June 2025
Journal of Clinical Oncology
3559 Background: Circulating tumor DNA (ctDNA) has shown potential in predicting the efficacy of neoadjuvant treatment for colorectal cancer (CRC). However, evidence is limited for patients with deficient mismatch repair (dMMR) / microsatellite instability-high (MSI-H) CRC, who respond well to immunotherapy. This study explores an optimal ctDNA-guided neoadjuvant immunotherapy (nIT) strategy in locally advanced CRC (LACRC) patients with dMMR/MSI-H. Methods: We conducted a single-center, open-label, phase 2 trial named WINDOW involving dMMR/MSI-H LACRC patients. Patients received the anti-PD-1 antibody tislelizumab every three weeks. ctDNA was monitored at baseline, after 2, 4, 5, or even 6 to 8 cycles, and during post-watch-and-wait (W&W) or post-surgery periods using the Signatera platform. Starting from the 4th cycle, patients with two consecutive ctDNA-negative results were eligible for the W&W approach. Those who did not achieve ctDNA-negative or turned positive during follow-up continued immunotherapy. Surgery was performed for patients not meeting W&W criteria by the 8th cycle. The primary endpoint was the complete response (CR) rate, including ctDNA-negative clinical complete response (cCR) and pathological complete response (pCR). The trial is registered at ClinicalTrials.gov (NCT06477991). Results: From January 2023 to May 2024, 24 patients with stage II-III dMMR/MSI-H CRC were enrolled, including 18 with colon cancer and 6 with rectal cancer. At baseline, all patients had detectable ctDNA. After nIT, 87.5% (21/24) achieved two consecutive ctDNA-negative results. Among the remaining three, two underwent surgery due to sustained ctDNA-positivity (both TRG3), while one showed continuous ctDNA decrease despite obstruction and achieved pCR. One patient required emergency surgery for perforation (confirmed as pCR), while 20 were managed with the W&W strategy. Of these, 90% (18/20) achieved ctDNA negativity after two cycles, and 95% (19/20) after five cycles of nIT. With a median follow-up of 20.3 months (range: 8.7–25.0 months), none experienced recurrence, resulting in an overall CR rate of 91.7% (22/24). Notably, if ctDNA remained positive after the 5th cycle, the CR rate was only 25% (1/4), while it reached 100% (20/20) if ctDNA became negative. From the perspective of organ preservation, only 45.5% (10/22) of patients avoided surgery based on imaging alone; however, with ctDNA-guided management, 83.3% (20/24) avoided surgery. Conclusions: NIT demonstrates high efficacy in dMMR/MSI-H LACRC. The ctDNA-guided W&W strategy significantly improves organ preservation rates. Monitoring ctDNA negativity after the 5th cycle of nIT is a crucial marker of high CR rates, suggesting this cycle may represent the optimal monitoring window during nIT. Clinical trial information: NCT06477991 .