations and additional studies such as biopsies and sometimes even surgeries to obtain conclusive results. And the fact is that there is a mortality rate – although an extremely low one – associated with the screening process. And that would be the worst possible outcome: to die from an early diagnosis screening. All screening programmes have this problem, but it is particularly acute in the case of lung cancer.
Be that as it may, the validity of such a screening approach has already been amply demonstrated and all that is now left to do is to implement it. If everything goes well, we at the Champalimaud Foundation are going to launch a screening programme in the next few months. On the other hand, we also have an early diagnosis appointment that is beginning to detect patients who are then referred for early diagnosis of lung cancer.
By the way, ASCO (the American Society of Clinical Oncology), which produces an annual report of the year’s news, has noted that in the last few years, cancer mortality rates have decreased mostly due to lung cancer. Why? Probably because we are starting to become more aware of the relevance of earlier diagnosis.
The lung cancer screening dogma is very much centered on smokers, but is also starting to value family history. But even without any known mutation for this disease that passes on from one generation to the next (except for that case in Japan, which was in fact rather special), we manage to recognise many familial cases, which pickup smokers and non-smokers.
Yes. There are families with several cases of lung cancer, in smokers and non-smokers. So we know there is a familial risk, but there is no known specific mutation. If a family has several cases, that means there is a risk, however intangible. And families are starting to perceive this and asking to be screened.
At this point, the problem we face is how we are going to be able to avoid the risks of the screening itself. Precisely with this in mind, we are developing a study, called VOX-PULMO, whose aim is to try to identify chemical profiles of odors which are specific to lung cancer (assuming that diseases have odors of their own), and ultimately to obtain what experts call a “respiratory biopsy”.
This method would only entail a non-invasive examination, be easy to perform, not very costly, and could also provide a direct response to that growing group of people who, not being smokers, do not belong in the classic risk groups eligible for CT scan screening. Performing such a biological screening first, and obtaining a suspicious profile, could then lead to a CT scan – and avoid performing “I don’t know how many” useless CT scans. Because it is also a fact that most non-smokers will never get lung cancer.
The majority of lung cancer patients who are diagnosed at an early stage do not have any symptoms. Their cancer is detected, for instance, because they had to do a preoperative X-ray for another condition and there is a lesion of the lung that is visible there, or because they did a CT scan for some other reason. Nowadays, cardiac CT scans are performed, and they show part of the lungs – and sometimes, something is detected. But basically, these patients are “caught” by chance.
Yes. We have progressed substantially. Clearly, the best treatment is surgery, but only 25 to 30 percent of the patients are eligible for surgery. The others already have the disease either in an advanced local stage or with metastases when diagnosed. And obviously, the probability of a cure for a metastasised disease is almost zero. I say “almost” because some of those patients can benefit from treatment, mostly based on immunotherapy, which can have fabulous responses in the long term – though we still don’t know for how long.
No. Those patients never undergo surgery. Sometimes, when the disease is localised in the thorax, but already in an advanced stage, we treat them with a combination of chemotherapy and immunotherapy. We are talking about non-small cell lung cancers.
Yes. Small cell cancers are very aggressive, very systemic right from the start, so only about 20 percent of this type of cancer is localised. Almost all of them are treated with chemotherapy.
The treatment I use for patients with small cell lung cancer is still the same we used 40 years ago. And if in 40 years the first-line treatment for small cell cancer hasn’t changed, that’s very bad indeed. It’s somewhat devastating, don’t you think?
Yes, but that number concerns patients with mutations. Those that do not have mutations that can be treated specifically – and they are currently the majority – nowadays are already being treated with chemotherapy and immunotherapy simultaneously, because immunotherapy brings a plus to what chemotherapy does.
Just to give you an idea, if a patient with a Stage IV lung adenocarcinoma at the time of diagnosis – that is, without any specific mutation – were not to be treated at all, the median survival would be four months. If he or she is treated with one of the most recent chemotherapies (so-called third-generation), the median survival goes up to between nine and eleven months. But today, by adding immunotherapy, we are probably reaching a median survival of 22 months. We are gaining time…
That’s a difficult question, because it’s obvious that quality of life should be a sine qua non condition. It seems to me that treating a living dead person is useless and ethically objectionable. The question is how we define quality of life.
I’d like to give the example of a patient I treated many years ago, who had a small cell carcinoma of the lung in an advanced stage. The treatment – the same we have used for 40 years, as I already mentioned – is cisplatin, a derivative of platine.
Cisplatine is toxic and has many side effects. One of them is nausea and subsequent vomiting. By chance, I happened to catch the transition from those ineffective drugs against vomiting to what is one the greatest innovations in oncology: drugs that really work against this side effect of treatment. One of our major gains, I think.
So with this patient, I started giving him all the medication, and each time I saw him I made a check-list of all the side effects. To see if we could minimise them, right? Because with these treatments there is no such thing as a “free lunch”, there’s always a price to pay. And I asked him: “how are we in terms of nausea and vomiting?”. To which he replied: “I’ve been vomiting, really vomiting a lot”. Me: “Are you taking the drugs I gave you?” The patient: “No”. Me again: “Doesn’t that make you suffer, harm you?”. And then he said: “No. And you know why? Because I feel that when I vomit, I’m expelling the bad chemicals from my body”. On the scale of quality of life, vomiting is a bad thing; for him, it was a good thing.
But of course we are always trying to improve quality of life as much as possible. I think we’ve gotten substantially better at that, and, overall, we can say that having chemotherapy today is no longer the ordeal patients had to go through to achieve a gain that wasn’t always substantial… Moreover, in lung cancer, we have one advantage, which is that few drugs result in hair loss, something I consider a very important hallmark of quality of life.
I think quality of life has improved and, if our main objective is to prolong life, we have been successful.
Yes, we have an appointment for smoking cessation. It would be ironic if we didn’t, wouldn’t it? But consider now this crazy idea: transforming tobacco into a therapeutic agent... But that’s another story.
Let’s turn the question around: how many patients directly ask us how much time they have left to live? I would say they are more or less 10 percent. I already had several patients who told me, “look, when this becomes really bad, don’t tell me”.
On the other hand, I never use the word “cure” with patients. There is a definition of cure by an English oncologist that I use a lot: “‘Cure’”, he says, “is the word you use on the day a cancer patient dies, but not from cancer”.
What I usually say to them is that this disease, based on our current medical knowledge, is not curable, but is usually treatable. And if they ask me what awaits them in terms of the time they have left (even though few of them ask this), I tell them I don’t know. It’s not meant to avoid the question, it’s precisely because I have already seen so many things happen that I think anything can happen. Medicine tends to consider median survivals, but the statistical distribution of the survival times is a so-called bell-curve (because of its shape, with the median at its peak). But that doesn’t mean the patient won’t die in the next few days or live for many years.
Yes, but if the probability of being alive a year from now is 99 percent, I can be part of the one percent that doesn’t get there or that exceeds expectations.
No. When I acknowledge that I do not know what to say I am being fair to the truth. Because I don’t know what the individual truth is. I only know the statistical truth. But you have to understand that statistical truth always refers to the past; it’s not the truth of the present moment, which is when I need to make a decision. What I mean to say by this is that all statistical curves concern old, past studies, and can have changed in the meantime.
I do. Doctors have to have the courage to cry, if necessary.
So many times!
With the patient. Afterward, a lot more. I normally try to somewhat avoid these situations, though, because the patient expects me to be his rock, someone that is not going to break down into pieces. But sometimes, it’s just impossible to avoid. For me, at least [his face expresses deep sadness].
They are. To the point that we have already had training about how to communicate bad news. Including role-playing, precisely to train what we are not taught at medical school, which is how to communicate. Communication skills are essential, and not only to make the relationship with the patient as good as possible. They are also essential for the teams and even for our personal happiness. When I was at medical school, no time at all was devoted to communication skills. They are hugely needed.
Most patients, maybe by virtue of survival instinct, usually have a certain motivation to do something about their disease. Only a few say, from the start, that they are not interested in doing anything. But it’s true that there are some.
How do we give hope where there is apparently none? Actually, hope can be about other things. What we do is to focus hope on secondary, achievable objectives, and not on the main one. The patient will have ample time to digest that one – survival – and to feel whether he will achieve it or not.
But if, for example, we focus on an objective such as telling them that what we are going to do will improve shortness of breath and fatigue from effort, those can be achievable objectives – and they are legitimate, fair, adequate and true. That can shift the focus from the future to something that is immediately tangible and give the patient greater motivation to face whatever comes next.
Understanding what the patient wants requires several moments; it rarely happens the first time we see the patient. And during those various moments, we need to ask open questions so that the patient can express what he or she wants. If we repeat those questions at different times and get the same answer every time, we can then understand the patient’s wishes – knowing all along that the patient might change his or her mind at any time, just because. Without understanding why.
It’s still there. In the media, they always say “prolonged illness”, not cancer.
I think they almost always do. Even if they don’t say it out loud, that thought is always there.
True. Some cancers have a much better prognosis than lung cancer. That definition of cure I quoted, of dying without a tumour, is becoming more and more reachable. But the word “cancer” continues to be associated with death and suffering. And not only for patients, but also for their families. A cancer diagnosis is not just for one person, it is for a whole family.