Background
Phase 3 open-label RCT. 224 patients with extensive-stage SCLC who had any response to platinum-based doublet chemotherapy AND had confirmed absence of brain metastases on MRI. Enrolled at 47 Japanese institutions (April 2009–July 2013). Key distinction from EORTC 08993: MRI was required before enrollment and serial MRI surveillance was performed every 3 months during follow-up.
Interventions and follow up
Arm A: PCI 25 Gy in 10 daily fractions of 2.5 Gy (N=113)
Arm B: Observation + serial brain MRI every 3 months (N=111)
Primary endpoint: OS
mFollow up: Trial stopped early for futility after interim analysi
Arm B: Observation + serial brain MRI every 3 months (N=111)
Primary endpoint: OS
mFollow up: Trial stopped early for futility after interim analysi
Results
Median OS: 11.6 months (PCI) vs 13.7 months (observation), HR 1.27 (95% CI 0.96–1.68), P=.094 — not significant (PCI trend toward WORSE OS)
Early stopping: Bayesian predictive probability of PCI being superior: 0.011% → trial stopped for futility
Early stopping: Bayesian predictive probability of PCI being superior: 0.011% → trial stopped for futility
Adverse events
Main adverse events: Most common grade ≥3 AEs at 3 months: anorexia (6% PCI vs 2% observation), malaise (3% vs 1%), muscle weakness lower limb (1% vs 5%). No treatment-related deaths. PCI group had expected neurocognitive adverse effects not formally quantified in this analysis.
Conclusions
PCI did not improve and may have worsened OS compared to MRI surveillance in extensive SCLC patients with confirmed MRI-negative brains (HR 1.27, p=0.094). This trial substantially challenged the practice of routine PCI for extensive SCLC when MRI surveillance is available, directly contradicting the EORTC 08993 conclusion.
Key Limitations
Key Limitations: Open-label design — knowledge of treatment may influence management decisions. Conducted exclusively in Japan — survival outcomes (median OS 13.7 months for observation arm) were notably better than Western studies, possibly reflecting population-level differences in chemotherapy response, performance status, or healthcare access. The MRI surveillance strategy in the observation arm identified brain metastases early, allowing effective local treatment — this may have offset the PCI benefit seen in EORTC 08993. The trial may have been underpowered to detect a modest OS benefit (stopped early). Generalizability to non-Japanese populations and to healthcare systems without routine MRI access is uncertain.
Clinical Context
Takahashi 2017 fundamentally changed the PCI landscape for extensive SCLC. Current NCCN and ESMO guidelines now offer two acceptable approaches for ES-SCLC responders: (1) PCI (consistent with EORTC 08993) or (2) close MRI surveillance every 2–3 months without PCI (consistent with Takahashi). Patient preference, access to MRI surveillance, and neurocognitive concerns should guide shared decision-making. PCI remains standard of care for limited SCLC after complete remission (Auperin meta-analysis) — this trial applies only to extensive SCLC.
References