Presenter(s)
O. M. Oderinde1,2, R. Chimmula3, O. DeHaven4, G. Durm5, P. A. Anthony6, T. Perekattu Kuruvilla6, G. E. Sandusky4, and B. Molden4; 1Department of Radiation Oncology, Indiana University School of Medicine, Indianapolis, IN, 2Advanced Molecular Imaging in Radiotherapy (AdMIRe) Research Laboratory, School of Health Sciences, Purdue University, West Lafayette, IN, 3Advanced Molecular Imagiing in Radiotherapy (AdMIRe) Research Laboratory, School of Health Sciences, Purdue University, West Lafayette, IN, 4Department of Pathology, Indiana University School of Medicine, Indianapolis, IN, Indianapolis, IN, 5Department of Medical Oncology, Indiana University School of Medicine, Indianapolis, IN, 6Department of Radiation Oncology, Indiana University School of Medicine, Indianapolis,IN, Indianapolis, IN
Purpose/Objective(s): Biopsy-derived immune biomarkers are increasingly used to stratify patients for radiotherapy (RT) and RT–immunotherapy combinations in head and neck squamous cell carcinoma (HNSCC). However, the spatial heterogeneity of the tumor microenvironment raises uncertainty regarding whether diagnostic biopsies adequately represent whole-tumor immune architecture. We evaluated biopsy–resection concordance and investigated how microanatomic compartment composition influences immune sampling adequacy.
Materials/Methods: Four patients with HNSCC who underwent pre-treatment biopsy followed by surgical resection were included. Hematoxylin and Eosin (H&E) blocks for these patients were retrieved, and immunohistochemical stains were performed using Lymphocytes (CD8), and macrophages (CD163) markers, and subsequently digitized. Tumor, stromal, and necrotic compartments were manually annotated by an anatomic pathologist, followed by automated compartment-specific immune cell detection on the digitized H&E slides. Immune density (cells/mm²), and percent positivity were quantified. Percent differences between biopsy and resection values were calculated. Tumor-adjusted immune density was computed to account for differences in tumor fraction.
Results: Biopsy specimens contained higher tumor fractions (40–90%) compared with resections (10–50%), demonstrating substantial microanatomic composition differences. Raw CD8 density showed moderate concordance (median absolute deviation 13.4%; range 3.7–57.1%), with 3 of 4 patients demonstrating =25% deviation. In contrast, CD163 density exhibited greater variability (median deviation 90.9%; range 8.2–177.9%). After tumor-fraction normalization, CD8 concordance improved in 3 of 4 patients, including near-equivalent tumor-adjusted densities in one case (18,294 vs 19,410 cells/mm²). CD163 variability paralleled stromal and non-tumor compartment differences, suggesting spatial clustering of macrophage populations.
Conclusion: Biopsy adequacy for immune profiling is marker- and compartment-dependent. CD8 infiltration demonstrated relative spatial stability despite microanatomic differences, whereas CD163 quantification was strongly influenced by stromal and non-tumor heterogeneity. These findings suggest biopsy-based immune stratification in RT research may be reliable for lymphocyte-driven phenotypes but may underrepresent compartment-sensitive myeloid states. This study provides a compartment-aware framework for evaluating biopsy adequacy in spatially heterogeneous tumors. More cases are needed to make a definitive conclusion.