Annals of Oncology Advance Access originally published online on August 11, 2008
Annals of Oncology 2009 20(1):27-33; doi:10.1093/annonc/mdn544
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
urogenital tumours |
Circulating tumour cell (CTC) counts as intermediate end points in castration-resistant prostate cancer (CRPC): a single-centre experience
1 Drug Development Unit, The Royal Marsden NHS Foundation Trust, Sutton
2 Section of Medicine, The Institute of Cancer Research, Sutton
3 ICR Clinical Trials and Statistical Unit, Section of Clinical Trials, The Institute of Cancer Research, Sutton
4 Academic Urology Unit, The Institute of Cancer Research and The Royal Marsden NHS Foundation Trust, Sutton, UK
5 Immunicon Corporation, Huntington Valley, PA, USA
* Correspondence to: Dr Johann de Bono, Royal Marsden Hospital, Downs Road, Sutton, Surrey SM2 5P2, UK. Tel: +44-208-6426011; Fax: +44-208-6427979; E-mail: johann.de-bono{at}icr.ac.uk
 |
Abstract |
|---|
 Top Abstract introductionmethodsresultsdiscussionfundingAcknowledgementsReferences |
|---|
Background: The purpose of this study was to evaluate the association of circulating tumour cell (CTC) counts, before and after commencing treatment, with overall survival (OS) in patients with castration-resistant prostate cancer (CRPC).
Experimental design: A 7.5 ml of blood was collected before and after treatment in 119 patients with CRPC. CTCs were enumerated using the CellSearch®System.
Results: Higher CTC counts associated with baseline characteristics portending aggressive disease. Multivariate analyses indicated that a CTC 
5 was an independent prognostic factor at all time points evaluated. Patients with baseline CTC 5 had shorter OS than those with <5 [median OS 19.5 versus >30 months, hazard ratio (HR) 3.25, P = 0.012]; patients with CTC >50 had a poorer OS than those with CTCs 5–50 (median OS 6.3 versus 21.1 months, HR 4.1, P < 0.001). Patients whose CTC counts reduced from 5 at baseline to <5 following treatment had a better OS compared with those who did not. CTC counts showed a similar, but earlier and independent, ability to time to disease progression to predict OS.
Conclusion: CTC counts predict OS and provide independent prognostic information to time to disease progression; CTC dynamics following therapy need to be evaluated as an intermediate end point of outcome in randomised phase III trials.
Key words: biomarkers, castration-resistant prostate cancer, circulating tumour cell
|
introduction |
|---|
|
TopAbstractintroductionmethodsresultsdiscussionfundingAcknowledgementsReferences |
|---|
Castration-resistant prostate cancer (CRPC) is a heterogeneous disease with prognoses varying significantly between patients. In addition, the assessment of treatment response in CRPC remains a major challenge impacting not only routine clinical care but also anticancer drug development. Post-therapy declines in prostate-specific antigen (PSA) have been commonly utilised to identify antitumour activity [1] as conventional radiological assessments have severe limitations. First, commonly used criteria such as response evaluation criteria in solid tumours (RECIST) [2] are not useful in a large proportion of patients [3]. Secondly, changes in radionuclide bone scans are difficult to quantify objectively and reproducibly [4]. However, it has long been recognised that changes in PSA often do not accurately reflect disease activity [4, 5], and PSA decline posttreatment is not a reliable intermediate end point of overall survival (OS). There is an urgent need to develop biomarkers for this disease and to assess antitumour activity of novel therapies.
The isolation, separation and enumeration of circulating tumour cells (CTCs) have been reported by several groups, but controversy exists on the optimal approach to enumerate them, although only the CellSearchTM assay (Immunicon, Huntingdon, PA) is Food and Drug Administration (FDA) cleared. In particular, a recent report, by Nagrath et al. [6], has suggested a new highly sensible assay for CTC isolation. However, different methodologies have also used different criteria (presence of 4',6-diaminidine-2-phenylindole (DAPI) staining, size and morphology, antibodies used for identification) to measure these CTC counts and have led to substantial variations in the numbers of cells; it remains difficult to directly compare these methodologies at this time. Nonetheless, the CellSearchTM assay (Immunicon) has been shown to be reproducible with low intrapatient and interlaboratory variability [7]. Using this technique, high CTC counts have been detected only in patients with cancer and may represent the haematogenous spread of tumours. CTC counts have additionally been found to have prognostic utility in different tumour types including advanced breast and colorectal cancer with high CTC counts correlating with poorer prognosis [7–11]. Specifically, the presence of more than five CTCs in 7.5 ml of blood has been associated with poor OS in patients with metastatic CRPC [12]. Furthermore, a fall in CTC count early after treatment for breast cancer correlates with longer progression-free survival and OS [8–10]. In fact, the correlation of CTC counts with OS was superior to that between time to radiological progression with OS in advanced breast cancer [13]. We therefore elected to examine the prognostic and therapeutic values of CTC counts before and after treatment in the population of CRPC patients treated on clinical trials of novel anticancer drugs at the Royal Marsden Hospital.
|
methods |
|---|
|
TopAbstractintroductionmethodsresultsdiscussionfundingAcknowledgementsReferences |
|---|
study design and patient selection
Patients with biochemically or histologically confirmed progressive metastatic CRPC and castrate (<50 ng/dl) levels of testosterone treated on phase I or II clinical trials at the Royal Marsden NHS Foundation Trust were eligible. Patients gave informed consent for the collection and analysis of CTC. A medical history was taken from all patients, including details of the initial treatments and all subsequent therapeutic interventions. A physical examination, including Eastern Collaborative Oncology Group (ECOG) performance status (PS) [14] and laboratory studies, including full blood count, routine biochemistry, albumin, alkaline phosphatase (ALP), lactate dehydrogenase (LDH) and PSA, were carried out at baseline. Patients were deemed to have progressive disease if they had either a rise in PSA with a minimum of three rising levels at least 1 week apart or radiological progression by RECIST criteria. All clinical trials were approved by the Royal Marsden NHS Foundation Trust Research Ethics Committee and deliberately included trials in both the pre- and postdocetaxel space with novel molecularly targeted drugs.
CTC enumeration
CTC isolation and enumeration were carried out using the CellSearchTM system (Immunicon) as described previously [7] Blood samples were drawn into CellSaveTM tubes (Immunicon) at baseline before patients were started on trial medication. Further samples were collected before each cycle whenever possible. All samples were kept at room temperature and processed within 72 h of collection. To calculate the CTC count, 7.5 ml of blood was mixed with magnetic particles coated with anti-EpCAM antibodies. After immunomagnetic enrichment, cells were fluorescently labelled and individually captured using a four-colour semiautomated fluorescent microscope. The captured images were then presented to trained operators, blinded to patient outcome, who selected cells that met the definition of CTC. Criteria used to define a CTC include round to oval morphology, size >5 µm, a visible nucleous (4',6-diamidino-2-phenylindole positive), positive staining for cytokeratins 8,18 and/or 19 (phycoerythrin) and negative staining for CD45 (allophycocyanin). Results were expressed as the number of cells per 7.5 ml of blood.
statistical analysis
This analysis was carried out using SPSS 14.0 (SPSS Inc., Chicago, IL). OS was defined as the time between the blood draw and either the date of death or the last follow-up (if death was not observed during the follow-up period). Time to disease progression was defined as the time between the start of treatment and progression of disease on the basis of a composite end point comprising (i) PSA progression [1], (ii) radiological progression [1, 2], (iii) clinical progression requiring a new therapeutic intervention and/or (iv) death of any cause. Associations of CTC count with baseline patient characteristics were analysed using Mann–Whitney U or Kruskal–Wallis tests; for ordinal variables, Spearman rank correlation was assessed. Subsequently, significant characteristics associated (P value <0.05) with higher CTC counts in the univariate analysis were analysed in multivariate analyses using a logistic regression model. A threshold of 5 CTC/7.5 ml, which has been shown to be prognostic in a number of breast cancer trials [8, 9] and in a recent communication of a prostate cancer trial [15], was used for OS analysis at each of the blood draw time points. Moreover, thresholds of 10–100 were systematically correlated in our series in an attempt to identify groups with different prognosis for OS. Median OS and the 95% confidence intervals (CIs) were determined with the Kaplan–Meier method and OS curves were compared using the log-rank test. A multivariate analysis for OS using a Cox regression model was carried out to determine the subset of baseline characteristics that provided independent prognostic information in our series. In addition, a Cox regression model with a time-dependent covariable was used to compare the independent prognostic value of CTC count at different time points. All P values reported were two sided. Time to disease progression and CTC counts were compared by calculating the proportion of the variation in OS that could be explained by these parameters using the method of Royston [16] employing the Stata 9.0 programme (StataCorp, College Station, TX) CTC counts were treated as a log-transformed continuous variable in this context.
|
results |
|---|
|
TopAbstractintroductionmethodsresultsdiscussionfundingAcknowledgementsReferences |
|---|
patient characteristics
In all, 119 CRPC patients were treated in 14 different phase I and II trials from January 2005 to July 2007 (Table 1). The median age of patients included in this analysis was 67.5 years (range 48.2–85.5). A total of 94.1% of patients had ECOG PS of zero to one at baseline. The pattern of metastatic spread included disease limited to the lymph nodes without bone metastases in 11 patients (9.2%) and bone metastases in 108 patients (90.8%). In all, 87 patients (73%) were chemotherapynaive (previous hormone therapies median 2, range 2–4). Overall, 32 patients (27%) received previous chemotherapy (median range 1–4). A total of 76 patients were treated with a molecularly targeted drug only, 35 patients received docetaxel plus/minus a targeted agent. Eight patients received single-agent mitoxantrone. After a median follow-up time of 15.3 months (range 1.3–33.9), the median OS was 26.4 months (95% CI 20.0–32.7).
|
CTC counts
The median CTC count at baseline before starting trial treatment was six CTC per 7.5 ml of blood (range 0–545). In all, 59 patients (49.6%) had a CTC count <5, while 38 patients (31.9%) had a CTC count 5–50 and 22 patients (18.5%) had a CTC count >50. Overall, 101 patients had CTC counts measured following the first course of treatment at 3–4 weeks, and 98 patients had CTC counts measured following the second course of treatment at 6–8 weeks. The median CTC counts following the first and second courses of treatment were three CTC (range 0–1317) and one CTC (range 0–1144), respectively.
baseline CTC count correlation with patients’ characteristics
The correlation of CTC count distribution and baseline characteristics are shown in Table 2. Multivariate analysis revealed that higher CTC counts were associated with: ALP >upper normal limit (P = 0.0004), haemoglobin level <12 g/dl (P = 0.027), PSA >150 ng/ml (P < 0.035) and prior cytotoxic chemotherapy administration (P = 0.008). The univariate analysis revealed that presence of bone metastases was correlated with higher CTC counts (P = 0.013). Patients with metastatic disease confined exclusively to lymph nodes had a median of zero CTC at baseline and all time points on study. Patients with bone involvement and concomitant lymph node metastases had a higher median CTC count than patients with solely bone metastases (10 CTC versus seven CTC, P = NS).
|
multivariate analyses indicate that CTC counts at baseline are an independent predictor of OS
Multivariate analysis demonstrated that patients with a CTC count
5 at baseline had a shorter OS [19.5 months, 95% CI 8.9–30.1, hazard ratio (HR) 3.25, P = 0.005] compared with patients with a CTC count <5 (>30 months; Table 3). Apart from CTC count 5, LDH >UNL was also independently associated with a poor OS (HR 2.44, 95% CI 1.2–4.9, P = 0.012). Additionally, we categorised patients into three groups of CTC counts (<5, 5–50 and >50). This categorisation enabled us to show that patients with a CTC count >50 had a poorer OS compared with patients with 5–50 CTCs and CTCs <5 at baseline (6.3 versus 21.1 versus >30 months, HR 4.1, P < 0.001). These categories continued to show significant different outcomes at all CTC time points following treatment (Figure 1).
|
|
CTC count dynamics predict OS
To investigate whether a change in CTC level from baseline predicts a change in the initial prognosis for survival, we compared changes in the level between baseline and after the first cycle and the second cycle (Figure 2). Four different groups of patients were compared—group 1: patients with <5 CTC/7.5 ml at every blood drawn time points; group 2: patients with
5 CTC/7.5 ml before the initiation of therapy who decreased to <5 CTC/7.5 ml after therapy; group 3: patients with <5 CTC/7.5 ml who increased to 5 after treatment and group 4: patients with 5 CTC/7.5 ml at all blood drawn time points. After treatment, 114 patients were eligible for this analysis. Patients with 5 CTC/7.5 ml at all time points (group 4, n = 29) had the shortest median OS of 9.2 months (95% CI 19.6–25.0), which was significantly worse compared with group 1 (n = 51, median OS > 30 months, P < 0.0001) and with group 2 (n = 28, median OS 21.4 months, 95% CI 19.6–25.0, P = 0.0055). Furthermore, patients in group 3 (n = 6) showed a shorter OS (11.2 months, 95% CI 5.6–16.8), and this also differed from both group 1 (P < 0.0001) and group 2 (P = 0.048). However, there was no significant difference between group 3 and group 4 (P = 0.49).
|
To evaluate CTC dynamics as a continuous variable, we were able to demonstrate that a drop of
30% from baseline CTC count during the first two treatment cycles was associated with improved OS in all patients with a CTC count 5 (Figure 3). Among the 48 patients eligible for this analysis, patients with a drop of <30% from baseline had a poorer OS (5.9 months, 95% CI 5.3–6.5, P < 0.0001) compared with patients with a drop of 30% (21.1 months, 95% CI 19.4–22.8).
|
time to disease progression and CTC counts
After a median follow-up time of 15.3 months (range 1.3–33.9), the median time to disease progression, on the basis of a composite of PSA, radiology and symptoms, was 5.3 months (95% CI 4.5–6.1). In our analysis, we applied time-dependent covariates to evaluate the use of time to disease progression. In this model, R2 values were used to estimate the proportion of variation in OS explained by variation in the intermediate end points (in this case CTC counts and time to disease progression). CTC counts at baseline and after one or two treatments (after 3–8 weeks) showed R2 values of 52%, 55% and 55%, meanwhile time to disease progression showed a R2 of 39%. The R2 values were compared by observing that the simulation-based 95% confidence interval for the first R2 value was 52% (95% CI 32% to 69%), supporting the fact that CTC was independent but not necessarily superior to time to disease progression in predicting OS, although the CTC counts are earlier predictors of outcome.
|
discussion |
|---|
|
TopAbstractintroductionmethodsresultsdiscussionfundingAcknowledgementsReferences |
|---|
Several methods to isolate and evaluate CTC counts have been reported [6, 17, 18]. A noncytometric methodological approach, on the basis of the detection of telomerase activity in by real time polymerase chain reaction, has been reported to be very sensitive for CTC detection in the context of prostate cancer; however, this kind of molecular approach does not allow the individual CTC identification. Previous reported cytometric methodologies [6, 17] appear to have been less restrictive in requiring nuclear staining by DAPI and in mandating other critical cytometric features key to cell identification such as cell size and shape resulting in apparently higher CTC counts. Direct comparisons between these methods are now urgently required. Nevertheless, the current gold standard remains CTC enumeration by the CellSearchTM system which is now FDA cleared as a prognostic factor for breast, colorectal and most recently CRPC on the basis of the data acquired from several large international studies in breast [8], colorectal [11] and prostate cancer (de Bono JS et al., 2008, In press). This validation study supports these data and extends the studies to patients treated with noncytotoxic agents before chemotherapy as well as evaluating falls in CTC counts following treatment as predictors of outcome. Overall, these findings encourage further research of CTCs in CRPC and show that CTCs are correlated with several factors of poor prognosis and tumour burden.
Higher baseline CTC counts were associated with features portending aggressive disease, in keeping with previously published reports [18]. Elevated ALP and LDH, a higher PSA, a PSADT <3 months and low haemoglobin and albumin were all correlated with high CTC counts. In addition, a higher percentage of patients who had bone metastases and who received prior chemotherapy had higher CTC counts. The fact that high CTC counts were found in patients with bone metastases and no CTCs were detectable in patients with solely lymph node disease supports the hypothesis of lymphatic drainage as the predominant route of metastatic spread in patients with lymph node alone disease, whereas bone metastases suggest haematogenous spread and therefore detectable CTC.
As reported in other diseases, the detection of 5 CTC at any time point was associated with a poorer survival. Furthermore, there seems to be an inverse relationship between CTC counts and survival with worsening prognoses as CTC counts increased in each of the categories we examined, i.e. >50, 5–50 and <5. This supports the hypothesis by Danila et al. [19] who argued for CTC counts being important as a continuous variable in predicting outcome.
This is one of the first studies in this disease to examine early changes in CTC counts as a reflection of treatment benefit. We showed that changes in CTC counts from one category to another could have a role in predicting clinical outcome. As shown in Figure 2, patients who remained below the threshold of five CTCs during the initial two cycles of treatment had the best clinical outcome, whereas patients in whom the CTC count decreased from 5 to <5 had a better outcome compared with patient who remained in the category 5 throughout treatment. Interestingly, patients in whom the CTC count increased from <5 to 5 had a similar outcome as those patients whose count remained 5, suggesting the presence of treatment-resistant disease.
In addition, we demonstrated that other ways of characterising CTC changes are also useful. For instance, patients with a proportional fall in CTC count of 30% had an improved prognosis compared with those whose CTC count remained stable or increased. This supports the prospective exploration of CTC changes as a proportion of change and possibly as a continuous variable.
Moreover, in this series, prediction of OS using CTC counts was as effective as that on the basis of time to progression, but CTC levels are available earlier and could decrease the use of ineffective treatments. Changes in levels of LDH may also offer additional prognostic information to that offered by CTC counts because they have been shown to have independent prognostic relevance in our series. Change in CTC count, perhaps in combination with other parameters in a composite model, is therefore a strong candidate for use as an intermediate end point. Further exploration of CTC counts in this context should be supported.
The present series is heterogeneous, but multivariate analysis revealed CTC counts to be an independent factor. Significantly, we maintain that this is a true reflection of the CRPC setting in daily clinical practice.
In conclusion, our data suggest that CTC may provide a more sensitive marker in monitoring disease status during treatment, especially in early-stage disease, and give an important indication of long-term outcome. In order to validate these techniques prospectively, large randomised trials are under way, in which predictive models on the basis of CTC counts would be validated. This method may additionally have potential advantages in accessing tissue for molecular analysis. CTCs thus hold immense potential as improved biomarker of response and to accelerate evaluation of emerging novel therapies.
|
funding |
|---|
|
TopAbstractintroductionmethodsresultsdiscussionfundingAcknowledgementsReferences |
|---|
2006–2007 ESMO fellowship for Circulating Tumor Cell Research in Prostate Cancer to DO.
|
Acknowledgements |
|---|
|
TopAbstractintroductionmethodsresultsdiscussionfundingAcknowledgementsReferences |
|---|
Parts of this work were presented at the 2008 Genitourinary Cancers Symposium and at the 2008 AACR Annual Meeting and has been awarded with: The 2008 Genitourinary Cancer Symposium Merit Award and The 2008 GSK-AACR Scholarship.
Received for publication April 4, 2008. Revision received June 18, 2008. Revision received June 26, 2008. Accepted for publication July 4, 2008.
|
References |
|---|
|
TopAbstractintroductionmethodsresultsdiscussionfundingAcknowledgementsReferences |
|---|
1. Bubley GJ, Carducci M, Dahut W, et al. Eligibility and response guidelines for phase II clinical trials in androgen-independent prostate cancer: recommendations from the Prostate-Specific Antigen Working Group. J Clin Oncol (1999) 17:3461–3467.
2. Therasse P, Arbuck SG, Eisenhauer EA, et al. New guidelines to evaluate the response to treatment in solid tumors. J Natl Cancer Inst (2000) 92:205–216.
3. Scher HI, Morris MJ, Kelly WK, et al. Prostate cancer clinical trial end points: "RECIST"ing a step backwards. Clin Cancer Res (2005) 11:5223–5232.
4. Scher HI, Mazumdar M, Kelly WK. Clinical trials in relapsed prostate cancer: defining the target. J Natl Cancer Inst (1996) 88:1623–1634.
5. Collette L, Burzykowski T, Carroll K, et al. Is prostate specific antigen a valid surrogate end-point for survival in hormonally treated patients with metastatic prostate cancer. J Clin Oncol (2005) 23:6139–6148.
6. Nagrath S, Sequist LV, Maheswaran S, et al. Isolation of rare circulating tumour cells in cancer patients by microchip technology. Nature (2008) 450:1235–1239.
7. Allard WJ, Matera J, Miller MC, et al. Tumor cells circulate in the peripheral blood of all major carcinomas but not in healthy subjects or patients with nonmalignant diseases. Clin Cancer Res (2004) 10:6897–6904.
8. Cristofanilli M, Budd GT, Ellis MJ, et al. Circulating tumor cells, disease progression, and survival in metastatic breast cancer. New Engl J Med (2004) 351:781–791.
9. Cristofanilli M, Hayes DF, Budd GT, et al. Circulating tumor cells: a novel prognostic factor for newly diagnosed metastatic breast cancer. J Clin Oncol (2005) 23:1420–1430.
10. Hayes DF, Cristofanilli M, Budd GT, et al. Circulating tumor cells at each follow-up time point during therapy of metastatic breast cancer patients predict progression-free and overall survival. Clin Cancer Res (2006) 12:4218–4224.
11. Cohen S, Punt C, Ianott N, et al. Relationship of circulating tumour cells to tumour response, progression free survival and overall survival in patients with metastatic colorectal cancer. J Clin Oncol (2008) 26:3213–3221.
12. Moreno JG, Miller MC, Gross S, et al. Circulating tumor cells predict survival in patients with metastatic prostate cancer. Urology (2005) 65:713–718.[CrossRef][Web of Science][Medline]
13. Budd GT, Cristofanilli M, Ellis MJ, et al. Circulating tumor cells versus imaging—predicting overall survival in metastatic breast cancer. Clin Cancer Res (2006) 12:6403–6409.
14. Oken MM, Creech RH, Tormey DC, et al. Toxicity and response criteria of the Eastern-Cooperative-Oncology-Group. Am J Clin Oncol (1982) 5:649–655.[Web of Science][Medline]
15. de Bono JS, Scher H, Montgomery B, et al. Circulating tumor cells (CTC) predict survival benefit from treatment in metastatic castration resistant prostate cancer (CRPC). Clin Can Res (2008) In press.
16. Royston P. Explained variation for survival models. Stata J (2006) 6:83–96.
17. Pachmann K, Camara O, Kavallaris A, et al. Monitoring the response of circulating epithelial tumor cells to adjuvant chemotherapy in breast cancer allows detection of patients at risk of early relapse. J Clin Oncol (2008) 26:1208–1215.
18. Fizazi K, Morat L, Chauveinc L, et al. High detection of circulating tumor cells in blood of patients with prostate cancer using telomerase activity. Ann Oncol (2007) 18:518–521.
19. Danila DC, Heller G, Gignac GA, et al. Circulating tumor cell number and prognosis in progressive castration-resistant prostate cancer. Clin Cancer Res (2007) 13:7053–7058.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  G. Attard, A. H.M. Reid, R. A'Hern, C. Parker, N. B. Oommen, E. Folkerd, C. Messiou, L. R. Molife, G. Maier, E. Thompson, et al. Selective Inhibition of CYP17 With Abiraterone Acetate Is Highly Active in the Treatment of Castration-Resistant Prostate Cancer J. Clin. Oncol., August 10, 2009; 27(23): 3742 - 3748. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Massard, A. Chauchereau, and K. Fizazi The quest for the 'bony Grail' of detecting circulating tumour cells in patients with prostate cancer Ann. Onc., February 1, 2009; 20(2): 197 - 199. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||