We are pleased to present highlights of the latest advances in immunotherapy emerging from the 2026 ASCO Annual Meeting, June 2, 2026.
(Thanks to Chinmay Jani, MD and Terri Holzen, PhD for compiling these scientific highlights)
Fitness-enhanced CEA-targeting CAR T cells for advanced colorectal cancer with peritoneal metastases
3514. Phase 1 study of intraperitoneal fast-manufactured IL-9-secreting CEACAM5-targeted CAR T cells in advanced colorectal cancer with peritoneal metastases
Hangyu Zhang (The First Affiliate Hospital, Zhejiang University School of Medicine, Hangzhou, China) presented safety and preliminary efficacy data from an open-label single-arm Phase 1 dose escalation study to evaluate optimized CAR T cells targeting carcinoembryonic antigen-related cell adhesion molecule 5 (CEACAM5) in patients with colorectal cancer with peritoneal metastasis (CRCPM). An earlier generation of CEACAM5-targeting CAR T cells exhibited limited efficacy due to decreased T cell fitness and T cell exhaustion (NCT05396300). The CAR includes a hypoxia-responsive element to increase CAR expression and decrease oxidative metabolism in hypoxic tumor microenvironments, reducing CAR T cell exhaustion and preserving an effector memory phenotype. Novel Th9/Tc9 polarization of CEACAM5-targeting CAR T cells occurred during manufacturing to enhance CAR T proliferation, reduce exhaustion, and increase persistence. 15 patients with measurable CEACAM5-expressing CRCPM that progressed after at least two prior therapies participated in the study. Patients underwent apheresis followed by lymphodepletion and CAR T cell infusion, which was administered intraperitoneally. Three CAR T cell dose levels were evaluated using a 3+3 dose escalation design: dose level 1 (DL1, n=3) was 2.0 x 10^5 cells/kg, dose level 2 (DL2, n=6) was 3.0 x 10^5 cells/kg, and dose level 3 (DL3, n=6) was 4.0 x 10^5 cells/kg. The primary objectives were safety and tolerability, and the secondary objectives included disease control rate (DCR) and cellular kinetics. The overall patient population received a median of three prior lines of therapy (range 1 – 6), and all patients had peritoneal involvement. Median age was 45 (27-68) with 6 patients (40%) having KRAS mutation, 2 (13.3%) having BRAF mutation. The safety profile was acceptable. The most common grade 3 or higher treatment-emergent adverse event (TEAE) was lymphopenia, which was experienced by all patients. Diarrhea was the most common non-hematologic grade 3 or higher TEAE, which occurred in 53% of patients. Cytokine release syndrome (CRS) occurred in all patients, and only one case was grade 3 or above. One patient receiving DL2 experienced a dose-limiting toxicity consisting of grade 4 pneumonia and CRS, which improved with corticosteroids and infliximab. No treatment-related deaths, perforation, obstruction, or surgical intervention occurred. Preliminary treatment responses are encouraging. The objective response rate was 57.1% in the evaluable patient population, with 8 partial responses. Responses occurred across all three dose levels. The DCR was 100%, and tumor regression was observed in 85.7% (12/14) of patients with a maximal tumor shrinkage of 85.6%. Median progression-free survival (PFS) was 4.4 months, and median overall survival (OS) was 13.4 months. 100% of patients exhibited reduced serum tumor marker s after CAR T infusion, and high Cmax values during cellular expansion were associated with improved PFS and OS. Although more time and larger patient populations are needed to validate these results, these data indicate that fitness-enhanced, regionally delivered Th9/Tc9-like CEACAM5-targeting CAR T cell therapy is a promising strategy for treating CRCPM and support additional studies on optimal dosing, T cell persistence, and strategies for combination therapy.
ctDNA dynamics and real-world treatment patterns and outcomes in patients receiving immune checkpoint inhibitors as first-line treatment of microsatellite instability-high metastatic colorectal cancer
3516. Real-world outcomes and circulating tumor DNA dynamics (ctDNA) in patients with microsatellite instability-high (MSI-H) metastatic colorectal cancer (mCRC) treated with immune checkpoint inhibitors (ICI)
Hannah Ruth Robinson (University of Colorado Anschutz School of Medicine, Aurora, CO, USA) reported an analysis of real-world clinical treatment patterns and outcomes and longitudinal circulating tumor DNA (ctDNA) dynamics in patients who received immune checkpoint inhibitors (ICI) for microsatellite instability-high (MSI-H) metastatic colorectal cancer (mCRC). Patients with MSI-H mCRC treated with first-line ICI were identified from Natera’s real-world database, which includes linked longitudinal ctDNA results. Further data were obtained by matching information from insurance claims databases (Forian’s hybrid data ecosystem, CHRONOS) along with mortality data obtained from Veritas. 561 patients were identified, representing 4774 ctDNA timepoints, with a median of 7 timepoints per patient (range 1-58). Patients were on ICI therapy for a median of 240 days. 70.2% of patients received pembrolizumab as their first-line therapy, 16.6% received ipilimumab combined with nivolumab, and 10.9% received nivolumab. The primary tumor site was colon in 64.5% of patients, rectal in 3.4%, and the rest had colorectal not otherwise specified. 28.3% (n=159) of patients also had Lynch syndrome. Among patients who were ctDNA positive prior to ICI therapy (n=108), 81% achieved ctDNA clearance at any time on or after receiving ICI. 55% of patients who achieved ctDNA clearance did so in the first 4 months of ICI treatment. 83% of patients who achieved ctDNA clearance maintained it for at least 6 months. ctDNA clearance occurred in 80.5% of patients on a single ICI regimen and in 84.6% of patients receiving a doublet regimen, and this difference was not statistically significant. ctDNA negativity at the first time point after beginning ICI was associated with significantly improved overall survival (OS), compared to ctDNA positivity, with 3-year OS of 93.1% and 76.1% (HR 3.4; 95% CI 1.8 – 6.3; p<0.0001), although there was a high level of variation in when the first ctDNA sample was obtained during the course of treatment. When analysis was limited to patients whose first ctDNA timepoint was obtained within 4 months of beginning ICI (n=180), OS was higher among patients who were ctDNA negative, compared to those who were ctDNA positive, but the difference was not significant (3-year OS 89.3% and 78.8%, respectively; p=0.263). Among patients who achieved ctDNA negativity at any time on or after ICI, OS was significantly improved compared with patients who remained ctDNA positive, with 3-year OS of 95.7% and 68.6% (HR 8.3; 95% CI 4.4 – 15.5; p<0.0001). ctDNA status at the first time point after starting ICI was also associated with distinct treatment patterns. 32% of patients who were ctDNA positive at the first timepoint received second-line therapy, compared with 6.6% of patients who were ctDNA negative. Among all patients who needed second-line therapy, patients who were ctDNA positive received second-line therapy after a median of 302 days, compared with 683 days for patients who were ctDNA negative. Patients who were consistently ctDNA positive after ICI discontinuation had significantly worse outcomes compared with patients who were consistently ctDNA negative, with 3-year OS of 58.2% and 96.6%, respectively (HR 9.8; 95% CI 4.3 – 22.9; p<0.0001). Patients who experienced delayed ctDNA clearance (i.e., achieved ctDNA clearance after discontinuing ICI) exhibited similar survival patterns to patients who were consistently ctDNA negative, with 3-year OS rates of 100% and 96.9%, respectively (p=0.95). Patients who experienced delayed ctDNA clearance spent a median of 84 days on first-line ICI, compared with 234 days for patients who were persistently ctDNA negative and 190 days for patients who were persistently ctDNA positive, suggesting patients with delayed ctDNA clearance may have been more likely to discontinue ICI due to toxicity. This analysis indicates that ctDNA dynamics strongly predict outcomes in patients receiving ICI as first-line treatment for MSI-H mCRC. These findings support the use of ctDNA to inform real-time treatment decisions in this patient population and support additional prospective studies of using ctDNA to monitor response and guide treatment in other patient populations.
To learn about a radiomics-based model to predict response to neoadjuvant ICI for colorectal cancer, read this paper from Zhang et al in JITC.
First clinical results from a phase 1 study of a GPC x 4-1BB bispecific for advanced or metastatic hepatocellular carcinoma
3016. A phase 1 study of BGB-B2033 (GPC x 4-1BB bispecific antibody) monotherapy in patients with selected advanced or metastatic solid tumor: First disclosure of clinical data
Hong Jae Chon (CHA Bundang Medical Center, Bundang, Republic of Korea) presented the first report of clinical data from a phase 1a dose escalation cohort investigating BGB-B2033, a first-in-class bispecific antibody targeting glypican-3 (GPC3) and the costimulatory T cell receptor 4-1BB/CD137, in patients with advanced or metastatic hepatocellular carcinoma (HCC). GPC3 is a tumor-associated antigen highly expressed in multiple tumor types, including approximately 90% of HCC tumors. BGB-B2033 activates 4-1BB on CD8+ T cells only in the presence of GPC3-expressing tumor cells, enabling targeted anti-tumor activity. The Fc region of BGB-B2033 is engineered to extend pharmacokinetic exposure while maintaining localized immune activation. 61 patients with advanced GPC3-expressing HCC treated with at least one prior therapy and one measurable lesion by RECIST 1.1 participated in the study. Patients received BGB-B2033 monotherapy at doses ranging from 1 mg to 1000 mg every three weeks. The primary endpoints are safety, recommended dose for expansion, and maximum tolerated dose (MTD). Secondary endpoints include objective response rate (ORR), pharmacokinetics/pharmacodynamics, and immunogenicity. The median follow-up was 4.8 months. 60 patients had HCC, and one patient had gastric cancer. 91.8% of patients were Asian, and 86.9% of patients tested positive for the hepatitis B virus. 96.7% of patients had prior anti-PD(L)-1 therapy, 78.7% of patients had prior exposure to tyrosine kinase inhibitor (TKI), and 75.4% of patients had prior exposure to both anti-PD(L)-1 therapy and TKI. Doses of BGB-B2033 over 300 mg exhibited linear pharmacokinetics, and higher dose levels of BGB-B2033 were associated with increased levels of soluble 4-1BB. Treatment-related adverse events (TRAE) occurred in 47.5% of the patient population; 8.2% of patients experienced grade 3 TRAEs, and no grade 4 or 5 TRAEs were observed. Immune-mediated adverse events occurred in 6.6% of patients, none of which were grade 3 or higher. 2 patients discontinued treatment due to adverse events, and one dose-limiting toxicity occurred. At doses of 300 mg or higher, the confirmed ORR was 28.9%, and the unconfirmed ORR was 31.6%. BGB-B2033 exhibits compelling anti-tumor activity and a favorable safety profile in heavily pretreated patients with advanced or metastatic HCC. Dose expansion studies of BGB-B2033 and triple dose escalation studies of BGB-B2033 with tislelizumab and bevacizumab are ongoing, and registration enabling clinical trials are being planned.