Endocrine Abstracts

Background: Peptide receptor radionuclide therapy with ¹⁷⁷Lu-DOTATATE (PRRT) improves outcomes in well-differentiated neuroendocrine tumor (NETs). However, heterogeneous responses and relapse suggest that features within the tumor microenvironment (TME) might mediate treatment resistance. We investigated immune- and stromal remodeling pre- and post PRRT across metastatic NET lesions, correlating findings with clinical response and genomic features to identify determinants of PRRT efficacy.

Methods: We analyzed 34 tumor samples from 12 patients with pancreatic (PNET, n=5) and small bowel NETs (SBNETs, n=7) treated with PRRT that underwent subsequent surgical debulking. Tumor samples included matched pre/post-PRRT tissues from liver metastases categorized as responding, stable, or progressing. Multiplex immunofluorescence profiling enabled TME phenotyping, spatial analysis, and quantification of immune-stromal interactions. Ki-67 proliferation indices and OncoPlus next-generation sequencing were also performed to evaluate somatic mutations and tumor mutational burden (TMB).

Results: Median time between PRRT and surgical debulking was 10.5 months. Four patients had progression in some, 5 patients in all and 3 patients in none of the liver tumors after PRRT and prior to debulking. Across all samples, PRRT increased CD8⁺ T cell infiltration (20%, P = 3.9×10⁻³¹), CD31⁺ vascular remodeling (18%, P < 1×10⁻¹⁰), and PD-L1 upregulation (5%, P = 1.3×10⁻⁴). Spatial analysis confirmed post-PRRT CD8⁺ proximity to PD-L1⁺, FoxP3⁺, and CD31⁺ cells, indicating immune exclusion.Progressing PNET liver metastases showed an immune-silent baseline with CD4⁺ dominance (90%, P < 1×10⁻¹⁰) and minimal CD8⁺ (5%), CD31⁺ (2%), and FoxP3⁺ (0%) cells. In contrast, responding PNET liver metastases exhibited elevated CD8⁺ (20%), CD31⁺ (18%), and PD-L1⁺ (5%) infiltration, while exhausted CD8⁺PD-L1⁺ subsets remained unchanged (3%, P = 0.50). Stable SBNET liver metastases had elevated FoxP3⁺ (16%) and CD68⁺PD-L1⁺ (1%) macrophages while progressing SBNETs showed higher CD8⁺ (56%), PD-L1⁺ (9%), and CD8⁺PD-L1⁺ (2.2%) infiltration. Genomically, PNETs harbored higher Ki-67 indices (~11.33% vs. ~8.9%), a greater number of somatic mutations (average 2.8 vs. 1.5 mutations), and higher TMB (4.54 vs. 3.44 mutations/Mb).

Conclusions: PRRT induces CD8⁺ infiltration and angiogenic remodeling but is limited by immunosuppressive features. SBNETs exhibiting cytotoxic infiltration and PD-L1 expression after PRRT may benefit from immune checkpoint blockade. In contrast, PNETs characterized by CD4⁺ and CD68⁺ dominance with minimal checkpoint expression may require priming with multikinase inhibitors to remodel the TME prior to checkpoint inhibition to enhance PRRT efficacy.

Abstract ID #33448