- Email: [email protected]
Contemporary Population-Based Comparison of Localized Ductal Adenocarcinoma and High-Risk Acinar Adenocarcinoma of the Prostate
Oncology Contemporary Population-Based Comparison of Localized Ductal Adenocarcinoma and High-Risk Acinar Adenocarcinoma of the Prostate Vignesh T. Packiam, Sanjay G. Patel, Joseph J. Pariser, Kyle A. Richards, Adam B. Weiner, Gladell P. Paner, David J. VanderWeele, Gregory P. Zagaja, and Scott E. Eggener OBJECTIVE MATERIALS AND METHODS
RESULTS
CONCLUSION
To compare pathological characteristics, treatment patterns, and survival in patients with ductal adenocarcinoma (DC) compared to those with acinar adenocarcinoma (AC). Using the National Cancer Database, we identified patients diagnosed with clinically localized (cN0, cM0) pure DC (n ¼ 1328) and AC (n ¼ 751,635) between 1998 and 2011. High-risk AC was defined as Gleason 8-10. Demographic, treatment, pathological, and survival characteristics of patients were compared. Compared to patients with Gleason 8-10 AC, those with DC presented with lower mean prostate-specific antigen (10.3 vs 16.2 ng/mL, P <.001), had similar rates (11.7% vs 11.5%, P ¼ .8) of clinical extra-capsular extension (stage cT3), and were more likely to undergo prostatectomy (54% vs 36%, P <.001). Compared to patients with Gleason 8-10 AC undergoing prostatectomy, those with DC had more favorable pathology: stage T3 (39% vs 52%, P <.001), fewer positive lymph nodes (4% vs 11%, P <.001), and fewer positive margins (25% vs 33%, P <.001). On Kaplan-Meier analysis, patients with DC had similar 5-year survival (75.0%, 95% confidence interval [CI] [71.7-78.9]) compared to those with Gleason 8-10 AC (77.1%, 95% CI [76.6%-77.6%], P ¼ .2). On Cox multivariable analysis, patients with Gleason 8-10 AC had a similar risk of death compared to those with DC (hazards ratio ¼ 0.92, 95% CI [0.69-1.23], P ¼ 6). In this large contemporary population-based series, patients with DC of the prostate presented with lower prostate-specific antigen, had more favorable pathological features, and similar overall survival compared to men with Gleason 8-10 AC. UROLOGY 86: 777e782, 2015. 2015 Elsevier Inc.
A
cinar adenocarcinoma (AC) represents the vast majority of prostate cancers with the remaining 5%-10% comprising a variety of other histologic subtypes. Ductal carcinoma (DC) comprises w5% of all prostate cancers.1 Some initial reports considered DC to portend a more favorable prognosis.2 Multiple small series subsequently demonstrated more aggressive features associated with DC.3-6 In an update to the Gleason scoring system in 2010, Epstein favored defining these Financial Disclosure: The authors declare that they have no relevant financial interests. From the Section of Urology, Department of Surgery, University of Chicago, Chicago, IL; the Department of Urology, University of Wisconsin-Madison, Madison, WI; the Pritzker School of Medicine, University of Chicago, Chicago, IL; the Department of Pathology, University of Chicago, Chicago, IL; and the Department of Medicine, University of Chicago, Chicago, IL Address correspondence to: Vignesh T. Packiam, M.D., The University of Chicago, 5841 South Maryland Ave. MC-6038, Chicago, IL 60637. E-mail: vignesh.packiam@ uchospitals.edu Submitted: April 14, 2015, accepted (with revisions): July 3, 2015
ª 2015 Elsevier Inc. All Rights Reserved
tumors as Gleason 4þ4 while retaining the nomenclature of ductal to highlight differences compared to conventional acinar tumors.7 More recently, 2 large studies of the SEER database demonstrated DC is associated with advanced disease at presentation, increased risk of mortality when compared to all patients with AC, and has a similar prognosis to high-risk AC.8,9 Although there is general consensus that DC is more aggressive than AC, there is not necessarily agreement pertaining to recommended treatment modalities.10 Studies have shown varying efficacy for hormone therapy (HT),2,10-12 radiation treatment (RT),10,13,14 and radical prostatectomy (RP).3,4 These discrepancies are likely to be influenced both by differences in delivery of treatment between older and more recent literatures, as well as the small size of most series. The controversies surrounding optimal treatment for this rare histologic subtype were a primary incentive for utilizing a large series http://dx.doi.org/10.1016/j.urology.2015.07.009 0090-4295/15
777
Table 1. Demographic and clinical characteristics of patients by tumor type Parameter Number of patients Age (y) <65 65-75 >75 Ethnicity White Black Other Unknown Charlson comorbidity index None 1 2 Distance from hospital Same zip code <60 miles >60 miles >120 miles Geographic location Northeast South/Southeast Midwest West Location Metro Urban Rural Year of diagnosis 1998-1999 2000-2003 2004-2007 2008-2011 PSA at diagnosis* <10 10-20 >20 Clinical T stage cT1 cT2 cT3 cT4
DC (%)
AC Gleason 6-7 (%)
AC Gleason 8-10 (%)
1328 (0.2)
664,987 (88.3)
86,648 (11.5)
349,788 (53) 252,504 (38) 62,695 (9)
29,374 (34) 36,322 (42) 20,952 (24)
543,242 89,188 17,271 15,286
69,995 12,013 3041 1599
523 (39) 504 (38) 301 (23) 1096 156 53 23
(82.5) (11.8) (4.0) (1.7)
875 (82.4) 160 (15.1) 27 (2.5)
(81.7) (13.4) (2.6) (2.3)
(80.8) (13.9) (3.5) (1.8)
570,996 (85.9) 81,809 (12.3) 12,182 (1.8)
72,451 (83.6) 11,840 (13.7) 2357 (2.7)
42 1043 91 152
(3.2) (78.5) (6.9) (11.4)
18,838 549,718 38,690 57,741
(2.8) (82.7) (5.8) (8.7)
2873 73,173 4254 6348
(3.3) (84.5) (4.9) (7.3)
276 322 325 405
(20.8) (24.2) (24.5) (30.5)
142,837 188,927 177,143 156,080
(21.5) (28.4) (26.6) (23.5)
18,060 22,998 23,340 22,250
(20.8) (26.6) (26.9) (25.7)
1045 (82.4) 186 (14.7) 37 (2.9)
518,498 (81.7) 100,723 (15.9) 14,944 (2.4)
67,079 (80.9) 13,673 (16.5) 2116 (2.6)
(6.0) (19.5) (31.3) (43.2)
0y 0y 338,871 (51.0) 326,116 (49.0)
0y 0y 42,666 (49.2) 43,982 (50.8)
595 (74.1) 125 (15.6) 83 (10.3)
478,706 (81.3) 67,478 (11.5) 42,327 (7.2)
44,430 (57.2) 17,118 (22.0) 16,106 (20.8)
581 377 89 38
416,825 160,962 11,142 536
37,685 30,219 7690 1125
80 258 416 574
(53.5) (34.8) (8.2) (3.5)
(70.7) (27.3) (1.9) (0.1)
P-Value <.001
<.001
<.001
<.001
<.001
<.001
<.001
<.001
<.001
(49.1) (39.4) (10.0) (1.5)
AC, acinar adenocarcinoma; DC, ductal adenocarcinoma. * Mean, median, min, max, interquartile range were: 10.3, 6.3, 0.5, 74.7, [4.3-10.3] for ductal; 9.1, 5.7, 0.4, 73.5, [4.3-8.4] for Gleason 6-7; 16.1, 8.5, 0.5, 98, [5.3-16.9] for Gleason 8-10. y Gleason score was not available for patient diagnosed prior to 2004.
in our study. We utilized the National Cancer Database (NCDB) to describe patient characteristics, patterns of treatment, and clinical outcomes in the largest contemporary cohort to date of DC of the prostate.
MATERIALS AND METHODS We utilized the participant use file (PUF) for prostate cancer to identify our study cohort. Institutional review board exemption was not required since no patient, physician, or hospital identifiers were examined. The database utilized was developed by the NCDB, which is managed by the American College of Surgeons’ Commission on Cancer and the American Cancer Society. The NCDB includes data from over 1400 approved cancer programs and captures information on approximately 70% of all cancers diagnosed in the United States. 778
We identified all patients diagnosed with prostate cancer on prostate biopsy between 1998 and 2011, and only included those who were diagnosed with pure, nonmetastatic (American Joint Committee on Cancer stage cM0 and cN0/cNx) DC and Gleason 6-10 AC of the prostate. Patients were stratified by histology as DC, AC Gleason 6-7, and AC Gleason score 8-10 to allow for comparison across demographic, treatment, and survival characteristics. For patients undergoing RP, it was unknown whether pathology data reflected prostate biopsy or final surgical pathology. The NCDB comprises data initially entered at participating hospital-based cancer registries, which includes patient demographics (age, race, metropolitan urban and/or rural home county, distance traveled to cancer facility, year of diagnosis and Deyo-Charlson Comorbidity Index [CCI]),15 cancer facility (type and location), cancer characteristics (prostate-specific antigen [PSA] and American Joint Committee on Cancer UROLOGY 86 (4), 2015
Table 2. Treatment of patients by tumor type Parameter Radical prostatectomy Yes No Radiation Yes No Local therapy No local therapy RP RT RP þ RT
DC (%)
AC Gleason 6-7 (%)
AC Gleason 8-10 (%)
713 (53.8) 612 (46.2)
346,452 (52.2) 316,697 (47.8)
31,147 (36.1) 55,158 (63.9)
370 (28.1) 948 (71.9)
239,315 (36.2) 421,353 (63.8)
44,618 (51.8) 41,461 (48.2)
306 647 308 61
88,627 334,064 229,479 9537
16,092 25,602 39,308 5238
P-value <.001
(23.2) (48.9) (23.3) (4.6)
(13.4) (50.5) (34.7) (1.4)
<.001 <.001
(18.6) (29.7) (45.6) (6.1)
RP, Radical Prostatectomy; RT, Radiation Treatment; other abbreviations as in Table 1.
Table 3. Demographic, clinical, pathologic characteristics of patients undergoing RP by tumor type Parameter Number of patients Age (y) <65 65-75 >75 PSA at diagnosis* <10 10-20 >20 Clinical T stage cT1 cT2 cT3 cT4 Pathological T stage pT1 pT2 pT3 pT4 Unknown Lymph nodes None Dissection Unknown Node status Positive Negative Margin status Positive Negative
AC Gleason 6-7 (%)
AC Gleason 8-10 (%)
647 (0.2)
DC (%)
334,064 (92.7)
25,602 (7.1)
370 (57.2) 253 (39.1) 24 (3.7)
233,170 (69.8) 97,819 (29.3) 3075 (0.9)
14,109 (55.1) 10,779 (42.1) 714 (2.8)
370 (81.3) 54 (11.9) 31 (6.8)
248,628 (85.3) 25,175 (8.7) 14,468 (5.0)
15,764 (70.1) 4166 (18.5) 2571 (11.4)
328 162 32 2
(62.6) (30.9) (6.1) (0.4)
192,132 77,586 5128 119
(69.9) (28.2) (1.9) (0.04)
11,360 7717 1610 59
(54.7) (37.2) (7.8) (0.3)
1 356 236 14 37
(0.2) (55.3) (36.6) (2.2) (5.7)
341 261,525 48,087 934 20,747
(0.1) (78.8) (14.5) (0.3) (6.3)
21 11,035 12,690 435 1245
(0.1) (43.4) (49.9) (1.7) (4.9)
133 (20.6) 484 (74.8) 30 (4.6)
141,593 (42.4) 191,090 (57.2) 1381 (0.4)
4371 (17.1) 21,141 (82.6) 91 (0.36)
24 (4.4) 522 (95.6)
2414 (1.0) 230,198 (99.0)
2330 (10.7) 19,389 (89.3)
157 (24.6) 480 (75.4)
60,893 (18.4) 269,968 (81.6)
8276 (32.7) 17,019 (67.3)
P-Value <.001
<.001
<.001
<.001
<.001
<.001 <.001
Abbreviations as in Tables 1 and 2. * Mean, median, min, max, interquartile range were: 8.8, 5.9, 0.5, 92.6, [4.2-8.3] for ductal; 7.8, 5.3, 0.4, 98, [4.1-7.4] for Gleason 6-7; 11.2, 6.8, 0.5, 98, [4.8-11.3] for Gleason 8-10.
clinical and/or pathological stage), and treatment modalities (RP, RT, HT, and chemotherapy). A patient’s setting is defined as their hometown’s population categorized as metropolitan (>250,000), urban (>2500), or rural (<2500). Hospital variables include geographic location and program type. Programs are categorized based on the volume of patients per year: comprehensive (>500), community (100-500), and academic (>500). Academic programs are required to provide graduate medical programs in at least four different areas. Patients were further categorized as receiving local therapy if they underwent RP, RT, or RP þ RT. The RP þ RT group UROLOGY 86 (4), 2015
included patients who underwent RP followed by adjuvant or salvage RT. The dosage of RT was not specified in the NCDB, and may have ranged between 65 and 80 Gy over the duration of this series. There was inconsistent use and documentation of androgen deprivation in NCDB, and thus this was not included in analysis. The NCDB recorded PSA starting in 2004 per collaborative staging stipulations as the highest value prior to the diagnostic procedure or, if that value is unavailable, the earliest pretreatment but postdiagnostic value. Statistical analysis was performed using Stata, version 12 (College Station, TX). Categorical variables are presented as 779
whole numbers and percentages, which were compared utilizing the Fisher’s exact and Mann-Whitney U tests. Kaplan-Meier survival analysis and stratified log-rank test were utilized to compare overall survival stratified by histology. Overall survival was estimated as time from the diagnosis of prostate cancer to death. Patients known to be alive were censored at the last follow-up evaluation or the last contact by correspondence. Patients were only included if they had follow-up time of at least 5 years. A multivariable Cox proportional hazards model was performed to identify independent predictors of overall survival. For all statistical measures, a 2-sided P-value <.05 was considered statistically significant.
RESULTS We identified and analyzed 1328 (0.2%) and 715,635 (99.8%) patients with nonmetastatic DC and AC of the prostate, respectively. Of the patients with AC, 664,987 (88.5%) had Gleason 6-7 disease and 86,648 (11.5%) had Gleason 8-10 disease. The median follow-up for the entire cohort was 81 months. The demographic and clinical characteristics of patients with DC, Gleason 6-7 AC, and Gleason 8-10 AC are summarized in Table 1. There was a similar proportion of young (<65 years old) patients presenting with DC compared to patients with Gleason 8-10 AC (39% vs 34%, P <.001). Patients with DC presented with similar mean PSA compared to all patients with AC (10.3 vs 10.0 ng/mL, P ¼ .6). However, patients with DC had higher mean PSA compared to patients with Gleason 6-7 AC (9.1 ng/mL, P <.001) and lower mean PSA compared to patients with Gleason 8-10 AC (16.1 ng/mL, P <.001). Patients with DC presented with similar rate of clinical extra-capsular extension, cT3 (11.7% vs 11.5%, P ¼ .8) as patients with Gleason 8-10 AC. Patients with DC were more likely to undergo surgical treatment compared to patients with Gleason 8-10 AC (54% vs 36%, P <.001) but less likely to undergo local RT with or without surgery (28% vs 52%, P <.001) (Table 2). Compared to patients with Gleason 8-10 AC undergoing RP, those with DC had more favorable pathology with a smaller percentage, with stage T3 (39% vs 52%, P <.001), positive lymph nodes (4% vs 11%, P <.001), and positive margins (25% vs 33%, P <.001) (Table 3). The demographic and clinical characteristics of patients undergoing XRT are summarized in Supplementary Table 1. For patients undergoing RT, the rate of clinically localized disease (cT2) was higher for Gleason 6-7 AC (98%) than DC (83%) or Gleason 8-10 AC (87%) (all P <.001). Regardless of histology, patients with PSA below 10 ng/mL were more likely to undergo RP compared to RT (81% vs 71% for DC, 85% vs 77% for Gleason 6-7 AC, and 70% vs 54% for Gleason 8-10 AC). Patients with DC had similar 5-year overall survival (75.0%, 95% confidence interval [CI] [71.7%-78.9%]) compared to patients with Gleason 8-10 AC (77.1%, 95% CI [76.6%-77.6%], P ¼ .2; Supplementary Figure 1), and both had worse survival compared with Gleason 6-7 780
Table 4. Multivariable Cox proportional hazards analysis of overall mortality Factor
HR
Age (y) <65 1 (referent) 65-75 1.71 >75 3.44 Charlson comorbidity None 1 (referent) 1 1.57 2 2.70 PSA at diagnosis <10 1 (referent) 10-20 1.37 >20 1.69 Histology DC 1 (referent) AC (Gleason 0.46 6-7) AC (Gleason 0.92 8-10) Local therapy No local 1 (referent) therapy RP 0.26 RT 0.53 RP þ RT 0.34
Upper 95% CI
Lower 95% CI
P-Value
1.66 3.32
1.76 3.56
<.001 <.001
1.52 2.55
1.62 2.86
<.001 <.001
1.33 1.64
1.41 1.74
<.001 <.001
0.34
0.61
<.001
0.69
1.23
.583
0.25 0.51 0.31
0.27 0.54 0.38
<.001 <.001 <.001
CI, confidence interval; HR, hazard ratio; PSA, prostate-specific antigen; other abbreviations as in Tables 1-3.
AC (92.5%, 95% CI [92.4%-92.7%], P <.001). Considering only patients who underwent RP, patients with DC had 5-year overall survival of 92.0% (95% CI [87.6%94.9%]), patients with Gleason 8-10 AC had 91.7% (95% CI [91.0%-92.3%]), and patients with Gleason 6-7 AC had 97.1% (95% CI [97.0%-97.2%]). After adjusting for age, PSA, treatment modality, and CCI, patients with Gleason 6-7 AC had decreased risk of death compared to DC (hazards ratio [HR] 0.46, 95% CI [0.34-0.61], P <.001) while Gleason 8-10 AC (HR 0.92, 95% CI [0.69-1.23], P ¼ .6) had similar risk of death (Table 4).
COMMENT With the rise in detection and treatment of prostate cancer in the past decade, DC has been increasingly identified on prostate biopsy, with a recent update by Epstein designating DC as being similar in presentation and prognosis to Gleason 8 AC prostate cancer.7 Several large series have noted pathologically aggressive features associated with DC.8,9 Utilizing the NCDB, our series represents the largest and most contemporary study to date reporting the comparative analysis of treatment trends and survival between DC and AC of the prostate. Our results suggest that: (1) patients with DC presented with lower mean PSA compared to patients with Gleason 8-10 AC, (2) patients with DC who underwent radical prostatectomy had poorer pathological features compared to patients with Gleason 6-7 AC, yet better pathological UROLOGY 86 (4), 2015
features compared to patients with Gleason 8-10 AC, (3) patients with DC were more likely to undergo radical prostatectomy compared to patients with Gleason 8-10 AC, and (4) patients with DC have similar mortality rates compared to patients with Gleason 8-10 AC. DC differentiates itself from AC on both the morphologic and molecular levels. Recent data demonstrate differing expression levels of receptors and enzymes between DC and high-grade AC, suggesting a separation in terms of their development pathways.16 Histologically, DC has distinct features such as papillary architecture and nuclear elongation. However, there are several features that overlap with AC, leading to difficulties identifying and diagnosing postoperative DC specimens.17 Efforts to improve preoperative diagnoses with biopsy specimens include the use of the biomarkers expressed more frequently in DC (chromogranin A, EGFR, and p53), although this has not been applied to a large, practical setting.18 Imaging using T2 weighted magnetic resonance imaging has not proven fruitful as DC largely resembles Gleason 6 AC.19 Nearly three-quarters of patients with localized DC present with pretreatment PSA values below 10 ng/mL. Although PSA values in localized DC tend to be modestly greater than those of Gleason 6-7 AC, these values are on average much less than those of Gleason 8-10 AC tumors. This finding complements previous work by showing that DC produces lower PSA values compared to that of AC in the setting of advanced metastatic disease.20 Multiple SEER studies have shown that patients with DC produce lower PSA values than those of patients with AC,6,8,9 although one surgical series reports higher PSA values in patients with a DC component.6 These relative differences in mean PSA are likely due to differing composition of higher Gleason stage AC in these studies. By directly comparing high-risk AC to DC, our study resolves the inconsistency in previous studies and affirms localized DC tends to produce lower PSA values than Gleason 8-10 AC, which is considered similar to DC in terms of disease aggressiveness. Our results suggest patients with DC who underwent radical prostatectomy had poorer pathological features compared to patients with Gleason 6-7 AC, with respect to positive nodes, positive margins, and extraprostatic extension (EPE). However, these factors were slightly better compared to patients with Gleason 8-10 AC. Samaratunga et al6 reported that the presence of DC on final pathology after radical prostatectomy was an independent predictor of EPE (73% vs 33%, respectively) even after controlling for tumor volume and Gleason score. A 58 patient series by Brinker et al4 revealed 63% EPE and also showed an increased number of positive biopsy scores for DC was correlated with a significantly higher rate of positive surgical margins and increased likelihood of clinical progression. However, in contrast to our series, they did not have any patients with positive lymph nodes. A recent large series by Seipel et al21 assessed 1051 radical prostatectomy specimens and UROLOGY 86 (4), 2015
found that the presence of ductal component was predictive for EPE (67% vs 42%) and seminal vesicle invasion (13% vs 5%), although this ultimately was not predictive for biochemical recurrence. The previous studies showed more aggressive features when comparing patients with DC to all patients with AC while our study stratifies AC by Gleason stage, likely accounting for differing results. Recommendations for treatment of DC are scarce in the literature and have varied considerably over time. HT was initially believed to be ineffective, stemming from the belief that ductal tumors were histologically similar to endometrial tumors.22 Early reports also showed poor efficacy of HT.11 However, more recent literature suggests patients with pure DC and metastatic DC are sensitive to androgen ablation with a mean survival greater than 5 years.2,10 RT similarly has been shown to be more efficacious in recent studies, although this may be tied to use of combined radiation and hormone treatment as well as more refined radiation administration techniques in contemporary series.10,12-14 Surgical treatment has historically been shown to have poor outcomes, although the series demonstrating this were performed in the prePSA era.3,4 Similar studies have also shown RP to portend a high risk of local recurrence and poor survival, as low as 24% at 5 years, although these authors notably still promote surgery as the greatest chance for long-term survival.23-25 Our study shows 54% of patients with DC undergo RP as one of their treatment modalities, compared with 36% of patients with Gleason 8-10 AC. We speculate that this may be due to lower preoperative PSA in the DC group, especially given that lower PSA was associated with higher rates of RP compared to RT in our series. The improved survival seen in the surgical cohort of patients in our series must be interpreted with caution due to possible selection bias. Another confounding factor is that some patients may have been misdiagnosed with DC on prostate biopsy, which could affect true rates of surgical treatment for DC and subsequent survival. Correlation of histologic composition and survival can yield prognostic data and guide treatment recommendations. Historic series showed DC has favorable prognosis.2,22,26 Conversely, there are several reports showing that DC has poorer overall prognosis compared to AC.3-6 An interesting study by Tu et al12 reported pure DC results in greater risk of local recurrence compared to mixed ductal histology. Overall, DC is traditionally approached as having similar prognostic significance with Gleason 8 AC. Marcus et al27 assessed incidence and survival of rare histologic variants of prostate cancer in the United States from 1973 to 2008 and found 5-year overall survival of 662 patients with DC to be 62%, compared to 75% in our more modern dataset, which may reflect a lead-time bias for PSA-detected DC. Our study demonstrates DC mortality mirrors Gleason 8-10 AC, with similar overall survival on both Kaplan-Meier and multivariable Cox proportional hazard analysis. 781
This is valuable when considering treatment recommendations for DC. There are several limitations to this study. There was no central pathologic review, and it is unclear how patients with mixed ductal and acinar histology were classified. Therefore, similar to other large series of DC, there is an inherent possibility of misdiagnosis. While our database did define preoperative clinical staging, it lacks granularity that may influence aggressiveness of local therapy, such as family history, detailed digital rectal exam, body mass index, and imaging findings. In patients receiving RP and RT, it is unclear whether RT was given as adjuvant or salvage therapy. Androgen deprivation was not consistently used or clearly described in the NCDB, which affects analysis of treatment patterns. There is likely confounding by indication (treatment bias) since there was no randomization of treatment, and therefore generalization about its impact on survival must be interpreted with caution. Finally, our series encompasses a wide time course during which Gleason staging was modified and thus can affect comparison of stages of AC with DC. Despite these limitations, the dataset allowed for characterization of patient characteristics, treatments and clinical outcomes across the United States in a large sample size of this rare histological variant of prostate cancer.
CONCLUSION In the largest contemporary population-based series of DC, patients with DC presented with lower PSA and were more likely to undergo surgery than patients with Gleason 8-10 AC. Although DC had more favorable pathological features compared to men with Gleason 8-10 AC, there was no difference observed in overall survival. Further studies are needed to determine whether different and perhaps more optimal treatment strategies would result in improved outcomes for patients with DC. References 1. Bock BJ, Bostwick DG. Does prostatic ductal adenocarcinoma exist? Am J Surg Pathol. 1999;23:781-785. 2. Millar EK, Sharma NK, Lessells AM. Ductal (endometrioid) adenocarcinoma of the prostate: a clinicopathological study of 16 cases. Histopathology. 1996;29:11-19. 3. Christensen WN, Steinberg G, Walsh PC, et al. Prostatic duct adenocarcinoma. Findings at radical prostatectomy. Cancer. 1991; 67:2118-2124. 4. Brinker DA, Potter SR, Epstein JI. Ductal adenocarcinoma of the prostate diagnosed on needle biopsy: correlation with clinical and radical prostatectomy findings and progression. Am J Surg Pathol. 1999;23:1471-1479. 5. Rubin MA, de La Taille A, Bagiella E, et al. Cribriform carcinoma of the prostate and cribriform prostatic intraepithelial neoplasia: incidence and clinical implications. Am J Surg Pathol. 1998;22: 840-848. 6. Samaratunga H, Duffy D, Yaxley J, et al. Any proportion of ductal adenocarcinoma in radical prostatectomy specimens predicts extraprostatic extension. Hum Pathol. 2010;41:281-285.
782
7. Epstein JI. An update of the Gleason grading system. J Urol. 2010; 183:433-440. 8. Meeks JJ, Zhao LC, Cashy J, et al. Incidence and outcomes of ductal carcinoma of the prostate in the USA: analysis of data from the Surveillance, Epidemiology, and End Results program. BJU Int. 2012;109:831-834. 9. Morgan TM, Welty CJ, Vakar-Lopez F, et al. Ductal adenocarcinoma of the prostate: increased mortality risk and decreased serum prostate specific antigen. J Urol. 2010;184:2303-2307. 10. Orihuela E, Green JM. Ductal prostate cancer: contemporary management and outcomes. Urol Oncol. 2008;26:368-371. 11. Greene LF, Farrow GM, Ravits JM, et al. Prostatic adenocarcinoma of ductal origin. J Urol. 1979;121:303-305. 12. Tu SM, Lopez A, Leibovici D, et al. Ductal adenocarcinoma of the prostate: clinical features and implications after local therapy. Cancer. 2009;115:2872-2880. 13. Bostwick DG, Kindrachuk RW, Rouse RV. Prostatic adenocarcinoma with endometrioid features. Clinical, pathologic, and ultrastructural findings. Am J Surg Pathol. 1985;9:595-609. 14. Tannenbaum M. Transitional cell carcinoma of prostate. Urology. 1975;5:674-678. 15. Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40:373-383. 16. Jardel P, Debiais C, Godet J, et al. Ductal carcinoma of the prostate shows a different immunophenotype from high grade acinar cancer. Histopathology. 2013;63:57-63. 17. Seipel AH, Delahunt B, Samaratunga H, et al. Diagnostic criteria for ductal adenocarcinoma of the prostate: interobserver variability among 20 expert uropathologists. Histopathology. 2014;65:216-227. 18. Tarjan M, Chen HH, Tot T, et al. Improved differentiation between ductal and acinar prostate cancer using three-dimensional histology and biomarkers. Scand J Urol Nephrol. 2012;46:258-266. 19. Schieda N, Coffey N, Gulavita P, et al. Prostatic ductal adenocarcinoma: an aggressive tumour variant unrecognized on T2 weighted magnetic resonance imaging (MRI). Eur Radiol. 2014;24:13491356. 20. Leibovici D, Spiess PE, Agarwal PK, et al. Prostate cancer progression in the presence of undetectable or low serum prostatespecific antigen level. Cancer. 2007;109:198-204. 21. Seipel AH, Wiklund F, Wiklund NP, et al. Histopathological features of ductal adenocarcinoma of the prostate in 1,051 radical prostatectomy specimens. Virchows Arch. Int J Pathol. 2013;462: 429-436. 22. Melicow MM, Pachter MR. Endometrial carcinoma of proxtatic utricle (uterus masculinus). Cancer. 1967;20:1715-1722. 23. Colpaert C, Gentens P, Van Marck E. Ductal (“endometrioid”) adenocarcinoma of the prostate. Acta Urol Belg. 1998;66:29-32. 24. Perrapato SD, Shah PC, Huben RP, et al. Locally recurrent endometrioid adenocarcinoma of the prostate after radical prostatectomy. J Urol. 1991;145:373-375. 25. Catalona WJ, Kadmon D, Martin SA. Surgical considerations in treatment of intraductal carcinoma of the prostate. J Urol. 1978;120: 259-261. 26. Dube VE, Farrow GM, Greene LF. Prostatic adenocarcinoma of ductal origin. Cancer. 1973;32:402-409. 27. Marcus DM, Goodman M, Jani AB, et al. A comprehensive review of incidence and survival in patients with rare histological variants of prostate cancer in the United States from 1973 to 2008. Prostate Cancer Prostatic Dis. 2012;15:283-288.
APPENDIX SUPPLEMENTARY DATA
Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.urology. 2015.07.009
UROLOGY 86 (4), 2015
RESULTS
CONCLUSION
To compare pathological characteristics, treatment patterns, and survival in patients with ductal adenocarcinoma (DC) compared to those with acinar adenocarcinoma (AC). Using the National Cancer Database, we identified patients diagnosed with clinically localized (cN0, cM0) pure DC (n ¼ 1328) and AC (n ¼ 751,635) between 1998 and 2011. High-risk AC was defined as Gleason 8-10. Demographic, treatment, pathological, and survival characteristics of patients were compared. Compared to patients with Gleason 8-10 AC, those with DC presented with lower mean prostate-specific antigen (10.3 vs 16.2 ng/mL, P <.001), had similar rates (11.7% vs 11.5%, P ¼ .8) of clinical extra-capsular extension (stage cT3), and were more likely to undergo prostatectomy (54% vs 36%, P <.001). Compared to patients with Gleason 8-10 AC undergoing prostatectomy, those with DC had more favorable pathology: stage T3 (39% vs 52%, P <.001), fewer positive lymph nodes (4% vs 11%, P <.001), and fewer positive margins (25% vs 33%, P <.001). On Kaplan-Meier analysis, patients with DC had similar 5-year survival (75.0%, 95% confidence interval [CI] [71.7-78.9]) compared to those with Gleason 8-10 AC (77.1%, 95% CI [76.6%-77.6%], P ¼ .2). On Cox multivariable analysis, patients with Gleason 8-10 AC had a similar risk of death compared to those with DC (hazards ratio ¼ 0.92, 95% CI [0.69-1.23], P ¼ 6). In this large contemporary population-based series, patients with DC of the prostate presented with lower prostate-specific antigen, had more favorable pathological features, and similar overall survival compared to men with Gleason 8-10 AC. UROLOGY 86: 777e782, 2015. 2015 Elsevier Inc.
A
cinar adenocarcinoma (AC) represents the vast majority of prostate cancers with the remaining 5%-10% comprising a variety of other histologic subtypes. Ductal carcinoma (DC) comprises w5% of all prostate cancers.1 Some initial reports considered DC to portend a more favorable prognosis.2 Multiple small series subsequently demonstrated more aggressive features associated with DC.3-6 In an update to the Gleason scoring system in 2010, Epstein favored defining these Financial Disclosure: The authors declare that they have no relevant financial interests. From the Section of Urology, Department of Surgery, University of Chicago, Chicago, IL; the Department of Urology, University of Wisconsin-Madison, Madison, WI; the Pritzker School of Medicine, University of Chicago, Chicago, IL; the Department of Pathology, University of Chicago, Chicago, IL; and the Department of Medicine, University of Chicago, Chicago, IL Address correspondence to: Vignesh T. Packiam, M.D., The University of Chicago, 5841 South Maryland Ave. MC-6038, Chicago, IL 60637. E-mail: vignesh.packiam@ uchospitals.edu Submitted: April 14, 2015, accepted (with revisions): July 3, 2015
ª 2015 Elsevier Inc. All Rights Reserved
tumors as Gleason 4þ4 while retaining the nomenclature of ductal to highlight differences compared to conventional acinar tumors.7 More recently, 2 large studies of the SEER database demonstrated DC is associated with advanced disease at presentation, increased risk of mortality when compared to all patients with AC, and has a similar prognosis to high-risk AC.8,9 Although there is general consensus that DC is more aggressive than AC, there is not necessarily agreement pertaining to recommended treatment modalities.10 Studies have shown varying efficacy for hormone therapy (HT),2,10-12 radiation treatment (RT),10,13,14 and radical prostatectomy (RP).3,4 These discrepancies are likely to be influenced both by differences in delivery of treatment between older and more recent literatures, as well as the small size of most series. The controversies surrounding optimal treatment for this rare histologic subtype were a primary incentive for utilizing a large series http://dx.doi.org/10.1016/j.urology.2015.07.009 0090-4295/15
777
Table 1. Demographic and clinical characteristics of patients by tumor type Parameter Number of patients Age (y) <65 65-75 >75 Ethnicity White Black Other Unknown Charlson comorbidity index None 1 2 Distance from hospital Same zip code <60 miles >60 miles >120 miles Geographic location Northeast South/Southeast Midwest West Location Metro Urban Rural Year of diagnosis 1998-1999 2000-2003 2004-2007 2008-2011 PSA at diagnosis* <10 10-20 >20 Clinical T stage cT1 cT2 cT3 cT4
DC (%)
AC Gleason 6-7 (%)
AC Gleason 8-10 (%)
1328 (0.2)
664,987 (88.3)
86,648 (11.5)
349,788 (53) 252,504 (38) 62,695 (9)
29,374 (34) 36,322 (42) 20,952 (24)
543,242 89,188 17,271 15,286
69,995 12,013 3041 1599
523 (39) 504 (38) 301 (23) 1096 156 53 23
(82.5) (11.8) (4.0) (1.7)
875 (82.4) 160 (15.1) 27 (2.5)
(81.7) (13.4) (2.6) (2.3)
(80.8) (13.9) (3.5) (1.8)
570,996 (85.9) 81,809 (12.3) 12,182 (1.8)
72,451 (83.6) 11,840 (13.7) 2357 (2.7)
42 1043 91 152
(3.2) (78.5) (6.9) (11.4)
18,838 549,718 38,690 57,741
(2.8) (82.7) (5.8) (8.7)
2873 73,173 4254 6348
(3.3) (84.5) (4.9) (7.3)
276 322 325 405
(20.8) (24.2) (24.5) (30.5)
142,837 188,927 177,143 156,080
(21.5) (28.4) (26.6) (23.5)
18,060 22,998 23,340 22,250
(20.8) (26.6) (26.9) (25.7)
1045 (82.4) 186 (14.7) 37 (2.9)
518,498 (81.7) 100,723 (15.9) 14,944 (2.4)
67,079 (80.9) 13,673 (16.5) 2116 (2.6)
(6.0) (19.5) (31.3) (43.2)
0y 0y 338,871 (51.0) 326,116 (49.0)
0y 0y 42,666 (49.2) 43,982 (50.8)
595 (74.1) 125 (15.6) 83 (10.3)
478,706 (81.3) 67,478 (11.5) 42,327 (7.2)
44,430 (57.2) 17,118 (22.0) 16,106 (20.8)
581 377 89 38
416,825 160,962 11,142 536
37,685 30,219 7690 1125
80 258 416 574
(53.5) (34.8) (8.2) (3.5)
(70.7) (27.3) (1.9) (0.1)
P-Value <.001
<.001
<.001
<.001
<.001
<.001
<.001
<.001
<.001
(49.1) (39.4) (10.0) (1.5)
AC, acinar adenocarcinoma; DC, ductal adenocarcinoma. * Mean, median, min, max, interquartile range were: 10.3, 6.3, 0.5, 74.7, [4.3-10.3] for ductal; 9.1, 5.7, 0.4, 73.5, [4.3-8.4] for Gleason 6-7; 16.1, 8.5, 0.5, 98, [5.3-16.9] for Gleason 8-10. y Gleason score was not available for patient diagnosed prior to 2004.
in our study. We utilized the National Cancer Database (NCDB) to describe patient characteristics, patterns of treatment, and clinical outcomes in the largest contemporary cohort to date of DC of the prostate.
MATERIALS AND METHODS We utilized the participant use file (PUF) for prostate cancer to identify our study cohort. Institutional review board exemption was not required since no patient, physician, or hospital identifiers were examined. The database utilized was developed by the NCDB, which is managed by the American College of Surgeons’ Commission on Cancer and the American Cancer Society. The NCDB includes data from over 1400 approved cancer programs and captures information on approximately 70% of all cancers diagnosed in the United States. 778
We identified all patients diagnosed with prostate cancer on prostate biopsy between 1998 and 2011, and only included those who were diagnosed with pure, nonmetastatic (American Joint Committee on Cancer stage cM0 and cN0/cNx) DC and Gleason 6-10 AC of the prostate. Patients were stratified by histology as DC, AC Gleason 6-7, and AC Gleason score 8-10 to allow for comparison across demographic, treatment, and survival characteristics. For patients undergoing RP, it was unknown whether pathology data reflected prostate biopsy or final surgical pathology. The NCDB comprises data initially entered at participating hospital-based cancer registries, which includes patient demographics (age, race, metropolitan urban and/or rural home county, distance traveled to cancer facility, year of diagnosis and Deyo-Charlson Comorbidity Index [CCI]),15 cancer facility (type and location), cancer characteristics (prostate-specific antigen [PSA] and American Joint Committee on Cancer UROLOGY 86 (4), 2015
Table 2. Treatment of patients by tumor type Parameter Radical prostatectomy Yes No Radiation Yes No Local therapy No local therapy RP RT RP þ RT
DC (%)
AC Gleason 6-7 (%)
AC Gleason 8-10 (%)
713 (53.8) 612 (46.2)
346,452 (52.2) 316,697 (47.8)
31,147 (36.1) 55,158 (63.9)
370 (28.1) 948 (71.9)
239,315 (36.2) 421,353 (63.8)
44,618 (51.8) 41,461 (48.2)
306 647 308 61
88,627 334,064 229,479 9537
16,092 25,602 39,308 5238
P-value <.001
(23.2) (48.9) (23.3) (4.6)
(13.4) (50.5) (34.7) (1.4)
<.001 <.001
(18.6) (29.7) (45.6) (6.1)
RP, Radical Prostatectomy; RT, Radiation Treatment; other abbreviations as in Table 1.
Table 3. Demographic, clinical, pathologic characteristics of patients undergoing RP by tumor type Parameter Number of patients Age (y) <65 65-75 >75 PSA at diagnosis* <10 10-20 >20 Clinical T stage cT1 cT2 cT3 cT4 Pathological T stage pT1 pT2 pT3 pT4 Unknown Lymph nodes None Dissection Unknown Node status Positive Negative Margin status Positive Negative
AC Gleason 6-7 (%)
AC Gleason 8-10 (%)
647 (0.2)
DC (%)
334,064 (92.7)
25,602 (7.1)
370 (57.2) 253 (39.1) 24 (3.7)
233,170 (69.8) 97,819 (29.3) 3075 (0.9)
14,109 (55.1) 10,779 (42.1) 714 (2.8)
370 (81.3) 54 (11.9) 31 (6.8)
248,628 (85.3) 25,175 (8.7) 14,468 (5.0)
15,764 (70.1) 4166 (18.5) 2571 (11.4)
328 162 32 2
(62.6) (30.9) (6.1) (0.4)
192,132 77,586 5128 119
(69.9) (28.2) (1.9) (0.04)
11,360 7717 1610 59
(54.7) (37.2) (7.8) (0.3)
1 356 236 14 37
(0.2) (55.3) (36.6) (2.2) (5.7)
341 261,525 48,087 934 20,747
(0.1) (78.8) (14.5) (0.3) (6.3)
21 11,035 12,690 435 1245
(0.1) (43.4) (49.9) (1.7) (4.9)
133 (20.6) 484 (74.8) 30 (4.6)
141,593 (42.4) 191,090 (57.2) 1381 (0.4)
4371 (17.1) 21,141 (82.6) 91 (0.36)
24 (4.4) 522 (95.6)
2414 (1.0) 230,198 (99.0)
2330 (10.7) 19,389 (89.3)
157 (24.6) 480 (75.4)
60,893 (18.4) 269,968 (81.6)
8276 (32.7) 17,019 (67.3)
P-Value <.001
<.001
<.001
<.001
<.001
<.001 <.001
Abbreviations as in Tables 1 and 2. * Mean, median, min, max, interquartile range were: 8.8, 5.9, 0.5, 92.6, [4.2-8.3] for ductal; 7.8, 5.3, 0.4, 98, [4.1-7.4] for Gleason 6-7; 11.2, 6.8, 0.5, 98, [4.8-11.3] for Gleason 8-10.
clinical and/or pathological stage), and treatment modalities (RP, RT, HT, and chemotherapy). A patient’s setting is defined as their hometown’s population categorized as metropolitan (>250,000), urban (>2500), or rural (<2500). Hospital variables include geographic location and program type. Programs are categorized based on the volume of patients per year: comprehensive (>500), community (100-500), and academic (>500). Academic programs are required to provide graduate medical programs in at least four different areas. Patients were further categorized as receiving local therapy if they underwent RP, RT, or RP þ RT. The RP þ RT group UROLOGY 86 (4), 2015
included patients who underwent RP followed by adjuvant or salvage RT. The dosage of RT was not specified in the NCDB, and may have ranged between 65 and 80 Gy over the duration of this series. There was inconsistent use and documentation of androgen deprivation in NCDB, and thus this was not included in analysis. The NCDB recorded PSA starting in 2004 per collaborative staging stipulations as the highest value prior to the diagnostic procedure or, if that value is unavailable, the earliest pretreatment but postdiagnostic value. Statistical analysis was performed using Stata, version 12 (College Station, TX). Categorical variables are presented as 779
whole numbers and percentages, which were compared utilizing the Fisher’s exact and Mann-Whitney U tests. Kaplan-Meier survival analysis and stratified log-rank test were utilized to compare overall survival stratified by histology. Overall survival was estimated as time from the diagnosis of prostate cancer to death. Patients known to be alive were censored at the last follow-up evaluation or the last contact by correspondence. Patients were only included if they had follow-up time of at least 5 years. A multivariable Cox proportional hazards model was performed to identify independent predictors of overall survival. For all statistical measures, a 2-sided P-value <.05 was considered statistically significant.
RESULTS We identified and analyzed 1328 (0.2%) and 715,635 (99.8%) patients with nonmetastatic DC and AC of the prostate, respectively. Of the patients with AC, 664,987 (88.5%) had Gleason 6-7 disease and 86,648 (11.5%) had Gleason 8-10 disease. The median follow-up for the entire cohort was 81 months. The demographic and clinical characteristics of patients with DC, Gleason 6-7 AC, and Gleason 8-10 AC are summarized in Table 1. There was a similar proportion of young (<65 years old) patients presenting with DC compared to patients with Gleason 8-10 AC (39% vs 34%, P <.001). Patients with DC presented with similar mean PSA compared to all patients with AC (10.3 vs 10.0 ng/mL, P ¼ .6). However, patients with DC had higher mean PSA compared to patients with Gleason 6-7 AC (9.1 ng/mL, P <.001) and lower mean PSA compared to patients with Gleason 8-10 AC (16.1 ng/mL, P <.001). Patients with DC presented with similar rate of clinical extra-capsular extension, cT3 (11.7% vs 11.5%, P ¼ .8) as patients with Gleason 8-10 AC. Patients with DC were more likely to undergo surgical treatment compared to patients with Gleason 8-10 AC (54% vs 36%, P <.001) but less likely to undergo local RT with or without surgery (28% vs 52%, P <.001) (Table 2). Compared to patients with Gleason 8-10 AC undergoing RP, those with DC had more favorable pathology with a smaller percentage, with stage T3 (39% vs 52%, P <.001), positive lymph nodes (4% vs 11%, P <.001), and positive margins (25% vs 33%, P <.001) (Table 3). The demographic and clinical characteristics of patients undergoing XRT are summarized in Supplementary Table 1. For patients undergoing RT, the rate of clinically localized disease (cT2) was higher for Gleason 6-7 AC (98%) than DC (83%) or Gleason 8-10 AC (87%) (all P <.001). Regardless of histology, patients with PSA below 10 ng/mL were more likely to undergo RP compared to RT (81% vs 71% for DC, 85% vs 77% for Gleason 6-7 AC, and 70% vs 54% for Gleason 8-10 AC). Patients with DC had similar 5-year overall survival (75.0%, 95% confidence interval [CI] [71.7%-78.9%]) compared to patients with Gleason 8-10 AC (77.1%, 95% CI [76.6%-77.6%], P ¼ .2; Supplementary Figure 1), and both had worse survival compared with Gleason 6-7 780
Table 4. Multivariable Cox proportional hazards analysis of overall mortality Factor
HR
Age (y) <65 1 (referent) 65-75 1.71 >75 3.44 Charlson comorbidity None 1 (referent) 1 1.57 2 2.70 PSA at diagnosis <10 1 (referent) 10-20 1.37 >20 1.69 Histology DC 1 (referent) AC (Gleason 0.46 6-7) AC (Gleason 0.92 8-10) Local therapy No local 1 (referent) therapy RP 0.26 RT 0.53 RP þ RT 0.34
Upper 95% CI
Lower 95% CI
P-Value
1.66 3.32
1.76 3.56
<.001 <.001
1.52 2.55
1.62 2.86
<.001 <.001
1.33 1.64
1.41 1.74
<.001 <.001
0.34
0.61
<.001
0.69
1.23
.583
0.25 0.51 0.31
0.27 0.54 0.38
<.001 <.001 <.001
CI, confidence interval; HR, hazard ratio; PSA, prostate-specific antigen; other abbreviations as in Tables 1-3.
AC (92.5%, 95% CI [92.4%-92.7%], P <.001). Considering only patients who underwent RP, patients with DC had 5-year overall survival of 92.0% (95% CI [87.6%94.9%]), patients with Gleason 8-10 AC had 91.7% (95% CI [91.0%-92.3%]), and patients with Gleason 6-7 AC had 97.1% (95% CI [97.0%-97.2%]). After adjusting for age, PSA, treatment modality, and CCI, patients with Gleason 6-7 AC had decreased risk of death compared to DC (hazards ratio [HR] 0.46, 95% CI [0.34-0.61], P <.001) while Gleason 8-10 AC (HR 0.92, 95% CI [0.69-1.23], P ¼ .6) had similar risk of death (Table 4).
COMMENT With the rise in detection and treatment of prostate cancer in the past decade, DC has been increasingly identified on prostate biopsy, with a recent update by Epstein designating DC as being similar in presentation and prognosis to Gleason 8 AC prostate cancer.7 Several large series have noted pathologically aggressive features associated with DC.8,9 Utilizing the NCDB, our series represents the largest and most contemporary study to date reporting the comparative analysis of treatment trends and survival between DC and AC of the prostate. Our results suggest that: (1) patients with DC presented with lower mean PSA compared to patients with Gleason 8-10 AC, (2) patients with DC who underwent radical prostatectomy had poorer pathological features compared to patients with Gleason 6-7 AC, yet better pathological UROLOGY 86 (4), 2015
features compared to patients with Gleason 8-10 AC, (3) patients with DC were more likely to undergo radical prostatectomy compared to patients with Gleason 8-10 AC, and (4) patients with DC have similar mortality rates compared to patients with Gleason 8-10 AC. DC differentiates itself from AC on both the morphologic and molecular levels. Recent data demonstrate differing expression levels of receptors and enzymes between DC and high-grade AC, suggesting a separation in terms of their development pathways.16 Histologically, DC has distinct features such as papillary architecture and nuclear elongation. However, there are several features that overlap with AC, leading to difficulties identifying and diagnosing postoperative DC specimens.17 Efforts to improve preoperative diagnoses with biopsy specimens include the use of the biomarkers expressed more frequently in DC (chromogranin A, EGFR, and p53), although this has not been applied to a large, practical setting.18 Imaging using T2 weighted magnetic resonance imaging has not proven fruitful as DC largely resembles Gleason 6 AC.19 Nearly three-quarters of patients with localized DC present with pretreatment PSA values below 10 ng/mL. Although PSA values in localized DC tend to be modestly greater than those of Gleason 6-7 AC, these values are on average much less than those of Gleason 8-10 AC tumors. This finding complements previous work by showing that DC produces lower PSA values compared to that of AC in the setting of advanced metastatic disease.20 Multiple SEER studies have shown that patients with DC produce lower PSA values than those of patients with AC,6,8,9 although one surgical series reports higher PSA values in patients with a DC component.6 These relative differences in mean PSA are likely due to differing composition of higher Gleason stage AC in these studies. By directly comparing high-risk AC to DC, our study resolves the inconsistency in previous studies and affirms localized DC tends to produce lower PSA values than Gleason 8-10 AC, which is considered similar to DC in terms of disease aggressiveness. Our results suggest patients with DC who underwent radical prostatectomy had poorer pathological features compared to patients with Gleason 6-7 AC, with respect to positive nodes, positive margins, and extraprostatic extension (EPE). However, these factors were slightly better compared to patients with Gleason 8-10 AC. Samaratunga et al6 reported that the presence of DC on final pathology after radical prostatectomy was an independent predictor of EPE (73% vs 33%, respectively) even after controlling for tumor volume and Gleason score. A 58 patient series by Brinker et al4 revealed 63% EPE and also showed an increased number of positive biopsy scores for DC was correlated with a significantly higher rate of positive surgical margins and increased likelihood of clinical progression. However, in contrast to our series, they did not have any patients with positive lymph nodes. A recent large series by Seipel et al21 assessed 1051 radical prostatectomy specimens and UROLOGY 86 (4), 2015
found that the presence of ductal component was predictive for EPE (67% vs 42%) and seminal vesicle invasion (13% vs 5%), although this ultimately was not predictive for biochemical recurrence. The previous studies showed more aggressive features when comparing patients with DC to all patients with AC while our study stratifies AC by Gleason stage, likely accounting for differing results. Recommendations for treatment of DC are scarce in the literature and have varied considerably over time. HT was initially believed to be ineffective, stemming from the belief that ductal tumors were histologically similar to endometrial tumors.22 Early reports also showed poor efficacy of HT.11 However, more recent literature suggests patients with pure DC and metastatic DC are sensitive to androgen ablation with a mean survival greater than 5 years.2,10 RT similarly has been shown to be more efficacious in recent studies, although this may be tied to use of combined radiation and hormone treatment as well as more refined radiation administration techniques in contemporary series.10,12-14 Surgical treatment has historically been shown to have poor outcomes, although the series demonstrating this were performed in the prePSA era.3,4 Similar studies have also shown RP to portend a high risk of local recurrence and poor survival, as low as 24% at 5 years, although these authors notably still promote surgery as the greatest chance for long-term survival.23-25 Our study shows 54% of patients with DC undergo RP as one of their treatment modalities, compared with 36% of patients with Gleason 8-10 AC. We speculate that this may be due to lower preoperative PSA in the DC group, especially given that lower PSA was associated with higher rates of RP compared to RT in our series. The improved survival seen in the surgical cohort of patients in our series must be interpreted with caution due to possible selection bias. Another confounding factor is that some patients may have been misdiagnosed with DC on prostate biopsy, which could affect true rates of surgical treatment for DC and subsequent survival. Correlation of histologic composition and survival can yield prognostic data and guide treatment recommendations. Historic series showed DC has favorable prognosis.2,22,26 Conversely, there are several reports showing that DC has poorer overall prognosis compared to AC.3-6 An interesting study by Tu et al12 reported pure DC results in greater risk of local recurrence compared to mixed ductal histology. Overall, DC is traditionally approached as having similar prognostic significance with Gleason 8 AC. Marcus et al27 assessed incidence and survival of rare histologic variants of prostate cancer in the United States from 1973 to 2008 and found 5-year overall survival of 662 patients with DC to be 62%, compared to 75% in our more modern dataset, which may reflect a lead-time bias for PSA-detected DC. Our study demonstrates DC mortality mirrors Gleason 8-10 AC, with similar overall survival on both Kaplan-Meier and multivariable Cox proportional hazard analysis. 781
This is valuable when considering treatment recommendations for DC. There are several limitations to this study. There was no central pathologic review, and it is unclear how patients with mixed ductal and acinar histology were classified. Therefore, similar to other large series of DC, there is an inherent possibility of misdiagnosis. While our database did define preoperative clinical staging, it lacks granularity that may influence aggressiveness of local therapy, such as family history, detailed digital rectal exam, body mass index, and imaging findings. In patients receiving RP and RT, it is unclear whether RT was given as adjuvant or salvage therapy. Androgen deprivation was not consistently used or clearly described in the NCDB, which affects analysis of treatment patterns. There is likely confounding by indication (treatment bias) since there was no randomization of treatment, and therefore generalization about its impact on survival must be interpreted with caution. Finally, our series encompasses a wide time course during which Gleason staging was modified and thus can affect comparison of stages of AC with DC. Despite these limitations, the dataset allowed for characterization of patient characteristics, treatments and clinical outcomes across the United States in a large sample size of this rare histological variant of prostate cancer.
CONCLUSION In the largest contemporary population-based series of DC, patients with DC presented with lower PSA and were more likely to undergo surgery than patients with Gleason 8-10 AC. Although DC had more favorable pathological features compared to men with Gleason 8-10 AC, there was no difference observed in overall survival. Further studies are needed to determine whether different and perhaps more optimal treatment strategies would result in improved outcomes for patients with DC. References 1. Bock BJ, Bostwick DG. Does prostatic ductal adenocarcinoma exist? Am J Surg Pathol. 1999;23:781-785. 2. Millar EK, Sharma NK, Lessells AM. Ductal (endometrioid) adenocarcinoma of the prostate: a clinicopathological study of 16 cases. Histopathology. 1996;29:11-19. 3. Christensen WN, Steinberg G, Walsh PC, et al. Prostatic duct adenocarcinoma. Findings at radical prostatectomy. Cancer. 1991; 67:2118-2124. 4. Brinker DA, Potter SR, Epstein JI. Ductal adenocarcinoma of the prostate diagnosed on needle biopsy: correlation with clinical and radical prostatectomy findings and progression. Am J Surg Pathol. 1999;23:1471-1479. 5. Rubin MA, de La Taille A, Bagiella E, et al. Cribriform carcinoma of the prostate and cribriform prostatic intraepithelial neoplasia: incidence and clinical implications. Am J Surg Pathol. 1998;22: 840-848. 6. Samaratunga H, Duffy D, Yaxley J, et al. Any proportion of ductal adenocarcinoma in radical prostatectomy specimens predicts extraprostatic extension. Hum Pathol. 2010;41:281-285.
782
7. Epstein JI. An update of the Gleason grading system. J Urol. 2010; 183:433-440. 8. Meeks JJ, Zhao LC, Cashy J, et al. Incidence and outcomes of ductal carcinoma of the prostate in the USA: analysis of data from the Surveillance, Epidemiology, and End Results program. BJU Int. 2012;109:831-834. 9. Morgan TM, Welty CJ, Vakar-Lopez F, et al. Ductal adenocarcinoma of the prostate: increased mortality risk and decreased serum prostate specific antigen. J Urol. 2010;184:2303-2307. 10. Orihuela E, Green JM. Ductal prostate cancer: contemporary management and outcomes. Urol Oncol. 2008;26:368-371. 11. Greene LF, Farrow GM, Ravits JM, et al. Prostatic adenocarcinoma of ductal origin. J Urol. 1979;121:303-305. 12. Tu SM, Lopez A, Leibovici D, et al. Ductal adenocarcinoma of the prostate: clinical features and implications after local therapy. Cancer. 2009;115:2872-2880. 13. Bostwick DG, Kindrachuk RW, Rouse RV. Prostatic adenocarcinoma with endometrioid features. Clinical, pathologic, and ultrastructural findings. Am J Surg Pathol. 1985;9:595-609. 14. Tannenbaum M. Transitional cell carcinoma of prostate. Urology. 1975;5:674-678. 15. Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40:373-383. 16. Jardel P, Debiais C, Godet J, et al. Ductal carcinoma of the prostate shows a different immunophenotype from high grade acinar cancer. Histopathology. 2013;63:57-63. 17. Seipel AH, Delahunt B, Samaratunga H, et al. Diagnostic criteria for ductal adenocarcinoma of the prostate: interobserver variability among 20 expert uropathologists. Histopathology. 2014;65:216-227. 18. Tarjan M, Chen HH, Tot T, et al. Improved differentiation between ductal and acinar prostate cancer using three-dimensional histology and biomarkers. Scand J Urol Nephrol. 2012;46:258-266. 19. Schieda N, Coffey N, Gulavita P, et al. Prostatic ductal adenocarcinoma: an aggressive tumour variant unrecognized on T2 weighted magnetic resonance imaging (MRI). Eur Radiol. 2014;24:13491356. 20. Leibovici D, Spiess PE, Agarwal PK, et al. Prostate cancer progression in the presence of undetectable or low serum prostatespecific antigen level. Cancer. 2007;109:198-204. 21. Seipel AH, Wiklund F, Wiklund NP, et al. Histopathological features of ductal adenocarcinoma of the prostate in 1,051 radical prostatectomy specimens. Virchows Arch. Int J Pathol. 2013;462: 429-436. 22. Melicow MM, Pachter MR. Endometrial carcinoma of proxtatic utricle (uterus masculinus). Cancer. 1967;20:1715-1722. 23. Colpaert C, Gentens P, Van Marck E. Ductal (“endometrioid”) adenocarcinoma of the prostate. Acta Urol Belg. 1998;66:29-32. 24. Perrapato SD, Shah PC, Huben RP, et al. Locally recurrent endometrioid adenocarcinoma of the prostate after radical prostatectomy. J Urol. 1991;145:373-375. 25. Catalona WJ, Kadmon D, Martin SA. Surgical considerations in treatment of intraductal carcinoma of the prostate. J Urol. 1978;120: 259-261. 26. Dube VE, Farrow GM, Greene LF. Prostatic adenocarcinoma of ductal origin. Cancer. 1973;32:402-409. 27. Marcus DM, Goodman M, Jani AB, et al. A comprehensive review of incidence and survival in patients with rare histological variants of prostate cancer in the United States from 1973 to 2008. Prostate Cancer Prostatic Dis. 2012;15:283-288.
APPENDIX SUPPLEMENTARY DATA
Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.urology. 2015.07.009
UROLOGY 86 (4), 2015