Neoadjuvant chemotherapy with or without camrelizumab in resectable esophageal squamous cell carcinoma: the randomized phase 3 ESCORT-NEO/NCCESO1 trial

Esophageal squamous cell carcinoma (ESCC) is a major global health burden, with high incidence and mortality, particularly in China. In East Asia, neoadjuvant chemotherapy or chemoradiotherapy is standard for resectable locally advanced ESCC (LA-ESCC), but the optimal neoadjuvant strategy remains unsettled. While trials such as CROSS and NEOCRTEC5010 demonstrated survival benefits for chemoradiotherapy over surgery alone, JCOG9907 showed superior overall survival with neoadjuvant chemotherapy versus adjuvant therapy. Recent phase 3 trials (CMISG1701, JCOG1109) and a network meta-analysis have not shown a clear overall survival advantage for chemoradiotherapy over chemotherapy in LA-ESCC, leaving room for improved systemic approaches. Immune checkpoint inhibitors, including camrelizumab, have shown efficacy in advanced ESCC and in early-phase neoadjuvant settings with chemotherapy, yielding high pathological complete response (PCR) rates. Additionally, nab-paclitaxel may have advantages over solvent-based paclitaxel when combined with immunotherapy. The ESCORT-NEO/NCCES01 trial was designed to test whether adding camrelizumab to neoadjuvant chemotherapy (with either paclitaxel or nab-paclitaxel) improves PCR and event-free survival compared with chemotherapy alone in resectable LA-ESCC.

Key prior evidence includes: (1) CROSS and NEOCRTEC5010 trials showing neoadjuvant chemoradiotherapy improves outcomes over surgery alone; (2) JCOG9907 showing neoadjuvant chemotherapy superior to adjuvant chemotherapy; (3) recent phase 3 trials (CMISG1701, JCOG1109) and a network meta-analysis failing to demonstrate clear overall survival superiority of chemoradiotherapy over chemotherapy for LA-ESCC; (4) Phase 1b/2 neoadjuvant studies of camrelizumab plus chemotherapy in LA-ESCC reporting PCR rates of ~17.6–39.2%; (5) retrospective analyses suggesting neoadjuvant immunochemotherapy may improve 3-year OS and DFS versus chemoradiotherapy and that nab-paclitaxel-based regimens may yield higher PCR than solvent-based paclitaxel when combined with immunotherapy. These data motivated a phase 3, randomized assessment of neoadjuvant immunochemotherapy in ESCC and an exploration of paclitaxel formulation choice.

Design: Multicenter, randomized, open-label, phase 3 trial across 24 centers in China (ChiCTR2000040034). Patients: Adults (18–75) with histologically/cytologically confirmed thoracic LA-ESCC, staged T1b–3N1–3M0 or T3N0M0 (AJCC 8th), ECOG PS 0–1, candidates for R0 resection, treatment-naive, with adequate organ function. Main exclusions included gross invasion of adjacent organs, supraclavicular nodal metastasis, uncontrolled effusions, poor nutrition (BMI<18.5 unless corrected), autoimmune disease, recent immunosuppression or live vaccines, active infections (HBV/HCV/HIV/tuberculosis), significant cardiac disease, recent serious infection, other malignancies within 5 years, pregnancy/lactation, or other safety/compliance concerns. Randomization and interventions: Patients were randomized 1:1:1, stratified by clinical stage (I/II, III, IVA), to two 3-week cycles of neoadjuvant therapy: (1) Cam+nab-TP: camrelizumab 200 mg day 1; nab-paclitaxel 125 mg/m² days 1 and 8; cisplatin 75 mg/m² day 1. (2) Cam+TP: camrelizumab 200 mg day 1; paclitaxel 175 mg/m² day 1 (with standard hypersensitivity prophylaxis including dexamethasone, diphenhydramine, H2 blocker); cisplatin 75 mg/m² day 1 (hydration/diuresis; no prophylactic antiemetic corticosteroids). (3) TP: paclitaxel 175 mg/m² day 1; cisplatin 75 mg/m² day 1. Surgery (recommended McKeown esophagectomy with two-field lymphadenectomy; minimally invasive or open) was planned 4–6 weeks post-therapy. Patients in Cam arms received adjuvant camrelizumab 200 mg every 3 weeks for up to 15 cycles. Assessments: Pathology was centrally reviewed by a blinded independent review committee (Mandard TRG). PD-L1 expression was centrally assessed on pretreatment biopsies (22C3) as CPS and TPS. Safety was graded per NCI CTCAE v5.0; postoperative complications per ECCG and Clavien-Dindo. Endpoints: Dual primary endpoints were PCR (TRG 1 with node-negative) and investigator-assessed event-free survival (EFS). Secondary endpoints included major pathological response (MPR, <10% residual viable tumor), R0 resection rate, ypTNM stage, disease-free survival (DFS), overall survival (OS), adverse events (AEs), and surgical complications. Statistics: Superiority of Cam-containing arms versus TP was tested for PCR and EFS with a one-sided familywise α=0.025 using the Maurer-Bretz graphical approach (initial α 0.005 for PCR, 0.02 for EFS; hierarchical testing). Sample size assumed PCR rates of 30% (Cam+nab-TP), 25% (Cam+TP), 9% (TP); 111 per arm provided ≥93% and ≥75% power for the two camrelizumab comparisons; 130 per arm planned allowing 15% dropout. For EFS, an HR of 0.67 versus TP and 228 events were targeted for ≥80% power. Analyses were ITT for efficacy; safety set included treated patients. PCR/MPR 95% CIs used Clopper-Pearson; between-group differences and ORs used Mantel-Haenszel stratified by clinical stage; Cochran-Mantel-Haenszel tests compared groups. SAS 9.4 used for analyses.

- Enrollment and populations: 411 screened; 391 randomized (ITT: Cam+nab-TP n=132, Cam+TP n=130, TP n=129). Safety set: 132, 130, 125. Median follow-up 8.2 months (IQR 3.5–15.6). Baseline characteristics balanced. - Treatment and surgery: Completed two cycles: Cam+nab-TP 129/132; Cam+TP 125/130; TP 122/125. Esophagectomy rates (ITT-based): 86.4% (114/132), 89.2% (116/130), 79.8% (103/129). Median time last dose to surgery: 5.9, 5.7, 5.4 weeks. McKeown procedure used in ~92–94%. Median lymph nodes harvested: 34, 37, 32. - Primary endpoint (PCR, ITT): 28.0% (Cam+nab-TP) vs 4.7% (TP); difference 23.5% (95% CI 15.1–32.0); OR 8.11 (95% CI 3.28–20.06); P<0.0001. 15.4% (Cam+TP) vs 4.7% (TP); difference 10.9% (95% CI 3.7–18.1); OR 3.81 (95% CI 1.48–9.80); P=0.0034. EFS not mature. - Secondary pathological outcomes: MPR rates 59.1% (95% CI 50.2–67.6), 36.2% (27.9–45.0), 20.9% (14.3–29.0) for Cam+nab-TP, Cam+TP, TP. R0 resection: 99.1% (113/114), 95.7% (111/116), 92.2% (95/103). Post-neoadjuvant stage I achieved in 50.9%, 39.7%, 26.2%. Median residual viable tumor: 1% (IQR 0–20), 15% (1–70), 50% (8–80). - Safety: Any surgical complication: 34.2%, 38.8%, 32.0%; Clavien-Dindo ≥3: 6.1%, 12.1%, 6.8%. 30-day postoperative mortality: 0.9%, 1.7%, 1.0%; 90-day mortality: 1.8%, 1.7%, 1.0%. Any treatment-related AE (neoadjuvant): 93.9%, 83.1%, 83.2%; grade ≥3 TRAEs: 34.1%, 29.2%, 28.8%. Common grade ≥3 TRAEs: neutrophil count decreased, white blood cell count decreased. TRAE-related chemo discontinuation: 3.0%, 3.8%, 0.8%. Camrelizumab discontinuation due to TRAEs: 0.8% in each Cam arm. Immune-related AEs: 27.3% (Cam+nab-TP), 24.6% (Cam+TP); grade ≥3 irAEs: 4.5%, 3.8%; most common irAE was reactive cutaneous capillary endothelial proliferation (grade 1–2). A grade 5 TRAE (acute liver failure) occurred in the Cam+TP arm on day 1 of neoadjuvant therapy.

Adding camrelizumab to neoadjuvant chemotherapy significantly increased PCR versus chemotherapy alone in resectable LA-ESCC, supporting the strategy of perioperative immunochemotherapy to enhance systemic tumor control beyond locoregional approaches. The magnitude of PCR improvement (28.0% and 15.4% vs 4.7%) compares favorably with historical neoadjuvant chemotherapy and approaches chemoradiotherapy PCR benchmarks, suggesting potential for long-term benefit pending mature EFS/OS. Safety was manageable, with comparable surgical complication rates and perioperative mortality across arms, and expected hematologic toxicities and immune-related events under camrelizumab. The numerically higher PCR with nab-paclitaxel versus solvent-based paclitaxel may reflect steroid premedication effects with paclitaxel, higher dose intensity of nab-paclitaxel, and/or intrinsic pharmacologic advantages, aligning with prior studies in other cancers. Exploratory analyses indicated greater PCR benefit in patients with higher PD-L1 expression, consistent with prior camrelizumab data, though chemotherapy backbones may attenuate strict PD-L1 dependence. Overall, the trial provides confirmatory phase 3 evidence that neoadjuvant immunochemotherapy improves pathologic responses without compromising operability, laying groundwork for potential survival gains.

Neoadjuvant camrelizumab plus chemotherapy, followed by adjuvant camrelizumab, significantly improved pathological complete response compared with chemotherapy alone in resectable LA-ESCC, with acceptable perioperative and treatment-related safety. These results support immunochemotherapy as a promising neoadjuvant strategy in ESCC. Ongoing follow-up will clarify effects on EFS, DFS, and OS. Future research should directly compare nab-paclitaxel versus solvent-based paclitaxel backbones with ICIs, evaluate head-to-head versus chemoradiotherapy, refine biomarker strategies (including PD-L1 and others), and validate generalizability beyond the Chinese population.

- The trial did not prespecify or power a direct comparison between the two camrelizumab-based chemotherapy backbones (nab-paclitaxel vs paclitaxel), so observed differences are exploratory. - No direct comparison with neoadjuvant chemoradiotherapy, a standard option in LA-ESCC, was performed. - The second primary endpoint (event-free survival) and longer-term outcomes (DFS/OS) are immature. - Conducted exclusively in China, potentially limiting generalizability to other populations. - Open-label design may introduce bias despite blinded independent pathology review. - Limited biomarker analyses and a notable proportion of unknown PD-L1 status constrain predictive insights.