After a CellSearch-processed circulating tumor cell (CTC) sample is imaged, a segmentation algorithm selects nucleic acid positive (DAPI+), cytokeratin-phycoerythrin expressing (CK-PE+) events for further review by an operator. Failures in this segmentation can result in missed CTCs. The CellSearch segmentation algorithm was not designed to handle samples with high cell density, such as diagnostic leukapheresis (DLA) samples. Here, we evaluate deep-learning-based segmentation method StarDist as an alternative to the CellSearch segmentation. CellSearch image archives from 533 whole blood samples and 601 DLA samples were segmented using CellSearch and StarDist and inspected visually. In 442 blood samples from cancer patients, StarDist segmented 99.95% of CTC segmented by CellSearch, produced good outlines for 98.3% of these CTC, and segmented 10% more CTC than CellSearch. Visual inspection of the segmentations of DLA images showed that StarDist continues to perform well when the cell density is very high, whereas CellSearch failed and generated extremely large segmentations (up to 52% of the sample surface). Moreover, in a detailed examination of seven DLA samples, StarDist segmented 20% more CTC than CellSearch. Segmentation is a critical first step for CTC enumeration in dense samples and StarDist segmentation convincingly outperformed CellSearch segmentation.

Mammalian cells release extracellular vesicles (EVs) into their microenvironment that travel the entire body along the stream of bodily fluids. EVs contain a wide range of biomolecules. The transported cargo varies depending on the EV origin. Knowledge of the origin and chemical composition of EVs can potentially be used as a biomarker to detect, stage, and monitor diseases. In this paper, we demonstrate the potential of EVs as a prostate cancer biomarker. A Raman optical tweezer was employed to obtain Raman signatures from four types of EV samples, which were red blood cell- and platelet-derived EVs of healthy donors and the prostate cancer cell lines- (PC3 and LNCaP) derived EVs. EVs’ Raman spectra could be clearly separated/classified into distinct groups using principal component analysis (PCA) which permits the discrimination of the investigated EV subtypes. These findings may provide new methodology to detect and monitor early stage cancer.

Background: Circulating tumour cells (CTCs) in blood associate with overall survival (OS) of cancer patients, but they are detected in extremely low numbers. Large tumour-derived extracellular vesicles (tdEVs) in castration-resistant prostate cancer (CRPC) patients are present at around 20 times higher frequencies than CTCs and have equivalent prognostic power. In this study, we explored the presence of tdEVs in other cancers and their association with OS. Methods: The open-source ACCEPT software was used to automatically enumerate tdEVs in digitally stored CellSearch® images obtained from previously reported CTC studies evaluating OS in 190 CRPC, 450 metastatic colorectal cancer (mCRC), 179 metastatic breast cancer (MBC) and 137 non-small cell lung cancer (NSCLC) patients before the initiation of a new treatment. Results: Presence of unfavourable CTCs and tdEVs is predictive of OS, with respective hazard ratios (HRs) of 2.4 and 2.2 in CRPC, 2.7 and 2.2 in MBC, 2.3 and 1.9 in mCRC and 2.0 and 2.4 in NSCLC, respectively. Conclusions: tdEVs have equivalent prognostic value as CTCs in the investigated metastatic cancers. CRPC, mCRC, and MBC (but not NSCLC) patients with favourable CTC counts can be further prognostically stratified using tdEVs. Our data suggest that tdEVs could be used in clinical decision-making.

Circulating tumor cells, tumor-derived extracellular vesicles and plasma cytokeratins in castration-resistant prostate cancer patients Circulating tumor cells, tumor-derived extracellular vesicles and plasma cytokeratins in castration-resistant prostate cancer patients

Background: Tumor-derived extracellular vesicles (tdEVs) and circulating tumor cells (CTCs) in the blood of metastatic cancer patients associate with poor outcomes. In this study, we explored the human epidermal growth factor receptor 2 (HER2) expression on CTCs and tdEVs of metastatic breast cancer patients. Methods: Blood samples from 98 patients (CLCC-IC-2006-04 study) were originally processed with the CellSearch® system using the CTC kit and anti-HER2 as an additional marker in the staining cocktail. CTCs and tdEVs were automatically enumerated from the generated CellSearch images using the open-source ACCEPT software. Results: CTCs and tdEVs were subdivided based on their cytokeratin (CK) and HER2 phenotype into CK+HER2−, CK−HER2+, and CK+HER2+. The inclusion of anti-HER2 increased the percentage of informative samples with ≥ 1 detectable CTC from 89 to 95%. CK− CTCs and tdEVs correlated equally well with the clinical outcome as CK+ CTCs and tdEVs. Inter- and intra-patient heterogeneity was found for the CTC/tdEV phenotypes, and the presence of 2 or 3 classes of CTCs/tdEVs was associated with worse prognosis compared to a uniform CTC/tdEV phenotype present (1 class). The use of ≥ 7% HER2+CK+ tdEVs can predict HER2 expression of the tissue with 74% sensitivity and specificity using the HER2 amplification status of the primary tumor as a classification variable. Conclusions: HER2 can be detected on CTCs and tdEVs not expressing CK, and these CK− CTCs/tdEVs have similar clinical relevance to CTCs and tdEVs expressing CK. tdEVs perform better than CTCs in predicting the HER2 status of the primary tissue. CTC and tdEV heterogeneity in the blood of patients is inversely associated with overall survival.