Background
Individual patient data meta-analysis (PCI Overview Collaborative Group). N=987 SCLC patients in complete remission from 7 randomized trials of prophylactic cranial irradiation (PCI) vs no PCI. Limited or extensive stage, CR after chemotherapy; included trials used various PCI doses and schedules.
Interventions and follow up
Arm A: Prophylactic cranial irradiation (PCI; various doses/schedules across trials)
Arm B: No PCI
Primary endpoint: Overall survival; incidence of brain metastases
mFollow up: 3 yr from randomization
Arm B: No PCI
Primary endpoint: Overall survival; incidence of brain metastases
mFollow up: 3 yr from randomization
Results
OS relative risk of death (PCI vs no PCI): 0.84 (95% CI 0.73-0.97), P=.01
3-yr OS improvement: 5.4% (20.7% with PCI vs 15.3% without)
Disease-free survival: RR 0.75 (95% CI 0.65-0.86), P<.001
Brain metastasis RR: 0.46 (95% CI 0.38-0.57), P<.001 — 54% reduction
Dose-response (brain mets): Higher dose lower risk (P for trend=.02)
Timing trend: Earlier PCI after induction lower brain met risk (P=.01)
3-yr OS improvement: 5.4% (20.7% with PCI vs 15.3% without)
Disease-free survival: RR 0.75 (95% CI 0.65-0.86), P<.001
Brain metastasis RR: 0.46 (95% CI 0.38-0.57), P<.001 — 54% reduction
Dose-response (brain mets): Higher dose lower risk (P for trend=.02)
Timing trend: Earlier PCI after induction lower brain met risk (P=.01)
Adverse events
Reporting: Not the primary focus of the meta-analysis
Neurocognitive: Individual trials raised neurocognitive toxicity concerns with PCI, esp. at higher doses; not systematically analyzed as a meta-analysis endpoint
Neurocognitive: Individual trials raised neurocognitive toxicity concerns with PCI, esp. at higher doses; not systematically analyzed as a meta-analysis endpoint
Conclusions
PCI in SCLC patients achieving complete remission reduces brain metastasis risk by 54% and improves 3-yr OS by 5.4 percentage points. PCI is established as standard of care after CR in limited-stage SCLC and, at the time, was also considered for extensive-stage responders.
Key Limitations
Heterogeneous PCI doses/schedules across pooled trials; neurocognitive toxicity not systematically assessed; pre-MRI staging era (occult brain mets may be present at randomization); applicability to extensive-stage SCLC later questioned (Takahashi 2017 showed no OS benefit with baseline MRI surveillance).
Clinical Context
Landmark evidence establishing PCI after CR in limited-stage SCLC. Modern practice increasingly favors hippocampal-avoidance PCI to limit neurocognitive toxicity, and MRI surveillance as an alternative to PCI in extensive-stage SCLC.