The clinical relevance of circulating tumor cell clusters (CTC-clusters) in breast cancer (BC) has been mostly studied using the CellSearch^®, a marker-dependent method detecting only epithelial-enriched clusters. However, due to epithelial-to-mesenchymal transition, resorting to marker-independent approaches can improve CTC-cluster detection. Blood samples collected from healthy donors and spiked-in with tumor mammospheres, or from BC patients, were processed for CTC-cluster detection with 3 technologies: CellSearch^®, CellSieve™ filters, and ScreenCell^® filters. In spiked-in samples, the 3 technologies showed similar recovery capability, whereas, in 19 clinical samples processed in parallel with CellSearch^® and CellSieve™ filters, filtration allowed us to detect more CTC-clusters than CellSearch^® (median number = 7 versus 1, p = 0.0038). Next, samples from 37 early BC (EBC) and 23 metastatic BC (MBC) patients were processed using ScreenCell^® filters for attaining both unbiased enrichment and marker-independent identification (based on cytomorphological criteria). At baseline, CTC-clusters were detected in 70% of EBC cases and in 20% of MBC patients (median number = 2, range 0–20, versus 0, range 0–15, p = 0.0015). Marker-independent approaches for CTC-cluster assessment improve detection and show that CTC-clusters are more frequent in EBC than in MBC patients, a novel finding suggesting that dissemination of CTC-clusters is an early event in BC natural history.

Although the detection of CTCs expressing HER2 at low intensity (HER2-low CTCs) has been shown to have a negative prognostic value in metastatic breast cancer (MBC) patients, the biological intrinsic nature of HER2-low CTCs remains unexplored. Considering the technical challenges behind the selective collection of immunophenotype-specific CTCs, we developed a pipeline to individually capture HER2-low CTCs. Four different breast cancer cell lines (MDA-MB-231, T47D, MDA-MB-453, and SKBR3), that are known to express HER2 at different immunohistochemistry levels (respectively classified as 0, 1+, 2+, and 3+), were spiked in healthy donor blood tubes (7.5 mL) and processed with the CellSearch^® (Menarini Silicon Biosystems, Bologna, Italy) for enrichment and the DEPArray NxT^™ for single cell selection. The HER2 signal-intensities of each cell line was compared using the nonparametric Mann–Whitney U test. The optimal cut-offs to distinguish HER2 1+ from 0 and 2+ cells were calculated performing the Receiver operating characteristic (ROC) curve. Median HER2 signal-intensities detected with the DEPArray NxT^™ were: 2.59 (0), 3.58 (1+), 5.23 (2+) and 38.37 (3+). DEPArray NxT efficiently differentiated each single cell line (p < 0.001). The area under the ROC curve was 0.69 and 0.70 (respectively 0 vs. 1+ and 1+ vs. 2+) and the optimal calculated cut-offs were 2.85 (lower) and 4.64 (upper). HER2-low CTCs can be detected and separately collected using predetermined intensity cut-offs. This study will allow standardized single-cell or pooled collection of HER2-low CTCs for downstream molecular analyses.