Comparison of combined positive score, Salgado TIL score, immunofluorescence and single cell RNA sequencing for predicting response to therapy

• Published in: Gynecologic oncology Vol. 162; p. S117
• Main Authors: Johnson, Andrea, Chandra, Shruthi, Shetty, Mihir, Cepela, Jason, Talukdar, Shobhana, Mattson, Jordan, Macchietto, Marissa, Henzler, Christine, Nelson, Andrew, Wang, Jinhua, Winterhoff, Boris, Mullany, Sally, Starr, Timothy
• Format: Journal Article
• Language: English
• Published: Elsevier Inc 01.08.2021
• ISSN: 0090-8258; 1095-6859
• DOI: 10.1016/S0090-8258(21)00864-7

Immune cell infiltration is associated with positive ovarian cancer prognosis and is used to predict response to emerging immunotherapies. Multiple methods are used both clinically and in research to quantify immune infiltration, but there is a lack of consensus on the best method. The objective of this study was to compare four methods of measuring immune infiltration: immunohistochemistry for PDL1 (Combined Positive Score - CPS), H&E analysis of immune infiltration (Salgado TIL score), multiparameter immunofluorescence (IHC), and single cell RNA sequencing (scRNAseq). This exploratory work compares these four methods and calculates associations with disease outcomes. Thirty biopsies were collected. Sections were stained with anti-PDL1 or H&E and scored by a molecular pathologist to generate CPS and Salgado TIL scores. Additional slides were stained with antibodies for CD3, CD8, CD56, CD19, FOXP3, cytokeratin, and DAPI. Whole slides were analyzed using CD3 and CD8 density to quantify infiltration. Single cell gene expression was measured using the 10x Genomics Single Cell 3’ Protocol. Sequencing was performed to a depth of at least 50,000 paired-end reads per cell. Cells were annotated using the Seurat and SingleR R packages. Correlation coefficients of immune cell infiltration estimates by the four methods were generated and visualized on scatter plots. In evaluating immune cell infiltration, there was a moderate correlation between Salgado TIL scores and IHC with an R-squared value of 0.59. In evaluating T cell populations, CPS and scRNASeq did not correlate, with R-squared value of 0.0167. In our dataset, a higher CPS score was positively associated with platinum sensitivity. Single methodologies to estimate immune infiltration may not be sufficient to capture the complexity of the tumor microenvironment. Even though the correlation between methods was not high, it is possible that each method measures a different aspect of immune cell infiltration and the combination of methods may improve our ability to predict response to immunotherapies. Further analyses will be required to determine the biological properties that result in the different estimates.