Fax: (304) 243-5047
Primary Gleason pattern does not impact survival after permanent interstitial brachytherapy for Gleason score 7 prostate cancer
Version of Record online: 4 JUN 2007
Copyright © 2007 American Cancer Society
Volume 110, Issue 2, pages 289–296, 15 July 2007
How to Cite
Merrick, G. S., Galbreath, R. W., Butler, W. M., Waller, K. E., Allen, Z. A., Lief, J. and Adamovich, E. (2007), Primary Gleason pattern does not impact survival after permanent interstitial brachytherapy for Gleason score 7 prostate cancer. Cancer, 110: 289–296. doi: 10.1002/cncr.22793
- Issue online: 29 JUN 2007
- Version of Record online: 4 JUN 2007
- Manuscript Accepted: 21 MAR 2007
- Manuscript Revised: 15 MAR 2007
- Manuscript Received: 7 FEB 2007
- prostate cancer;
- Gleason score 7;
- cause-specific survival;
- overall survival;
The impact of primary Gleason pattern was determined on cause-specific (CSS), biochemical progression-free (bPFS), and overall survival (OS) after brachytherapy for Gleason score 7 prostate cancer.
From April 1995 to October 2003, 530 patients underwent brachytherapy for Gleason score 3+4 (n = 300) or Gleason 4+3 (n = 230) prostate cancer. All patients underwent brachytherapy more than 3 years before analysis. The median follow-up was 5.7 years. Of the 530 patients, 412 (77.7%) received supplemental external beam radiation therapy (XRT) and 177 (33.4%) received androgen deprivation therapy. bPFS was defined by a prostate-specific antigen (PSA) ≤0.40 ng/mL after nadir. Multiple parameters were evaluated as predictors of CSS, bPFS, and OS.
At 10 years, Gleason 3+4 versus 4+3 did not predict for CSS (96.7% vs 93.3%, P = .506), bPFS (97.0% vs 92.9%, P = .085), or OS (77.0% vs 78.0%, P = .933). Cox linear regression analysis demonstrated that clinical stage and radiation dose (D90) predicted for CSS, whereas pretreatment PSA, clinical stage, and prostate size predicted for bPFS. Patient age, diabetes, and tobacco were the strongest predictors for OS. To date, 57 patients have died, with 80.7% due to cardiovascular/pulmonary events or secondary malignancies. Five patients have died of prostate cancer.
The primary Gleason pattern did not impact CSS, bPFS, or OS in Gleason score 7 prostate cancer. Deaths from cardiovascular/pulmonary disease and second malignancies were 9.6 times more common than death from prostate cancer. Cancer 2007. © 2007 American Cancer Society.
Gleason score is 1 of the most important prognosticators for predicting biochemical progression and cause-specific survival after the definitive treatment of prostate cancer.1–8 Gleason score 7 is especially important, because it is 1 of the most common criteria for assigning patients a higher risk of treatment failure. Among patients with Gleason score 7, controversy exists regarding the importance of the primary pattern. Greater rates of extracapsular extension without significant differences in seminal vesicle or pelvic lymph node involvement have been demonstrated when clinical stage and pretreatment prostate-specific antigen (PSA) are controlled for in patients Gleason 4+3 versus 3+4.9 Within the uro-oncology community the perception exists that biochemical outcome for Gleason 4+3 is inferior to Gleason 3+4. However, conflicting conclusions have accumulated after radical prostatectomy1, 3–5, 8, 10–12 and brachytherapy,7, 13 whereas external beam radiation therapy (XRT) series have consistently reported inferior biochemical outcomes for Gleason 4+3 histology.6, 14
In a large recently published radical prostatectomy series from the Mayo Clinic, higher rates of biochemical failure, systemic recurrences, and cause-specific death were reported for Gleason 4+3 versus 3+4 patients.8 To date, the impact of Gleason 4+3 versus 3+4 on cause-specific and overall survival after brachytherapy has not been reported. Accordingly, we chose to evaluate the impact of primary Gleason pattern in Gleason score 7 patients on cause-specific (CSS), biochemical progression-free (bPFS), and overall survival (OS) after permanent interstitial brachytherapy with or without supplemental therapies.
MATERIALS AND METHODS
Between April 1995 and October 2003, 530 consecutive patients with Gleason score 7 prostate cancer underwent permanent interstitial brachytherapy by a single brachytherapist (G.S.M.). All patients underwent brachytherapy more than 3 years before analysis. Before the formulation of a treatment plan, all biopsy slides were reviewed by a single pathologist (E.A.). The primary Gleason pattern (3 vs 4) was defined by involvement of at least 51% of the specimen. Our preplanning technique, intraoperative philosophy, and dosimetric evaluation have been described in detail.15, 16 Calculation algorithms and seed parameters used in preplanning and postoperative dosimetry were those recommended by the American Association of Physicists in Medicine Task Group No. 43 (TG-43).17
Patients were clinically staged by medical history, physical examination, and serum PSA. A radiographic work-up consisting of a bone scan and computed tomography (CT) of the pelvis were routinely obtained for high-risk (Gleason score 7 with PSA >10 ng/mL and/or clinical stage ≥T2c) but not for intermediate risk (Gleason score 7, PSA ≤10 ng/mL and ≤ clinical stage T2b) disease. No patient underwent seminal vesicle or pathologic lymph node staging. Table 1 summarizes the clinical, treatment, and dosimetric parameters of the patient population stratified by primary Gleason pattern. Of the 530 patients, 300 were diagnosed with Gleason 3+4 and 230 with Gleason 4+3 histology.
|Gleason 3 + 4, n = 300||Gleason 4 + 3, n = 230||Total, n = 530|
|Age at implant, y||67.0||66.2||±7.1||67.0||66.1||±7.7||.953||67.0||66.1||±7.4|
|Percentage positive biopsies||36.9||43.4||±24.4||50.0||50.2||±25.4||.002||41.7||46.3||±25.1|
|ADT (mo, all patients)||0.0||1.9||±4.5||0.0||3.4||±6.8||.005||0.0||2.5||±5.7|
|ADT (mo, ADT patients)||4.0||6.3||±6.4||4.0||9.0||±8.5||.017||4.0||7.6||±7.6|
|Most recent PSA||<0.04||0.04||±0.07||<0.04||0.05||±0.09||.180||<0.04||0.04||±0.08|
|No. (%)||No. (%)||P||No. (%)|
|Intermediate||237 (79.0)||152 (66.1)||389 (73.4)|
|High||63 (21.0)||78 (33.9)||141 (26.6)|
|T1b-T2b||283 (94.3)||201 (87.4)||484 (91.3)|
|T2c-T3a||17 (5.7)||29 (12.6)||46 (8.7)|
|Pd-103||278 (92.7)||216 (93.9)||494 (93.2)|
|I-125||22 (7.3)||14 (6.1)||36 (6.8)|
|No||68 (22.7)||50 (21.7)||118 (22.3)|
|Yes||232 (77.3)||180 (78.3)||412 (77.7)|
|None||209 (69.7)||144 (62.6)||353 (66.6)|
|<6 mo||73 (24.3)||51 (22.2)||124 (23.4)|
|>6 mo||18 (6.0)||35 (15.2)||53 (10.0)|
|Yes||143 (47.7)||119 (51.7)||262 (49.4)|
|No||157 (52.3)||111 (48.3)||268 (50.6)|
|Yes||29 (9.7)||35 (15.2)||64 (12.1)|
|No||271 (90.3)||195 (84.8)||466 (87.9)|
|Never||108 (36.0)||90 (39.1)||198 (37.4)|
|Former||133 (44.3)||102 (44.3)||235 (44.3)|
|Current||59 (19.7)||38 (16.5)||97 (18.3)|
The brachytherapy target volume consisted of the prostate gland and periprostatic region with a resultant planning volume approximately 1.75 × the ultrasound-determined volume.15, 18 The minimal peripheral dose (mPD) was prescribed to the target volume with margin.
Of the 530 patients, 494 (93.2%) were implanted with Pd-103 and 36 (6.8%) with I-125. For monotherapeutic approaches the mPD was 125 Gy (National Institute of Standards and Technologies [NIST-99]) for Pd-103 and 145 Gy (TG-43) for I-125. In general, the brachytherapy boost doses were 90 Gy mPD (NIST-99 for Pd-103) and 110 Gy mPD (TG-43) for I-125. For patients receiving 20 Gy of supplemental XRT, all patients received a Pd-103 boost of 115 mPD. The brachytherapy procedure was performed with preloaded 18G needles using transverse and sagittal ultrasonography with fluoroscopic confirmation. At implantation, the prostate gland, periprostatic region, and base of the seminal vesicles were implanted.16
Of the 530 patients, 412 (77.7%) received supplemental XRT. Two hundred and ninety-three received 45 Gy in 1.8-Gy fractions utilizing 15–18 MV photons delivered via a 3D conformal technique utilizing 4 fields (opposed laterals and AP/PA) with custom treatment devices. For those patients receiving XRT with <10% incidence of pelvic lymph node involvement,9 the target volume consisted of the prostate gland/seminal vesicles with a 2.0 cm superior margin and inclusion of the pelvic side wall. For patients with a pelvic lymph node risk >10%,9 the target volume consisted of the prostate gland and pelvic lymph nodes with the superior border at the L5-S1 interface. The remaining 119 supplemental XRT patients received 20 Gy as part of a prospective randomized trial. As per protocol, the target volume consisted of the prostate gland and seminal vesicles with 2.0 cm margins in all directions except posteriorly (1.0 cm margin). In all cases, supplemental XRT was delivered before brachytherapy.
Androgen deprivation therapy (ADT) was administered for suboptimal geometry or poor prognosticators. Of the 530 patients, 177 (33.4%) received ADT, with 124 (23.4% of all patients) receiving cytoreductive ADT (≤6 months), whereas 53 (10.0% of all patients) received prolonged ADT (>6 months) primarily for poor prognosticators. ADT was initiated 3 months before implantation and consisted of a luteinizing hormone releasing hormone (LHRH) agonist and an antiandrogen. In hormonally manipulated patients, the mean and median duration of ADT was 7.6 ± 7.6 months and 4.0 months, respectively (range, 3–36 months).
After brachytherapy, patients were monitored by physical examination, including digital rectal examination and PSA determinations. Patients were initially evaluated 3 months after brachytherapy and then every 6 months thereafter. The endpoint of the analysis was CSS, bPFS (definition: PSA ≤0.40 ng/mL after nadir), and OS.19 No patient underwent routine postimplant biopsy. Clinical, treatment, and dosimetric parameters evaluated for CSS, bPFS, and OS included primary Gleason pattern, patient age, clinical T-stage, pretreatment PSA, risk group, percent positive biopsies, body mass index (BMI), perineural invasion, isotope, supplemental XRT, ADT, duration of ADT, prostate volume, percentage of the target volume receiving 100% of the prescribed dose (V100) minimum percentage of the dose covering 90% of the target volume (D90), tobacco status, hypertension, and diabetes.
Cause of death was determined for each deceased patient. Patients with metastatic prostate cancer and/or hormone refractory disease without obvious metastases who died of any cause were classified as dead of prostate cancer. All other deaths were attributed to the immediate cause of death.
Clinical and treatment variables that were continuous were compared between the 2 groups using an independent T-test. Categorical variables were compared using a chi-square analyses. Gleason 4+3 versus 3+4 was also stratified by categorical variables and then CSS, bPFS, and OS were determined using Kaplan-Meier analyses. Univariate Cox regression analysis was used to determine if any of the clinical or treatment variables predicted for failure. Those variables with P < .10 were then entered into a multivariate, forward conditional Cox regression. For all tests, P ≤ .05 was considered significant. Statistical analysis was performed with SPSS v. 13.0 software (SPSS, Chicago, Ill).
Table 1 summarizes the clinical, treatment, and dosimetric parameters for the study population stratified by primary Gleason pattern. The mean and median follow-up for the entire study population was 6.0 ± 2.3 years and 5.7 years, respectively. Gleason 4+3 histology was associated with relatively small clinical but statistically significant differences in pretreatment PSA, percent positive biopsies, BMI, risk group, clinical stage, and the incidence of prolonged (>6 months) ADT. No correlation was discerned between the dominant histologic pattern and patient age, follow-up, dosimetric quality, isotope, supplemental XRT, hypertension, diabetes, or tobacco usage. The most recent median posttreatment PSA in biochemically controlled patients was <0.04 ng/mL for both the Gleason 3+4 and 4+3 cohorts.
Figure 1 illustrates CSS, bPFS, and OS for the entire study population. At 10 years, CSS, bPFS, and OS were 95.2%, 95.2%, and 77.3%, respectively. Figure 2 shows CSS, bPFS, and OS stratified by Gleason 3+4 versus 4+3. No statistical differences were identified between Gleason 3+4 or 4+3 for CSS (96.7% vs 93.3%, P = .056), bPFS (97.3% vs 92.9%, P = .085), or OS (77.0% vs 78.0%, P = .933). In Table 2, univariate and multivariate Cox regression analyses determined predictors for CSS, bPFS, and OS. Delivered dose (D90) and clinical stage were the best predictors for CSS, whereas pretreatment PSA, larger prostate volume, and lower clinical stage were predictors for bPFS. Patient age, diabetes, and tobacco were the strongest predictors for OS. Statistically, Gleason 3+4 versus 4+3 did not impact any of the survival parameters.
|Cause-specific survival||Biochemical progression-free survival||Overall survival|
|% Positive biopsies||.486||.005||1.022||.063||.093||.160|
|Body mass index||.874||.145||.099||.458|
|Risk (int. vs high)*||.090||.436||<.001||6.436||.099||.696|
|Never vs Former||.032||2.258|
|Never vs Current||.010||3.080/|
To date, 57 patients have died (Table 3). The 29 deaths from cardiovascular/pulmonary disease (5.5% of the study population) and 17 deaths from second malignancies (3.2% of the study population) accounted for 80.7% of all deaths, whereas only 5 patients died from metastatic prostate cancer (0.9% of the study population). Deaths from cardiovascular/pulmonary disease and second malignancies were 9.6 times more common than death from prostate cancer. For all nonprostate cancer deaths, the mean and median time to death was 4.00 ± 2.46 years and 3.56 years, respectively, after brachytherapy. The mean and median time to death for prostate cancer death was 6.26 ± 3.25 years and 6.42 years, respectively
|3 + 4 n = 300 no. (%)||4 + 3 n = 230 no. (%)||Total n = 530 no. (%)|
|Prostate cancer||2 (0.7)||3 (1.3)||5 (0.9)|
|Myocardial infarction||11 (3.7)||10 (4.3)||21 (4.0)|
|CVA/neurologic||1 (0.3)||3 (1.3)||4 (0.8)|
|Pulmonary||2 (0.7)||2 (0.9)||4 (0.8)|
|Lung cancer||3 (1.0)||2 (0.9)||5 (0.9)|
|GI malignancy||5 (1.7)||1 (0.4)||6 (1.1)|
|Leukemia||2 (0.7)||1 (00.4)||3 (0.6)|
|Head/neck cancer||1 (0.3)||0 (0.0)||1 (0.2)|
|GBMF||2 (0.7)||0 (0.0)||2 (0.4)|
|GI bleed||0 (0.0)||1 (0.4)||1 (0.2)|
|Sepsis||1 (0.3)||0 (0.0)||1 (0.2)|
|Alzheimer||1 (0.3)||1 (0.4)||2 (0.4)|
|Trauma||1 (0.3)||1 (0.4)||2 (0.4)|
|Total deaths||32 (10.7)||25 (10.9)||57 (10.8)|
The histologic assessment of Gleason score represents an important prognostic factor of biologic aggressiveness, with higher Gleason scores negatively influencing outcome.1–8 Although the perception exists that Gleason score 7 can be stratified into prognostic categories by the dominant histologic grade, conflicting results have been published in the radical prostatectomy and brachytherapy literature.1, 3–5, 7, 8, 10–13 A Mayo Clinic Gleason score 7 radical prostatectomy series reported increased bPFS (48% vs 38%, P < .001), fewer distant metastases (8% vs 15%, P < .001), and a higher CSS (97% vs 93%, P = .013) at 10 years for Gleason score 7 patients with primary pattern-3 histology.8 The CSS was virtually identical in the Mayo Clinic RP series and in our current report (Fig. 2A), with absolute differences in 10-year CSS of 4% at the Mayo Clinic and 3.4% at our institution. In contrast to radical prostatectomy, the dominant histologic pattern did not predict for any of the survival variance after brachytherapy (Table 2). The cause for the discrepancy could reflect differences in patient selection, pathologic evaluation, or treatment effect.
Despite an identical definition of PSA success (<0.40 ng/mL), bPFS after brachytherapy was substantially higher than radical prostectomy (Fig. 2). Brachytherapy is possibly more resilient to extracapsular extension than radical prostatectomy and less dependent on tumor bulk than XRT because of generous periprostatic treatment margins (approximately 6 mm) and intraprostatic dose escalation.15, 16, 18 It is possible that the ability to eradicate extracapsular extension could mute the difference in cancer eradication rates for patients with Gleason 3+4 versus 4+3 cancers. Patients at significant risk for extracapsular extension but at a low risk for pelvic lymph node involvement/distant metastases should benefit from an aggressive locoregional approach. A review of the Partin tables illustrates that the vast majority of our patients possessed a substantial risk of extracapsular extension but a low risk of seminal vesicle and/or pelvic lymph node involvement.9
Consistent with our bPFS findings, Herman et al.1 failed to identify the primary histologic grade in Gleason score 7 as a significant predictor of disease progression (P = .76). Groll et al.11 reported that stratification of biopsy Gleason score 7 by primary pattern 3 or 4 resulted in no significant differences in the incidence of extracapsular extension, positive surgical margins, or seminal vesicle involvement. Recently, Vira et al.12 reported that biochemical outcome in Gleason 4+3 patients was comparable to Gleason 3+4 as long as the percent positive biopsies were <50%. In our series, the mean percent positive biopsies were statistically greater but clinically comparable between the 3+4 and 4+3 cohorts (43.4% vs 50.2%).
For Gleason score 7 prostate cancer, significant controversy exists regarding patterns of failure when stratified by the dominant histologic grade. Chan et al.4 reported 10-year radical prostatectomy bPFS rates of 56% and 37% for Gleason 3+4 versus 4+3, respectively (P < .0001), with the conclusion that Gleason 4+3 resulted in greater biochemical progression independent of stage and margin status, suggesting that failure was the result of subclinical distant metastatic disease. Although only 1.5% of patients in the Mayo Clinic series had positive pelvic lymph nodes, a greater risk of distant metastatic disease was noted in patients with Gleason 4+3.8 These findings regarding patterns of failure in Gleason score 7 patients are inconsistent with our results and those of others. Aggressive locoregional approaches have resulted in improvements in bPFS, decreased distant metastases, and improved CSS and OS.14, 20–23 In our current study, 10-year CSS was highly dependent on the delivered radiation dose (D90; Table 2). Consistent with a prior report,24 larger prostate size predicted for a more favorable biochemical outcome.
Shortcomings of our study include the retrospective, single-institutional nature of the evaluation, the administration of supplemental XRT and ADT for multiple reasons, and reliance on biopsy findings rather than whole section prostate specimens. The nonexpirative nature of brachytherapy has been used to criticize Gleason score results because of the inability to evaluate the entirety of the prostate gland and assign a pathologic Gleason score. However, the Shared Equal Access Regional Cancer Hospital (SEARCH) database reported that biopsy Gleason score was strongly correlated with disease progression even when the pathologic Gleason score was available.25 Although supplemental XRT doses of 20 Gy and 45 Gy were used in the current study, in a prospective randomized trial our group demonstrated that supplemental XRT dose does not impact bPFS.26 In addition, we previously reported that the use of ADT does not impact brachytherapy-related CSS or OS.27 This is in contrast to a prior brachytherapy series that reported an inverse relation between ADT use and OS.28 Potentially, the conflicting conclusions are due to the fact that patients in our series were substantially younger and probably healthier as witnessed by substantial differences in 10-year OS.27, 28 Consistent with our previous findings, neither supplemental XRT or androgen deprivation therapy statistically impacted CSS, bPFS, or OS (Table 2).
Part of the lack of primary Gleason pattern effect in our patient sample is likely due to the relatively small number of prostate cancer-related deaths. In our series cardiovascular/pulmonary disease and second malignancies accounted for 46 of the 57 deaths (80.7%), with only 5 of the deaths (8.8%) attributable to prostate cancer (0.9% of all patients). Nonprostate cancer deaths were nearly 10 times more likely than prostate cancer deaths. Changes in lifestyle to improve cardiovascular health and cessation of tobacco should improve OS in patients with clinically localized prostate cancer. In our clinic, after completion of treatment, all patients are encouraged to begin physical rehabilitation to maximize cardiovascular heath to include daily aerobic exercise, resistance training 3 times per week for the prevention of osteoporosis and maintenance of muscle strength, a high fiber/low fat diet, maintenance of ideal body weight, cessation of all tobacco products, alcohol in moderation, routine colonoscopy, and screening for hypertension, diabetes, and hypercholesterolemia.29, 30
After brachytherapy the dominant histologic pattern in Gleason score 7 prostate cancer did not impact CSS, bPFS, or OS. Death as a result of cardiovascular or pulmonary disease and second malignancies accounted for almost 10 times more deaths than prostate cancer.