Circulating tumor cells in metastatic breast cancer patients treated with immune checkpoint inhibitors - a biomarker analysis of the ALICE and ICON trials

Abstract

Immune checkpoint inhibitors (ICIs) have been introduced in breast cancer (BC) treatment and better biomarkers are needed to predict benefit. Circulating tumor cells (CTCs) are prognostic in BC, but knowledge is limited on CTCs in the context of ICI therapy. In this study, serial sampling of CTCs (CellSearch system) was evaluated in 82 patients with metastatic BC enrolled in two randomized trials investigating ICI plus chemotherapy. Programmed death-ligand 1 (PD-L1) expression on CTCs was also measured. Patients with ≥ 2 CTCs per 7.5 mL at baseline had gene expression profiles in tumor suggestive of increased T-cell activity, including increased tumor inflammation signature (TIS) in both triple-negative (P = 0.010) and hormone receptor-positive (P = 0.024) disease. Patients with luminal A BC had higher CTC levels. The association between CTC status and outcome was most apparent 4 weeks into therapy. PD-L1 expression in CTCs was observed in 6/17 CTC-positive patients and was associated with inferior survival. In conclusion, our study indicates that CTC numbers may inform on tumor immune composition, as well as prognosis. These findings suggest a potential of using CTCs as an accessible biomarker source in BC patients treated with immunotherapy.

Keywords: breast cancer; circulating tumor cells; immune checkpoint inhibitors; liquid biopsy.

Fig. 1

Overview of treatment cohorts and CTC sampling. For the purpose of the current study, CTC data from the 6 months of sampling in the main treatment cohorts have not been included in this analysis. atezo, atezolizumab; CTCs, circulating tumor cells; EOT, end of treatment; HER2⁻, human epidermal growth factor receptor 2‐negative; HR⁺, hormone receptor‐positive; ipi, ipilimumab; mBC, metastatic breast cancer; nivo, nivolumab; TNBC, triple‐negative breast cancer.

Fig. 2

Baseline CTC distribution by histological and molecular subtype. Panel (A) presents CTC counts at baseline in the TNBC (n = 21; left) and HR⁺ (n = 34; right) population. Panel (B) presents the distribution of patients in each cohort relative to the ≥5 CTCs per 7.5 mL cutoff at baseline. The distribution of CTC counts by the molecular subtype is presented in panel (C). Subtype classification was based on gene expression analysis of archival FFPE samples. In patients with more than one sample typed, the most recent sample available was used for classification. The boxes in (A) and (C) represent the middle half of the data between the 25th and 75th percentiles, and the line represents the median. CTCs, circulating tumor cells; HER2E, HER2 enriched; HR⁺, hormone receptor‐positive; TNBC, triple‐negative breast cancer.

Fig. 3

Baseline CTC counts vs. tumor tissue PD‐L1 status and TIL score. Associations between the baseline CTC count (≥ 2 CTCs per 7.5 mL) and tissue PD‐L1 status by IHC and TIL score were analyzed. The bar chart in panel (A) shows the distribution of PD‐L1 status by the SP142 assay in the CTC high and low group for the TNBC (left) and HR⁺ (right) population. The bar chart in panel (B) presents the distribution of TIL score high and low vs. CTC counts for both breast cancer subtypes. PD‐L1 status was analyzed in the most recent archival sample available. The TIL score was assessed in a study screening biopsy if sufficient material, otherwise in the most recent archival biopsy available. CTCs, circulating tumor cells; IHC, immunohistochemistry; PD‐L1, programmed death‐ligand 1; TILs, tumor‐infiltrating lymphocytes.

Fig. 4

Baseline CTC counts vs. immune‐related gene expression signatures. Panels (A–D) presents the most differentially expressed (P < 0.05) immune gene signatures from the NanoString BC360 assay in patients with low CTCs (n = 15) vs. high CTCs (n = 5) at baseline (≥ 2 CTCs per 7.5 mL cutoff) within the TNBC population. Panels (E–J) presents the most differentially expressed signatures in CTC low (n = 15) vs. CTC high (n = 18) patients within the HR⁺ population. The boxes represent the middle half of the data between the 25th and 75th percentiles, and the line represents the median. P values are calculated with the t‐test. CTCs, circulating tumor cells; HR⁺, hormone receptor‐positive; IFN‐g, interferon gamma; PD‐L2, programmed cell death ligand‐2; TGF‐beta, transforming growth factor‐β; TIS, tumor inflammation signature; TNBC, triple‐negative breast cancer.

Fig. 5

Overall survival by baseline CTC count. The figure presents Kaplan–Meier plots of OS by the ≥ 2 CTCs per 7.5 mL cutoff in each of the four treatment cohorts in (A–D). OS by the ≥ 5 CTCs per 7.5 mL cutoff in each cohort is presented in (E–H). atezo, atezolizumab; CTCs, circulating tumor cells; HR, hazard ratio; HR⁺, hormone receptor‐positive; ipi, ipilimumab; nivo, nivolumab; OS, overall survival; PFS, progression‐free survival; TNBC, triple‐negative breast cancer.

Fig. 6

Survival outcomes by week 4 CTC count (≥ 2 CTCs per 7.5 mL). The figure presents Kaplan–Meier plots of survival outcomes by CTCs measured at week 4 by the ≥ 2 CTCs per 7.5 mL cutoff in each of the four treatment cohorts. PFS is presented in panel (A–D), and OS is presented in panel (E–H). atezo, atezolizumab; CTCs, circulating tumor cells; HR, hazard ratio; HR⁺, hormone receptor‐positive; ipi, ipilimumab; nivo, nivolumab; OS, overall survival; PFS, progression‐free survival; TNBC, triple‐negative breast cancer; w4, week 4.

Fig. 7

CTC dynamics and survival outcomes in patients receiving ipi/nivo‐only. The diagram in (A) presents the observed CTC dynamics in the HR⁺ ipi/nivo‐only cohort (n = 12) with CTC measurements at baseline, week 4, 6 months, and the time of progression. Cutoff for positivity was ≥ 1 CTC. Three patients had measurable target lesion reduction, with two of the three developing durable responses. Panel (B, C) presents Kaplan–Meier plots of PFS and OS in the ipi/nivo‐only cohort by baseline CTCs by ≥ 2 CTCs per 7.5 mL. Panels (D, E) presents Kaplan–Meier plots of PFS and OS by CTC measured 4 weeks into therapy by the ≥ 2 CTCs per 7.5 mL cutoff. CTCs, circulating tumor cells; HR, hazard ratio; HR⁺, hormone receptor‐positive; ipi, ipilimumab; nivo, nivolumab; OS, overall survival; PFS, progression‐free survival; w4, week 4.

Fig. 8

CTC PD‐L1 expression. The diagram in panel (A) presents the 31 patients with baseline samples analyzed with the D8T4X/CXC assay. OS by baseline CTC PD‐L1 status regardless of treatment cohort is presented in (B). The bar chart in panel (C) presents the tumor PD‐L1 status by IHC with the SP142 assay in CTC PD‐L1‐negative (n = 11) vs. PD‐L1‐positive (n = 6) patients at baseline. Panel (D) presents the proportion with high or low tumor‐infiltrating lymphocytes by CTC PD‐L1 status at baseline. The box plots in (E, F) present the tumor expression levels of TIS and PD‐L1 gene expression signature by CTC PD‐L1 status. The boxes represent the middle half of the data between the 25th and 75th percentiles, and the line represents the median. CTCs, circulating tumor cells; HR, hazard ratio; OS, overall survival; PD‐L1, programmed death‐ligand 1; TILs, tumor‐infiltrating lymphocytes; TIS, tumor inflammation signature.