After completing my unpacking from ASCO and beginning to see a path of light as I dig out from under the backup of work that accumulated, I wanted to begin covering a few of the most relevant content from ASCO. Though covering only preliminary work, an abstract presented by radiation oncologist Daniel Gomez from MD Anderson Cancer Center was among those that I think could have sweeping implications. But for now, I think it’s also important to provide some context so that the results are not misinterpreted as practice changing yet.
We know that patients with good initial disease control from metastatic lung cancer are most likely to demonstrate progression at sites where they already have residual disease. This led to the question of whether there is a value of “local therapy” (surgery or radiation to treat a specific area of disease) in patients with limited residual disease after at least 4 cycles of initial chemotherapy or, for patients with an EGFR mutation or ALK rearrangement, at least 3 months of targeted therapy. Patients could not have had a malignant pleural effusion (cancer in fluid outside of the lung, which is a common complication in lung cancer) and had to have no more than 3 sites of visible disease on scans. After first line therapy, patients with this definition of limited residual disease were eligible for randomization to one of two arms:
1) standard systemic therapy maintenance or surveillance, at the discretion of the treating physician, or
2) local consolidation therapy with any combination of radiation, surgery, or other ablative therapy followed by standard systemic therapy maintenance or surveillance, at the discretion of the treating physician.
For those patients assigned to the consolidation therapy arm, the trial allowed for any of a multitude of sites to be treated with any combination of local interventions, potentially radiation to some sites along with resection of others, or even chemo/radiation.
Importantly, patients assigned to the standard arm without local therapy were permitted to cross over to receive local therapy after progression. The “primary endpoint” being evaluated was progression-free survival (PFS), the time between randomization and imaging revealing significant progression.
The study was conducted at three large cancer centers, but all but six of the treated patients were enrolled from MD Anderson. Notably, they enrolled just 74 patients from these three centers over three years for starting first line systemic therapy, with 49 of them eligible for subsequent randomization. The trial was reported at ASCO after it was stopped early by MD Anderson’s Data Safety Monitoring Committee, which did an early analysis that revealed a significant improvement in PFS.
The difference was certainly impressive, with the median PFS (the time when half of the patients have shown progression, and the other half have not) of 11.9 months for the local consolidation therapy arm, compared with just 3.9 months in the arm assigned to standard systemic therapy maintenance therapy or surveillance. Subset analyses showed that patients with just one residual area of disease did significantly better than those with 2 or 3 areas of disease. In addition, the 8 patients with an EGFR mutation or ALK rearrangement had a longer PFS than the patients without a driver mutation.
While these are very intriguing results, we must be very cautious in interpreting them. As noted above, the final randomized trial population was fewer than 50 patients, which is only a tiny fraction of the patients with metastatic NSCLC who had been seen and treated at these three participating centers over a three year period. This suggests that the patients studied here are not broadly representative of most patients and are far more likely to represent a very unusual group that has an unusually small amount of disease. With all but six of the patients coming from MD Anderson, we must view these results as being close to a single center trial, which is always a red flag for “selection bias” (good results are likely to come from the enrolled patients not representing the broader patient population but a very cherry-picked population that isn’t like most others).
But the more important issue is the very problem of using PFS when the treatment itself undermines the ability to assess that endpoint because you’re resecting or radiating the disease that is most prone to progress. Once half of the patients have had their visible cancer resected or ablated, only the standard treatment arm without local therapy is now at risk for local progression. We wouldn’t be surprised to learn that prophylactic appendectomy significantly reduces the risk of subsequent appendicitis; we need to show that the intervention does more than just making the scans look better, just as if we had used Adobe Photoshop to erase the remaining lesions and just pretended they weren’t there.
Of course, seeing a survival benefit would be very powerful, but the results are not mature, with too many patients on both arms still alive, and the survival analysis compromised by the cross-over of patients on the standard systemic therapy arm to local therapy at the time of progression. But short of that, it would be helpful to see that local therapy changed a measure of the disease process beyond the areas directed treated with surgery or radiation. Toward that end, the investigators did an exploratory analysis that assessed the “time to new site failure” (TNSF) and found that the patients on the consolidation arm also had a significantly longer TNSF, 11.9 vs. 5.7 months (P<0.0497).
Though the trial enrolls far too small a population to say anything definitive and the choice of PFS as the primary endpoint was a major shortcoming, the finding of a significantly longer TNSF with local therapy provides a very important proof of principle if it holds up. Beyond the unsurprising finding that surgery and radiation can make scans look nicer for a while and lead patients and physicians to congratulate ourselves about that, the exploratory analysis of TNSF, which is not a definitive answer to the question, indicates that treating areas of “oligo-residual disease” (areas of disease when there are just one or a very few sites of visible disease after first line systemic therapy) may kill enough cancer to significantly postpone development of diffuse disease, potentially even improving overall survival in the process.
For now, it is critical to remember that this was a very small population of patients winnowed from a much larger one, highlighting that this concept doesn’t apply to patients with disease progression, more than 3 areas of residual disease, or a pleural effusion. Even within the tested population, the results showed that survival was far better for the subset of enrolled patients with just one residual spot. In other words, patients with more than even just one area of remaining disease may not have “oligo-metastatic” or “oligo-residual” disease (“oligo” meaning few), but actually conventional “poly-metastatic” disease (“poly” meaning many), for which systemic therapy may well be the only modality that offers a benefit.
This work deserves to be followed up by a large trial across North America, as is currently being developed, so that we can clarify if it is truly the case that local treatment can alter the disease trajectory of systemic disease and change outcome beyond just the appearance of the scan. Importantly, however, if it is too broadly inclusive and allows patients with more than 1-2 areas of residual disease, it will have a far greater chance of undermining its ability to demonstrate a benefit because the benefit is much less likely to be real for all but a highly selected subgroup of patients with advanced NSCLC.