While PCI is a recommended component of aggressive multimodality treatment  for LD-SCLC, PCI does not at this time have any clear role in the treatment of NSCLC.  To be potentially valuable, the risk of brain metastases needs to be high enough to justify to potential side effects, time, and expense of PCI, and is likely to be of value only if the risk of cancer recurrence elsewhere is low enough to make risk of brain recurrence a limiting factor in survival.  With stage I and II disease, the risk of brain recurrence is not especially high compared with the risk of disease in the rest of the body, and in stage IV disease, control outside of the brain is usually a more pressing issue.  But in stage III, or locally advanced NSCLC, as we have gotten better at controlling cancer in the chest, dealing with the potential for cancer in the brain has become enough of a problem that PCI for stage III disease is now a very timely question.    How big an issue is recurrence in the brain after treatment for stage III NSCLC?  Many of the recent trials show recurrence rates in the brain first, or brain only, to be in the 25-40% rate.   For example, Harvey Mamon and colleagues from the Dana Farber Cancer Institute in Boston reported on their review of 177 patients with stage IIIA NSCLC treated with induction therapy (abstract here) and found that the risk of brain recurrence as first site of relapse was 34%, and 40% of patients had relapse in the brain at some point.  Patients with non-squamous lung cancer had a rate of brain recurrence of 53% (adenocarcinomas consistently have a higher risk of distant spread than squamous lung cancers).  In the multi-center setting, the Southwest Oncology Group performed a very influential trial called SWOG 9504 (abstract here) with 83 patients that led to the common adoption of concurrent chemoradiation followed by consolidation taxotere, in which there was a 29% long-term survival, but one third of the recurrences were in the brain first (and 62.5% of the recurrences outside of the chest).   Multiple trials have demonstrated the consistent result that as we have gotten better at treating the chest, the brain remains a sanctuary site at risk for distant failure.

   However, PCI can potentially have detrimental effects on cognitive function.  This has not been studied well in NSCLC, but a small study (abstract here) of PCI in NSCLC found changes in attention and visual memory, as well as significant changes on MRI scans of 2/9 patients who received PCI and none of 4 patients who didn’t receive PCI.   Other studies (Johnson abstract here; Komaki abstract here) have also demonstrated subtle neurocognitive impairments in long-term survivors of SCLC, with a suggestion of a higher risk in recipients of PCI, although it is debated whether the impairments may be due to other causes. 

   PCI has been studied in multiple small NSCLC trials, and in pretty much all of them it has shown a decrease in the likelihood of subsequent brain metastases, but no significant impact on survival.  Many of these trials have been done in an an era well before we were effectively controlling disease outside of the chest.  You can see from the following table that the later trial by Stuschke and colleagues, as the only one of those listed from a more ”modern” era, shows the new particularly high risk in the brain as we control disease elsewhere (patients were primarily failing in the rest of the body before this time):

 (click to enlarge)

  So right now we have a compelling reason for considering PCI, but in light of the lack of clear survival benefit and the potential for long-term adverse effects from PCI, I would not recommend it outside of a clinical trial.  However, I am VERY supportive of an important trial being run by the Radiation Therapy Oncology Group (RTOG) that is widely available throughout North America, known as RTOG 0214, that randomizes patients with stage III NSCLC who have received all of their planned treatment, potentially including surgery or just aggressive chemo and radiation, to be closely observed or receive PCI with the same approach used in the Stuschke trial (30 Gy over 3 weeks):

 (click to enlarge)

All patients receive regular MRI scans and cognitive testing to detect any detrimental neurological effects of PCI. 

    We have found it challenging to run trials designed as “treatment vs. no treatment” in the US, because patients may have an idea of what they want, and/or doctors may have their own idea of what approach they would prefer.  Some may see the rationale for PCI and pursue it, but we’ll never learn the true value, carefully assessing both the benefits and the risks, of PCI in NSCLC if we don’t complete the studies designed to answer the key questions.

   This is a controversial topic, so feel free to provide your own perspective, questions, comments, and objections.