Why this matters
SRS has replaced whole-brain radiotherapy for limited brain metastases from most solid tumors because it preserves cognition and quality of life while maintaining effective intracranial treatment.
Small-cell lung cancer has remained an important exception.
Patients with SCLC were excluded from the landmark randomized SRS trials, and concerns about rapid intracranial dissemination have kept whole-brain radiotherapy as the traditional standard.
NRG-CC009 is the first phase III randomized trial directly comparing SRS with a modern whole-brain strategy, hippocampal-avoidant WBRT plus memantine, in patients with SCLC brain metastases.
The primary cognitive endpoint was negative. But the trial produced an unexpected and potentially important survival signal favoring SRS.
Study design
NRG-CC009 randomized patients with brain metastases from SCLC to:
- SRS
- hippocampal-avoidant WBRT plus memantine
Eligible brain metastases had to be at least 5 mm from the hippocampi, no larger than 4 cm individually, and have a total intracranial metastatic volume of no more than 30 cm3.
The protocol initially limited the number of metastases to 10, but this restriction was later removed. The median number of brain metastases was 2, with a range of 1-26.
Patients were stratified by:
- DS-GPA
- number of brain metastases
The primary endpoint was time to neurocognitive failure.
Neurocognitive testing included:
- Hopkins Verbal Learning Test-Revised
- Controlled Oral Word Association
- Trail Making Test
Secondary endpoints included:
- overall survival
- intracranial disease progression
- neurologic mortality
- salvage therapy
- treatment-related adverse events
A total of 151 patients were randomized.
Median follow-up for surviving patients was 9.5 months.
Key results
The primary endpoint was negative.
There was no significant difference in neurocognitive failure between treatment arms.
The 6-month neurocognitive failure rate was approximately:
- 61.7% with SRS
- 59.2% with HA-WBRT plus memantine
Gray's test:
- p=0.66
Adjusted analysis similarly showed no significant difference:
- HR 0.83
- 95% CI 0.53-1.28
- p=0.39
Therefore, SRS did not demonstrate superior neurocognitive preservation compared with modern HA-WBRT plus memantine.
Overall survival
The major unexpected result was overall survival.
Median OS was:
- 17.4 months with SRS
- 8.6 months with HA-WBRT plus memantine
The difference was statistically significant.
Adjusted analysis also favored SRS:
- HR 0.60
- 95% CI 0.39-0.91
- p=0.016
The magnitude of the survival difference is substantial and persisted after adjustment for baseline prognostic factors.
However, OS was a secondary endpoint and should therefore be interpreted with greater caution than the primary neurocognitive result.
Intracranial disease control
The intracranial control story is more nuanced.
Crude cumulative incidence analysis showed more intracranial failure after SRS.
At 12 months:
- 27.9% with SRS
- 12.6% with HA-WBRT plus memantine
However, after cause-specific multivariable adjustment, the difference in intracranial failure was no longer statistically significant.
Local brain failure also did not differ significantly between treatment groups.
At 1 year:
- 15.3% with SRS
- 12.6% with HA-WBRT plus memantine
- p=0.13
There was also no significant difference in leptomeningeal disease.
Neurologic mortality
Neurologic mortality was not significantly different between the two strategies.
At 1 year:
- 4.6% with SRS
- 9.4% with HA-WBRT plus memantine
At 2 years:
- 11.5% with SRS
- 20.6% with HA-WBRT plus memantine
The difference was not statistically significant.
This is important because one of the main concerns with using SRS in SCLC has been the possibility that untreated microscopic intracranial disease could lead to more neurologic deaths.
That signal was not observed.
Toxicity
Treatment-related adverse events were numerically more common with HA-WBRT plus memantine.
No treatment-related adverse events were reported in:
- 60.9% with SRS
- 38.8% with HA-WBRT plus memantine
Grade 3-4 adverse events occurred in:
- 8.7% with SRS
- 14.9% with HA-WBRT plus memantine
The difference was not statistically significant.
Interpretation
NRG-CC009 changes the level of evidence supporting SRS for SCLC brain metastases.
The trial did not show what might have been expected: better cognition with SRS.
Modern HA-WBRT plus memantine performed well enough that the primary neurocognitive endpoint was similar between the two arms.
But the survival result shifts the clinical interpretation.
Patients randomized to SRS had a median survival of 17.4 months compared with 8.6 months after HA-WBRT plus memantine.
That difference is striking.
At the same time, it should not be overinterpreted. Overall survival was a secondary endpoint, and a local CNS treatment strategy producing such a large survival difference is biologically difficult to explain from intracranial control alone.
In fact, HA-WBRT produced better crude intracranial control, yet this did not translate into lower neurologic mortality or better survival.
The mechanism behind the OS difference therefore remains uncertain.
What the randomized data do establish more confidently is that SRS did not result in worse neurocognitive outcomes, adjusted intracranial control, neurologic mortality, or severe toxicity.
Taken together with the survival signal, these findings strongly support SRS as a standard treatment option for appropriately selected patients with SCLC brain metastases.
Limitations
The primary neurocognitive endpoint was negative.
Overall survival was a secondary endpoint, and the magnitude of the observed difference requires cautious interpretation.
Median follow-up among surviving patients was relatively short at 9.5 months.
The trial included patients with a broad range in the number of brain metastases, although total intracranial disease volume was limited.
Crude intracranial failure favored HA-WBRT, even though the difference was not significant after multivariable adjustment.
Longer-term quality-of-life and salvage-treatment analyses are still important for understanding the full clinical tradeoff between the two strategies.