In my last post I covered much of the controversy about whether patients with stage IIIA, N2-node positive NSCLC should be treated with induction therapy (chemotherapy or chemo/radiation) followed by surgery, or an alternative approach of chemo along with radiation delivered at a definitive dose (curative, not just the supplemental, lower doses used in induction therapy). In fact, while there are some flaws in the analysis that suggests that patients who require a lobectomy do better with surgery vs. without it, I really didn't disagree with Dr. Swisher, the surgeon who provided the pro-surgical view, that a subset of stage IIIA patients are particularly well served by receiving the most aggressive approach, including surgery.

My main counterpoints were these:

1) The patients viewed as most likely to benefit from surgery are currently the ones who respond very well to induction chemoradiation or chemo, specifically the ones who clear their mediastinal lymph nodes. I've described this issue in a prior post, and some of the highlight slides/figures are reproduced here:

(click to enlarge; mediastinal clearance after chemo alone was as important as whether the tumor was completely resected in predicting overall survival)

(in another trial of induction chemoradiotherapy, survival was remarkably better for the patients who had mediastinal sterilization before surgery).

So we consistently see that patients are likely to do much better if they demonstrate a good response to their initial non-surgical treatment. But why would we automatically say that these people should then have surgery, if they've shown that they're doing extremely well with chemo, with or without radiation? These are exactly the patients with responsive cancers who are likely to do very well without surgery, and perhaps surgery is adding more side effects and risk than benefit for these people, who seem to have their favorable prognosis conferred before they go to surgery.

However, I also recognize that more people want surgery if they can get it. In the poll I've been conducting here, about 2/3 of the respondents have said that they'd want surgery if a surgical and non-surgical approach were both available and appeared to provide similar outcomes. I'm pretty confident that people outside of this website feel the same way.

But there are other points to make.

2) Because the patients who are put on surgical trials have less disease and are healthier than the patients put on non-surgical trials, the only good way to compare these strategies is to directly compare them in the same trial with the same population. We already talked about the Intergroup 0139 trial that I described in detail previously, which, at the end of the day, didn't show an overall survival difference favoring surgery. I have also described in another prior post the results of a trial from Europe that showed that chemo followed by surgery didn't do any better than the folks who received chemo followed by radiation. Moreover, the non-surgical group on the European trial had the disadvantage of receiving sequential rather than concurrent chemo and radiation (consistently shown to be associated with modestly lower long-term survival), and on 40% of the patients the surgical arm also received post-operative radiation, so for nearly half of the surgery group, it wasn't a test of either/or, but radiation vs. surgery AND radiation for local control. So the head to head trials of surgery vs. no surgery fail to show that the overall population does better with surgery.

3) The most centrial issue is that while there is a subset of patients who probably can tolerate and do very well with surgery along with chemo and radiation, they are clear minority, and most surgeons cannot reliably limit their surgical plans to just those appropriate patients. In fact, the study showing the heterogeneity of stage IIIA patients treated surgically actually showed that only 1/3 of them were in the group who were likely to do well:

Moreover, these were patients who were analyzed after already undergoing surgery, so if we were to look at the whole population of stage IIIA patients, including those who didn't go to surgery, the subset well-suited for surgery would be smaller.

On the intergroup trial that was designed for surgical patients, 1/4 of the patients assigned to surgery couldn't get it or had an inadequate surgery. On the European trial, an astounding 47% of patients had the more dangerous pneumonectomy, and as in the US-based trial, those getting a pneumonectomy did far worse than those getting a lobectomy. Moreover, half of the patients on the European trial had a resection where there were positive surgical margins or visible tumor left behind. This means that the surgeons put patients through a pretty much useless surgery half of the time.

In truth, surgeons can't reliably predict who will need a pneumonectomy and who won't. On the Intergroup 0139, about half of the pneumonectomies were done in patients who actually had no evidence of residual disease after induction chemo and radiation (this fact hasn't been published but has been relayed to me directly by one of the surgical leaders of the trial). So a life-endangering surgery was performed on people with no viable tumor after non-surgical therapy. Why did the surgeon need to remove a whole lung? I'll mention that it takes less skill to do a pneumonectomy than to do a lobectomy, and that in a prior post I reviewed results showing that the experience and skill of the surgeon matters a lot, generally associated with lower pneumonectomy rates (sometimes required but also sometimes more than needed).

THE TAKE HOME MESSAGE

This is a lot of information, and in fact I'm skipping over some of the complexities because it would become excruciating (or more excruciating, if you're already well beyond your limit now), but the main point I would make is that induction chemo or chemoradiation followed by surgery is an aggressive option that is probably the best choice for the folks who are quite fit and who have a very limited amount of mediastinal lymph node involvement (ideally one node or nodal area, and not enlarged more than about 1.5 - 2 cm). But that's not the majority of patients, and it's actually closer to 20-30% of the patients with what now fits into stage IIIA N2 NSCLC. The problem is the other 70-80%, since maybe half of those patients are clearly not good candidates for surgery due to bulk and/or extent of mediastinal disease, or lack of adequate health reserves to undergo such aggressive treatment. But then the other 35-40% or so are likely not well served by surgery, but most prefer it, and too many surgeons doing lung resections don't have the level of specialized thoracic training, and are performing too many surgeries that shouldn't be done. These people would be recommended by the more expert surgeons to not pursue surgery, but when they go to surgeons who don't acknowledge their limitations and are more inclined to push the limits of appropriate resection because their job is to perform surgery, this leads to ill-advised and unhelpful surgeries for cancer that could have arguably been better treated by saving the patient's strength for chemo/radiation.

There's an old adage "Never ask a barber if you need a haircut". That applies to all sorts of settings, including surgeons, who originally started centuries ago as a profession of barber surgeons, after all. My main point would be that the best surgeons can distinguish between the patients who should and should not have surgery, but too many are not making that distinction and performing ill-advised surgeries on patients who are all too eager to have a resection, even when it's not the best option for them.