One of the issues we struggle the most with, as oncologists, patients, and families, is how to choose a therapy that won’t make someone feel worse. There are so many things to factor into these decisions: what is the baseline function of the person, what comorbidities (other chronic illnesses) might interact or interfere, what side effects are acceptable or worth the risk, to what degree is the cancer interfering with their functioning and can this be reversed with chemo, and of course what does any individual patient want and expect from chemo? Not to mention the choices between a few different agents or combinations of agents!
The issue of “performance status” (PS) is fundamentally important in lung cancer, especially in non-small cell lung cancer (NSCLC). There are different scales we use to measure this, which have been posted and discussed previously. Most publications use the Eastern Cooperative Oncology Group (ECOG) scale – partly because it’s rated 0 – 4 which is easy to remember. Zero is normal, and 4 is very debilitated, essentially bed-bound. Perhaps it is somewhat surprising that despite all of the fancy molecular markers and blood tests we have, a person’s functional impairment of PS is one of the most important predictors of how “bad” a cancer is. The other two factors are stage (how much cancer is there?) and degree of weight loss. It is well described, and there are other posts on the site elaborating on this, that if someone is very sick either from the cancer or from other concurrent illnesses, standard chemotherapy will only make things worse. But there’s an “in between” group, with a performance status of 2, who may not be as well served by the more aggressive standard approaches but still may benefit from cancer treatment.
The great hope of the next generation of “targeted” therapy is that oncologists will be able to more accurately predict who will benefit from which drug and that the newer drugs will be less toxic and thus perhaps even better for people who are too sick to tolerate our current forms of chemotherapy.
So – it was with considerable surprise, quite frankly, when the results of Dr. Rogerio Lilenbaum’s phase II study of the epidermal growth factor receptor (EGFR) inhibitor tarceva versus he standard chemo combination carboplatin-taxol in people with a poor PS, was presented at ASCO last year and it was found that the erlotinib arm did considerably worse. The full and peer-reviewed results were recently published in the Journal of Clinical Oncology (abstract here) thus prompting me to write this post.
The study design was very straightforward: 103 patients with advanced NSCLC and ECOG PS of 2 were randomly assigned to tarceva 150 mg orally daily until progression or to carboplatin and taxol IV every three weeks (standard doses) for up to four cycles. Patients who experienced progression or did not tolerate or refused further chemotherapy were allowed to cross over to erlotinib (29 patients, 57% did cross over). Of note, PS 2 basically means that people are up and about for more than 50% of their waking hours and able to look after themselves but are unable to work.
Between 2002 – 2004, the enrolled patients were put on the study at 14 different centers. There were no major clinical differences between the 2 arms (chemo or tarceva). A summary of the results demonstrates that cancers were more likely to shrink by 50% or more with carbo-taxol (though not much more likely: 12% versus 4%) and both treatments were associated with stable disease in about 40%. The toxicities were what you might expect from each type of treatment. What was somewhat surprising was that the progression free survival (the time it takes for a cancer to grow from the start of treatment) was longer in the carbo-taxol arm: 3.5 months versus 1.9 months. And the median survival was also better in the carbo-taxol arm: 9.5 months with chemo versus 6.6 months with tarceva, a pretty big difference:
(Click on image to enlarge)
In addition, although there were some chemo-related toxicities like numbness/tingling with the taxol – the quality of life scores filled out by the patients were about the same if not slightly better on the carbo-taxol arm.
So – what is my take on this? I think that there are some things we have to be careful of with this study and things that I learned from it. First, it was not statistically powered to really compare the arms to each other; it is more of a “pick the winner in a race” design. But of course we all tend to compare them. Perhaps the newer generation of chemotherapies like tarceva and others are less toxic but if they don’t “work” as well, then the quality of life on carbo-taxol might be better or at least the same, i.e. maybe chemo isn’t so bad after all. This reinforces the previous finding that if you start out with a poor PS, it doesn’t bode well for the disease — so don’t delay too long if you are going to try chemotherapy. And finally – these new drugs need to be given to more selected patient population, and we’re going to need to do more research to find out who these people are. Until then, just because it’s less toxic doesn’t mean it’s better.
posted by Dr. West @ 8:18 pm link to this post
Loading ...
March 13th, 2008 at 3:35 am
Dr. Laskin,
Thanks very much for writing this very interesting piece.
bev
March 13th, 2008 at 10:44 pm
Was it known if any in the study had the EGFR mutation? or if they were non smokers and female?
March 14th, 2008 at 11:28 am
I am very surprised at these results. I only have my own personal experiences to compare. In March of ‘07 when I was first diagnosed with stage IV NSCLC, the first oncologist (local) I met with me told me my type of cancer does not respond well to standard chemo. I was told to go and get my affairs in order and enjoy the final ten months or so of my life.
When I traveled to Boston, I was told I was a prime candidate for a clinical trial with Tarceva. I started the drug on 3-26-07 and within two weeks my cough had disappeared. The Tarceva destroyed many cancer cells and I was very fortunate to have few side effects. One tumor in my right lung showed an increase in January 08 so I am now off Tarceva and have been on the xl 647. My cough had returned during the month of February. After being on the xl-647 for twenty days, my cough has disappeared again. My blood work looks fine, I am feeling well and hardly any side effects. My next scans are April 16th- I hope this new drug shows a decrease on that tumor.
I’m glad I had Tarceva as a first line of treatment because the quality of my life as been great this past year. I don’t think I would have been able to continue to teach if I had chemo and radiation. ( I have had one-half day out for not feeling well this past year.) Even though the cancer had spread to many areas of different bones, I never had any bone pain. I do have zometa every 28 days.
I tell everyone they should look into taking the Tarceva (if they are a non-smoker, female…) I know it doesn’t work for everyone but you know pretty quickly if it works. I would gladly do a commercial for this drug:-)
Sorry this is so lengthy-
Terri
March 16th, 2008 at 8:33 am
When I checked the abstract for how long patients were on Taxol and Carboplatin, I noticed it was four infusions. Was this done because so many people’s quality of life diminishes with the 5th and 6th infusion? Because my recovery after the lst and 2nd were relatively benigh, I think the difficulties with 3 and 4 were due to extenuating circumstances. But 5 and 6 were pretty rough and I’ve heard a number of people confirm theirs were as well.
The abstract also stated “Sex, histology, skin rash, and smoking history predicted outcome with erlotinib.” I’d be interested in more details about those factors.
March 16th, 2008 at 12:44 pm
Thanks for the comments and questions.
here are some of the details of the patient groups: There were slightly more females (56% v 45%) and adenocarcinoma with bronchoalveolar features (17% v 6%) in the erlotinib arm. However, there were more patients with nonadenocarcinoma histology in the erlotinib arm (50% v 38%). Smoking history was not different between the arms. 12% on the tarceva arm and 8% on the carbo-taxol arm were never smokers.
this paper looked at performace status as a factor in choosing chemotherapies. there are clearly some individuals who do very well with tarceva as a first line option, what we need to do is figure out who those people are, and that takes a lot of research.
the reason 4 cycles was given is because that is the standard number of cycles people receive. there is little evidence to suggest much benefit after 4 cycles, though if a cancer is still shrinking after the 4th cycle sometimes going to 6 is reasonable. not only are there studies demonstrating this, but in actual practice in 1000’s of patients on clinical trials 4 is the median number of cycles delivered. Usually the side effects start to outweigh any benefits by then, as you yourself point out, and particularly so in a population of people who are sicker (PS 2) to begin with.
March 21st, 2008 at 7:54 pm
Nomad,
I’m sorry we never answered your question about molecular profiles of patients, which was looked at in the small number of patients (19) who had tissue available to analyze. As luck would have it, all 5 of the EGFR mutations (associated with storng responses to EGFR inhibitors like tarceva) were in the patients who were assigned to chemo, and all three of the patients with ras mutations (associated with little or no likelihood of response to EGFR inhibitors) were assigned to the tarceva arm. While this would suggest that bad luck would have stacked the deck in favor of the chemo arm, the numbers with tissue are small enough that I don’t think we can say anything meaningful.
This situation highlights how an absence of tissue in many patients with advanced lung cancer is a major limitation in our understanding of the field. We need more tissue from patients so that we can do these types of molecular analyses. Historically, we have often obtained fine needle aspirations, with a very small needle, that collect a minimal number of cells. This is enough to often make a diagnosis (SCLC vs NSCLC, for instance) but not say what kind of NSCLC, or run molecular studies. I and other lung cancer experts feel that we need to get out of that habit and obtain more substantial tissue from a core biopsy. While it entails a bigger needle, the information from such tissue samples could very possible guide our better treatment of lung cancer patients once we get to the point of routinely tailoring therapy based on cancer subtype and molecular markers.
-Dr. West