There are many open questions in managing lung cancer, but one of our historical areas that has been especially challenging has been locally advanced/stage III NSCLC, which we most commonly treat with at least two different forms of therapy, such as chemotherapy followed by surgery, chemo and radiation followed by surgery, or (most commonly) chemotherapy and radiation without surgery. Why is it such a controversial area?

A key issue is the heterogeneity of both the cancer and patients who are part of this population, about 1/3 of all NSCLC patients, who are initially diagnosed with stage III disease.  A patient can have locally advanced NSCLC with just microscopic involvement of their cancer in a single mediastinal (mid-chest) lymph node that isn't enlarged on scans, while others can have many bulky, enlarged nodes on both sides of the mid-chest, or lymph node involvement above the collarbone in the lower neck, or a large tumor that may be growing into important chest structures like the trachea, heart, or aorta.  Stage III NSCLC can be caused by an advanced tumor, extensive regional lymph node involvement, or a combination of these. 

Not surprisingly, patients with very minimal disease burden have a much better prognosis than patients with larger and/or more extensive tumor burden.  Patients with a single non-enlarged mediastinal lymph node involved with cancer can be cured with surgery +/- chemotherapy about 1/3 of the time, but the prognosis is much less favorable in patients with several lymph nodes involved or those with abnormally enlarged lymph nodes.  Yet when we do studies of stage III NSCLC, we pool many different patients and cancers that start with a very different prognosis.

We also know that some cancers are far more sensitive to our treatments than others. Unfortunately, we only learn how sensitive or resistant a patient's cancer is by looking in hindsight. In 2015, we aren't yet able to predict which cancers will be far more or less responsive to chemotherapy and/or radiation.

Another critical variable is the underlying health of the patient.  Some are very fit, have a good reserve of lung function, and can tolerate rigorous treatment very well. On the other end of the spectrum, we also have many patients who have compromised lung function from chronic smoking, emphysema, and other issues, who may be frail from other medical conditions, and are not likely to tolerate very aggressive treatments without developing debilitating side effects or even possibly die from the treatment. In most trials of multimodality treatment of stage III NSCLC, about 5% of patients die from "treatment-related mortality", meaning that they died from the treatment rather than the disease.

This leaves us with the difficult situation of a cancer that may require a lot of intensive treatment to cure the cancer, often just at the upper limit of what a patient can tolerate, while other cancers may require more intensive treatment than it's possible to safely deliver to that patient without the treatment causing more harm than benefit as the treatment becomes increasingly intensive. 

For different patients, the intensity of what is required vs. what is tolerable will vary.  In many cases, it isn't possible to deliver a treatment intensity sufficient to eradicate the cancer without going beyond the point of safety. That may be because the cancer is more resistant to treatment (and a higher level of intensity is required), because a patient is frail and can't tolerate a standard level of treatment intensity (so the patient is still harmed before getting into the range), or both of these.

Today, we have no evidence that more intensive treatment is better. Many clinical trials have added chemotherapy or targeted therapy after the standard 6-7 weeks of concurrent chemotherapy and chest irradiation, and the evidence has shown that escalating the chemo adds side effects but doesn't improve overall survival. We have evidence that escalating the radiation dose is associated with worse survival.

In this last figure, I've been uncharitable in having no overlap of the green bar with the yellow in the second condition of a fit patient with a large tumor burden or more resistant cancer. This implies that there is not a point where more intensive treatment (compared with the typical standard treatment) is sufficient to eradicate the cancer but still be within the range of what is tolerable for a patient.  Instead, let's look at the figure below and envision a situation in which the treatment intensity to eradicate cancer is higher than the best case but still overlaps with the yellow of tolerability but falls short of the red, "danger" zone.

Here, maybe if we give more treatment, perhaps we can cure some who wouldn't be cured with standard treatment.  

If we acknowledge that our population of patients with stage III NSCLC include a wide range of patients, some of whom fit, some much less able to tolerate rigorous treatment, some with more sensitive disease, some with resistant disease, some with minimal disease that still falls just into the range of stage III, and others with very bulky disease and a large tumor burden that barely fits within the confines of stage III disease, we might expect that these populations aren't likely to all be optimally treated the exact same way.  And I think that's why it's been so hard to show that anything beyond the initial 7 weeks of chemotherapy and chest irradiation improves survival. For some patients, it's very helpful, and perhaps the critical difference that might cure then. For others, however, it provides little or no benefit and may escalate the intensity of their treatment into a range that harms rather than helps them. And when you pool these many patients into the same trial results, you end up with a wash.

But just as we've seen when looking at groups and then identify important molecularly-based subgroups, I believe the optimal therapy may well be different for these stage III subgroups -- though we don't have evidence to speak to that speculation.  Today, many lung cancer specialists and general oncologists alike feel that we just haven't studied the question well enough to conclusively say that stage III patients can't/won't benefit from additional treatment before or, more commonly, after the initial chemo/radiation, and we may be inclined to recommend additional treatment for some of our patients.  

My strategy is to be selective in how I approach patients with locally advanced NSCLC, favoring consolidation chemotherapy for the patients who I feel are likely to tolerate a more rigorous treatment, especially if they have unfavorable features of their cancer (larger tumor, poorly differentiated, etc.), but the absence of any proven benefit for the "more is better" concept makes me very comfortable stopping after completing the core 7 weeks of chemo/radiation for patients who I suspect are more likely to be harmed than helped by additional treatment without evidence of a survival benefit.

Overall, then, even though we've done studies for decades trying to identify the best way to treat locally advanced NSCLC, it's still a setting in which there's plenty of room for judgment and interpreting how a person is doing over the course of treatment.