Abstract
Introduction: The immune tumor microenvironment (iTME) of soft-tissue sarcomas (STS) is not well characterized. The use of tissue microarray (TMA) methods is an attractive means of high-throughput immunohistochemistry-based immune profiling, Typically, between 1-3 tumor cores per case are included on conventional TMAs that are well validated for many protein markers across multiple tumor types. However, the reproducibility/representativeness of TMA for sarcomas is not fully established, and a significant factor is the marked intratumoral morphologic heterogeneity in STS. Here, we examine the variability of lymphoid infiltrate between different regions of the same tumor and how this might be optimally represented on a TMA.
Methods: Multiple FFPE blocks, representing spatially discrete tumor areas, were selected from primary surgical specimen of 46 cases of localized leiomyosarcoma (LMS) and 9 cases of localized synovial sarcoma (SS) identified from institutional diagnostic archive. Following sectioning and immunohistochemistry for lymphocyte markers, average tumor-infiltrating lymphocytes (TIL) number per section was manually assessed and scored as a numerical variable. Variability of TIL numbers within and between cases was then statistically assessed through two-way ANOVA and intraclass correlation
To determine the optimal number of 1-mm TMA cores that need to be sampled to provide representative overview of immune microenvironment, we performed iterative sampling of 1 to 20 virtual cores from digital microscopy images taken from representative blocks from 46 LMS cases. TILs within each virtual core area were digitally counted, with the average TIL number from the combination of increasing number of cores compared to the tumor overall TIL value.
Results: We found a dynamic range in TIL numbers between individual cases of LMS but little variability in low TIL numbers seen between SS cases. Focusing on infiltrating T-lymphocytes in LMS, we found marked variation of TIL numbers between discrete areas of highly infiltrated tumors. However, 2-way ANOVA showed that intratumoral variance only accounted for 0.2% of variation in TILs, whilst intertumoral variation accounted for 54.1% of variance, suggesting greater intertumor than intratumor heterogeneity in lymphocyte infiltration in the LMS cohort.
In our virtual core experiment, many more TMA cores (median = 11) than the conventional 1-3 were required in a sample to provide an accurate picture of the degree of an individual tumor's lymphocyte infiltration in an individual tumor (within +/-20% of mean TIL number). However, when the 46-case cohort was divided into high or low lymphocyte-infiltrated tumors around the cohort median TIL value, a single core was sufficient to correctly identify a tumor as being TIL high or TIL low in the majority of cases.
Conclusion: The variability in TIL number between LMS cases (intertumor variation) in our cohort was much greater than the variability between different tissue blocks from the same individual tumor (intratumor variation), indicating that choice of tumor-containing block from those available for a given specimen may not be an important source of variation when studying the immune microenvironment of STS cohorts. We found that using a conventional number of TMA cores was sufficient to distinguish between tumors with high and low degrees of lymphocyte infiltration. However, in many cases, far more than 3 cores were required to accurately recapitulate the precise degree of infiltration in a tumor. Researchers investigating the iTME of STS should be wary of the apparent limitations of conventional TMA approaches in this field. The extent and biologic significance in the variability of lymphocyte distribution in STS lesions requires further investigation.
Citation Format: Alex Lee, Khin Thway, Seth Pollack, Ian Judson, Paul Huang, Robin Jones. The accuracy of tissue microarrays in the study of the sarcoma immune microenvironment is dependent on the number of sampled cores [abstract]. In: Proceedings of the AACR Conference on Advances in Sarcomas: From Basic Science to Clinical Translation; May 16-19, 2017; Philadelphia, PA. Philadelphia (PA): AACR; Clin Cancer Res 2018;24(2_Suppl):Abstract nr B28.