Hormone-driven expression of the ERG oncogene after fusion with TMPRSS2 occurs in 30% to 70% of therapy-naive prostate cancers. Its relevance in castration-resistant prostate cancer (CRPC) remains controversial as ERG is not expressed in some TMPRSS2-ERG androgen-independent xenograft models. However, unlike these models, CRPC patients have an increasing prostate-specific antigen, indicating active androgen receptor signaling. Here, we collected blood every month from 89 patients (54 chemotherapy-naive patients and 35 docetaxel-treated patients) treated in phase I/phase II clinical trials of an orally available, highly specific CYP17 inhibitor, abiraterone acetate, that ablates the synthesis of androgens and estrogens that drive TMPRSS2-ERG fusions. We isolated circulating tumor cells (CTC) by anti–epithelial cell adhesion molecule immunomagnetic selection followed by cytokeratin and CD45 immunofluorescence and 4′,6-diamidino-2-phenylindole staining. We used multicolor fluorescence in situ hybridization to show that CRPC CTCs, metastases, and prostate tissue invariably had the same ERG gene status as therapy-naive tumors (n = 31). We then used quantitative reverse transcription–PCR to show that ERG expression was maintained in CRPC. We also observed homogeneity in ERG gene rearrangement status in CTCs (n = 48) in contrast to significant heterogeneity of AR copy number gain and PTEN loss, suggesting that rearrangement of ERG may be an earlier event in prostate carcinogenesis. We finally report a significant association between ERG rearrangements in therapy-naive tumors, CRPCs, and CTCs and magnitude of prostate-specific antigen decline (P = 0.007) in CRPC patients treated with abiraterone acetate. These data confirm that CTCs are malignant in origin and indicate that hormone-regulated expression of ERG persists in CRPC. [Cancer Res 2009;69(7):2912–8]

Background: Presence of five or more circulating tumor cells (CTC) in patients with metastatic carcinomas is associated with poor survival. Although many objects positive for epithelial cell adhesion molecules and cytokeratin (EpCAM+CK+) are not counted as CTC, they may be an important predictor for survival. We evaluated the association between these objects and survival in patients with prostate cancer. Patients and methods: Included in this follow-up study were 179 patients with castration-resistant prostate cancer. CellSearch was used to isolate EpCAM+ objects and to stain DNA, cytokeratin and CD45. All EpCAM+CK+ objects were subdivided into seven classes on the basis of predefined morphological appearance in 63 independent samples. Association of each class with survival was studied using Kaplan–Meier and Cox regression analyses. Results: Each EpCAM+CK+CD45− class showed a strong association with overall survival (P < 0.001). This included small tumor microparticles (S-TMP), which did not require a nucleus and thus are unable to metastasize. A higher number of objects in any class was associated with decreased survival. A good prediction model included large tumor cell fragments (L-TCF), age, hemoglobin and lactate dehydrogenase. Models with S-TMP or CTC instead of L-TCF performed similarly. Conclusion: EpCAM+CK+CD45− that do not meet strict definitions for CTC are strong prognostic markers for survival.

Purpose: Presence of circulating tumor cells (CTC) in metastatic carcinoma is associated with poor survival. Phenotyping and genotyping of CTC may permit “real-time” treatment decisions, provided CTCs are available for examination. Here, we investigate what is needed to detect CTC in all patients. Experimental Design: CTCs enumerated in 7.5 mL of blood together with survival from 836 patients with metastatic breast, colorectal, and prostate cancer were used to predict the CTC concentration in the 42% of these patients in whom no CTCs were found and to establish the relation of concentration of CTCs with survival. Influence of different CTC definitions were investigated by automated cell recognition and a flow cytometric assay without an enrichment or permeabilization step. Results: A log-logistic regression of the log of CTC yielded a good fit to the CTC frequency distribution. Extrapolation of the blood volume to 5 L predicted that 99% of patients had at least one CTC before therapy initiation. Survival of patients with EpCAM+, cytokeratin+, CD45− nucleated CTCs is reduced by 6.6 months for each 10-fold CTC increase. Using flow cytometry, the potential three-fold recovery improvement is not sufficient to detect CTC in all patients in 7.5 mL of blood. Conclusions: EpCAM+, cytokeratin+, CD45− nucleated CTCs are present in all patients with metastatic breast, prostate, and colorectal cancer and their frequency is proportional to survival. To serve as a liquid biopsy for the majority of patients, a substantial improvement of CTC yield is needed, which can only be achieved by a dramatic increase in sample volume. Clin Cancer Res; 18(20); 5711–8. ©2012 AACR.

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.

The presence of circulating tumor cells (CTCs) in the blood of cancer patients may guide the use of therapy. We investigated how to evaluate a reduction in the number of CTCs after administration of therapy. CTCs were enumerated with the CellSearch system in 111 metastatic breast and 185 metastatic prostate cancer patients before start of a new line of chemotherapy and after initiation of therapy. Different means to express changes in CTC counts were evaluated with respect to overall survival (OS). A static CTC cutoff is the best method to determine whether a therapy is effective. This is exemplified by the highest Cox hazard ratio of 2.1 for OS; three methods to express relative differences performed worse. A lookup table is provided from which the significance of a change in CTCs can be derived. The aim of therapy should be the elimination of all CTCs. A period of 10 to 12 weeks of therapy is needed to reach the treatment effect on CTCs.

Circulating tumour cells (CTC) in patients with metastatic carcinomas are associated with poor survival and can be used to guide therapy. Classification of CTC however remains subjective, as they are morphologically heterogeneous. We acquired digital images, using the CellSearch™ system, from blood of 185 castration resistant prostate cancer (CRPC) patients and 68 healthy subjects to define CTC by computer algorithms. Patient survival data was used as the training parameter for the computer to define CTC. The computer-generated CTC definition was validated on a separate CRPC dataset comprising 100 patients. The optimal definition of the computer defined CTC (aCTC) was stricter as compared to the manual CellSearch CTC (mCTC) definition and as a consequence aCTC were less frequent. The computer-generated CTC definition resulted in hazard ratios (HRs) of 2.8 for baseline and 3.9 for follow-up samples, which is comparable to the mCTC definition (baseline HR 2.9, follow-up HR 4.5). Validation resulted in HRs at baseline/follow-up of 3.9/5.4 for computer and 4.8/5.8 for manual definitions. In conclusion, we have defined and validated CTC by clinical outcome using a perfectly reproducing automated algorithm.