Oncology · Medical Literature
[Articles] Neoadjuvant toripalimab plus celecoxib versus toripalimab monotherapy for mismatch repair-deficient or microsatellite instability-high, locally advanced colorectal cancer (PICC-2): an open-label, multicentre, randomised, phase 2 trial
Abstract
In patients with dMMR or MSI-H locally advanced colorectal cancer, neoadjuvant toripalimab plus celecoxib significantly increased the proportion of patients attaining pathological complete response compared with toripalimab monotherapy, with a similar safety profile. These findings support further investigation of this combination strategy in larger phase 3 trials.
Clinical relevance
In patients with dMMR or MSI-H locally advanced colorectal cancer, does 12 cycles of neoadjuvant toripalimab (anti-PD-1) plus celecoxib (COX-2 inhibitor) achieve a superior rate of pathological complete response compared with 12 cycles of toripalimab monotherapy? __SNAPSHOT__ {"studyDesign":"Open-label, multicentre, randomised, phase 2 trial (PICC-2 cohort of the investigator-initiated PICC platform trial, NCT03926338)","setting":"Three academic hospitals in China","sampleSize":"n=110 (toripalimab plus celecoxib: n=55; toripalimab monotherapy: n=55); enrolled May 5, 2022 – January 20, 2025; data cutoff October 10, 2025","population":"Adults aged 18–75 years with histologically confirmed dMMR (by IHC) or MSI-H (by PCR) colorectal cancer, clinical stage cT3–T4 or any cT with node-positive disease (cN+), ECOG PS 0–1, adequate end-organ function. Median age 50 years; 65 males; all patients Chinese; 88 colon/22 rectal; 76 cT4 (25 cT4b); 105 clinically node-positive; Lynch syndrome in 37 patients. Key exclusions: distant metastases, prior systemic colorectal therapy, long-term NSAID/COX-2 inhibitor or aspirin use, active autoimmune disease, systemic immunosuppression.","intervention":"Toripalimab 3 mg/kg IV on Day 1 of each 14-day cycle for 12 cycles PLUS celecoxib 200 mg orally twice daily on Days 1–14 of each 14-day cycle; followed by surgery with curative intent (or non-operative management if complete clinical response at restaging)","comparator":"Toripalimab 3 mg/kg IV on Day 1 of each 14-day cycle for 12 cycles (monotherapy); followed by surgery with curative intent","primaryOutcome":"Pathological complete response (pCR), defined as no residual viable tumour in the primary tumour and all sampled lymph nodes at surgery; assessed by central blinded independent pathological review in the intention-to-treat population","followUp":"Median follow-up 18.1 months at data cutoff (October 10, 2025); immature for survival endpoint assessment","fundingSource":"Not explicitly reported for PICC-2 in available sources; the broader PICC platform was supported by government sources (National Key R&D Program of China, National Natural Science Foundation of China) and a non-profit oncology research fund (Chinese Society of Clinical Oncology–Junshi Biosciences Oncology Immunity Research Fund); Junshi Biosciences is the manufacturer of toripalimab — potential industry-affiliated funding conflict should be noted"}
Key findings
Primary outcome: pCR was achieved in 49/55 patients (89%) in the toripalimab plus celecoxib arm versus 38/55 patients (69%) in the toripalimab monotherapy arm — an absolute difference of 20 percentage points (calculated) — which was statistically significant (p=0.014). Confidence intervals for individual group pCR rates were not reported in available sources. The NNT to achieve one additional pCR with the addition of celecoxib is approximately 5 (calculated: 1/0.20). Secondary efficacy endpoints: major pathological response (MPR) occurred in 93% vs. 85%; tumour regression grade 0 (complete regression) in 91% vs. 71%; pathological response in the primary tumour in 96% vs. 91%. In the per-protocol population, pCR was reported in 92% of the combination arm (monotherapy rate not reported in available sources). Surgery was performed in 53/55 combination patients and 51/55 monotherapy patients; all who underwent surgery achieved R0 resection. Treatment completion: 49/55 patients in each group completed all 12 planned cycles. Safety: Treatment-related adverse events (TRAEs) of any grade occurred in 38% (combination) vs. 33% (monotherapy). Grade 3 TRAEs occurred in 3 patients (combination: tumour perforation, bowel obstruction) and 4 patients (monotherapy: rash, tumour perforation, increased aminotransferases, hypothyroidism). No grade 4 or 5 TRAEs were reported in either arm. Discontinuation due to TRAEs: 7% (combination) vs. 5% (monotherapy). No TRAE led to surgical cancellation. Surgery-related adverse events of any grade: 11% (combination) vs. 8% (monotherapy). Survival data are immature at a median follow-up of 18.1 months; OS, DFS, and EFS were not reported as primary results for this cohort. Patient-reported outcomes and quality-of-life data were not collected.
Practice implications
PICC-2 is the first randomised trial to demonstrate that the addition of a COX-2 inhibitor (celecoxib) to neoadjuvant PD-1 blockade significantly improves pCR rate in dMMR/MSI-H locally advanced colorectal cancer, achieving an 89% pCR vs. 69% with toripalimab alone (p=0.014), with an acceptable and similar safety profile. The biological rationale — that COX-2-driven prostaglandin E2 (PGE2) signalling suppresses antitumour T-cell immunity, and that celecoxib can restore this — is mechanistically plausible and supported by preclinical data. However, several limitations prevent immediate practice change: (1) pCR is a surrogate endpoint; no mature survival data (OS, DFS, EFS) are yet available from PICC-2, and the relationship between pCR and long-term survival in this specific molecular subtype with neoadjuvant immunotherapy has not been firmly established; (2) median follow-up is only 18.1 months — insufficient to detect differences in recurrence or survival; (3) the study enrolled an entirely Chinese population with a relatively young median age of 50 years and a high proportion (80%) of colon (vs. rectal) tumours, limiting generalisability to Western populations, older patients, or rectal cancer-predominant cohorts; (4) the lack of an upfront-surgery control arm precludes conclusions about the absolute benefit of neoadjuvant immunotherapy over surgery alone; (5) celecoxib's cardiovascular and gastrointestinal risks (particularly relevant in older or cardiovascular-comorbid patients) were not systematically characterised at this follow-up duration; (6) quality-of-life data are absent; and (7) the open-label design introduces performance and detection bias risk. The findings are hypothesis-confirming and signal-generating, and appropriately support a phase 3 trial, which is the authors' stated conclusion. This summary is based on secondary reporting of the abstract and trial registry data, as the full article was not directly accessible; clinicians should review the complete publication for additional methodological detail, confidence intervals, and subgroup analyses. __LIMITATIONS__ Pathological complete response is a surrogate endpoint; survival benefit (OS, DFS, EFS) remains unproven at current follow-up of 18.1 months,Small-to-moderate sample size for a phase 2 trial (n=110); insufficient power to detect differences in survival endpoints,Open-label design introduces performance bias and potential ascertainment bias in outcome assessment, partially mitigated by central blinded pathological review,Entirely Chinese population with young median age (50 years) and predominantly colon cancer (80%) limits generalisability to older Western patients and to rectal cancer,No upfront-surgery control group; cannot determine superiority of neoadjuvant immunotherapy over surgery alone,Potential industry-affiliated co-funding (Junshi Biosciences, manufacturer of toripalimab) introduces a conflict-of-interest risk for efficacy estimates,Short follow-up precludes characterisation of late toxicities (e.g., cardiovascular events with celecoxib, late immune-related adverse events),Patient-reported outcomes and quality-of-life data not collected,Biological interaction between celecoxib and toripalimab on immune pathways is mechanistically proposed but not causally proven in this study,Only the abstract and secondary reporting were available for this summary; full confidence intervals, stratification factors, and subgroup data were not accessible __BOTTOM_LINE_CATEGORY__ promising_but_preliminary __BOTTOM_LINE__ In PICC-2, adding celecoxib to neoadjuvant toripalimab produced a clinically and statistically significant improvement in pCR (89% vs. 69%, p=0.014) in dMMR/MSI-H locally advanced colorectal cancer, with no increase in grade ≥3 toxicity. However, pCR is a surrogate endpoint and survival data are immature (median follow-up 18.1 months), so this combination should not replace current standard approaches outside of a clinical trial. These results are compelling enough to support accrual to a confirmatory phase 3 trial, and clinicians caring for dMMR/MSI-H patients should be aware of this signal.