Background
Phase III RCT (RTOG 9811). 682 patients with carcinoma of the anal canal (T2–T4 or any N+). Compared MMC-based concurrent CRT vs cisplatin-based CRT (with 2-cycle induction). MMC arm: 5-FU 1000 mg/m²/d × 4d + MMC 10 mg/m² on days 1 and 29 + 45–59 Gy RT. Cisplatin arm: induction 5-FU 1000 mg/m²/d × 4d + CDDP 75 mg/m² (2 cycles), then 5-FU + CDDP + same RT. RT dose: 45 Gy pelvis + 14-Gy boost for T2 or 10–14 Gy boost for T3–T4 nodal disease. This long-term results report (Gunderson 2012) followed the initial publication showing no significant difference at 2 years.
Interventions and follow up
Arm A: 5-FU + mitomycin C + RT (45–59 Gy) — standard CRT
Arm B: Induction 5-FU + cisplatin (×2 cycles), then 5-FU + cisplatin + RT (same dose)
Primary endpoint: Disease-free survival
mFollow up: 8.2 year
Arm B: Induction 5-FU + cisplatin (×2 cycles), then 5-FU + cisplatin + RT (same dose)
Primary endpoint: Disease-free survival
mFollow up: 8.2 year
Results
5-year DFS: 67.8% (MMC) vs 57.8% (cisplatin), P=.006
5-year OS: 78.3% (MMC) vs 70.7% (cisplatin), P=.026
5-year colostomy rate: 10% (MMC) vs 19% (cisplatin), P=.02
Local-regional failure: 25% (MMC) vs 33% (cisplatin), P=.08
5-year OS: 78.3% (MMC) vs 70.7% (cisplatin), P=.026
5-year colostomy rate: 10% (MMC) vs 19% (cisplatin), P=.02
Local-regional failure: 25% (MMC) vs 33% (cisplatin), P=.08
Adverse events
Main adverse events: Grade 3–4 hematologic toxicity: 62% (MMC) vs 42% (cisplatin), P<.001. Grade 3–4 non-hematologic toxicity: similar between arms. Late toxicity: comparable. Treatment-related deaths: 2 (MMC arm) vs 1 (cisplatin arm).
Conclusions
At long-term follow-up, MMC-based CRT was superior to cisplatin induction followed by cisplatin-concurrent CRT for both DFS and OS in anal canal carcinoma, confirming 5-FU + MMC + RT as the standard of care.
Key Limitations
Key Limitations: The cisplatin arm used an induction-then-concurrent approach (2 cycles induction + concurrent), not a direct MMC-versus-cisplatin-concurrent comparison — making it difficult to determine whether induction was responsible for inferior outcomes. Initial 2-year results (JAMA 2008) showed no significant difference, and long-term superiority of MMC emerged later. Higher hematologic toxicity in the MMC arm remains a clinical concern, driving interest in capecitabine substitution.
Clinical Context
RTOG 9811 long-term data definitively confirmed 5-FU + MMC + RT as the standard; cisplatin-based induction followed by concurrent CRT cannot replace MMC-based concurrent CRT. ACT-II (James Lancet Oncol 2013) subsequently demonstrated that MMC and cisplatin are equivalent when both are used as concurrent (not induction) agents, with MMC having less hematologic toxicity.