Prognostic Factors and Risk Stratification in Patients With Castration-Resistant Prostate Cancer Receiving Docetaxel-based Chemotherapy

Shimpei Yamashita; Yasuo Kohjimoto; Takashi Iguchi; Hiroyuki Koike; Hiroki Kusumoto; Akinori Iba; Kazuro Kikkawa; Yoshiki Kodama; Nagahide Matsumura; Isao Hara


BMC Urol. 2016;16(13) 

In This Article


Two randomized trials demonstrated significant survival improvements in CRPC patients treated with docetaxel-based chemotherapy,[11,12] and it is widely used, including in Japan. Recently, several novel agents, including enzaltamide and abiraterone acetate, were shown to prolong overall survival in CRPC patients and have appeared on the market.[5,8] However, docetaxel still remains a standard option, especially in CRPC patients with evidence of progressive disease. Accordingly, this provides clinicians with a wide choice between docetaxel and novel agents.[14] Therefore, it is crucial to identify prognostic factors and develop a risk stratification for CRPC patients treated with docetaxel.

Many prognostic models in patients with CRPC have been developed using pre and post chemotherapeutic factors. Although post chemotherapeutic factors, including PSA decline, tumor response, and pain response, are often used to evaluate patient's prognosis,[15,16] it might be more beneficial for CRPC patients to predict the response and outcome before the initiation of chemotherapy. Armstrong et al. reported that four independent risk factors, including visceral metastases, bone scan progression, significant pain, and anemia (hemoglobin level < 13 g/dL), predicted OS well, and they developed a risk stratification model for CRPC patients treated with docetaxel.[16] This Armstrong risk classification (ARC) is highly reliable because it was developed from 656 CRPC patients treated with docetaxel and was also internally validated in 333 CRPC patients who were administered mitoxantrone from among the 1006 CRPC patients in the TAX 327 study.[11] However, few reports have externally validated the ARC in CRPC patients who were treated with docetaxel. Nakano et al. reported that the ARC could predict the prognosis in docetaxel-treated CRPC patients.[17] On the other hand, Shiota et al. reported that the ARC failed to stratify the cohort satisfactorily.[18] They showed that ALP value, visceral metastasis, and duration from initiation of hormone treatment were independent prognostic factors on multivariate analysis. Indeed, the reported prognostic factors for CRPC patients have differed widely among studies. The prognostic factors for CRPC patients reported in previous studies are shown in Table 3. Other than the above, Bamias et al. reported that baseline PSA >100, pain, weight loss, and simultaneous extraosseous and bone disease were associated with worse prognosis in CRPC patients.[19] Matsuyama et al. reported that a decrease of ≥ 50 % in the PSA, serum markers at the start of docetaxel therapy (PSA, ALP, and CRP), and the number of docetaxel courses were independent predictors of OS, and ALP, hemoglobin, and age at the start of docetaxel therapy were useful for deciding the duration of docetaxel therapy in CRPC patients.[20]

The pre-chemotherapeutic prognostic factors are divided broadly into two categories. The first category reflects cancer progression. PSA value, PSADT, bone scan progression, and significant pain are included in this category. The second category represents the patient's general condition. Age, ECOG PS, and CRP are included in the second category. The magnitude of the benefit provided by each factor has varied among studies.

In this retrospective study, prognostic factors were identified using only prechemotherapeutic factors in a single-institute Japanese cohort. The serum PSA level at the start of chemotherapy and anemia have been reported to be prognostic factors for CRPC patients in previous studies,[16,21,22] and these factors were shown to be independent prognostic factors for CRPC patients treated with docetaxel in the present study. In addition, this study showed that age was also an independent prognostic factor and suggested that ECOG PS and the CRP-to-albumin ratio could be prognostic factors for CRPC patients. Accumulating evidence has indicated that cancer and inflammation are linked,[23] and inflammation-based prognostic scores, including the NLR, Glasgow Prognostic Score (GPS), and Prognostic Nutritional Index (PNI), have been reported to have prognostic value in patients with various types of cancer in the last few years.[24–28] It has been reported that CRP and the NLR are useful clinical markers of the systemic inflammatory response in predicting OS in patients with CRPC treated with docetaxel.[21,22] However, Kinoshita et al. reported that the CRP-to-albumin ratio might be an independent prognostic marker in patients with hepatocellular carcinoma and may have comparable prognostic ability to other established inflammation-based prognostic scores.[29] The present study suggested that this novel inflammation-based prognostic score could be a valuable prognostic marker in CRPC patients.

Our risk stratification could effectively stratify CRPC patients treated with docetaxel in terms of OS. Most of the prognostic factors included in the ARC, such as visceral metastases, bone scan progression, and significant pain, reflect the metastases of cancer. On the other hand, the prognostic factors included in the present risk stratification reflect the patient's general condition. This difference was likely caused by the differences in the baseline characteristics of target patients. In the TAX 327 study, the median serum PSA level at chemotherapy initiation of the patients was over 100 ng/mL.[11] Bone metastases were observed in more than 90 % of patients, and visceral metastases were seen in about 20 %. However, in the present study, the median serum PSA level at chemotherapy initiation was 43.2 ng/mL, and the frequency of bone metastases or visceral metastases was less. Therefore, in the present study, a PSA response of ≥ 50 % from baseline was observed in about 70 % patients, and this PSA response rate was better than in previous clinical trials performed in Western countries, including the TAX327 trial.[11,12] Moreover, the median OS duration in this cohort was 22.5 months, while the median OS duration in the TAX327 cohort was <20 months. The difference of disease burden between our cohort and the TAX327 cohort caused these differences of PSA response and OS. Considering these facts, it appears that factors representing disease progression might be associated with OS in patients with progressive disease, while factors reflecting the patient's general condition might be correlated with OS in patients with less progressive disease – namely, prostate cancer with low serum PSA level at chemotherapy initiation or without bone metastases and visceral metastases. Lower disease burden in our cohort might be one of the reasons that significant pain, included in ARC, was not a significant factor in this study.

With the advent of novel agents, the demand for docetaxel in CRPC patients might slightly decrease. However, the results of the CHAARTED trial indicate the new effect of docetaxel for prostate cancer, and the importance of docetaxel will never be lost. Our study focused on the prognostic factors in CRPC patients receiving docetaxel-based chemotherapy, but the study about the prognostic factors in hormone sensitive prostate cancer (HSPC) patients receiving docetaxel-based chemotherapy will be needed.

This study had some limitations. First, the sample size was small, and the observation period might not have been long enough to determine the actual OS. Second, this study was a retrospective study using data from a single institution. Thirdly, many patients in our cohort received estramustine. It is less likely to combine estramustine to docetaxel based chemotherapy in Europe and United States. In fact, EAU and NCCN guidelines do not recommend to combine estramustine with docetaxel based chemotherapy. However, serious adverse events of estrogen, such as embolism, are reported to be rather rare in Japanese patients.[30] Finally, most patients required docetaxel dose reductions because of adverse events. Additional larger confirmatory studies are warranted to validate our results.