Spatial distribution of tumor infiltrating lymphocytes (TILs) and cancer cells in the tumor microenvironment (TME) along with tumor gene mutation status are of vital importance to the guidance of cancer immunotherapy and prognoses. In this work, we developed a deep learning-based computational framework, termed ImmunoAIzer, which involves: (1) the implementation of a semi-supervised strategy to train a cellular biomarker distribution prediction network (CBDPN) to make predictions of spatial distributions of CD3, CD20, PanCK, and DAPI biomarkers in the tumor microenvironment with an accuracy of 90.4%; (2) using CBDPN to select tumor areas on hematoxylin and eosin (H&E) staining tissue slides and training a multilabel tumor gene mutation detection network (TGMDN), which can detect APC, KRAS, and TP53 mutations with area-under-the-curve (AUC) values of 0.76, 0.77, and 0.79. These findings suggest that ImmunoAIzer could provide comprehensive information of cell distribution and tumor gene mutation status of colon cancer patients efficiently and less costly; hence, it could serve as an effective auxiliary tool for the guidance of immunotherapy and prognoses. The method is also generalizable and has the potential to be extended for application to other types of cancers other than colon cancer.

Squamous cell carcinomas (SCCs) are common and aggressive malignancies. Immune check point blockade (ICB) therapy using PD-1/PD-L1 antibodies has been approved in several types of advanced SCCs. However, low response rate and treatment resistance are common. Improving the efficacy of ICB therapy requires better understanding of the mechanism of immune evasion. Here, we identify that the SCC-master transcription factor TP63 suppresses interferon-γ (IFNγ) signaling. TP63 inhibition leads to increased CD8⁺ T cell infiltration and heighten tumor killing in in vivo syngeneic mouse model and ex vivo co-culture system, respectively. Moreover, expression of TP63 is negatively correlated with CD8⁺ T cell infiltration and activation in patients with SCC. Silencing of TP63 enhances the anti-tumor efficacy of PD-1 blockade by promoting CD8⁺ T cell infiltration and functionality. Mechanistically, TP63 and STAT1 mutually suppress each other to regulate the IFNγ signaling by co-occupying and co-regulating their own promoters and enhancers. Together, our findings elucidate a tumor-extrinsic function of TP63 in promoting immune evasion of SCC cells. Over-expression of TP63 may serve as a biomarker predicting the outcome of SCC patients treated with ICB therapy, and targeting TP63/STAT/IFNγ axis may enhance the efficacy of ICB therapy for this deadly cancer.

Research on patients with advanced esophageal squamous cell carcinoma (ESCC) who have progressed on immunotherapy remains limited. BL-B01D1 is a first-in-class antibody–drug conjugate consisting of an EGFR–HER3 bispecific antibody bound to a topoisomerase I inhibitor (Ed-04) payload via a cleavable linker. Here, we present safety and efficacy data from a phase 1 study of BL-B01D1, in 82 patients previously treated for ESCC. The primary endpoint was the recommended phase 2 dose. Administered doses were 2.0 (n = 22) and 2.5 (n = 60) mg kg⁻¹ D1D8 infusion every 3 weeks (Q3W). The confirmed objective response rate (cORR) was 29.3% (24 of 82) in all patients and 32.9% (24 of 73) among evaluable patients. For patients dosed at 2.5 mg kg⁻¹, cORR was 39.6% (21 of 53) and disease control rate was 79.2% (42 of 53). In the 2.0 mg kg⁻¹ group, cORR was 15.0% (3 of 20), and the disease control rate was 50.0% (10 of 20). The phase 2 dose was established at 2.5 mg kg⁻¹ D1D8 Q3W. The incidence of G3 treatment-related adverse events at 2.5 mg kg⁻¹ was 63.3%; most common adverse events were anemia (28.3%), leukopenia and thrombocytopenia (18.3%, each), and neutropenia (16.7%). Two cases of ≥G3 interstitial lung disease were observed. Overall, BL-B01D1 demonstrated promising efficacy and manageable safety in patients with metastatic ESCC. A further phase 3 clinical trial has already been initiated. ClinicalTrials.gov registration: NCT05262491.

PURPOSE: Although multiple agents targeting PD-1 have been approved as second-line treatment for esophageal squamous cell carcinoma (ESCC), only a fraction of patients derive long-term survival. Hence, reliable predictive biomarkers are urgently needed. METHODS: Comprehensive tumor genomic profiling and transcriptome sequencing were performed on samples from the RATIONALE-302 study. We also conducted single-cell RNA sequencing analysis on Notch1 knockdown ESCC murine models to further explore the potential molecular mechanisms underlying anti–PD-1 benefit. RESULTS: We identified NOTCH1 mutation as a potential predictive biomarker for longer overall survival (OS) with tislelizumab versus chemotherapy (18.4 months v 5.3 months; hazard ratio, 0.35 [95% CI, 0.17 to 0.71]). At the transcriptional level, type I IFN (IFN-I)/toll-like receptor expression signatures were positively associated with OS benefit of tislelizumab, whereas B-cell and neutrophil signatures predicted unfavorable OS. Exploratory analyses showed that the presence of NOTCH1 mutation correlated with enrichment of IFN-I signatures and reduced infiltration of B cells and neutrophils. In murine models, comparative single-cell transcriptome analyses further revealed that Notch1 deficiency facilitated a more immunologically activated tumor microenvironment which potentiated anti–PD-1 treatment. CONCLUSION: Our data provide novel insights for anti–PD-1 treatment selection using NOTCH1 mutations and may provide a rationale for combination therapy in ESCC.

Immune checkpoint inhibitors (ICIs) have become one of the most promising approaches in cancer immunotherapy.^1,2 In the field of gastrointestinal (GI) cancer, ICIs have made great breakthroughs in advanced esophageal, gastric, and deficient mismatch repair (dMMR) colorectal cancers, showing encouraging efficacy and bringing new hope for patients with cancer.^3-6