Prophylactic cranial irradiation, or PCI, for SCLC, usually limited disease (LD-SCLC), remains a controversial issue, although this is generally recommended for patients with LD-SCLC who have a complete response to treatment (no evidence of disease). However, the idea of radiating the brain of someone who has no evidence of cancer there and may never get it is something that many patients and also some oncologists (radiation oncologists and medical oncologists) may not embrace. So how did we get to a point where we standardly recommend radiation to prevent brain metastases from developing?

Well, as I mentioned in previous posts, the brain is remarkably fertile soil for brain metastases for SCLC, which has a consistent propensity to spread there. In some studies, up to two thirds of patients with SCLC who don't receive PCI develop brain metastases within two years. Several small trials in the 1970s and 1980s consistently showed a reduced risk of developing brain metastases but no clear improvement in survival from PCI, although these trials were really too small to show any significant benefits.

Two larger trials tested PCI in greater detail. The first, by Arriagada and colleagues (abstract here), randomized 300 patients with SCLC (163 with LD-SCLC) to PCI or no PCI. The PCI dose and schedule used were less than optimal by current standards (24 Gray over only 8 fractions, vs. a more common recommendation today for something like 30 Gy over 15 fractions). There was a significant reduction in the risk of developing brain metastases (45% vs. 19% at two years, after which only 2% of patients had them appear as a first recurrence). Importantly, there were NO significant differences in cognitive/neuropsychiatric function, which was built into the study design. There was a trend toward a better two-year survival in the group that received PCI (29% vs. 22%).

The second trial, by Gregor and colleagues (abstract here), attempted to test two different dose schedules (36 Gray over 18 fractions, or 24 Gray over 12 fractions) of PCI vs. no radiation. The follow-up at two years revealed that the risk of disease in the brain was nearly halved with PCI (52% vs. 29%, p < 0.0002). The group that received 24 Gray in just 12 fractions had much less benefit than the patients who received the higher dose. As in the other trial, detailed neurocognitive testing showed no significant decline in the patients who received PCI. There was a modest but not significant improvement in overall survival among the folks who received PCI, but this trial was also too small to detect a difference in survival of 5-7%.

With so many smaller trials showing the same trend of benefit that is not statistically significant, investigators pooled the results of these trials into a meta-analysis of 7 trials of PCI, including a total of 987 patients with SCLC (primarily LD-SCLC) who had a complete response to therapy. In keeping with the results from the individual trials, the overall survival was now significantly improved by about 5% in absolute terms, from 15.3% to over 20.7%. The disease-free survival was also significantly improved, and the risk of developing brain metastases and complications from them was about halved with PCI. Higher total doses were more effective than lower doses of PCI. Overall results are summarized in the slide below.

(click to enlarge)

So while the strength of the evidence is not overwhelming, it has consistently been shown that PCI very signficantly reduces risk of subsequent brain metastases and appears to confer a modest improvement in overall cure rate. Keep in mind, however, that the 5% improvement in survival with PCI is the same degree of cure rate improvement seen from adding chest radiation to chemotherapy for LD-SCLC, which is now pretty much accepted as a clear standard of care. Nevertheless, there is a real risk of at least some degree of neurologic compromise, which is higher if PCI is given at the same time as chemo (it is recommended that PCI NOT be given with concurrent chemotherapy), and by giving higher doses per fraction (treatment session) over fewer sessions. So giving a lower dose per fraction over a longer period of time maximizes the long-term safety. A current international trial being conducted now randomizes patients with a complete response after treatment to PCI with any of three different schedules, seeking the best combination of disease control/survival and minimal side effects:

(click to enlarge)

PCI is still an important and individualized decision, and I would not recommend it for patients with marginal performance status (activity level and ability to remain independent), significant other medical problems, or impaired mental functions from a stroke or any other cause. I would be cautious about recommending PCI in elderly patients (how old is elderly? you get different answers depending on how old the person answering is, but usually 65 or 70 is a turning point), because the risk of neurologic side effects is likely greater in older patients. And we often will discuss and cautiously recommend it for patients with extensive disease SCLC (ED-SCLC) who have a complete response, because the same trends seem to apply here as well, although this population has not been as well studied with PCI.

So it's not right for everybody, but PCI is really an appropriate thing for many patients who have a very good response to initial treatment for SCLC. THe favorable results have also led us to consider whether it may also be beneficial to incorporate it in treatment for locally advanced NSCLC, where we see a distressingly high frequency of recurrences in the brain as well. I'll discuss that topic next.