(P 100) HISTOLOGY SPECIFIC INTRATUMORAL B-CELL BIOLOGY IN ALVEOLAR SOFT-PART SARCOMAS (ASPS) IN RESPONSE TO DURVALUMAB-TREMELIMUMAB (D-T)

Objective: ASPS is characterized by the ASPSCR1-TFE3 translocation and exceptional response to immunotherapy as demonstrated in the recent phase II clinical trial in advanced sarcoma patients with select histotypes, including those with ASPS. In other tumor types (undifferentiated pleomorphic sarcomas [UPS], melanoma, renal cell carcinoma), intratumoral B-cells and tertiary lymphoid structures (TLS) are important predictive biomarkers of response to immunotherapy. Therefore, we sought to investigate the intratumoral B-cell biology of ASPS patients compared to other patients included in this trial.

Methods: This trial enrolled 62 sarcoma patients (10 ASPS, 6 liposarcoma, 5 angiosarcoma, 5 leiomyosarcoma, 5 UPS, 5 synovial sarcoma, 5 osteosarcoma, 5 chordomas, 11 other). The primary endpoint of this trial was progression-free rate at 12 weeks. Longitudinal biospecimen collection included biopsies at baseline (n=56; ASPS n=8), at week 6 (optional; n=43; ASPS n=8), and at progression (optional, n=2). Whole hematoxylin and eosin slides were evaluated for presence of lymphoid aggregates (LA) by an experienced sarcoma pathologist, which was used as a surrogate marker for TLS and reported as present or absent. Biopsies were stained for CD20 by immunohistochemistry (IHC). The number of CD20 positive cells quantified and reported as cell densities. RNA sequencing (RNAseq) was performed and immune cell populations were inferred using MCPcounter immune deconvolution method and reported as z-scores. Intratumoral B-cell receptor (BCR) repertoire at baseline was evaluated by bulk tumor RNAseq using the TRUST4 algorithm and BCR repertoire diversity and clonal distribution were assessed by inverse Simpson index and Gini coefficient, respectively. Comparisons of continuous variables were performed using Wilcoxon Rank-Sum test and categorical variables using Fisher’s exact test.

Results: ASPS samples showed no increase in the proportion of samples with presence of LA (p > 0.9) whereas non-ASPS samples showed a significant increase in the number of samples with presence of LA with D-T (p = 0.038; Figure 1A). ASPS samples showed a non-significant decrease in CD20 densities with D-T (p = 0.22), whereas non-ASPS samples had an increase in CD20 positive cell densities with D-T (p = 0.012; Figure 1B). At baseline, ASPS had higher densities of CD20 positive cells compared to non-ASPS (p = 0.025; Figure 2A) which was confirmed with RNAseq (p < 0.001; Figure 2B). Concerning the BCR repertoire, there were no significant differences in diversity between ASPS and non-ASPS patients, neither for the heavy nor light chains. However, there was a non-significant trend for greater clonality in ASPS compared to non-ASPS both in the heavy and light chains at baseline and after D-T (Figure 3).

Conclusion: The way in which the B-cells in ASPS may contribute to the exceptional responses seen with D-T may be related more to the pre-existing numbers of B-cells and independent of lymphoid aggregation. Further work in this area should focus on the mechanisms underlying this ASPS-specific intratumoral B-cell and to enhance presence of B-cells in other sarcoma types. Additional analyses on these trial samples are ongoing to further delineate the role B-cells in ASPS response to D-T.