- Email: [email protected]
Low-dose-rate prostate brachytherapy: 4–8 week postimplant prostate-specific antigen a novel predictor of biochemical failure-free survival
Brachytherapy
-
(2017)
-
Low-dose-rate prostate brachytherapy: 4e8 week postimplant prostate-specific antigen a novel predictor of biochemical failure-free survival Cameron M. Callaghan1,*, Lin Wang1, Abhishek Alluri2, Andrew Lauve3, Cynthia Boyer3, William Russell3 1 Tulane University School of Medicine, New Orleans, LA Department of Internal Medicine, Baton Rouge General Hospital, Baton Rouge, LA 3 Department of Radiation Oncology, Pennington Cancer Center, Baton Rouge, LA
2
ABSTRACT
PURPOSE: The purpose of this study was to determine the relationship between patient, disease, and treatment variables and biochemical failure-free survival (bFFS) following low-dose-rate prostate brachytherapy (LDR-BT). METHODS AND MATERIALS: Data from 624 consecutive patients who received LDR-BT for localized prostate cancer between 2002 and 2012 at a single institution were collected for various patient, disease, and treatment characteristics including a 4e8 week postimplant PSA (4e8wkPSA). Subgroup analysis was stratified by risk category and treatment regimen. Analysis was performed using KaplaneMeier survival curves, Cox proportional hazard ratios (HRs), and receiver-operator characteristic curves. RESULTS: A total of 624 consecutive patients were included with followup time of 4.0 3.1 years. Predictors of bFFS included PSA nadir and 4e8wkPSA (HR 5 2.48, p 5 0.000 and HR 5 1.24, p 5 0.000, respectively) for total population. Diabetes mellitus ( p 5 0.026), chronic obstructive pulmonary disease ( p 5 0.000), alcohol use ( p 5 0.024), and age ( p 5 0.002) were predictors for specific subgroups. Receiver-operator characteristic curves 4e8wkPSA were found to be significant ( p 5 0.036). CONCLUSION: 4e8wkPSA is a novel predictor of bFFS for patients receiving LDR-BT across several risk categories and treatment regimens with potential clinical utility as a prognostic indicator. Certain comorbidities and exposure histories also demonstrated significant relationships with bFFS including chronic obstructive pulmonary disease, diabetes mellitus, age, alcohol history, proton pump inhibitor use, PSA nadir, and PSA density. Ó 2017 American Brachytherapy Society. Published by Elsevier Inc. All rights reserved.
Keywords:
Prostate cancer; Brachytherapy; PSA; Predictors; Biochemical failure-free survival
Introduction Low-dose-rate prostate brachytherapy (LDR-BT), alone or in combination with external beam radiation therapy (EBRT) and/or androgen deprivation therapy (ADT), has been shown to be a cost-effective and safe definitive therapy for prostate cancer with complication and failure rates Received 5 May 2017; received in revised form 27 July 2017; accepted 27 July 2017. * Corresponding author. Tulane University School of Medicine, 4117 Vista Lago Cir. Apt. 102, Kissimmee, FL 34741. Tel.: þ1 (504) 373-8328; fax: þ1 (225) 381 - 6578. E-mail address: [email protected] or [email protected] (C.M. Callaghan).
comparable to other local therapies (1e3). Other studies have also investigated possible pretreatment predictors of urinary and erectile function following LDR-BT (4e6) focusing on dosimetry and disease characteristics but have not extensively studied the interaction of various comorbidities or patient characteristics. Many aspects of prostate-specific antigen (PSA) have been examined for their relation to biochemical failure-free survival (bFFS) including PSA nadir, PSA density, PSAvelocity, and varying definitions of biochemical recurrence (BCR) (7). However, to our knowledge, there has not been a study of PSA measured at 4e8 weeks postimplantation. This study aims to characterize the patient population and to identify pretreatment and posttreatment predictors of bFFS.
1538-4721/$ - see front matter Ó 2017 American Brachytherapy Society. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.brachy.2017.07.016
2
C.M. Callaghan et al. / Brachytherapy
Methods and materials Study design and patient population A total of 624 consecutive patients with localized biopsyproven adenocarcinoma of the prostate presented between October 2002 and July 2012 at a single institution. Patients were stratified by NCCN risk categories (8) and by treatment modality (LDR-BT monotherapy vs. LDR-BT þ EBRT vs. LDR-BT þ ADT vs. triple therapy) for subgroup analysis. Subgroups with !45 cases were not included for meaningful regression or survival curve analysis. Retrospective data were collected from electronic medical records per internal review board protocol for patient, disease, treatment, and posttreatment variables including continuous variables (age, initial PSA [iPSA], pretreatment clinician estimated [Clin] and ultrasound [US] prostate volume, PSA density [iPSA/US prostate volume], radiation dose delivered to 90% of the prostate [D90], volume of prostate receiving 100% of the prescription dose [V100], number of needles per implant [# needles], number of seeds per implant [# seeds], prescription dose [Gy of implant], 4e8 week postimplant PSA [4e8wkPSA], time to followup, and PSA nadir), ordinal variables (alcohol, family history of prostate cancer, transient ischemic attack/cerebrovascular accident, Gleason’s Score total [Gleason’s], and T-stage), and categorical variables (smoking, chronic obstructive pulmonary disease [COPD], coronary artery disease, diabetes mellitus [DM], hypertension, peripheral vascular disease, gastroesophageal reflux disease, Gout, hyperlipidemia, perineural invasion, sourcedIodine-125 vs. Palladium-103 [source 125I/103Pd], proton pump inhibitor [PPI] use, and BCR). 4e8wkPSA was defined as any PSA measured 28e56 days postimplant. If multiple PSA measurements were taken during this period, the measurement most proximal to 36 days postimplant was utilized.
-
(2017)
-
neoadjuvant ADT was followed by EBRT and then prostate seed implant. Postimplant pelvic CT scans were performed 1e2 months after implant for dosimetry with goals of D90 5 90e130% of prescription dose and V100 O90% of prostate volume. LDR-BT, EBRT, and/or ADT treatments for later salvage were not included. Outcomes/followup The primary end point in this study was BCR and bFFS. BCR was defined using the Phoenix criteria (9) and recorded as a binary categorical variable without a time component. Definition of bFFS was time from date of implant to date of BCR or date of last followup without BCR. Rising PSA that fulfilled criteria for BCR but fell back to nadir or below without active clinical intervention was reclassified as a PSA bounce and not recorded as BCR unless the patient subsequently failed (10e13). Statistical analysis Total study population as well as subgroups containing O45 cases were analyzed. Descriptive statistics were reported as percentages for categorical variables and mean SDs for continuous variables. Variables were compared between subgroups using c2, KruskaleWallis, and analysis of variance for categorical, ordinal, and continuous variables, respectively. The KaplaneMeier method was utilized to analyze bFFS with log-rank tests between levels. A Cox proportional hazard regression analysis was performed to obtain hazard ratios (HRs). Receiver-operator characteristic (ROC) curves were generated for 4e8wkPSA and BCR. All statistics and figures were generated using IBM SPSS, version 24.0 (SPSS Inc, Chicago, IL), using two-sided statistical testing with a ! 0.05 significance level. Results
Treatment
Patient demographics
LDR-BT implants utilized as monotherapy utilized 40e182 seeds of either 125I or 103Pd in 6e41 needles. Implants were prescribed to doses of 144 Gy for 125I and 125 Gy for 103Pd implants. When used in combination with EBRT (LDR-BT þ EBRT), 3-D conformal or intensitymodulated radiation therapy was used to deliver total median dose of 45e46 Gy at 1.8e2.0 Gy/fraction administered 5 days per week for 5 weeks to the prostate and seminal vesicles with elective coverage of the pelvic lymph nodes to 50 Gy in 25 fractions. Implants were prescribed to a dose of 110 Gy and 90 Gy for 125I and 103Pd sources, respectively. When combined with hormonal therapy (LDR-BT þ ADT), ADT was typically administered neoadjuvantly with a short course of peripheral androgen blockade followed by LHRH agonist for approximately 3 months to reduce prostate volume prior to implant. Long-term ADT was utilized for a median of 2e3 years. When triple-modality therapy was employed (LDR-BT þ EBRT þ ADT), typically, 3 months of
A total of 624 patients underwent LDR-BT from 2002 to 2012. Mean time to followup was 4.0 3.1 years (mean SD). Three subgroups contained $45 cases including low-risk patients treated with LDR-BT monotherapy, intermediate-risk patients treated with LDR-BT monotherapy, and high-risk patients treated with LDR-BT þ EBRT þ ADT. Excluding variables that risk categories were based on (iPSA, Gleason’s total, and T-stage), significant differences between subgroups were found for alcohol use, clinical estimate of prostate volume, PSA density, D90, source 125I/103Pd, 4e8wkPSA, and PSA nadir ( p 5 0.002, 0.039, 0.000, 0.000, 0.000, 0.000, and 0.015 respectively) (Table 1). Biochemical failure-free survival KaplaneMeier curves of bFFS demonstrated PSA nadir and 4e8wkPSA to be significant within each subgroup at
C.M. Callaghan et al. / Brachytherapy
-
(2017)
-
3
Table 1 Patient demographics LDR-BT þ EBRT þ ADT
LDR-BT Variable (statistic) Patient Age (mean SD) Smoking (%) Alcohol (%) Mild/abstinence Moderate Heavy FHxPCa (%) None Moderate Significant COPD (%) CAD (%) DM (%) HTN (%) PVD (%) GERD (%) Gout (%) HLD (%) TIA/CVA (%) None TIA CVA Disease iPSA (ng/mL) (mean SD) Clin. prostate volume (cc) (mean SD) US prostate volume (cc) (mean SD) PSA density (ng/mL)/(cc) (mean SD) Gleason’s total (%) #6 7 8 9 T-Stage (%) T1c T2a T2b T2c $T3a PNI (%) Treatment D90 (mean SD) V100 (mean SD) # Needles (mean SD) # Seeds (mean SD) Source (I/Pd) PPI (%) Posttreatment 4e8wkPSA (mean SD) PSA nadir (mean SD) BCF (%) Followup, years (mean SD)
Low risk (n 5 241)
Intermediate risk (n 5 192)
High risk (n 5 48)
Total (n 5 624)
65.2 7.9 59
66.4 7.9 62
68.0 7.2 68
66.4 7.8 60
46.1 44 6.6
44.8 45.3 8.3
20.8 54.2 20.8
45 45 10
38.2 35.3 22.4 3 18 23 64 3 12 7 32
36.5 32.3 27.6 5 20 20 59 6 6 7 21
37.5 25 31.3 2 23 26 55 4 13 11 28
37 32 24 4 19 22 63 5 8 7 30
91.3 2.9 3.3
94.3 2.1 3.6
87.5 4.2 6.3
93.5 3.0 3.5
6.6 3.6 36.8 11.3 35.7 14.7 0.21 0.14
41.0 143.0 37.3 5.4 36.5 16 1.7 6.6
9.2 22 35.5 12 38.0 22 0.3 2.0
100 d d d
34.8 65.2 d d
10.4 12.5 45.8 29.8
58 32 6 4
93.4 2.1 d d d 25
59.9 1.6 29.7 8.3 d 36
62.5 2.1 8.3 10.4 13.1 57
72 2 13 5 3 15
d
148 18.0 92.0 6.1 23 4 86 19 191/50 17
145 19.6 92.6 4.3 24 5 87 21 115/77 10
112 16.8 93.5 3.7 24 5 89 26 27/21 13
141 23 92 5 24 5 87 22 435/189 14
0.000 d d d 0.000 d
2.92 2.6 0.48 0.73 8 4.4 3.3
2.2 1.8 0.52 0.95 7 3.8 3.0
0.42 0.78 0.12 0.28 15 3.9 2.4
2.25 2.2 0.44 0.8 9 4.0 3.10
0.000 0.015 d d
p-value d d 0.002
d
5.07 32.8 38.7 0.15
2.2 11 28.0 0.08
d d d d d d d d d
0.000 0.039 0.000 0.000
0.000
ANOVA 5 analysis of variance; PSA 5 prostate-specific antigen; PPI 5 proton pump inhibitor; PNI 5 perineural invasion; US 5 ultrasound; iPSA 5 initial PSA; CVA 5 cerebrovascular accident; TIA 5 transient ischemic attack; HLD 5 hyperlipidemia; GERD 5 gastroesophageal reflux disease; PVD 5 peripheral vascular disease; HTN 5 hypertension; DM 5 diabetes mellitus; CAD 5 coronary artery disease; COPD 5 chronic obstructive pulmonary disease; LDR-BT 5 low-dose-rate prostate brachytherapy; EBRT 5 external beam radiation therapy; ADT 5 androgen deprivation therapy; 4e8wkPSA 5 4e8 week postimplant PSA; BCF 5 biochemical failure. p-values for ANOVA, KruskaleWallis, and c2 test for continuous, ordinal, and categorical variables between subgroups, respectively.
4
C.M. Callaghan et al. / Brachytherapy
-
(2017)
-
Table 2 KaplaneMeier curve log-rank tests for bFFS p-value LDR-BT þ EBRT þ ADT
LDR-BT Variable
Levels
PSA nadir 4e8wkPSA COPD DM PPI Alcohol Age T-stage PNI PSA density iPSA
$0.25 vs. !0.25 ng/mL $2.0 vs. !2.0 ng/mL COPD vs. no COPD DM vs. no DM PPI use vs. no PPI use Abstinence/mild vs. moderate vs. heavy !50, 50e59, 60e69, 70e79, 80e89 and $90 years old #T1c vs. $T2 PNI vs. no PNI $0.25 vs. !0.25 ng/mL/cm3 $5.0 vs. !5.0 ng/mL
Low risk (n 5 241)
Intermediate risk (n 5 192)
!0.001 0.003 !0.001 0.020 0.049 d d d d
0.005 0.047 d d d 0.043 d d d
d
d
High risk (n 5 48)
Total (n 5 624)
!0.001 0.011 d d d d 0.012 0.049 0.040 0.039 d
!0.001 d d d d d d d d 0.041 0.012
PSA 5 prostate-specific antigen; PPI 5 proton pump inhibitor; PNI 5 perineural invasion; iPSA 5 initial PSA; LDR-BT 5 low-dose-rate prostate brachytherapy; EBRT 5 external beam radiation therapy; ADT 5 androgen deprivation therapy; DM 5 diabetes mellitus; COPD 5 chronic obstructive pulmonary disease; bFFS 5 biochemical failure-free survival; 4e8wkPSA 5 4e8 week postimplant PSA; d 5 not significant. Variables with no significant findings not included.
cutoffs of 2.0 ng/mL and 0.25 ng/mL, respectively (Table 2, Fig. 1). PSA nadir was also found to be significant for total study population. Low-risk patients treated with LDR-BT also had significantly different survival
curves based on COPD, DM, and PPI use ( p 5 0.000, 0.020, and 0.049, respectively) (Fig. 2). For intermediaterisk patients treated with LDR-BT monotherapy, bFFS and alcohol use was significant ( p 5 0.043). For
Fig. 1. KaplaneMeier curves bFFS by PSA nadir and 4e8wkPSA with log-rank test. bFFS 5 biochemical failure-free survival; PSA 5 prostate-specific antigen; LDR-BT 5 low-dose-rate prostate brachytherapy; EBRT 5 external beam radiation therapy; ADT 5 androgen deprivation therapy; 4e8wkPSA 5 4e8 week postimplant PSA.
C.M. Callaghan et al. / Brachytherapy
-
(2017)
-
5
Figure 1. (continued).
high-risk patients treated with LDR-BT þ EBRT þ ADT, bFFS was significantly different based on age, T-stage, perineural invasion, and PSA density ( p 5 0.012, 0.049, 0.040, and 0.039, respectively). Total study population had significantly different bFFS based on iPSA and PSA density ( p 5 0.012 and 0.041, respectively). Cox proportional HRs of bFFS for low-risk patients treated with LDR-BT monotherapy showed significant PSA nadir, 4e8wkPSA, iPSA, COPD, and DM with HRs of 5.88, 1.66, 1.27, 50.02, and 2.83, respectively (Table 3). For intermediate-risk patients treated with LDR-BT monotherapy, bFFS and alcohol use had a significant inverse relationship with bFFS with an HR of 0.25. High-risk patients treated with LDR-BT þ EBRT þ ADT demonstrated PSA nadir, 4e8wkPSA, and age to have significant relationships with bFFS with HRs of 49.94, 2.56, and 0.82, respectively. Total study population found significant predictors in Gleason’s total, 4e8wkPSA, and PSA nadir with HRs of 2.48, 1.24, and 1.42, respectively.
4e8wkPSA ROC curves ROC curves of 4e8wkPSA were significant for low-risk patients treated with LDR-BT monotherapy (area under curve [AUC] 5 0.792, p 5 0.000), high-risk patients treated with LDR-BT þ EBRT þ ADT (AUC 5 0.904, p 5 0.005), and total study population (AUC 5 0.601, p 5 0.036) (Fig. 3). Discussion This study provides strong evidence that 4e8wkPSA in addition to a variety of pretreatment factors may offer valuable prognostic information early in the posttreatment course after LDR-BT depending on risk category and treatment modality. To our knowledge, this is the first study to explore this metrics utility. Other studies have demonstrated the prognostic utility of 36 and 48 month postimplant PSA (14e16). Lo et al. (16) demonstrated highly significant relationship between 48-month postimplant
6
C.M. Callaghan et al. / Brachytherapy
-
(2017)
-
Fig. 2. KaplaneMeier curves bFFS for significant variables with log-rank tests. (a) Low risk - LDR-BT and diabetes mellitus. (b) Intermediate risk - LDRBT and alcohol use. (c) High risk - LDR-BT þ EBRT þ ADT and age. (d) High risk - LDR-BT þ EBRT þ ADT and PSA density. (e) Total and PSA density. (f) Total and iPSA. bFFS 5 biochemical failure-free survival; LDR-BT 5 low-dose-rate prostate brachytherapy; DM 5 diabetes mellitus; EBRT 5 external beam radiation therapy; ADT 5 androgen deprivation therapy; iPSA 5 initial PSA; BCR 5 biochemical recurrence.
C.M. Callaghan et al. / Brachytherapy
-
(2017)
-
7
Fig. 3. Receiver operator characteristic curves of 4e8wkPSA for BCR. ROC curves for the following cohorts (a) low risk - LDR-BT, (b) intermediate risk LDR-BT, (c) high risk - LDR-BT þ EBRT þ ADT and (d) total study population. LDR-BT 5 low-dose-rate prostate brachytherapy; CI 5 confidence interval; EBRT 5 external beam radiation therapy; ADT 5 androgen deprivation therapy; BCR 5 biochemical recurrence; 4e8wkPSA 5 4e8 week postimplant PSA.
PSA and disease relapse in low- and intermediate-risk patients with ROC AUC of 93.6% on subset analysis of patients with followup durations $8 years. PSA nadir was also a strong prognostic indicator across subgroups occurring at a median of 2.6 years postimplant. Prior studies have shown the significance of achieving a PSA nadir of !0.5 or !0.2 ng/mL (17, 18). This study
confirmed these findings but found significance across more subgroups with cutoff value of !0.25 ng/mL. PSA density was also significant in high-risk patients treated with LDR-BT þ EBRT þ ADT and in total study population. Past studies have shown that having a low PSA density correlates with low rates of BCR even in the setting of suboptimal implantation (19).
8
C.M. Callaghan et al. / Brachytherapy
As subgroups were risk-stratified there were expected significant differences between groups based on iPSA, Gleason’s total, and T-stage (Table 1). There were also significant differences in D90, 4e8wkPSA, PSA nadir, source (125I vs. 103Pd), clinical estimate of prostate volume, and alcohol. This study found bFFS comparable to that of other study populations (1, 7, 16, 20, 21) (Table 3). KaplaneMeier curves seen in Fig. 2 demonstrate multiple factors significant for bFFS within subgroups and total study population. In the subgroup of low-risk patients treated with LDRBT, significant variables included DM, COPD, and PPI use. A relationship between DM and prostate cancer was noted in the REDUCE trial, and subsequent investigations have shown a complex relationship which may involve metformin, ion channels, interactions with androgens and their receptors, and reduced PSA levels (22e25). Reduction in PSA levels in patients with DM has been postulated to present a falsely decreased incidence of low-grade prostate cancer in this population. This study showed decreased levels of bFFS in patients with DM undergoing LDR-BT monotherapy for low-risk prostate cancer. COPD was highly predictive of bFFS in low-risk LDR-BT monotherapy subgroup. Prevalence of COPD was low in this population, resulting in a wide 95% confidence interval for the effect size. However, the lower bound of this interval maintained an impressive HR with highly significant p-value for both Cox regression model and log-rank test of KaplaneMeier curves for bFFS. Patients with COPD have been found to have a higher incidence of prostate cancer compared to the general population (26). One study found that hypoxemia during acute exacerbations of COPD increased serum-free PSA levels from baseline (27). Certain genes related to glutathione S-transferase enzyme family have been implicated in increasing risk of both prostate cancer and COPD (28). In addition, COPD has been identified as a source of independent other cause mortality risk in the setting of prostate cancer (29). The relationship between prostate
-
(2017)
-
cancer and COPD is likely complex and multifactorial, and additional prospective studies are warranted to determine any additional risk prostate cancer patients with comorbid COPD may carry. PPI use improved bFFS in the low-risk patients treated with LDR-BT on KaplaneMeier log-rank test but not on Cox regression. PPIs have been shown to decrease production of PSA in vitro, although using different compounds than those used clinically (30). This finding may simply represent correlates with degree of patient utilization of health care, gastroesophageal reflux disease, or alcohol use, interactions with the cytochrome P450 system and androgen metabolism, or alteration of tumor microenvironment (31e34). In intermediate-risk patients treated with LDR-BT monotherapy, alcohol use had a significant inverse relationship with bFFS with ‘‘abstinent/mild’’ drinkers having the worse bFFS compared to ‘‘heavy’’ drinkers. The role of alcohol in prostate cancer has been controversial. Alcohol use was identified in the REDUCE & the Prostate Cancer Prevention trials as possibly interfering with the effects of 5-alpha reductase inhibitors (35, 36). A 30-year prospective cohort study of Finnish twins demonstrated that while heavy drinkers were more likely to develop prostate cancer, abstainers were at higher risk of prostate cancerespecific mortality when compared to light drinkers (37). Some studies have found the role of alcohol may have more to do with lifetime alcohol consumption (38). However, while a meta-analysis in 2012 concluded that there was no relationship between alcohol and prostate cancerdthis study did not risk-stratify patients (39). The current study found alcohol use to be statistically significant only in the intermediate-risk subgroup. We hypothesize that this may be related to the livers function as metabolizer of androgens or its effect on the cytochrome P450 system (alcohol may have an inductive or inhibitory effect, depending on pattern of use) or androgen production directly. Alternatively, alcoholics may garner more health problems and therefore have more exposure to health care professionals; for
Table 3 bFFS Cox proportional hazard ratios LDR-BT þ EBRT þ ADT
LDR-BT Low risk (n 5 241)
Intermediate risk (n 5 192)
High risk (n 5 48)
Total (n 5 624)
Variable
p-value
HR
95% CI
p-value
HR
95% CI
p-value
HR
95% CI
p-value
HR
95% CI
PSA nadir 4e8wkPSA iPSA COPD DM Alcohol Age Gleason’s total
!0.001 0.001 0.020 !0.001 0.026 d d d
5.88 1.66 1.27 50.02 2.83 d d d
3.106 1.23e2.23 1.04e1.56 6.73e371.86 1.14e7.06 d d d
0.001 d d d d 0.024 d d
3.36 d d d d 0.251 d d
1.67e6.78 d d d d 0.076e0.834 d d
0.007 0.006 d d d d 0.002 d
49.94 2.56 d d d d 0.82 d
2.90e859.81 1.32e4.97 d d d d 0.72e0.93 d
!0.001 !0.001 d d d d d 0.027
2.48 1.24 d d d d d 1.42
2.02e3.06 1.11e1.39 d d d d d 1.04e1.94
PSA 5 prostate-specific antigen; iPSA 5 initial PSA; LDR-BT 5 low-dose-rate prostate brachytherapy; EBRT 5 external beam radiation therapy; ADT 5 androgen deprivation therapy; DM 5 diabetes mellitus; COPD 5 chronic obstructive pulmonary disease; bFFS 5 biochemical failure-free survival; HR 5 hazard ratio; CI 5 confidence interval; 4e8wkPSA 5 4e8 week postimplant PSA; d 5 not significant. Variables with no significant findings not included.
C.M. Callaghan et al. / Brachytherapy
example, portal hypertension-induced hemorrhoids may lead to earlier/more frequent digital rectal examinations compared to those who do not drink or drink in moderation. In the high-risk patients treated with LDRBT þ EBRT þ ADT, age was inversely related to bFFS. The relationship between age and aggressive clinical behavior of prostate cancer has been previously documented (40), and there is evidence that men !55 years old may have genetically distinct pathways of carcinogenesis (41). The lack of significance of Gleason’s total, iPSA, and T-stage on bFFS in subgroup analysis is not surprising as subgroups were already stratified on the basis of these variables. This study had several limitationsdmost notably the institution and selection bias inherent in any retrospective analysis. Studies using national databases are excellent for large study populations, detection of small effects, heterogeneous patient populations, and generalizability. However, they do have limitations in terms of uniformity of data collection methods and provider practices, which may confound results. All statistical tests in this study were sufficiently powered to detect significant effects.
Conclusion This study demonstrated that 4e8wkPSA was predictive of bFFS. Predictors of bFFS that were identified in this study included COPD, DM, and PPI use for lowrisk patients receiving LDR-BT monotherapy, alcohol use for intermediate-risk patients receiving LDR-BT monotherapy, and age for high-risk patients receiving LDR-BT þ EBRT þ ADT.
Acknowledgments The authors would like to thank all patients who participated in this study, the staff of the Pennington Cancer Center Department of Radiation Oncology, the John Hanna-Hoerner Scholarship for Cancer Research, and in particular Katie Vance PhD and Ato Wright MD, PhD, for all their contributions to this study. References [1] Smith GD, Pickles T, Crook J, et al. Brachytherapy improves biochemical failure-free survival in low- and intermediate-risk prostate cancer compared with conventionally fractionated external beam radiation therapy: A propensity score matched analysis. Int J Radiat Oncol Biol Phys 2015;91:505e516. [2] Laviana A, Ilg A, Veruttipong D, et al. Utilizing time-driven activitybased costing to understand the short-and long-term costs of treating localized, low-risk prostate cancer. Cancer 2016;122:447e455. [3] Rasmusson E, Gunnlaugsson A, Kjellen E. Low-dose rate brachytherapy with I-125 seeds has an excellent 5-year outcome with few side effects in patients with low-risk prostate cancer. Acta Oncol 2016;55:1016e1021. [4] Tanimoto R, Bekku K, Katayama N. Predictive factors for acute and late urinary toxicity after permanent interstitial brachytherapy in Japanese patients. Int J Urol 2013;20:812e817.
-
(2017)
-
9
[5] Cesaretti JA, Kao J, Stone NN, et al. Effect of low dose-rate prostate brachytherapy on the sexual health of men with optimal sexual function before treatment: Analysis at $ 7 Years of follow-up. BJU Int 2007;100:362e367. [6] Mian OY, Robertson SP, Narang A, et al. Dosimetric predictors of sexual function decline following low-dose-rate brachytherapy for prostate cancer (PCA). Int J Radiat Oncol Biol Phys 2015;93:S123. [7] Warner A, Pickles T, Crook J, et al. Development of ProCaRS clinical Nomograms for biochemical failure-free survival following either low-dose rate brachytherapy or conventionally fractionated external beam radiation therapy for localized prostate cancer. Cureus 2015;7:e276. [8] Mohler J, Antonarakis E, Armstrong A. NCCN Guidelines, version 2.2017: Prostate Cancer Updates. 2017. [9] Abramowitz M, Li T, Buyyounouski M, et al. The Phoenix definition of biochemical failure predicts for overall survival in patients with prostate cancer. Cancer 2008;112:55e60. [10] Thompson A, Keyes M, Pickles T, et al. Evaluating the Phoenix definition of biochemical failure after (125)I prostate brachytherapy: Can PSA kinetics distinguish PSA failures from PSA bounces? Int J Radiat Oncol Biol Phys 2010;78:415e421. [11] Zwahlen DR, Smith R, Andrianopoulos N, et al. Prostate-specific antigen bounce after permanent iodine-125 prostate brachytherapyean Australian analysis. Int J Radiat Oncol Biol Phys 2011;79:179e187. [12] Mehta NH, Kamrava M, Wang P-C, et al. Prostate-specific antigen bounce after high-dose-rate monotherapy for prostate cancer. Int J Radiat Oncol Biol Phys 2013;86:729e733. [13] Hackett C, Ghosh S, Sloboda R, et al. Distinguishing prostate-specific antigen bounces from biochemical failure after low-dose-rate prostate brachytherapy. J Contemp Brachytherapy 2014;6:247e253. [14] Zelefsky MJ, Kuban DA, Levy LB, et al. Multi-institutional analysis of long-term outcome for stages T1-T2 prostate cancer treated with permanent seed implantation. Int J Radiat Oncol Biol Phys 2007;67:327e333. [15] Kuban DA, Levy LB, Potters L, et al. Comparison of biochemical failure definitions for permanent prostate brachytherapy. Int J Radiat Oncol Biol Phys 2006;65:1487e1493. [16] Lo A, Morris W, Lapointe V, et al. Prostate-specific antigen at 4 to 5 years after low-dose-rate prostate brachytherapy is a strong predictor of disease-free survival. Int J Radiat Oncol Biol Phys 2014;88:87e93. [17] Ko EC, Stone NN, Stock RG. PSA nadir of !0.5 ng/mL following brachytherapy for early-stage prostate adenocarcinoma is associated with freedom from prostate-specific antigen failure. Int J Radiat Oncol Biol Phys 2012;83:600e607. [18] Critz FA, Williams WH, Holladay CT, et al. Using Modern Techniques. Urology. 4295:6e9. [19] Benzaquen D, Delouya G, Menard C, et al. Prostate-specific antigen density is predictive of outcome in suboptimal prostate seed brachytherapy. Brachytherapy 2017;16:348e352. [20] Tran A, Mandall P, Swindell R, et al. Biochemical outcomes for patients with intermediate risk prostate cancer treated with I-125 interstitial brachytherapy monotherapy. Radiother Oncol 2013;109:235e240. [21] Logghe P, Verlinde R, Bouttens F. Long term outcome and side effects in patients receiving low-dose I125 brachytherapy: A retrospective analysis. Int Braz J Urol 2016;42:906e917. [22] Wu C, Moreira DM, Gerber L, et al. Diabetes and prostate cancer risk in the REDUCE trial. Prostate Cancer Prostatic Dis 2011;14:326e331. [23] Grossmann M, Wittert G. Androgens, diabetes and prostate cancer. Endocr Relat Cancer 2012;19:47e62. [24] Karabulut-Bulan O. New insight into metformin action: Diabetes, prostate cancer, ion channels. Marmara Pharm J 2016;20:216e223. [25] MacHiela MJ, Lindstr€om S, Allen NE, et al. Association of type 2 diabetes susceptibility variants with advanced prostate cancer risk in the breast and prostate cancer cohort consortium. Am J Epidemiol 2012;176:1121e1129. [26] Ho C-H, Chen Y-C, Wang J-J, et al. Incidence and relative risk for developing cancer among patients with COPD: A nationwide cohort study in Taiwan. BMJ Open 2017;7:e013195.
10
C.M. Callaghan et al. / Brachytherapy
[27] Ozge C, Bozlu M, Ozgur ES, et al. The impact of hypoxemia on serum total and free prostate-specific antigen levels in patients with chronic obstructive pulmonary disease. Med Oncol 2015;32:156. [28] Thakur H, Gupta L, Sobti RC, et al. Association of GSTM1T1 genes with COPD and prostate cancer in north Indian population. Mol Biol Rep 2011;38:1733e1739. [29] Daskivich TJ, Chamie K, Kwan L, et al. Improved prediction of longterm, other cause mortality in men with prostate cancer. J Urol 2011; 186:1868e1873. [30] Michel V, Licon-Munoz Y, Trujillo K, et al. Inhibitors of vacuolar ATPase proton pumps inhibit human prostate cancer cell invasion and prostatespecific antigen expression and secretion. Int J Cancer 2013;132:E1eE10. [31] Lee Y, Jeon H, Hong J, et al. Proton pump inhibitors enhance the effects of cytotoxic agents in chemoresistant epithelial ovarian carcinoma. Oncotarget 2015;6:35040e35050. [32] Papagerakis S, Bellile E, Peterson L, et al. Proton pump inhibitors and histamine 2 blockers are associated with improved overall survival in patients with head and neck squamous carcinoma. Cancer Prev Res(Phila) 2014;7:1258e1269. [33] Avnet S, Lemma S, Cortini M, et al. Altered pH gradient at the plasma membrane of osteosarcoma cells is a key mechanism of drug resistance. Oncotarget 2016;7:63408e63423. [34] Pereira C, Guedes J, Gonc¸alves M, et al. Lactoferrin selectively triggers apoptosis in highly metastatic breast cancer cells through
[35]
[36]
[37]
[38]
[39]
[40]
[41]
-
(2017)
-
inhibition of plasmalemmal V-Hþ-ATPase. Oncotarget 2016;7: 62144e62158. Fowke J, Howard L, Andriole G, et al. Alcohol intake increases high-grade prostate cancer risk among men taking dutasteride in the REDUCE trial. Eur Urol 2014;66:1133e1138. Gong Z, Kristal AR, Schenk JM, et al. Alcohol consumption, finasteride, and prostate cancer risk: Results from the prostate cancer prevention trial. Cancer 2009;115:3661e3669. Dickerman B, Markt S, Koskenvuo M. Alcohol intake, drinking patterns, and prostate cancer risk and mortality: A 30-year prospective cohort study of Finnish twins. Cancer Causes Control 2016; 27:1049e1058. McGregor SE, Courneya KS, Kopciuk KA, et al. Case-control study of lifetime alcohol intake and prostate cancer risk. Cancer Causes Control 2013;24:451e461. Rota M, Scotti L, Turati F, et al. Alcohol consumption and prostate cancer risk: A meta-analysis of the dose-risk relation. Eur J Cancer Prev 2012;21:350e359. Hamstra DA, Bae K, Pilepich MV, et al. Older age predicts decreased metastasis and prostate cancer-specific death for men treated with radiation therapy: Meta-analysis of radiation therapy oncology group trials. Int J Radiat Oncol Biol Phys 2011;81:1293e1301. Hussein S, Satturwar S, Van der Kwast T. Young-age prostate cancer. J Clin Pathol 2015;68:511e515.
-
(2017)
-
Low-dose-rate prostate brachytherapy: 4e8 week postimplant prostate-specific antigen a novel predictor of biochemical failure-free survival Cameron M. Callaghan1,*, Lin Wang1, Abhishek Alluri2, Andrew Lauve3, Cynthia Boyer3, William Russell3 1 Tulane University School of Medicine, New Orleans, LA Department of Internal Medicine, Baton Rouge General Hospital, Baton Rouge, LA 3 Department of Radiation Oncology, Pennington Cancer Center, Baton Rouge, LA
2
ABSTRACT
PURPOSE: The purpose of this study was to determine the relationship between patient, disease, and treatment variables and biochemical failure-free survival (bFFS) following low-dose-rate prostate brachytherapy (LDR-BT). METHODS AND MATERIALS: Data from 624 consecutive patients who received LDR-BT for localized prostate cancer between 2002 and 2012 at a single institution were collected for various patient, disease, and treatment characteristics including a 4e8 week postimplant PSA (4e8wkPSA). Subgroup analysis was stratified by risk category and treatment regimen. Analysis was performed using KaplaneMeier survival curves, Cox proportional hazard ratios (HRs), and receiver-operator characteristic curves. RESULTS: A total of 624 consecutive patients were included with followup time of 4.0 3.1 years. Predictors of bFFS included PSA nadir and 4e8wkPSA (HR 5 2.48, p 5 0.000 and HR 5 1.24, p 5 0.000, respectively) for total population. Diabetes mellitus ( p 5 0.026), chronic obstructive pulmonary disease ( p 5 0.000), alcohol use ( p 5 0.024), and age ( p 5 0.002) were predictors for specific subgroups. Receiver-operator characteristic curves 4e8wkPSA were found to be significant ( p 5 0.036). CONCLUSION: 4e8wkPSA is a novel predictor of bFFS for patients receiving LDR-BT across several risk categories and treatment regimens with potential clinical utility as a prognostic indicator. Certain comorbidities and exposure histories also demonstrated significant relationships with bFFS including chronic obstructive pulmonary disease, diabetes mellitus, age, alcohol history, proton pump inhibitor use, PSA nadir, and PSA density. Ó 2017 American Brachytherapy Society. Published by Elsevier Inc. All rights reserved.
Keywords:
Prostate cancer; Brachytherapy; PSA; Predictors; Biochemical failure-free survival
Introduction Low-dose-rate prostate brachytherapy (LDR-BT), alone or in combination with external beam radiation therapy (EBRT) and/or androgen deprivation therapy (ADT), has been shown to be a cost-effective and safe definitive therapy for prostate cancer with complication and failure rates Received 5 May 2017; received in revised form 27 July 2017; accepted 27 July 2017. * Corresponding author. Tulane University School of Medicine, 4117 Vista Lago Cir. Apt. 102, Kissimmee, FL 34741. Tel.: þ1 (504) 373-8328; fax: þ1 (225) 381 - 6578. E-mail address: [email protected] or [email protected] (C.M. Callaghan).
comparable to other local therapies (1e3). Other studies have also investigated possible pretreatment predictors of urinary and erectile function following LDR-BT (4e6) focusing on dosimetry and disease characteristics but have not extensively studied the interaction of various comorbidities or patient characteristics. Many aspects of prostate-specific antigen (PSA) have been examined for their relation to biochemical failure-free survival (bFFS) including PSA nadir, PSA density, PSAvelocity, and varying definitions of biochemical recurrence (BCR) (7). However, to our knowledge, there has not been a study of PSA measured at 4e8 weeks postimplantation. This study aims to characterize the patient population and to identify pretreatment and posttreatment predictors of bFFS.
1538-4721/$ - see front matter Ó 2017 American Brachytherapy Society. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.brachy.2017.07.016
2
C.M. Callaghan et al. / Brachytherapy
Methods and materials Study design and patient population A total of 624 consecutive patients with localized biopsyproven adenocarcinoma of the prostate presented between October 2002 and July 2012 at a single institution. Patients were stratified by NCCN risk categories (8) and by treatment modality (LDR-BT monotherapy vs. LDR-BT þ EBRT vs. LDR-BT þ ADT vs. triple therapy) for subgroup analysis. Subgroups with !45 cases were not included for meaningful regression or survival curve analysis. Retrospective data were collected from electronic medical records per internal review board protocol for patient, disease, treatment, and posttreatment variables including continuous variables (age, initial PSA [iPSA], pretreatment clinician estimated [Clin] and ultrasound [US] prostate volume, PSA density [iPSA/US prostate volume], radiation dose delivered to 90% of the prostate [D90], volume of prostate receiving 100% of the prescription dose [V100], number of needles per implant [# needles], number of seeds per implant [# seeds], prescription dose [Gy of implant], 4e8 week postimplant PSA [4e8wkPSA], time to followup, and PSA nadir), ordinal variables (alcohol, family history of prostate cancer, transient ischemic attack/cerebrovascular accident, Gleason’s Score total [Gleason’s], and T-stage), and categorical variables (smoking, chronic obstructive pulmonary disease [COPD], coronary artery disease, diabetes mellitus [DM], hypertension, peripheral vascular disease, gastroesophageal reflux disease, Gout, hyperlipidemia, perineural invasion, sourcedIodine-125 vs. Palladium-103 [source 125I/103Pd], proton pump inhibitor [PPI] use, and BCR). 4e8wkPSA was defined as any PSA measured 28e56 days postimplant. If multiple PSA measurements were taken during this period, the measurement most proximal to 36 days postimplant was utilized.
-
(2017)
-
neoadjuvant ADT was followed by EBRT and then prostate seed implant. Postimplant pelvic CT scans were performed 1e2 months after implant for dosimetry with goals of D90 5 90e130% of prescription dose and V100 O90% of prostate volume. LDR-BT, EBRT, and/or ADT treatments for later salvage were not included. Outcomes/followup The primary end point in this study was BCR and bFFS. BCR was defined using the Phoenix criteria (9) and recorded as a binary categorical variable without a time component. Definition of bFFS was time from date of implant to date of BCR or date of last followup without BCR. Rising PSA that fulfilled criteria for BCR but fell back to nadir or below without active clinical intervention was reclassified as a PSA bounce and not recorded as BCR unless the patient subsequently failed (10e13). Statistical analysis Total study population as well as subgroups containing O45 cases were analyzed. Descriptive statistics were reported as percentages for categorical variables and mean SDs for continuous variables. Variables were compared between subgroups using c2, KruskaleWallis, and analysis of variance for categorical, ordinal, and continuous variables, respectively. The KaplaneMeier method was utilized to analyze bFFS with log-rank tests between levels. A Cox proportional hazard regression analysis was performed to obtain hazard ratios (HRs). Receiver-operator characteristic (ROC) curves were generated for 4e8wkPSA and BCR. All statistics and figures were generated using IBM SPSS, version 24.0 (SPSS Inc, Chicago, IL), using two-sided statistical testing with a ! 0.05 significance level. Results
Treatment
Patient demographics
LDR-BT implants utilized as monotherapy utilized 40e182 seeds of either 125I or 103Pd in 6e41 needles. Implants were prescribed to doses of 144 Gy for 125I and 125 Gy for 103Pd implants. When used in combination with EBRT (LDR-BT þ EBRT), 3-D conformal or intensitymodulated radiation therapy was used to deliver total median dose of 45e46 Gy at 1.8e2.0 Gy/fraction administered 5 days per week for 5 weeks to the prostate and seminal vesicles with elective coverage of the pelvic lymph nodes to 50 Gy in 25 fractions. Implants were prescribed to a dose of 110 Gy and 90 Gy for 125I and 103Pd sources, respectively. When combined with hormonal therapy (LDR-BT þ ADT), ADT was typically administered neoadjuvantly with a short course of peripheral androgen blockade followed by LHRH agonist for approximately 3 months to reduce prostate volume prior to implant. Long-term ADT was utilized for a median of 2e3 years. When triple-modality therapy was employed (LDR-BT þ EBRT þ ADT), typically, 3 months of
A total of 624 patients underwent LDR-BT from 2002 to 2012. Mean time to followup was 4.0 3.1 years (mean SD). Three subgroups contained $45 cases including low-risk patients treated with LDR-BT monotherapy, intermediate-risk patients treated with LDR-BT monotherapy, and high-risk patients treated with LDR-BT þ EBRT þ ADT. Excluding variables that risk categories were based on (iPSA, Gleason’s total, and T-stage), significant differences between subgroups were found for alcohol use, clinical estimate of prostate volume, PSA density, D90, source 125I/103Pd, 4e8wkPSA, and PSA nadir ( p 5 0.002, 0.039, 0.000, 0.000, 0.000, 0.000, and 0.015 respectively) (Table 1). Biochemical failure-free survival KaplaneMeier curves of bFFS demonstrated PSA nadir and 4e8wkPSA to be significant within each subgroup at
C.M. Callaghan et al. / Brachytherapy
-
(2017)
-
3
Table 1 Patient demographics LDR-BT þ EBRT þ ADT
LDR-BT Variable (statistic) Patient Age (mean SD) Smoking (%) Alcohol (%) Mild/abstinence Moderate Heavy FHxPCa (%) None Moderate Significant COPD (%) CAD (%) DM (%) HTN (%) PVD (%) GERD (%) Gout (%) HLD (%) TIA/CVA (%) None TIA CVA Disease iPSA (ng/mL) (mean SD) Clin. prostate volume (cc) (mean SD) US prostate volume (cc) (mean SD) PSA density (ng/mL)/(cc) (mean SD) Gleason’s total (%) #6 7 8 9 T-Stage (%) T1c T2a T2b T2c $T3a PNI (%) Treatment D90 (mean SD) V100 (mean SD) # Needles (mean SD) # Seeds (mean SD) Source (I/Pd) PPI (%) Posttreatment 4e8wkPSA (mean SD) PSA nadir (mean SD) BCF (%) Followup, years (mean SD)
Low risk (n 5 241)
Intermediate risk (n 5 192)
High risk (n 5 48)
Total (n 5 624)
65.2 7.9 59
66.4 7.9 62
68.0 7.2 68
66.4 7.8 60
46.1 44 6.6
44.8 45.3 8.3
20.8 54.2 20.8
45 45 10
38.2 35.3 22.4 3 18 23 64 3 12 7 32
36.5 32.3 27.6 5 20 20 59 6 6 7 21
37.5 25 31.3 2 23 26 55 4 13 11 28
37 32 24 4 19 22 63 5 8 7 30
91.3 2.9 3.3
94.3 2.1 3.6
87.5 4.2 6.3
93.5 3.0 3.5
6.6 3.6 36.8 11.3 35.7 14.7 0.21 0.14
41.0 143.0 37.3 5.4 36.5 16 1.7 6.6
9.2 22 35.5 12 38.0 22 0.3 2.0
100 d d d
34.8 65.2 d d
10.4 12.5 45.8 29.8
58 32 6 4
93.4 2.1 d d d 25
59.9 1.6 29.7 8.3 d 36
62.5 2.1 8.3 10.4 13.1 57
72 2 13 5 3 15
d
148 18.0 92.0 6.1 23 4 86 19 191/50 17
145 19.6 92.6 4.3 24 5 87 21 115/77 10
112 16.8 93.5 3.7 24 5 89 26 27/21 13
141 23 92 5 24 5 87 22 435/189 14
0.000 d d d 0.000 d
2.92 2.6 0.48 0.73 8 4.4 3.3
2.2 1.8 0.52 0.95 7 3.8 3.0
0.42 0.78 0.12 0.28 15 3.9 2.4
2.25 2.2 0.44 0.8 9 4.0 3.10
0.000 0.015 d d
p-value d d 0.002
d
5.07 32.8 38.7 0.15
2.2 11 28.0 0.08
d d d d d d d d d
0.000 0.039 0.000 0.000
0.000
ANOVA 5 analysis of variance; PSA 5 prostate-specific antigen; PPI 5 proton pump inhibitor; PNI 5 perineural invasion; US 5 ultrasound; iPSA 5 initial PSA; CVA 5 cerebrovascular accident; TIA 5 transient ischemic attack; HLD 5 hyperlipidemia; GERD 5 gastroesophageal reflux disease; PVD 5 peripheral vascular disease; HTN 5 hypertension; DM 5 diabetes mellitus; CAD 5 coronary artery disease; COPD 5 chronic obstructive pulmonary disease; LDR-BT 5 low-dose-rate prostate brachytherapy; EBRT 5 external beam radiation therapy; ADT 5 androgen deprivation therapy; 4e8wkPSA 5 4e8 week postimplant PSA; BCF 5 biochemical failure. p-values for ANOVA, KruskaleWallis, and c2 test for continuous, ordinal, and categorical variables between subgroups, respectively.
4
C.M. Callaghan et al. / Brachytherapy
-
(2017)
-
Table 2 KaplaneMeier curve log-rank tests for bFFS p-value LDR-BT þ EBRT þ ADT
LDR-BT Variable
Levels
PSA nadir 4e8wkPSA COPD DM PPI Alcohol Age T-stage PNI PSA density iPSA
$0.25 vs. !0.25 ng/mL $2.0 vs. !2.0 ng/mL COPD vs. no COPD DM vs. no DM PPI use vs. no PPI use Abstinence/mild vs. moderate vs. heavy !50, 50e59, 60e69, 70e79, 80e89 and $90 years old #T1c vs. $T2 PNI vs. no PNI $0.25 vs. !0.25 ng/mL/cm3 $5.0 vs. !5.0 ng/mL
Low risk (n 5 241)
Intermediate risk (n 5 192)
!0.001 0.003 !0.001 0.020 0.049 d d d d
0.005 0.047 d d d 0.043 d d d
d
d
High risk (n 5 48)
Total (n 5 624)
!0.001 0.011 d d d d 0.012 0.049 0.040 0.039 d
!0.001 d d d d d d d d 0.041 0.012
PSA 5 prostate-specific antigen; PPI 5 proton pump inhibitor; PNI 5 perineural invasion; iPSA 5 initial PSA; LDR-BT 5 low-dose-rate prostate brachytherapy; EBRT 5 external beam radiation therapy; ADT 5 androgen deprivation therapy; DM 5 diabetes mellitus; COPD 5 chronic obstructive pulmonary disease; bFFS 5 biochemical failure-free survival; 4e8wkPSA 5 4e8 week postimplant PSA; d 5 not significant. Variables with no significant findings not included.
cutoffs of 2.0 ng/mL and 0.25 ng/mL, respectively (Table 2, Fig. 1). PSA nadir was also found to be significant for total study population. Low-risk patients treated with LDR-BT also had significantly different survival
curves based on COPD, DM, and PPI use ( p 5 0.000, 0.020, and 0.049, respectively) (Fig. 2). For intermediaterisk patients treated with LDR-BT monotherapy, bFFS and alcohol use was significant ( p 5 0.043). For
Fig. 1. KaplaneMeier curves bFFS by PSA nadir and 4e8wkPSA with log-rank test. bFFS 5 biochemical failure-free survival; PSA 5 prostate-specific antigen; LDR-BT 5 low-dose-rate prostate brachytherapy; EBRT 5 external beam radiation therapy; ADT 5 androgen deprivation therapy; 4e8wkPSA 5 4e8 week postimplant PSA.
C.M. Callaghan et al. / Brachytherapy
-
(2017)
-
5
Figure 1. (continued).
high-risk patients treated with LDR-BT þ EBRT þ ADT, bFFS was significantly different based on age, T-stage, perineural invasion, and PSA density ( p 5 0.012, 0.049, 0.040, and 0.039, respectively). Total study population had significantly different bFFS based on iPSA and PSA density ( p 5 0.012 and 0.041, respectively). Cox proportional HRs of bFFS for low-risk patients treated with LDR-BT monotherapy showed significant PSA nadir, 4e8wkPSA, iPSA, COPD, and DM with HRs of 5.88, 1.66, 1.27, 50.02, and 2.83, respectively (Table 3). For intermediate-risk patients treated with LDR-BT monotherapy, bFFS and alcohol use had a significant inverse relationship with bFFS with an HR of 0.25. High-risk patients treated with LDR-BT þ EBRT þ ADT demonstrated PSA nadir, 4e8wkPSA, and age to have significant relationships with bFFS with HRs of 49.94, 2.56, and 0.82, respectively. Total study population found significant predictors in Gleason’s total, 4e8wkPSA, and PSA nadir with HRs of 2.48, 1.24, and 1.42, respectively.
4e8wkPSA ROC curves ROC curves of 4e8wkPSA were significant for low-risk patients treated with LDR-BT monotherapy (area under curve [AUC] 5 0.792, p 5 0.000), high-risk patients treated with LDR-BT þ EBRT þ ADT (AUC 5 0.904, p 5 0.005), and total study population (AUC 5 0.601, p 5 0.036) (Fig. 3). Discussion This study provides strong evidence that 4e8wkPSA in addition to a variety of pretreatment factors may offer valuable prognostic information early in the posttreatment course after LDR-BT depending on risk category and treatment modality. To our knowledge, this is the first study to explore this metrics utility. Other studies have demonstrated the prognostic utility of 36 and 48 month postimplant PSA (14e16). Lo et al. (16) demonstrated highly significant relationship between 48-month postimplant
6
C.M. Callaghan et al. / Brachytherapy
-
(2017)
-
Fig. 2. KaplaneMeier curves bFFS for significant variables with log-rank tests. (a) Low risk - LDR-BT and diabetes mellitus. (b) Intermediate risk - LDRBT and alcohol use. (c) High risk - LDR-BT þ EBRT þ ADT and age. (d) High risk - LDR-BT þ EBRT þ ADT and PSA density. (e) Total and PSA density. (f) Total and iPSA. bFFS 5 biochemical failure-free survival; LDR-BT 5 low-dose-rate prostate brachytherapy; DM 5 diabetes mellitus; EBRT 5 external beam radiation therapy; ADT 5 androgen deprivation therapy; iPSA 5 initial PSA; BCR 5 biochemical recurrence.
C.M. Callaghan et al. / Brachytherapy
-
(2017)
-
7
Fig. 3. Receiver operator characteristic curves of 4e8wkPSA for BCR. ROC curves for the following cohorts (a) low risk - LDR-BT, (b) intermediate risk LDR-BT, (c) high risk - LDR-BT þ EBRT þ ADT and (d) total study population. LDR-BT 5 low-dose-rate prostate brachytherapy; CI 5 confidence interval; EBRT 5 external beam radiation therapy; ADT 5 androgen deprivation therapy; BCR 5 biochemical recurrence; 4e8wkPSA 5 4e8 week postimplant PSA.
PSA and disease relapse in low- and intermediate-risk patients with ROC AUC of 93.6% on subset analysis of patients with followup durations $8 years. PSA nadir was also a strong prognostic indicator across subgroups occurring at a median of 2.6 years postimplant. Prior studies have shown the significance of achieving a PSA nadir of !0.5 or !0.2 ng/mL (17, 18). This study
confirmed these findings but found significance across more subgroups with cutoff value of !0.25 ng/mL. PSA density was also significant in high-risk patients treated with LDR-BT þ EBRT þ ADT and in total study population. Past studies have shown that having a low PSA density correlates with low rates of BCR even in the setting of suboptimal implantation (19).
8
C.M. Callaghan et al. / Brachytherapy
As subgroups were risk-stratified there were expected significant differences between groups based on iPSA, Gleason’s total, and T-stage (Table 1). There were also significant differences in D90, 4e8wkPSA, PSA nadir, source (125I vs. 103Pd), clinical estimate of prostate volume, and alcohol. This study found bFFS comparable to that of other study populations (1, 7, 16, 20, 21) (Table 3). KaplaneMeier curves seen in Fig. 2 demonstrate multiple factors significant for bFFS within subgroups and total study population. In the subgroup of low-risk patients treated with LDRBT, significant variables included DM, COPD, and PPI use. A relationship between DM and prostate cancer was noted in the REDUCE trial, and subsequent investigations have shown a complex relationship which may involve metformin, ion channels, interactions with androgens and their receptors, and reduced PSA levels (22e25). Reduction in PSA levels in patients with DM has been postulated to present a falsely decreased incidence of low-grade prostate cancer in this population. This study showed decreased levels of bFFS in patients with DM undergoing LDR-BT monotherapy for low-risk prostate cancer. COPD was highly predictive of bFFS in low-risk LDR-BT monotherapy subgroup. Prevalence of COPD was low in this population, resulting in a wide 95% confidence interval for the effect size. However, the lower bound of this interval maintained an impressive HR with highly significant p-value for both Cox regression model and log-rank test of KaplaneMeier curves for bFFS. Patients with COPD have been found to have a higher incidence of prostate cancer compared to the general population (26). One study found that hypoxemia during acute exacerbations of COPD increased serum-free PSA levels from baseline (27). Certain genes related to glutathione S-transferase enzyme family have been implicated in increasing risk of both prostate cancer and COPD (28). In addition, COPD has been identified as a source of independent other cause mortality risk in the setting of prostate cancer (29). The relationship between prostate
-
(2017)
-
cancer and COPD is likely complex and multifactorial, and additional prospective studies are warranted to determine any additional risk prostate cancer patients with comorbid COPD may carry. PPI use improved bFFS in the low-risk patients treated with LDR-BT on KaplaneMeier log-rank test but not on Cox regression. PPIs have been shown to decrease production of PSA in vitro, although using different compounds than those used clinically (30). This finding may simply represent correlates with degree of patient utilization of health care, gastroesophageal reflux disease, or alcohol use, interactions with the cytochrome P450 system and androgen metabolism, or alteration of tumor microenvironment (31e34). In intermediate-risk patients treated with LDR-BT monotherapy, alcohol use had a significant inverse relationship with bFFS with ‘‘abstinent/mild’’ drinkers having the worse bFFS compared to ‘‘heavy’’ drinkers. The role of alcohol in prostate cancer has been controversial. Alcohol use was identified in the REDUCE & the Prostate Cancer Prevention trials as possibly interfering with the effects of 5-alpha reductase inhibitors (35, 36). A 30-year prospective cohort study of Finnish twins demonstrated that while heavy drinkers were more likely to develop prostate cancer, abstainers were at higher risk of prostate cancerespecific mortality when compared to light drinkers (37). Some studies have found the role of alcohol may have more to do with lifetime alcohol consumption (38). However, while a meta-analysis in 2012 concluded that there was no relationship between alcohol and prostate cancerdthis study did not risk-stratify patients (39). The current study found alcohol use to be statistically significant only in the intermediate-risk subgroup. We hypothesize that this may be related to the livers function as metabolizer of androgens or its effect on the cytochrome P450 system (alcohol may have an inductive or inhibitory effect, depending on pattern of use) or androgen production directly. Alternatively, alcoholics may garner more health problems and therefore have more exposure to health care professionals; for
Table 3 bFFS Cox proportional hazard ratios LDR-BT þ EBRT þ ADT
LDR-BT Low risk (n 5 241)
Intermediate risk (n 5 192)
High risk (n 5 48)
Total (n 5 624)
Variable
p-value
HR
95% CI
p-value
HR
95% CI
p-value
HR
95% CI
p-value
HR
95% CI
PSA nadir 4e8wkPSA iPSA COPD DM Alcohol Age Gleason’s total
!0.001 0.001 0.020 !0.001 0.026 d d d
5.88 1.66 1.27 50.02 2.83 d d d
3.106 1.23e2.23 1.04e1.56 6.73e371.86 1.14e7.06 d d d
0.001 d d d d 0.024 d d
3.36 d d d d 0.251 d d
1.67e6.78 d d d d 0.076e0.834 d d
0.007 0.006 d d d d 0.002 d
49.94 2.56 d d d d 0.82 d
2.90e859.81 1.32e4.97 d d d d 0.72e0.93 d
!0.001 !0.001 d d d d d 0.027
2.48 1.24 d d d d d 1.42
2.02e3.06 1.11e1.39 d d d d d 1.04e1.94
PSA 5 prostate-specific antigen; iPSA 5 initial PSA; LDR-BT 5 low-dose-rate prostate brachytherapy; EBRT 5 external beam radiation therapy; ADT 5 androgen deprivation therapy; DM 5 diabetes mellitus; COPD 5 chronic obstructive pulmonary disease; bFFS 5 biochemical failure-free survival; HR 5 hazard ratio; CI 5 confidence interval; 4e8wkPSA 5 4e8 week postimplant PSA; d 5 not significant. Variables with no significant findings not included.
C.M. Callaghan et al. / Brachytherapy
example, portal hypertension-induced hemorrhoids may lead to earlier/more frequent digital rectal examinations compared to those who do not drink or drink in moderation. In the high-risk patients treated with LDRBT þ EBRT þ ADT, age was inversely related to bFFS. The relationship between age and aggressive clinical behavior of prostate cancer has been previously documented (40), and there is evidence that men !55 years old may have genetically distinct pathways of carcinogenesis (41). The lack of significance of Gleason’s total, iPSA, and T-stage on bFFS in subgroup analysis is not surprising as subgroups were already stratified on the basis of these variables. This study had several limitationsdmost notably the institution and selection bias inherent in any retrospective analysis. Studies using national databases are excellent for large study populations, detection of small effects, heterogeneous patient populations, and generalizability. However, they do have limitations in terms of uniformity of data collection methods and provider practices, which may confound results. All statistical tests in this study were sufficiently powered to detect significant effects.
Conclusion This study demonstrated that 4e8wkPSA was predictive of bFFS. Predictors of bFFS that were identified in this study included COPD, DM, and PPI use for lowrisk patients receiving LDR-BT monotherapy, alcohol use for intermediate-risk patients receiving LDR-BT monotherapy, and age for high-risk patients receiving LDR-BT þ EBRT þ ADT.
Acknowledgments The authors would like to thank all patients who participated in this study, the staff of the Pennington Cancer Center Department of Radiation Oncology, the John Hanna-Hoerner Scholarship for Cancer Research, and in particular Katie Vance PhD and Ato Wright MD, PhD, for all their contributions to this study. References [1] Smith GD, Pickles T, Crook J, et al. Brachytherapy improves biochemical failure-free survival in low- and intermediate-risk prostate cancer compared with conventionally fractionated external beam radiation therapy: A propensity score matched analysis. Int J Radiat Oncol Biol Phys 2015;91:505e516. [2] Laviana A, Ilg A, Veruttipong D, et al. Utilizing time-driven activitybased costing to understand the short-and long-term costs of treating localized, low-risk prostate cancer. Cancer 2016;122:447e455. [3] Rasmusson E, Gunnlaugsson A, Kjellen E. Low-dose rate brachytherapy with I-125 seeds has an excellent 5-year outcome with few side effects in patients with low-risk prostate cancer. Acta Oncol 2016;55:1016e1021. [4] Tanimoto R, Bekku K, Katayama N. Predictive factors for acute and late urinary toxicity after permanent interstitial brachytherapy in Japanese patients. Int J Urol 2013;20:812e817.
-
(2017)
-
9
[5] Cesaretti JA, Kao J, Stone NN, et al. Effect of low dose-rate prostate brachytherapy on the sexual health of men with optimal sexual function before treatment: Analysis at $ 7 Years of follow-up. BJU Int 2007;100:362e367. [6] Mian OY, Robertson SP, Narang A, et al. Dosimetric predictors of sexual function decline following low-dose-rate brachytherapy for prostate cancer (PCA). Int J Radiat Oncol Biol Phys 2015;93:S123. [7] Warner A, Pickles T, Crook J, et al. Development of ProCaRS clinical Nomograms for biochemical failure-free survival following either low-dose rate brachytherapy or conventionally fractionated external beam radiation therapy for localized prostate cancer. Cureus 2015;7:e276. [8] Mohler J, Antonarakis E, Armstrong A. NCCN Guidelines, version 2.2017: Prostate Cancer Updates. 2017. [9] Abramowitz M, Li T, Buyyounouski M, et al. The Phoenix definition of biochemical failure predicts for overall survival in patients with prostate cancer. Cancer 2008;112:55e60. [10] Thompson A, Keyes M, Pickles T, et al. Evaluating the Phoenix definition of biochemical failure after (125)I prostate brachytherapy: Can PSA kinetics distinguish PSA failures from PSA bounces? Int J Radiat Oncol Biol Phys 2010;78:415e421. [11] Zwahlen DR, Smith R, Andrianopoulos N, et al. Prostate-specific antigen bounce after permanent iodine-125 prostate brachytherapyean Australian analysis. Int J Radiat Oncol Biol Phys 2011;79:179e187. [12] Mehta NH, Kamrava M, Wang P-C, et al. Prostate-specific antigen bounce after high-dose-rate monotherapy for prostate cancer. Int J Radiat Oncol Biol Phys 2013;86:729e733. [13] Hackett C, Ghosh S, Sloboda R, et al. Distinguishing prostate-specific antigen bounces from biochemical failure after low-dose-rate prostate brachytherapy. J Contemp Brachytherapy 2014;6:247e253. [14] Zelefsky MJ, Kuban DA, Levy LB, et al. Multi-institutional analysis of long-term outcome for stages T1-T2 prostate cancer treated with permanent seed implantation. Int J Radiat Oncol Biol Phys 2007;67:327e333. [15] Kuban DA, Levy LB, Potters L, et al. Comparison of biochemical failure definitions for permanent prostate brachytherapy. Int J Radiat Oncol Biol Phys 2006;65:1487e1493. [16] Lo A, Morris W, Lapointe V, et al. Prostate-specific antigen at 4 to 5 years after low-dose-rate prostate brachytherapy is a strong predictor of disease-free survival. Int J Radiat Oncol Biol Phys 2014;88:87e93. [17] Ko EC, Stone NN, Stock RG. PSA nadir of !0.5 ng/mL following brachytherapy for early-stage prostate adenocarcinoma is associated with freedom from prostate-specific antigen failure. Int J Radiat Oncol Biol Phys 2012;83:600e607. [18] Critz FA, Williams WH, Holladay CT, et al. Using Modern Techniques. Urology. 4295:6e9. [19] Benzaquen D, Delouya G, Menard C, et al. Prostate-specific antigen density is predictive of outcome in suboptimal prostate seed brachytherapy. Brachytherapy 2017;16:348e352. [20] Tran A, Mandall P, Swindell R, et al. Biochemical outcomes for patients with intermediate risk prostate cancer treated with I-125 interstitial brachytherapy monotherapy. Radiother Oncol 2013;109:235e240. [21] Logghe P, Verlinde R, Bouttens F. Long term outcome and side effects in patients receiving low-dose I125 brachytherapy: A retrospective analysis. Int Braz J Urol 2016;42:906e917. [22] Wu C, Moreira DM, Gerber L, et al. Diabetes and prostate cancer risk in the REDUCE trial. Prostate Cancer Prostatic Dis 2011;14:326e331. [23] Grossmann M, Wittert G. Androgens, diabetes and prostate cancer. Endocr Relat Cancer 2012;19:47e62. [24] Karabulut-Bulan O. New insight into metformin action: Diabetes, prostate cancer, ion channels. Marmara Pharm J 2016;20:216e223. [25] MacHiela MJ, Lindstr€om S, Allen NE, et al. Association of type 2 diabetes susceptibility variants with advanced prostate cancer risk in the breast and prostate cancer cohort consortium. Am J Epidemiol 2012;176:1121e1129. [26] Ho C-H, Chen Y-C, Wang J-J, et al. Incidence and relative risk for developing cancer among patients with COPD: A nationwide cohort study in Taiwan. BMJ Open 2017;7:e013195.
10
C.M. Callaghan et al. / Brachytherapy
[27] Ozge C, Bozlu M, Ozgur ES, et al. The impact of hypoxemia on serum total and free prostate-specific antigen levels in patients with chronic obstructive pulmonary disease. Med Oncol 2015;32:156. [28] Thakur H, Gupta L, Sobti RC, et al. Association of GSTM1T1 genes with COPD and prostate cancer in north Indian population. Mol Biol Rep 2011;38:1733e1739. [29] Daskivich TJ, Chamie K, Kwan L, et al. Improved prediction of longterm, other cause mortality in men with prostate cancer. J Urol 2011; 186:1868e1873. [30] Michel V, Licon-Munoz Y, Trujillo K, et al. Inhibitors of vacuolar ATPase proton pumps inhibit human prostate cancer cell invasion and prostatespecific antigen expression and secretion. Int J Cancer 2013;132:E1eE10. [31] Lee Y, Jeon H, Hong J, et al. Proton pump inhibitors enhance the effects of cytotoxic agents in chemoresistant epithelial ovarian carcinoma. Oncotarget 2015;6:35040e35050. [32] Papagerakis S, Bellile E, Peterson L, et al. Proton pump inhibitors and histamine 2 blockers are associated with improved overall survival in patients with head and neck squamous carcinoma. Cancer Prev Res(Phila) 2014;7:1258e1269. [33] Avnet S, Lemma S, Cortini M, et al. Altered pH gradient at the plasma membrane of osteosarcoma cells is a key mechanism of drug resistance. Oncotarget 2016;7:63408e63423. [34] Pereira C, Guedes J, Gonc¸alves M, et al. Lactoferrin selectively triggers apoptosis in highly metastatic breast cancer cells through
[35]
[36]
[37]
[38]
[39]
[40]
[41]
-
(2017)
-
inhibition of plasmalemmal V-Hþ-ATPase. Oncotarget 2016;7: 62144e62158. Fowke J, Howard L, Andriole G, et al. Alcohol intake increases high-grade prostate cancer risk among men taking dutasteride in the REDUCE trial. Eur Urol 2014;66:1133e1138. Gong Z, Kristal AR, Schenk JM, et al. Alcohol consumption, finasteride, and prostate cancer risk: Results from the prostate cancer prevention trial. Cancer 2009;115:3661e3669. Dickerman B, Markt S, Koskenvuo M. Alcohol intake, drinking patterns, and prostate cancer risk and mortality: A 30-year prospective cohort study of Finnish twins. Cancer Causes Control 2016; 27:1049e1058. McGregor SE, Courneya KS, Kopciuk KA, et al. Case-control study of lifetime alcohol intake and prostate cancer risk. Cancer Causes Control 2013;24:451e461. Rota M, Scotti L, Turati F, et al. Alcohol consumption and prostate cancer risk: A meta-analysis of the dose-risk relation. Eur J Cancer Prev 2012;21:350e359. Hamstra DA, Bae K, Pilepich MV, et al. Older age predicts decreased metastasis and prostate cancer-specific death for men treated with radiation therapy: Meta-analysis of radiation therapy oncology group trials. Int J Radiat Oncol Biol Phys 2011;81:1293e1301. Hussein S, Satturwar S, Van der Kwast T. Young-age prostate cancer. J Clin Pathol 2015;68:511e515.