Several members here have asked questions on the discussion/Q&A forum over the last several months about the potential value of continuing treatment, usually tarceva, in the setting of mild progression.  Our general rule in oncology is that if someone shows that a tumor is growing (in trials, generally defined as any new cancer lesions or a 25% increase in the size of measured lesions being followed for response) you stop that treatment and consider it ineffective. In actual clinical practice, we don’t usually make precise measurements and usually just discontinue a chemo regimen or other strategy any time the progression is convincing, not looking for a 25% increase.  But there are a few exceptions that we make routinely in oncology.  Patients with breast cancer who have tumors that overexpress HER2/neu, the protein target of the antibody Herceptin, usually continue on Herceptin from one line of chemo to another and another when they demonstrate progression on a chemo/Herceptin combination.  Men with prostate cancer who are on lupron, a hormone therapy that blocks testosterone production, are generally continued on lupron to reduce the rate of progression even after their PSA is consistently rising on lupron alone, generally adding new treatments to the lupron.  And patients on Gleevec for GI stromal tumors may respond initially to one dose, then become resistant, and again respond to a higher dose level of the same agent.  So there are exceptions to the rules, but we generally haven’t operated that way in lung cancer. 

   In addition to a presumed ongoing clinical benefit from continuing the treatment, these therapies need to be tolerable enough that people can continue them on a more chronic basis than we usually use for chemo, or at least combination chemotherapy.  Most people find platinum-based doublets (especially with cisplatin) challenging enough for a few months that the idea of continuing indefinitely is infeasible.  Many of our agents also have cumulative toxicities with ongoing treatment.  Cisplatin can lead to neuropathy (nerve damage, usually to the longest nerves of the body, going from the spinal cord to the tips of the fingers and toes, also known as a stocking-glove distribution), kidney damage, and hearing loss that all tend to worsen with cumulative treatment, and the taxanes also often lead to cumulative neuropathy.  The vast majority of our chemo drugs lead to drops in blood counts that can become more and more of a problem as the bone marrow becomes increasingly unable to bounce back from one cycle after another.

   But we do have a few agents that may be amenable to longer term use.  The EGFR inhibitors such as iressa and tarceva have been on a short list of anti-cancer agents that are often (but not always) well tolerated, and there are a fortunate minority who continue to show no progression for a year or more.  Almost inevitably, however, patients have become resistant to these agents.  So should we just abandon them in a patient who had a dramatic and brisk partial response to iressa or tarceva , tolerated it well for a 18 months, then demonstrates mild, asymptomatic progression with a new small lesion and slight re-growth of the primary cancer? 

   The thoracic oncology team at Memorial Sloan-Kettering Cancer Center (MSKCC) in New York noted, in fact, that some of these patients appeared to worsen quickly in the first few weeks after stopping the EGFR inhibitor that had served them well for a long while.  Rather than just leave it at that, they wrote a protocol that allowed them to characterize this phenomenon better (abstract here).  Enrolling patients who either had a documented objective response (meeting trial criteria for a 50% or greater reduction in tumor volume) or had an EGFR mutation, they had patients who showed progression undergo a CT and PET/CT fusion scan at the time of stopping iressa or tarceva, then again three weeks later at the time the investigators had them restart their EGFR inhibitor, and then one final time three weeks after restarting:

 (Click to enlarge)

They also assessed the patients in terms of how they felt, whether their cancer-related symptoms worsened.  With a total of just 13 patients, of whom only 10 completed the whole follow-up protocol with visits and scans, this isn’t the definitive work on the subject.  It’s a start of the dialogue, however, and it’s the best information we’ve got for an issue that many people here are very interested in.

   What they found was that the results were quite variable, but many patients had a perceived worsening of cancer-related symptoms when they came off of iressa or tarceva, with several improving after restarting:

And similar to the variability seen with symptoms, they saw that measurements of the tumors with repeat CT scans, or PET scan metabolic activity as measured by the maximum standard uptake value (SUV) also worsened suddenly after stopping treatment and then in some cases responded again, or at least stabilized, upon restarting the EGFR inhibitor.   In the following figures, each line represents one patient over time, with a line moving upward indicating progression, while downward movement from left to right signifies tumor shrinkage (and level lines show stable results):

You can see from the figures that there is no clear pattern, but there are clearly patients who appear to be doing far better when they’re on their EGFR inhibitor than when they’re off it; faulty brakes are better than no brakes. 

   When the folks at MSKCC looked at the medians (half of the results above the median number, and half below it), trying to make rhyme or reason from these varied results, they found that the median change in tumor diameter was +9% (i.e., worse) after stopping and -1% (i.e., slight shrinkage) after restarting EGFR TKIs; the median change in tumor volume was +50% after stopping and -1% after restarting EGFR TKIs; and the median change in maximum SUV was +23% after stopping and -11% after restarting EGFR TKIs.  The numbers of patients are small and don’t leave me with any firm conclusions except that some patients appear to do better staying on EGFR TKIs after progression.  Even then, I don’t think this is likely to continue forever, but rather may be a gradual transition from relative to complete resistance.

   There are also ideas about increasing dose or adding another agent to the EGFR TKI to restore activity.  Agents to potentially be combined would include concurrent chemo (an idea I’m not a fan of, except perhaps using pharmacodynamic separation of the EGFR inhibitor and the chemo), the monoclonal antibody against EGFR erbitux (cetuximab), cyclooxygenase-2 (COX-2) inhibitors like celebrex (celecoxib), avastin (bevacizumab), and various other targeted therapies.  But none of these has been shown to be better than doing just a new agent after stopping iressa or tarceva.   I don’t think anyone can really suggest a “best” option here for combinations, or whether it would just be better to stop the EGFR inhibitor and start something new.  Don’t forget that the people on the MSKCC trial were being followed off of these agents but not on anything else, just a “washout”/break after their iressa or tarceva.  The results might have looked different had everyone switched immediately to alimta or some other reasonable alternative.
 
    Putting all of this together, I’d say that there are no rules here, but that also means that our old rule about stopping treatment on which someone has progressed doesn’t necessarily apply.  I tend to consider the factors of how long a patient has shown benefit, how well they’re tolerating this treatment, and what alternatives remain after moving off of the current treatment.  This issue of continuing after mild progression can also apply to other therapies than just EGFR TKIs, such as alimta or other chemo agents if a patient isn’t experiencing cumulative side effects (I have had patients continue to have nearly stable disease for up to 6-8 months on second or third line chemo after demonstrating mild progression).   For a patient who is progressing on a toxic first-line doublet, there are several other options remaining, and I’d have a lower threshold to stop the first treatment and try an alternative approach that may be better tolerated and possibly more effective.  On the other hand, in someone with mild progression on well tolerated third or fourth line tarceva or single-agent chemo, I don’t have anything terrific to offer as an alternative, so I’d be inclined to squeeze every bit of benefit possible from this salvage treatment.

    While this moves away from any firm guidelines about best practices for treating advanced NSCLC, I hope it provides an understandable framework to justify why I think it’s worth individualizing my treatment recommendations more and more for patients as they move through the limited array of treatment options we have to offer.  Sometimes that involves re-evaluating some of the dogmas about oncology treatment, including the concept that an agent on which a patient has shown some progression is no longer of value to that person.