Fusion 3D gross sampling method to overcome heterogeneity in clear cell renal cell carcinoma (ccRCC) and grading angiogenic versus immune signatures.

Abstract

e16565 Background: Grading tissue signatures in ccRCC is a potential tool to improve patients’ selection for anti-angiogenic drugs and immunotherapies. After the molecular audit to the TCGA public platforms, we aimed to grade angiogenic and immune signatures in ccRCC using a new 3D next-generation gross sampling method to overcome intratumoral heterogeneity. Methods: 100 consecutive advanced ccRCCs (≥pT3a) were sampled. Paraffin-embedded blocks were obtained after mapping the position of each sample to the whole tumor, to allow the reconstruction of the entire 3D tumor mass ( fusion 3D). Multisite tumor sampling was performed to analyze the whole tumor. TCGA platforms were assessed for angiogenic and immune molecular signatures: CD31 and CD34 to evaluate the absolute count and density of vessels, while E1L3N and sp263 clones for PD-L1 expression in tumor cells (TCs). The digital analysis was performed with image processing: comparing each tissue block to whole 3D assessment, the coefficient of variation (CV) was the statistical measure of the dispersion of data points in the data series around the mean. CV < 0.2 defined the homogeneity of the assessment. Results: 656 gross photographs representative of the 3D tumoral masses were collected and 4231 paraffin blocks and tissue sections were stored. Matching gross photographs with tissue samples was performed. 6324 tissue cores were evaluated after combining standard routine sampling plus the 3D multisite sampling and tissue microarray cores. Only 18% of cores displayed homogeneous profile of angiogenesis (CV < 0.2) with two distinct patterns: homogeneous high level of angiogenesis (pattern A) (10% of cases) or homogeneous low level of angiogenesis (pattern B) (8% of cases). The heterogeneous profile of angiogenesis was more frequent than the homogeneous one and was characterized by zones with high and low density of angiogenesis (82% of cases) (pattern C). On the other hand, the homogeneity of PD-L1 expression was more frequently observed both as diffuse absence ( < 1% of TCs, grade 0) or high expression (≥50% of TCs, grade 2) compared to low PD-L1 expression (1-49% of TCs, grade 1) (60% plus 7% versus 33% of cases, respectively). After the comparison of grading angiogenic versus immune signatures, we observed that cases with low PD-L1 expression (grade 0/1) usually expressed high density of angiogenesis (pattern A). Conclusions: Grading angiogenic (pattern A, B and C) versus immune (grade 0, 1 and 2) signatures in ccRCCs can be performed using commonly available tissue vascular (CD31 or CD34) and immune (PD-L1) antibodies. We promoted a simple assay to perform fusion 3D gross sampling to reduce at minimum the bias of heterogeneity in RCCs analyses. Both angiogenic versus immune signature by using the grading systems may help treatment decision-making and response assessment in ccRCC patients.