Background The H&E stromal tumor-infiltrating lymphocyte (sTIL) score and programmed death ligand 1 (PD-L1) SP142 immunohistochemistry assay are prognostic and predictive in early-stage breast cancer, but are operator-dependent and may have insufficient precision to characterize dynamic changes in sTILs/PD-L1 in the context of clinical research. We illustrate how multiplex immunofluorescence (mIF) combined with statistical modeling can be used to precisely estimate dynamic changes in sTIL score, PD-L1 expression, and other immune variables from a single paraffin-embedded slide, thus enabling comprehensive characterization of activity of novel immunotherapy agents. Methods Serial tissue was obtained from a recent clinical trial evaluating loco-regional cytokine delivery as a strategy to promote immune cell infiltration and activation in breast tumors. Pre-treatment biopsies and post-treatment tumor resections were analyzed by mIF (PerkinElmer Vectra) using an antibody panel that characterized tumor cells (cytokeratin-positive), immune cells (CD3, CD8, CD163, FoxP3), and PD-L1 expression. mIF estimates of sTIL score and PD-L1 expression were compared to the H&E/SP142 clinical assays. Hierarchical linear modeling was utilized to compare pre- and post-treatment immune cell expression, account for correlation of time-dependent measurement, variation across high-powered magnification views within each subject, and variation between subjects. Simulation methods (Monte Carlo, bootstrapping) were used to evaluate the impact of model and tissue sample size on statistical power. Results mIF estimates of sTIL and PD-L1 expression were strongly correlated with their respective clinical assays (p < .001). Hierarchical linear modeling resulted in more precise estimates of treatment-related increases in sTIL, PD-L1, and other metrics such as CD8+ tumor nest infiltration. Statistical precision was dependent on adequate tissue sampling, with at least 15 high-powered fields recommended per specimen. Compared to conventional t-testing of means, hierarchical linear modeling was associated with substantial reductions in enrollment size required (n = 25➔n = 13) to detect the observed increases in sTIL/PD-L1. Conclusion mIF is useful for quantifying treatment-related dynamic changes in sTILs/PD-L1 and is concordant with clinical assays, but with greater precision. Hierarchical linear modeling can mitigate the effects of intratumoral heterogeneity on immune cell count estimations, allowing for more efficient detection of treatment-related pharmocodynamic effects in the context of clinical trials. Trial registration NCT02950259.

Metaplastic breast cancer is a rare and often chemo-refractory subtype of breast cancer with poor prognosis and limited treatment options. Recent studies have reported overexpression of programmed death ligand 1 (PD-L1) in metaplastic breast cancers, and there are several reports of anti-PD-1/L1 being potentially active in this disease. In this case series, we present 5 patients with metastatic metaplastic breast cancer treated with anti-PD-1-based therapy at a single center, with 3 of 5 cases demonstrating a response to therapy, and one of the responding cases being a metaplastic lobular carcinoma with low-level hormone receptor expression. Cases were evaluated for PD-L1 expression, tumor infiltrating lymphocytes (TILs), DNA mutations, RNA sequencing, and T-cell receptor sequencing. Duration of the response in these cases was limited, in contrast to the more durable responses noted in other recently published reports.

1015 Background: Atezolizumab (anti-PD-L1) plus nab-paclitaxel was shown to improve outcomes in mTBNC in a phase III clinical trial. Subjects were required to be > 12 months from curative-intent therapy in this trial. It remains unknown whether non-taxane chemo + anti-PD-1/L1 will be beneficial in mTNBC, or whether this approach is effective in rapidly-progressing patients ( < 12 mo from curative-intent therapy). Methods: mTNBC patients were enrolled in a phase Ib study of anti-PD-1 (pembro, 200mg IV q3w) plus physician’s choice chemo (cape: n = 14, 2000mg BID, 7d on/7d off; or taxol: n = 14, 80mg/m2 q1w). Primary/secondary objectives were to evaluate safety/tolerability (primary) and RECIST1.1 response (w12). The exploratory objective was to explore for differences in immunomodulation according to chemo choice. Mixed effects models were employed to compare the longitudinal effects of chemo on peripheral immune cells (flow cytometry) and T-cell diversity (Immunoseq assay). Results: Enrollment of the trial is complete (n = 28), with 100% of evaluable patients tolerating therapy (n = 22) as of 2/1/2019. Cape ORR was 43% (5 PR, 1 CR, 2 SD) with median PFS = 155d. Taxol ORR was 25% (1 CR, 1 PR, 3 SD) with median PFS = 99d. Subjects enrolled < 12 months from curative-intent therapy had numerically lower response (ORR = 27%, 1 CR, 2 PR, 3 SD) than subjects without rapid progression (ORR = 45%, 1 CR, 4 PR, 2 SD). No significant differences in immunomodulation were observed according to chemo type, however both cape & taxol were associated with declines in T-cell quantity (CD4 p < .02, CD8 p < .04) and Immunoseq T-cell fraction over time. Conclusions: Pembro plus cape or taxol is safe with encouraging efficacy, however activity may be lower in the setting of rapid progression following curative-intent chemo. Cape+pembro efficacy is favorable with no measurable differences in immunomodulation, and therefore cape may be preferred as a chemo backbone in selected patients. Both cape and taxol are associated with iatrogenic declines in T-cell quantity, which may explain the observed dropoff in anti-PD-1/L1 activity in later lines. Clinical trial information: NCT02734290.

BackgroundChemoimmunotherapy is a standard treatment for triple-negative breast cancer (TNBC), however, the impacts of different chemotherapies on T-cell populations, which could correlate with clinical activity, are not known. Quantifying T-cell populations with flow cytometry and T-cell receptor (TCR) immunosequencing may improve our understanding of how chemoimmunotherapy affects T-cell subsets, and to what extent clonal shifts occur during treatment. TCR immunosequencing of intratumoral T cells may facilitate the identification and monitoring of putatively tumor-reactive T-cell clones within the blood.MethodsBlood and tumor biopsies were collected from patients with metastatic TNBC enrolled in a phase Ib clinical trial of first or second-line pembrolizumab with paclitaxel or capecitabine. Using identical biospecimen processing protocols, blood samples from a cohort of patients treated for early-stage breast cancer were obtained for comparison. Treatment-related immunological changes in peripheral blood and intratumoral T cells were characterized using flow cytometry and TCR immunosequencing. Clonal proliferation rates of T cells were compared based on intratumoral enrichment.ResultsWhen combined with pembrolizumab, paclitaxel and capecitabine resulted in similar time-dependent lymphodepletions across measured peripheral T-cell subsets. Their effects were more modest than that observed following curative-intent dose-dense anthracycline and cyclophosphamide (ddAC) (average fold-change in CD3+ cells, capecitabine: −0.42, paclitaxel: −0.56, ddAC: −1.21). No differences in T-cell clonality or richness were observed following capecitabine or paclitaxel-based treatments. Regression modeling identified differences in the emergence of novel T-cell clones that were not detected at baseline (odds compared with ddAC, capecitabine: 0.292, paclitaxel: 0.652). Pembrolizumab with paclitaxel or capecitabine expanded T-cell clones within tumors; however, these clones did not always expand within the blood. Proliferation rates within the blood were similar between clones that were enriched and those that were not enriched within tumors.ConclusionChemoimmunotherapy for metastatic TNBC with pembrolizumab and capecitabine or paclitaxel resulted in similar peripheral T-cell subset lymphodepletion without altering T-cell clonal diversity. Regression modeling methods are applicable in immune monitoring studies, such as this to identify the odds of novel T-cell clones emerging during treatment, and proliferation rates of tumor-enriched T-cell clones.

TPS1106 Background: ICB (atezolizumab, anti-PD-L1) is known to improve survival when added to chemo, however only in PD-L1-positive, triple-negative MBC. ICB is less effective in hormone receptor positive (HR+) MBC, or when administered following palliative chemo. Novel approaches are required to broaden clinical benefit of ICB, particularly in PD-L1-negative, HR+, or chemo-experienced MBC. Dual ICB with anti-PD-1 (nivolumab) and anti-CTLA-4 (ipilimumab) is associated with enhanced activity in melanoma other malignancies, but has not been explored extensively in MBC. Androgen receptor (AR) blockade, in addition to known direct cytostatic effects in AR-expressing MBCs (50% of TNBC, > 75% of HR+ MBC), may also modulate immune response. AR blockade has been shown experimentally to stimulate thymic production of naïve T-cell clones, which in turn can facilitate de novo anti-tumor immune responses. Concurrent ICB can enhance the activity of these T-cell clones by interfering with PD-1-mediated peripheral tolerance. This combination approach is promising in MBC in light of known AR positivity, and the routine use of lymphodepleting chemo regimens in the curative-intent setting. Methods: This is a phase II trial of dual immune checkpoint blockade (nivolumab 240mg IV q2w; ipilimumab 1mg/kg IV q6w) plus AR blockade (bicalutamide, 150mg PO daily, dose reduction allowed) in triple-negative MBC (cohort A: AR-positive [ > 1% by IHC]; cohort B: AR-negative) or HR+ MBC (cohort C) in subjects who received 0/1 prior chemotherapies in the non-curative setting. Objectives include 24-week clinical benefit rate by iRECIST (primary), safety (CTCAE v4.0), and other response measures (RECIST1.1, PFS, OS). Efficacy for each cohort is defined as > 20% improvement in response over historical control (30% per EMBRACE clinical trial) employing a Simon 2-stage design to minimize futility (n = 46/cohort, stage I n = 15). Thymic generation of T-cells will be measured via quantitative deep sequencing of T-cell receptors (TcR, ImmunoSEQ assay) and TcR excision circles (TRECs), as well as real-time flow cytometry using surrogate cell surface markers of recent thymic emigration. Enrollment has commenced, sites: Earle A. Chiles Research Institute (Portland, OR), Memorial Sloan Kettering Cancer Center (New York, NY). Clinical trial information: NCT03650894.