C-reactive protein (CRP) has been associated with outcomes in patients with metastatic adenocarcinoma of the prostate. Associations between prostate adenocarcinoma-specific endpoints and CRP in patients who are treated for localized disease remain unknown.
In total, 206 patients who received radiation therapy for adenocarcinoma of the prostate had at least 1 CRP measured in follow-up and were analyzed. The primary outcome was biochemical failure-free survival. In addition, associations were examined between CRP and prostate-specific antigen (PSA).
On univariate analysis, higher CRP levels were associated significantly with shorter biochemical failure-free survival for patients who received radiation therapy after undergoing radical prostatectomy. For patients who were managed with definitive radiation therapy alone, higher CRP levels also were associated significantly with shorter biochemical failure-free survival on univariate and multivariable analyses (hazard ratio, 2.03; 95% confidence interval, 1.19-3.47; P = .009). In addition, CRP levels were associated significantly with PSA after radical prostatectomy for patients who had Gleason scores ≥8 (P = .037), for high-risk patients (P = .008), and for those with pretreatment PSA levels >20 ng/mL (P = .05). In patients who received definitive radiation therapy, CRP levels also were associated with PSA both for those with pretreatment PSA levels >20 ng/mL (P < .001), and for the intermediate-risk (P = .029) and high-risk (P = .009) subgroups.
Adenocarcinoma of the prostate remains the most common noncutaneous cancer in men and accounts for a large number of cancer-related deaths annually. Despite numerous large prospective trials, the optimal management of adenocarcinoma of the prostate remains controversial. Although the prostate-specific antigen (PSA) is a well established prognostic biomarker, its sensitivity and specificity remain less than optimal.[3, 4] The need for additional biomarkers to complement the PSA in adenocarcinoma of the prostate is profound.
The evidence that supports the role of systemic inflammation in the pathogenesis of numerous solid tumors is mounting. Several studies have linked the development of various chronic inflammatory states to carcinogenesis, including Helicobacter pylori-related gastric disease, inflammatory bowel disease, and chronic hepatitis.[6-8] It is believed that this is mediated through the paracrine actions of cytokines, adhesion molecules, and mediators of angiogenesis generated by the inflammatory response.[9, 10] The inflammatory response has also been increasingly associated with prostate carcinogenesis. Intriguing studies examining prostate autoantibodies have demonstrated that the inflammatory response a sensitive and specific test to distinguish between men who harbor adenocarcinoma of the prostate and those who do not. Furthermore, it is frequently observed that inflammatory T-cell infiltrates accompany adenocarcinoma of the prostate. It also has been hypothesized that chronic intraprostatic inflammation may contribute to the development of adenocarcinoma of the prostate.
C-reactive protein (CRP) is a well studied acute-phase reactant that is a sensitive and specific marker of tissue damage and inflammation. Circulating CRP levels have been increasingly associated with prognosis in a variety of solid tumors. Specifically, elevated CRP has been associated with shorter survival in renal cell carcinoma, ovarian cancer, endometrial cancer, cervical cancer, esophageal cancer, non-Hodgkin lymphoma, colorectal cancer, and melanoma. Moreover, a variety of urologic malignancies have been associated with systemic inflammation and CRP levels. Several series have examined CRP in metastatic adenocarcinoma of the prostate, and CRP recently was associated with lower overall survival (OS) in castrate-resistant adenocarcinoma of the prostate and also with a lower probability of PSA response to docetaxel-based chemotherapy.[24-26] Finally, a recent analysis in nonmetastatic adenocarcinoma of the prostate has been presented that associated CRP with OS and prostate adenocarcinoma-specific survival; however, that series has been criticized as low-quality evidence with critical limitations.
Although OS is an important endpoint for any prognostic biomarker, the well established association of CRP with other comorbidities, particularly cardiovascular disease, complicates its association with OS in localized adenocarcinoma of the prostate. In addition, the behavior and response of adenocarcinoma of the prostate that develops in the setting of chronic inflammation to ionizing radiation is unknown. In the current investigation, we explore the association of CRP with biochemical failure-free survival in patients with localized adenocarcinoma of the prostate who received radiation therapy.
MATERIALS AND METHODS
An institutional database of more than 1500 patients who attended radiation therapy consultation during the years 2002 through 2012 was analyzed. Patients who had CRP levels recorded in follow-up for any reason were included. Patients were excluded if they had absent demographic information, were missing radiation therapy or hormone therapy treatment information, did not have a CRP value and corresponding date recorded, or only had a recorded CRP value that was later than the date of biochemical failure if 1 occurred, only had a CRP recorded before treatment, did not receive radiation therapy of any kind, did not complete a full prescribed course of radiation therapy, had documented metastatic disease, or had missing biochemical failure-free survival information. The primary outcome was biochemical failure-free survival. The Phoenix consensus definition of biochemical failure was used in the analysis, and a biochemical failure was defined as the nadir PSA plus 2 ng/mL. Biochemical failure-free survival was defined as the time from the end of treatment to a Phoenix-defined biochemical failure or the date of last follow-up if there was no failure. Death was not accounted for as a competing risk, because the death rate was less than 1% for the cohort.
Computed tomography simulation data sets were used to design the treatment plans for each of the patients with either a volumetric-modulated arc therapy (VMAT) technique or an intensity-modulated radiation therapy (IMRT) technique. Some patients received either gold marker tracking or Calypso Beacons (Calypso, Seattle, Wash). All patients who did not receive Calypso Beacons had on-board imaging performed daily. The prostate, seminal vesicles, lymph nodes (when treated), and normal structures were manually contoured on the planning computed tomography image using Radiation Therapy Oncology Group guidelines. The VMAT and IMRT plans were designed using the Eclipse treatment planning system (Varian Medical Systems, Palo Alto, Calif). A 23EX, 21EX, or Trilogy linear accelerator (Varian Medical Systems) was used to deliver the VMAT treatment plans. This technique has been previously described. Either a single or double coplanar arc treatment plan was used for VMAT plans, and a 5-field or 7-field treatment plan was used for IMRT.
Patients who were receiving hormone therapy were allowed in the study. Hormone therapy was received by patients who had adverse prognostic factors and consisted of a combination of testosterone receptor antagonist (flutamide or bicalutamide) and a gonadotropin-releasing hormone agonist (leuprolide or goserelin); the duration of hormone therapy ranged from 6 months to 2 years, depending on the risk group. In general, patients received treatment for 6 months in the intermediate-risk group and for 2 years in the high-risk group. The presence or absence of hormone therapy use was known for each patient included in the study.
Patients who received brachytherapy alone received predominately iodine-125 to a prescription dose of approximately 144 grays (Gy) or palladium-103 to a prescription dose of approximately 125 Gy. If patients were treated with a combination of external-beam radiation therapy and brachytherapy, then they received a dose of approximately 45 Gy of external-beam radiation therapy followed by an iodine-125 brachytherapy boost of approximately 110 Gy.
Statistical analysis was conducted using the SAS statistical software package (version 9.3; SAS Institute, Inc., Cary, NC). Descriptive statistics for each variable were reported and included age, Gleason score, receipt of hormone therapy, risk group, race, T-classification, lymph node status, radiation therapy dose, pretreatment PSA level, and the number of positive cores. If patients had more than 1 CRP level, then the most recent CRP measurement was used for the analysis. Lymph node status and the number of positive cores were excluded from further analysis because of the large number of missing values. CRP and pretreatment PSA levels were naturally log-transformed before inclusion in the biochemical failure-free survival analysis because of their skewed distribution and outliers. The unadjusted association of each covariate with biochemical failure-free survival was derived from a Cox model. Multivariable biochemical failure-free survival analysis was conducted including CRP, risk group, and hormone therapy. The remaining covariates—race, age, and radiation dose—were entered into the model subject to a backward variable selection method with α = .20 removal criteria. Risk group was used in place of Gleason score, pretreatment PSA, and T classification.
In addition, each pair of post-treatment CRP and PSA measurements taken within 30 days of each other and before the occurrence of biochemical failure were examined for associations in several exploratory subgroups. Subgroups were defined based on Gleason score (≤6, 7, or ≥8), PSA level (<10 ng/mL, 10-20 ng/mL, or >20 ng/mL), and risk group (low, intermediate, or high). Because of skewed distributions and outliers, PSA and CRP were rank-transformed. Linear regression was conducted using PSA as the outcome and generalized estimating equations to cluster on patient, because 1 patient may have had multiple observations. A linear regression model was fit modeling the effect of CRP on PSA for the total cohort. In addition, to explore associations in different subgroups, models were fit that included the subgroup classification variable of interest and an interaction between that variable and CRP. Standardized regression coefficients for CRP were reported along with P values. Scatter plots of PSA versus CRP also were produced demonstrating the fitted regression lines.
In total, 206 patients were included in the final analysis, and patients were included only if they had a CRP level determined in follow-up after definitive treatment. Patients who underwent radical prostatectomy (RP) (n = 54) and patients who were managed with definitive radiation therapy (n = 152) were analyzed separately. Characteristics of the post-RP patients and the definitive radiation therapy patients are provided in Table 1. The median age was 61 years (range, 45-74 years) for the post-RP group and 66 years (range, 43-83 years) for the definitive radiation therapy group. The remaining baseline characteristics were as expected in this patient cohort.
Table 1. Characteristics of Patients in the Postradical Prostatectomy Group and the Definitive Radiotherapy Group
No. of Patients (%)
Post-RP Group, N = 54
Definitive RT Group, N = 152
Abbreviations: CRP, C-reactive protein; Gy, grays; PSA, prostate specific antigen; RP, radical prostatectomy; RT, radiotherapy; SD, standard deviation.
CRP: Median [Range], mg/L
Clinical tumor classification
Lymph node status
RT dose, Gy
Pretreatment PSA, ng/mL
No. of positive cores
CRP and PSA
Table 2 indicates that, for post-RP patients who had Gleason scores ≥8, pretreatment PSA levels >20 ng/mL, and those categorized as high risk, there was a statistical association between CRP and PSA. In the definitive radiation therapy group, there was also an association between CRP and PSA in those with pretreatment PSA levels >20 ng/mL and in those categorized as intermediate or high risk (Table 2). Scatter plots illustrating these associations are provided in Figures 1 and 2.
Table 2. Association of C-Reactive Protein With Prostate-Specific Antigen in the Postradical Prostatectomy Group and the Definitive Radiotherapy Group
The univariate association with biochemical failure-free survival for post-RP patients is indicated in Table 3. Although CRP did demonstrate some statistical significance on univariate analysis, this was not maintained in the multivariable analysis (Table 4). Data from the univariate analysis of patients who received definitive radiation therapy are provided in Table 3. The table indicates that CRP was significantly associated with biochemical failure-free survival. Furthermore, on multivariable analysis, CRP retained a significant association with biochemical failure-free survival (hazard ratio, 2.03; 95% confidence interval, 1.19-3.47; P = .009). Complete multivariable analyses for both the post-RP group and the definitive radiation therapy group are provided in Table 4.
Table 3. Univariate Association With Biochemical Failure-Free Survival in the Postradical Prostatectomy Group and the Definitive Radiotherapy Group
The need for additional biomarkers in adenocarcinoma of the prostate is profound. Evidence is mounting to support a role for chronic inflammation as a contributor to malignancy in a variety of disease sites, including adenocarcinoma of the prostate. Furthermore, CRP is a well established biomarker in other urologic malignancies and has been reliably associated with outcomes in both metastatic and castrate-resistant adenocarcinoma of the prostate.[23, 24] In addition, the response to ionizing radiation of adenocarcinoma of the prostate that develops in the setting of chronic inflammation is unknown. Given the differences in angiogenesis, cytokines, and adhesion molecules that contribute to malignancy progression in the setting of chronic inflammation, it is plausible that these cancers respond differently than those developed in a noninflammatory state. In the current series, we have demonstrated a statistical association between CRP and biochemical failure-free survival in patients with adenocarcinoma of the prostate who received radiation therapy.
To our knowledge, there have been a total of 6 series examining CRP specifically in adenocarcinoma of the prostate. All but 1 of those series examined CRP as it correlates with metastatic adenocarcinoma of the prostate or castrate-resistant prostate cancer treated with chemotherapy.[24-27, 31] There was a single prior series that examined CRP in the setting of localized adenocarcinoma of the prostate. The primary conclusions of that single series by McArdle et al were that the presence of a systemic inflammatory response independently predicted for a poor long-term outcome. Furthermore, those authors demonstrated that increased CRP was associated with prostate adenocarcinoma-specific mortality. However, small patient numbers, a heterogeneously treated population, and a failure to control for several known prognostic factors associated with CRP, such as cardiovascular disease, make drawing any firm conclusions from their series difficult. Moreover, although the prostate adenocarcinoma-specific mortality endpoint is intriguing to associate with CRP levels, the reliability of that endpoint is unclear when determined from a retrospective database like that in the series reported by McArdle et al. Evaluating such an endpoint in a retrospective database is difficult and prone to inaccuracy.
Current evidence for the association between CRP and patient outcomes in metastatic adenocarcinoma of the prostate is becoming increasingly robust. Beer et al presented some of the earliest and most convincing data to support the relation between CRP and patient outcomes in their secondary analysis of 160 patients enrolled in the prospective randomized ASCENT trial. In that series, the authors demonstrated that increasing levels of plasma CRP were a strong predictor of OS and a lower probability of PSA response to treatment. This was in a population of patients with castrate-resistant adenocarcinoma of the prostate who were receiving docetaxel-based chemotherapy. Supportive evidence also has been presented by Prins et al, who confirmed the findings of Beer et al in an independent data set. Finally, the same association was observed in other analyses of similar groups of patients, confirming again that the association of CRP with outcomes in metastatic adenocarcinoma of the prostate appears robust. Given the relation with metastatic and castrate-resistant adenocarcinoma of the prostate, it seems plausible that such a relation might exist in the nonmetastatic setting.
There are challenges with the association of CRP and OS in patients with localized adenocarcinoma of the prostate. Because CRP is a very nonspecific marker of systemic inflammation, it is elevated secondary to several other conditions, such as infection, cardiovascular disease, and concurrent illness. Furthermore, CRP is an independent predictor of increased cardiovascular risk and mortality. These factors introduce a large potential for confounders into any analysis associating OS with CRP levels. This confounder becomes particularly pronounced in nonmetastatic adenocarcinoma of the prostate secondary to the long natural history and low incidence of mortality in this malignancy. It is for these reasons that we sought to examine CRP as it associates with a prostate adenocarcinoma-specific endpoint: namely, biochemical failure-free survival.
In our heterogeneous population of 206 retrospectively analyzed patients, we have demonstrated an apparent association between increased CRP levels and lower biochemical failure-free survival. Furthermore, CRP was associated with PSA in several exploratory subgroups. This agrees with the conclusions drawn by McArdle et al; however, we believe that using both larger patient numbers and the endpoint of biochemical failure-free survival contributed to the strength of the current series.
In our smaller population of 54 patients who received radiation therapy after RP, univariate analysis demonstrated that increasing CRP does appear to be associated with worse biochemical failure-free survival. This finding was not maintained on multivariable analysis. Furthermore, CRP levels do appear to correlate with PSA levels in patients who have high-risk disease, higher Gleason score, and higher pretreatment PSA.
Weaknesses of the current series exist beyond its retrospective design. This was a heterogeneous group of patients who received treatment using a variety of techniques, which makes the results of the analysis difficult to interpret. Patients who were included in this series were selected based purely on the availability of CRP measurements in follow-up, which introduces the potential for bias. The study would have been considerably stronger if it had used a prospective cohort of patients with intentionally collected serum CRP measurements for the purpose of such an analysis. Additional strength would have been added had patients had a baseline CRP for comparison before their treatment. A tremendous effort was undertaken to account for and control for as many confounders as possible in this retrospective series, and several factors known to influence CRP were included in our analysis. It should be noted that hormone therapy, which is known to have a considerable effect on biochemical failure, does not have an impact on CRP levels. Moreover, the relation between radiation therapy and CRP level remains unclear. Our initial hope was to explore this association with a correlative analysis of radiation therapy treatment volumes and CRP levels; however, differences in the time interval between CRP collection and radiation therapy delivery introduced a prohibitive confounder into such an analysis. Notwithstanding the limitations of the study, to our knowledge, this is currently the largest series published to date examining CRP in patients with localized adenocarcinoma of the prostate who received radiation therapy. We were able to demonstrate a significant association with CRP and biochemical failure-free survival in the cohort of patients who had localized adenocarcinoma of the prostate treated with definitive radiation therapy.
The association with CRP, biochemical failure-free survival, and PSA requires further validation in a larger patient cohort, ideally as a secondary analysis from a prospective randomized trial. The biologic etiology for this correlation remains unclear and may warrant further mechanistic exploration and clinical validation. Given the controversy that exists surrounding the management of adenocarcinoma of the prostate, additional prognostic biomarkers for biochemical failure-free survival would be tremendously helpful for guiding treatment decisions. However, before this biomarker is ready for routine clinical use, its prognostic significance must be validated in a larger and prospectively enrolled patient cohort.
In conclusion, we have presented the results of a large, exploratory, retrospective analysis examining the association between CRP and biochemical failure-free survival in patients with nonmetastatic adenocarcinoma of the prostate who received treatment with radiation therapy. There appears to be a significant association between PSA, biochemical failure-free survival, and CRP levels for patients with adenocarcinoma of the prostate who receive radiation therapy. In those patients who received radiation therapy after RP, a significant association between CRP and biochemical failure-free survival was demonstrated on univariate analysis but was not maintained on multivariable analysis. Although they are intriguing, these findings require further validation in a larger and controlled cohort of patients. Given the importance of additional biomarkers in adenocarcinoma of the prostate it is the conclusion of the authors that CRP should be explored further and validated as a secondary endpoint in the setting of a prospective randomized trial.
This work was supported by a grant from the Georgia Cancer Coalition.
CONFLICT OF INTEREST DISCLOSURES
Dr. Jani has received a grant from the Georgia Cancer Coalition.