Spatial RNA–protein profiling of antibody–drug conjugate (ADC) targets in leiomyosarcoma (LMS).

Abstract

11509 Background: LMS remains a therapeutically challenging sarcoma with limited targeted options. Although ADCs are a promising strategy, none are FDA approved for sarcoma. Prior studies relying on bulk RNA or single-marker immunohistochemistry fail to capture spatial localization, co-expression, and intratumoral heterogeneity. Methods: LMS tissue microarrays were profiled using GeoMx DSP to generate spatial RNA and protein expression of ADC targets across multiple regions of interest (ROIs) per tumor. Cases were classified as soft-tissue (ST-LMS; n=23 patients; 94 ROIs) or uterine LMS (U-LMS; n=32 patients; 92 ROIs). A prespecified panel of 35 clinically relevant ADC targets was interrogated. Expression values were normalized to housekeeping genes with group-level scaling. Unsupervised clustering grouped ROIs by spatial expression. Intratumoral and inter-tumoral heterogeneity was quantified using coefficients of variance (CV) across ROIs. RNA–protein concordance and dual target co-expression were quantified using Spearman correlation (ρ). Survival differences between high- and low-expression groups, dichotomized by the median, were evaluated using Kaplan–Meier analysis. Results: Of the 35 ADC targets interrogated, 33 were evaluable at the RNA level and 10 at the protein level. Spatial profiling identified four distinct expression clusters. At the RNA level, F3, AXL, and PTK7 showed the highest expression, whereas CDH11, CD99, and ITGA11 exhibited lower expression. At the protein level, Tissue Factor, AXL, and CD70 showed the highest expression, while HER2, and CD99 were lower. Most ADC targets exhibited positive log₁₀ (CV) values, indicating substantial spatial variability. Across most targets, U-LMS showed greater intratumoral and intertumoral heterogeneity than ST-LMS. RNA–protein concordance was greatest for CD99, CD47, and Fibronectin/EDB (ρ≈0.44–0.69) and minimal for PDGFRA, AXL, ERBB2, EGFR, and MET (ρ≈-0.15-0.19). RNA co-expression identified seven significant pairs in ST-LMS (e.g., ROR2–FAP, ρ=0.54) and eighteen in U-LMS, including MRC2–CD248 (ρ=0.73). Dual protein co-expression revealed a strong ERBB2–EGFR correlation in both ST and U-LMS (ρ≈0.79–0.81), with moderate correlations of these two ADC targets to PDGFRA, AXL, and Tissue Factor (ρ≈0.53–0.70). High expression levels CD47 and AXL were associated with improved survival in ST-LMS (CD47 HR = 0.76, 95% CI 0.08–0.88, p = 0.021; AXL HR = 0.67, 95% CI 0.07–0.72, p = 0.0067). Conclusions: This study provides the first spatially resolved RNA–protein map of ADC targets in LMS, revealing differential expression, spatial heterogeneity, variable RNA–protein concordance, dual co-expression patterns, and survival associations. These findings establish a foundation for rational ADC development, biomarker-driven patient stratification, and dual ADC combination strategies in LMS.