Prostate Cancer Resource Center

Tumor Growth Kinetics Before and After First-line Chemotherapy in Metastatic Castration-resistant Prostate Cancer: A Prostate-specific Antigen-based Retrospective Analysis

American Journal of Clinical Oncology

Giuseppe Colloca, MD, Antonella Venturino, MD, and Domenico Guarneri, MD


Objectives: The role of the tumor growth fraction has been investigated poorly in metastatic castration-resistant prostate cancer (mCRPC). The aim of this study was to assess whether some prostate-specific antigen (PSA)-related variables of tumor cell kinetics predict the overall survival in early and late mCRPC, and to explore changes in the tumor growth fraction after chemotherapy.

Methods: A retrospective analysis of 3 tumor cell kinetic variables in patients with mCRPC receiving first-line chemotherapy has been performed. The PSA-related tumor growth rate, the log ratio, and the tumor response have been measured at 3 different times. A further analysis has been performed after stratification by the Gleason score and chemotherapy. Finally, tumor growth after progression to chemotherapy has been explored.

Results: G at castration resistance is significantly associated with survival after chemotherapy among patients with a low Gleason score (r=−0.650, P-value=0.022). At the time of first-line chemotherapy, both G and PSA response rates report a significant relationship with survival. At the time of postchemotherapy progression, only the G after 12 weeks of chemotherapy maintains a relationship with survival in patients with a low Gleason score (r=−0.483, P-value=0.023); in particular, a tumor growth rate <−0.5%/day appears to be associated with a poor postprogression survival. Despite the lack of correlation between postprogression G and postprogression survival, the response to chemotherapy defines 2 groups with different growth characteristics.

Conclusions: Among patients with mCRPC, tumor cell kinetics appears to be able to predict the outcome, especially in tumors with a low Gleason score.

In Europe, prostate cancer (PC) is the most common nonskin cancer among men.1 Although the 5-year relative survival increased from 1999-2001 to 2005-2007, from 73.4% to 81.7%,2 respectively, PC remains one of the most prevalent solid tumors. Death from PC occurs almost exclusively in the last stage of the disease, the metastatic castration-resistant prostate cancer (mCRPC).3 At this stage, the median survival has changed considerably over the past 10 years, after the introduction of a number of effective drugs, such as docetaxel, cabazitaxel, abiraterone, enzalutamide, and sipuleucel-T. The first of these molecules to enter in clinical practice was docetaxel,4,5 and many studies in recent years have been conducted on patients with mCRPC after progression to docetaxel.6–8 Consequently, the improvement of survival of mCRPC patients within these trials occurred mostly in a “late” mCRPC stage, a stage of disease after progression to docetaxel.

Neoplastic cell growth has been studied in preclinical models and clinical trials. Linear kinetics for tumor growth has been reported only for some phases of the growth, so that a multiphasic model has been hypothesized to explain these differences in the kinetics of tumor growth throughout the clinical course of the tumor. This is the Gompertzian model.9 It is a mathematical model where growth is the slowest at the start and the end of the tumor’s course. For mCRPC, a model of growth rate based on the size of metastases is not practical, due to the frequently bone-only and not measurable disease. In this setting, prostate-specific antigen (PSA) is frequently useful in clinical practice to monitor mCRPC growth and response to antineoplastic treatments. However, the PSA-related response rate (PSA-RR) did not appear as a suitable surrogate end point of the overall survival (OS) in 2 trials of first-line and one of second-line chemotherapy, whereas PSA-related progression-free survival (PSA-PFS) has been poorly studied.

Of the variables related to tumor cell kinetics, the specific growth rate (SGR) is one of the most interesting.10 It results from an equation that takes into account tumor cell growth and regression, both occurring simultaneously during cancer treatment. SGR can be reported in different ways: if the tumor dimensions grow by 1.1% in 1 day, this growth could be reported in a log10 scale with the corresponding value of −1.9502/day. However, in mCRPC, due to the prevailing bone-only spread of the disease, the calculation of SGR on the basis of the size of metastases appears to be feasible in a minority of the patients. For this reason, some authors calculated the SGR of patients with mCRPC using serial PSA determinations, and provided encouraging preliminary results in retrospective studies.11,12 In our previous experience of 49 patients with mCRPC receiving a docetaxel chemotherapy, a PSA-related SGR (G) of −2.4/day, or 0.4%/d, appeared to be the most appropriate cutoff to which the antineoplastic treatment was changed.12

The aim of this study was to assess the behavior of the PSA-related tumor growth along 3 different phases of the mCRPC: at castration resistance, during chemotherapy, and after progression to first-line chemotherapy. Further analyses were performed to evaluate whether any other tumor kinetic variable is able to predict patients’ outcome.


Patients were selected among men with mCRPC, who were treated at the Department of Oncology of the G. Borea Hospital in Sanremo, Italy, from March 2006 to March 2013. The date of death, or last censoring, was reported in the clinical record. Every patient signed a consent form on the management of his clinical data for research purposes. To be included in the study patients must have received at least one cycle of cytotoxic chemotherapy for mCRPC.

The OS was calculated from the start of chemotherapy (OS) until death or the last censoring, but the survival from the time of castration resistance (castration-resistance OS) and from the serologic postchemotherapy progression (postprogression OS) was also analyzed. PSA-PFS and PSA-RR were calculated from the start of chemotherapy by PCWG2 criteria.13

Kinetic measures were calculated at three moments of the course of the mCRPC: (1) from castration resistance to the start of first-line chemotherapy; (2) during chemotherapy; and (3) after progression to chemotherapy. For the pre-chemotherapy time, determinations of PSA included the PSA nadir after ADT, the first PSA above nadir, and the PSA at the beginning of chemotherapy; for the time during chemotherapy were included the PSA of 2 to 3 months before chemotherapy, the PSA at the beginning of chemotherapy, and the PSA after 12 weeks of chemotherapy; for the time following progression after chemotherapy, the PSA before the post-chemotherapy nadir, the PSA of the post-chemotherapy nadir after 12 weeks (or the value at 12 weeks in case of serologic progression), and the PSA at the beginning of the further medical treatment (or the last available PSA when another treatment was not started) were included.

The postchemotherapy progression was defined as the first date of disease progression as measured by PSA progression, radiologic tumor progression, or death; treatment discontinuation due to adverse events or patient decision was not considered as a progression. A RECIST evidence of radiologic progression for measurable disease, an increasing serum PSA concentration (at least 2 consecutive increases relative to a reference value measured at least a week apart), or the appearance of at least 1 new radiographic lesion were required to define progression.

All determinations of serum PSA were performed at the “Laboratorio Analisi” Department of the G. Borea Hospital. The upper value of the normal range was 4 ng/mL.

PSA-related kinetic variables were calculated according to the formulas reported for the SGR (G),10 the log ratio (LR),14 and the kinetic tumor response (TR).15 Similarly, the calculation of G at castration resistance and after disease progression to chemotherapy were performed with the same formula10 only for patients for whom all PSA determinations were available, and for postprogression G, irrespective of whether a serologic progression to chemotherapy occurred.

Relationships between OS and G, PSA-PFS, and PSA-RR were evaluated. Characteristics of postchemotherapy LR and TR as well as their relationships with OS, PSA-PFS, and PSA-RR were reported.

Relationships between variables were evaluated with the Spearman ρ correlation coefficient. Differences among groups were detected by the independent sample t test. Data were processed with the software SPSS v17.


From March 2006 to March 2013, 105 patients with mCRPC were evaluated at the Department of Oncology of the G. Borea Hospital. Among them, 79 patients received a first-line chemotherapy, and 61 were eligible for the present analysis. Fifty patients were treated with a 21-day regimen of docetaxel 75 mg/m2, 3 patients with a weekly schedule at 35 mg/m2, and 8 with a 21-day regimen of mitoxantrone 12 mg/m2. In all cases, chemotherapy was combined with oral prednisone at 5 mg bid. The median number of courses was 6 for the 21-day regimen, 24 for the weekly schedule of docetaxel, and 4 for the mitoxantrone regimen. No patient received vaccine therapy, before or after chemotherapy. Among the patients who received further treatments, 12 received cabazitaxel, 13 abiraterone, and 3 enzalutamide. Patient characteristics at the time of chemotherapy are listed in Table 1.

G and time ranges of sequential PSA determinations were collected, and are reported in Table 2, from which it appears that the median postprogression G is not different from the G at castration resistance (0.93%/d vs. 1.24%/d, P-value=0.677).



Correlation analyses between G and OS at the different time points, and with PSA-PFS and PSA-RR, are reported in Table 3. No relationship was apparent among the 3 Gs at the 3 different time points, and the correlation between G and OS, present at the time of castration resistance and after 12 weeks of chemotherapy, was lost after disease progression to chemotherapy.


The G at castration resistance appears to be related with castration-resistance OS, but not with OS. However, patients with a Gleason score <8 report a strong inverse relationship between the G at castration resistance and OS (r=−0.620, P-value=0.022), and a G at castration resistance >1.5%/d was significantly associated with OS (9.8 vs. 20.8 mo, t test=−2.298, P-value=0.027). The Gleason score modulates the relationship between G at castration resistance and OS, suggesting that the high growth fraction at castration resistance in mCRPC with a Gleason score <8 could predict a lower OS and the activity of chemotherapy.

Both G after 12 weeks of chemotherapy (r=−0.520, P-value <0.001) and PSA-RR (r=−0.502, P-value <0.001) reported a significant relationship with OS. In particular, a G after 12 weeks of chemotherapy <−0.5%/d appeared to predict a longer OS in 27 versus 32 patients (30.0 vs. 10.7 mo, t test=−4.326, P-value <0.001). Similarly, a PSA-RR, with a reduction of >30% from the baseline, was associated with a longer OS in 31 versus 28 patients (26.6 vs. 11.3 mo, t test=−3.240, P-value=0.002). As expected, after 12 weeks of chemotherapy a strong relationship has been reported between G and PSA-RR (r=0.983), between G and PSA-PFS (r=0.742), and between PSA-PFS and OS (r=0.565). A similar finding was made for the relationships between G at castration resistance with PSA-RR and with PSA-PFS, which are significant only among patients with a Gleason score <8.

At the time of postchemotherapy disease progression, only the G seems to be related with the postprogression OS (r=−0.346, P-value=0.007). Patients with a G after 12 weeks of chemotherapy <−0.5%/d, who were 27 versus 32, reported a postprogression OS of 19.0 versus 8.0 months, respectively (t test=−2.458, P=0.017). Despite the lack of correlation between the postprogression G and the postprogression OS, the response to chemotherapy defined 2 different groups with different characteristics, as reported in Table 4.


Table 5 shows the characteristics of LR and TR. Only LR>1.4 was related with a better OS, as it appeared from the comparison of the patients with a higher versus lower LR (29 vs. 30), who reported a median OS of 26.6 months versus 14.5 months (t-test=2.660, P-value=0.010). The relationship between LR and G was strong (r=0.801), similar to that between LR and PSA-RR (r=0.788). On the contrary, TR did not report any significant correlation with OS, but only a moderate association with G at castration resistance (r=0.440, P-value=0.006).


Finally, after stratification by the Gleason score at diagnosis in 3 different subgroups (<8, 8, and >8), whose characteristics are reported in Table 6, OS is moderately reduced in patients with a Gleason score >8 (10.5 vs. 19.7 mo, t test=−0.965, P-value=0.339), and no relationship resulted between the Gleason score and the PSA-RR, the PSA-PFS, and the time to castration resistance.



It is noteworthy that a G at castration resistance higher than 1.5%/d in mCRPC with a Gleason score <8 predicts survival after first-line chemotherapy. Larger studies should confirm this finding, because it could be a stratification factor for future trials of chemotherapy of mCRPC, in addition to a high Gleason score. A similar role can be hypothesized for G after 12 weeks of chemotherapy in predicting the postprogression survival of patients with late mCRPC.

In this paper, we used a simple calculation of G at different time points of the disease course, and it appeared to be reliable for the G at castration resistance and the postprogression G. After disease progression to chemotherapy, a variable change in the growth rate among patients probably occurs. It is interesting that chemotherapy-responsive patients have a longer postprogression OS and a higher postprogression growth fraction, as if the growth rate would be back on the Gompertzian curve, at an earlier and exponential stage after the effect of chemotherapy. Although the postprogression OS is not different in responsive versus unresponsive patients, the postprogression tumor growth fraction changes (t test=1.802, P-value=0.079).

Despite what is expected from models of linear cell kinetics and from previous studies, in our experience, we observed a slower growth fraction in late mCRPC than in early mCRPC (median 0.93 vs. 1.24%/d). However, this unexpected difference in the growth rate appears to be attributable to the subgroup of patients who did not report a previous response to chemotherapy, who have a very slow postprogression G (0.74 vs. 1.11%/d). Patients with a chemotherapy-responsive mCRPC appear to acquire a sustained cell growth after progression, whereas tumors of patients who did not respond to the first-line chemotherapy seemed to slow down, as if their mCRPC had reached the final plateau of the Gompertzian curve. Patients with a Gleason score >8 had lower values of postprogression G, suggesting that the late phase of cell growth of the more aggressive and chemoresistant mCRPCs is very slow.

The introduction of new effective drugs changed the scenario of mCRPC, and for future studies, it is increasingly important to find new criteria for defining the prognosis of late mCRPC. From a previous analysis of the curves of PSA-related cell kinetics of mCRPC, it has been concluded that the regression portion of the curve does not predict survival, whereas the growing fraction does.11 The PSA-related specific tumor growth is an indicator of the growth rate of PC and may be recalculated at every stage of disease. As such, its value could be considered as an ideal measure of progression/delay, as suggested by the criteria of PCWG2.13 To identify new prognostic factors for clinical trials of patients with mCRPC, new kinetic variables, such as the G at castration resistance and the G after 12 weeks of chemotherapy, with a cutoff of 1.5%/d and of −0.5%/d, respectively, are good candidates for predicting the prognosis of patients with early and late mCRPC, in particular those with a low Gleason score. The G after 12 weeks of chemotherapy needs prospective confirmation to be used as a surrogate endpoint of OS after first-line chemotherapy.

In conclusion, the results of our study suggest that the tumor growth fraction is a prognostic factor for early and late mCRPC. The analysis revealed that cell kinetics of mCRPC and OS are affected by what happened to the tumor after chemotherapy. Although chemotherapy does not appear to influence cell growth directly, it seems to affect the tumor growth fraction indirectly as a result of the reduction of the tumor volume. This phenomenon in turn could result in an increase in the kinetics of cell proliferation.


1. Ferlay J, Parkin DM, Steliarova-Foucher E. Estimates of cancer incidence and mortality in Europe in 2008. Eur J Cancer. 2008;46:765–781.

2. De Angelis R, Sant M, Coleman MP, et al. Cancer survival in Europe 1999-2007 by country and age: results of EUROCARE5—a population-based study. Lancet Oncol. 2014;15:23–34.

3. Scher HI, Heller G. 2000 Clinical states in prostate cancer: toward a dynamic model of disease progression. Urology. 2000;55:323–327.

4. Tannock IF, De Wit R, Berry WR, et al. Docetaxel plus prednisone or mitoxantrone plus prednisone for advanced prostate cancer. N Engl J Med. 2004;351:1502–1512.

5. Petrylak DP, Tangen CM, Hussain MHA, et al. Docetaxel and estramustine compared with mitoxantrone and prednisone for advanced refractory prostate cancer. N Engl J Med. 2004;351:1513–1520.

6. De Bono JS, Oudard S, Ozguroglu M, et al. Prednisone plus cabazitaxel or mitoxantrone for metastatic castration-resistant prostate cancer progressing after docetaxel treatment: a randomised open-label trial. Lancet. 2010;376:1147–1154.

7. De Bono JS, Logothetis CJ, Molina A, et al. Abiraterone and increased survival in metastatic prostate cancer. N Engl J Med. 2011;364:1995–2005.

8. Scher HI, Fizazi K, Saad F, et al. Increased survival with enzalutamide in prostate cancer after chemotherapy. N Engl J Med. 2012;367:1187–1197.

9. Norton L. A Gompertzian model of human breast cancer growth. Cancer Res. 1988;48:7067–7071.

10. Mehrara E, Forssell-Aronsson E, Ahlman H, et al. Quantitative analysis of tumor growth rate and changes in tumor marker level: specific growth rate versus doubling time. Acta Oncol. 2009;48: 591–597.

11. Stein WD, Gulley JL, Schlom J, et al. Tumor regression and growth rates determined in five intramural NCI prostate cancer trials: the growth rate constant as an indicator of therapeutic efficacy. Clin Cancer Res. 2011;17:907–917.

12. Colloca G, Venturino A, Addamo G, et al. Prostate-specific antigen growth rate constant after first-line cytotoxic chemotherapy in metastatic castration-resistant prostate cancer: a monoinstitutional experience. Urol Oncol. 2014;32:42.e1–42.e5.

13. Scher HI, Halabi S, Tannock I, et al. Design and end points of clinical trials for patients with progressive prostate cancer and castrate levels of testosterone: recommendations of the Prostate Cancer Clinical Trials Working Group. J Clin Oncol. 2008;26: 1148–1159.

14. Karrison TG, Maitland ML, Stadler WM, et al. Design of phase II cancer trials using a continuous endpoint of change in tumor size: application to a study of sorafenib and erlotinib in non-small-cell lung cancer. J Natl Cancer Inst. 2007;99:1455–1461.

15. Mehrara E, Forssell-Aronsson E, Bernhardt P. Objective assessment of tumour response to therapy based on tumour growth kinetics. Br J Cancer. 2011;105:682–686.