Background
Phase III RCT (POET — PreOperative chEmoradiation versus chemotherapy for resecTable esophagogastric junction adenocarcinoma). 126 patients (of 354 planned) with locally advanced (uT3-4NXM0) adenocarcinoma of the lower esophagus or gastric cardia. Closed early due to poor accrual. All patients underwent esophagectomy after treatment. Compared preoperative chemotherapy alone vs preoperative chemoradiotherapy.
Interventions and follow up
Arm A: Induction chemotherapy (15 weeks: 5-FU/leucovorin/cisplatin for first 3 mo, then cisplatin/etoposide × 2 cycles) → surgery
Arm B: Induction chemotherapy (12 weeks: same regimen) → CRT (3 weeks: 30 Gy in 15 fx + concurrent cisplatin/etoposide) → surgery
Primary endpoint: Overall survival
mFollow up: Median 46 month
Arm B: Induction chemotherapy (12 weeks: same regimen) → CRT (3 weeks: 30 Gy in 15 fx + concurrent cisplatin/etoposide) → surgery
Primary endpoint: Overall survival
mFollow up: Median 46 month
Results
3-year OS: 47.4% (chemoRT→surgery) vs 27.7% (chemo→surgery), log-rank P=.07 (not significant; HR 0.67, 95% CI 0.41–1.07)
pCR rate: 15.6% (chemoRT) vs 2.0% (chemo alone), P=.03
Node-negative at resection: 64.4% vs 37.7%
R0 resection rate: Not significantly different (69.5% vs 71.5%)
Postoperative mortality: 10.2% (chemoRT) vs 3.8% (chemo) — not significant
pCR rate: 15.6% (chemoRT) vs 2.0% (chemo alone), P=.03
Node-negative at resection: 64.4% vs 37.7%
R0 resection rate: Not significantly different (69.5% vs 71.5%)
Postoperative mortality: 10.2% (chemoRT) vs 3.8% (chemo) — not significant
Adverse events
Main adverse events: Postoperative mortality nonsignificantly higher with chemoRT (10.2% vs 3.8%). Chemoradiotherapy was feasible and well-tolerated. Hematologic and esophageal toxicities as expected with concurrent cisplatin/etoposide + RT.
Conclusions
Adding preoperative CRT to preoperative chemotherapy for adenocarcinoma of the EGJ significantly increased pCR rates (15.6% vs 2%) and lymph node clearance. The 3-year OS improvement (47% vs 28%) did not reach statistical significance (P=.07), primarily due to insufficient sample size. The 5-year OS in the long-term update (Stahl 2017) remained non-significantly different (28.5% chemoRT vs 26.2% chemo).
Key Limitations
Key Limitations: Grossly underpowered — enrolled only 36% of planned sample. The statistically non-significant P=.07 for OS improvement in a trial powered for 354 patients makes definitive conclusions impossible. Older chemotherapy regimen (not anthracycline-based; FLOT now standard). Short CRT course (30 Gy) compared to CROSS (41.4 Gy). Long-term update (Stahl 2017) showed 5-yr OS essentially equivalent (28.5% vs 26.2%), losing the initial trend. The trial predates modern peri-operative chemotherapy standards.
Clinical Context
POET has been superseded by the CROSS trial (carboplatin/paclitaxel + 41.4 Gy → surgery) as the standard neoadjuvant approach for esophageal/EGJ adenocarcinoma in many centers. The FLOT4 trial (perioperative FLOT chemotherapy) is an alternative ESMO-endorsed standard, particularly for gastric/GEJ adenocarcinoma. ESOPEC (2024) demonstrated FLOT superiority over CROSS specifically for esophageal adenocarcinoma, further redefining this field.