Ductal or endometrioid adenocarcinoma of the prostate may be a subtype of prostate cancer that is amenable to aggressive local therapeutic strategies. The authors of this report investigated the clinical outcome of patients who had prostate ductal adenocarcinoma after primary radical prostatectomy or radiotherapy.
The clinical features of 108 patients with locally confined or advanced prostate ductal adenocarcinoma who had undergone primary radical prostatectomy (surgical group, n = 76 men) or no surgery (nonsurgical group, n = 32 men) were evaluated retrospectively. Clinical records were reviewed, and Gleason scores, clinical/pathologic stages, and preoperative prostate-specific antigen levels were examined. The clinical features that were assessed included local recurrence, distant metastasis, and progression-free and overall survival after primary therapy.
In the surgical group, patients who had pure ductal prostate cancer survived longer (median, 13.8 years; 95% confidence interval [CI], from 13.8 years to not attained) than patients who had mixed ductal prostate cancer (median, 8.9 years; 95% CI, from 7.1 years to not attained; P = .05). In addition, the median time to local progression was shorter (2.8 years vs 4.9 years) and the median time to distant metastasis was longer (3.9 years vs 2.0 years) for patients who had pure ductal adenocarcinoma than for patients who had mixed ductal adenocarcinoma of the prostate after surgery, respectively.
Prostate cancer is a heterogeneous disease. The identification of unique subtypes of prostate cancer for a particular therapy may enhance the efficacy of treatment. It is well known that certain prostate cancers remain confined to the primary site for a prolonged period.1 In addition, approximately 25% patients with metastatic (D2) prostate cancer who were receiving androgen-deprivation therapy developed disease progression at the local site.2 In the long run, these locally advanced tumors cause debilitating symptoms and serious complications, such as intractable pelvic pain, recurrent urinary tract infections, and urinary and rectal bleeding or obstruction. Although these tumors can be controlled, it is unlikely that they will be eradicated by systemic treatments. Therefore, early identification of prostate cancer subtypes that pursue an indolent course and are likely to cause severe morbidity from local progression will be key for their optimal management.
Previously, we reported that patients with ductal or endometrioid adenocarcinoma of the prostate had a tendency to develop persistent or recurrent urinary obstructive symptoms after systemic therapy.3 Ductal adenocarcinoma is a unique subtype of prostate cancer characterized by the presence of tall, pseudostratified, columnar cells with abundant cytoplasm arranged in a papillary pattern.4 It often is diagnosed with a high Gleason score and at an advanced clinical stage.5 Ductal adenocarcinoma is rare and comprises between 0.13% to 6% of all prostate cancers.6, 7 Although prostate ductal adenocarcinoma has distinctive pathologic features, its biologic and clinical significance remain unknown.8, 9 We reasoned that ductal adenocarcinoma could be a subtype of prostate cancer that may be amenable to local therapeutic strategies. In this explorative, retrospective study, we evaluated the natural history of patients with nonmetastatic prostate ductal adenocarcinoma. We investigated their clinical outcomes after primary radical prostatectomy or radiotherapy.
MATERIALS AND METHODS
We identified patients who had histologically confirmed prostate ductal adenocarcinoma by searching patient records in the data and information system (GURU) of the Department of Genitourinary Medical Oncology at the University of Texas M. D. Anderson Cancer Center. Patients were eligible as long as 1 pertinent pathologic specimen had indicated ductal or endometrioid adenocarcinoma of the prostate. The “surgical” group included all patients who had undergone radical prostatectomy within 6 months after their diagnosis. The “nonsurgical” group included all patients who had localized prostate cancer but had not undergone radical prostatectomy as the primary treatment. Patients who had undergone salvage prostatectomy were included in the latter group. Patients with evidence of metastatic prostate cancer on any radiographic studies at the time of their diagnosis were excluded from this study. This study was approved by the Institutional Review Board of the M. D. Anderson Cancer Center (M. D. Anderson).
We reviewed the patients' clinical records and examined their Gleason score, clinical/pathologic stage, prostatectomy date, and preoperative prostate-specific antigen (PSA) level. In the surgical group, patients whose pathology report indicated pure or predominantly ductal adenocarcinoma (ie, >75% of the tumor contained a ductal component) were classified in the “pure pathology” group. The “mixed pathology” group comprised patients whose pathology reports indicated mixed ductal/acinar adenocarcinoma. Patients who had evidence of ductal adenocarcinoma and neuroendocrine, sarcomatoid, mucinous, or squamous differentiation were included in the mixed pathology group. The patients' clinical histories and laboratory results were collected from their medical records and from the M. D. Anderson data management computer system (NETPASS or ClinicStation). Survival data were obtained from the medical records or from the Social Security Death Index interactive search website (available at: http://ssdi.genealogy.rootsweb.com/ Accessed on July 15, 2007).
We also evaluated their clinical features, including local recurrence, distant metastasis, progression-free survival (PFS), and overall survival (OS). For local recurrence, we assessed the frequency and date of transurethral resection of the prostate (TURP) because of persistent or worsening urinary obstructive symptoms; the documentation of palpable local failure; the need to perform straight catheterization, placement of Foley catheter, suprapubic catheter, or percutaneous nephrostomy tubes as a result of hydronephrosis or renal insufficiency; and the occurrence of gross hematuria caused by tumor invasion into the bladder according to a cystoscopic examination. For distant metastasis, we assessed the frequency and time to development of metastases in visceral organs or bone. For survival, we assessed the time to local recurrence and distant metastasis. PFS was calculated from the diagnosis date to the time the patient had developed disease recurrence (considered an event). OS was computed as the time between the diagnosis date and the date of death. If a patient had not experienced a recurrence or remained alive at his last follow-up, then his data were censored, and his PFS and OS were calculated from the diagnosis date to the date of his last follow-up.
Fisher exact tests were performed to assess the association between the patients' characteristics (Gleason score, PSA level, and disease stage) and the pathology status (mixed or pure). Wilcoxon rank-sum tests were used to compare the patients' median ages by pathologic status in the surgical group. The Kaplan-Meier method was used to estimate survival distributions, and the log-rank test was used to compare the survival distributions between the pure and mixed ductal pathology status of patients in the surgical group. Univariate Cox proportional hazards regression models were used to assess the relations between age, Gleason score, PSA level, and disease stage with PFS and OS. All statistical analyses were performed using the software packages SAS version 9.2.0 (SAS Inc., Cary, NC) and SPLUS version 7.0 (Insightful Corp., Seattle, Wash).
By searching GURU, we identified the records of 108 patients who had been diagnosed with a locally confined or advanced prostate ductal adenocarcinoma from among 13,125 database entries for patients with prostate cancer who had been evaluated at M. D. Anderson from January 1, 2000 through January 29, 2007. Seventy-six of these 108 patients had undergone primary radical prostatectomy within 6 months of their diagnosis (ie, the surgical group). Most of the remaining 32 patients who had not undergone primary radical prostatectomy had received primary radiation therapy (ie, the nonsurgical group).
The 76 patients who had undergone radical prostatectomy had received their initial cancer diagnosis between July 1, 1985 and February 8, 2006. One of these patients had no follow-up records available for review and, thus, was excluded from our analysis, which left 75 patients in the surgical group. The median follow-up for this group was 4.9 years (range, 0.4-15 years). It is noteworthy that 79% of the men with ductal adenocarcinoma had been diagnosed in a the prostatectomy specimen (76%) or in a subsequent biopsy specimen (3%) rather than in an initial biopsy specimen of the prostate. Their clinical characteristics are summarized in Table 1.
Table 1. Clinical Characteristics of Patients With Localized Prostate Ductal Cancer Who Underwent Radical Prostatectomy (the Surgical Group; n=75)
We distinguished pure from mixed ductal prostate cancer in the surgical group. According to the pathology reports, 10 of the 75 patients (13%) had pure ductal prostate cancer, and 15 patients (20%) had predominantly ductal prostate cancer (pure pathology group), whereas the other 50 patients (67%) had mixed ductal and acinar prostate cancer (mixed pathology group). Seven of the 50 patients (14%) of in the mixed pathology group also had evidence of neuroendocrine, sarcomatoid, mucinous, or squamous differentiation. Greater than 90% of the ductal adenocarcinomas were located in the peripheral zone of the prostate. Intraductal spread of tumor was noted in <10% of patients.
No statistically significant difference was observed between patients in the pure and mixed pathology groups with regard to their age, PSA level, Gleason score, or clinical/pathologic stage (Table 2). A large proportion of patients (44%) in both groups had not developed any events (local recurrence or distant metastasis). Table 3 lists the median times to the development of local recurrence and distant metastasis after surgery for the pure and mixed pathology groups.
Table 2. Association Between Pathology Subgroup and Clinical Characteristics in Patients Who Underwent Radical Prostatectomy (the Surgical Group; n=75)
Table 3. Time to Progression (in Years) by Pathology Status in Patients Who Underwent Surgery (Surgery Group)
Time to Progression, y
No. of Patients
Pure pathology group
Local progression only
Metastatic progression only
Local and metastatic progression
Mixed pathology group
Local progression only
Metastatic progression only
Local and metastatic progression
By using each of the prognostic factors of interest in a univariate Cox model for OS, we observed that a higher clinical tumor (T) classification (ie, T3 or T4) predicted a worse prognosis for prostate ductal adenocarcinoma in the surgical group (P = .02). Multivariate modeling was not performed because of the small number of events.
Table 4 lists the median PFS and OS times and the estimated proportion of patients that survived at 5 years for each pathologic subgroup after surgery. Kaplan-Meier curves for PFS and OS by pathologic group are illustrated in Figures 1 and 2, respectively. The median OS was 13.8 years (95% confidence interval [CI], from 13.8 years to not attained) and 8.9 years (95% CI, from 7.1 years to not attained) for the pure and mixed pathology groups, respectively, and the OS was marginally longer in the pure pathology group (P = .05).
Table 4. Kaplan-Meier Estimates of Median Survival and Probability of Survival at 5 Years by Pathology Subgroup in Patients Who Underwent Radical Prostatectomy (the Surgical Group; n=75)
PFS indicates progression-free survival; OS, overall survival; CI, confidence interval; NA, not attained or not available.
No. of patients
No. of recurrences or deaths
Median survival (95% CI), y
6.9 (3.9 to NA)
13.8 (13.83 to NA)
8.9 (7.13 to NA)
Probability of 5-y survival (95% CI)
P (log-rank test)
Clinical and pathologic features
The 32 patients who had localized prostate ductal adenocarcinoma but did not undergo radical prostatectomy had received their initial diagnosis between October 20, 1989 and October 17, 2005. The median follow-up was 4.8 years (range, 1-16 years). The majority of these ductal adenocarcinomas (59%) had been diagnosed in a TURP specimen. Of the 16 patients whose medical records revealed some details about their primary radiotherapy, the median radiation dose given was 70 Gray (Gy) (range, 61-78 Gy). Their clinical characteristics are summarized in Table 5.
Table 5. Clinical Characteristics of Patients With Localized Prostate Ductal Cancer Who Did Not Undergo Radical Prostatectomy (the Nonsurgical Group; n=32)
To this day, the clinical and prognostic significance of prostate ductal adenocarcinoma remains controversial. Some investigators have reported that ductal adenocarcinoma is a more virulent form of prostate cancer,10-12 whereas others have reported that it is a more indolent form.4, 6, 13 In the current study, we observed that, in the surgical group, the patients with pure ductal prostate cancer survived longer than the patients with mixed ductal prostate cancer (median, 13.8 years vs 8.9 years, respectively; P = .05). It is noteworthy that the median time to local recurrence was shorter (2.8 years vs 4.9 years, respectively) and the median time to distant metastasis was longer (3.9 years vs 2.0 years, respectively) after surgery for patients who had pure prostate cancer than for patients who had mixed ductal prostate cancer. The results of this explorative study suggest that 1) pure ductal prostate cancer tends to pursue an indolent clinical course and carries an increased risk for local recurrence, and 2) local control (particularly by radical prostatectomy) may improve the clinical outcome of patients with pure ductal adenocarcinoma of the prostate.
Prostate ductal adenocarcinoma may be more prevalent than commonly believed. Although we examined the “classic” cases of prostate ductal adenocarcinoma in which the diagnoses were considered indisputable by the pathologists at M. D. Anderson, there may have been additional cases that could not be identified easily according to current histologic and immunohistochemical criteria. For example, it may be impossible to distinguish ductal adenocarcinoma with a solid pattern from a poorly differentiated carcinoma of the prostate.14 Half of the ductal adenocarcinomas may stain only focally or weakly for PSA or prostate acid phosphatase, especially after androgen-deprivation therapy. It is noteworthy that ductal adenocarcinoma may not be recognized easily or diagnosed readily in an initial transrectal ultrasound (TRUS)-guided biopsy of the prostate: In our surgical group, 79% of ductal adenocarcinomas had been diagnosed in the prostatectomy specimen or in a subsequent biopsy specimen; whereas, in our nonsurgical group, 59% had been diagnosed in a TURP specimen. Therefore, most prostate ductal adenocarcinomas would have been missed in an initial TRUS-guided biopsy.
In the current study, for the first time to our knowledge, we discerned that pure and mixed ductal prostate cancers might be unique and separate clinical entities. Only when complete specimens from prostatectomy were available for review did we detect a clinical distinction between these 2 pathologic entities. Nevertheless, the origin or nature of this putative distinction requires confirmation and deserves further investigation. Because the patients with pure ductal prostate adenocarcinoma had high Gleason scores from 8 to 10 (87% in the pure pathology group) (Table 2) but could survive for a prolonged period, we suspect that Gleason score may not be a useful prognostic marker for such patients. We anticipate that the discovery of novel, specific markers for prostate ductal adenocarcinoma will facilitate the diagnosis of equivocal or unconventional cases, confirm its clinical significance, and clarify its prognostic implications.
Unfortunately, prostate ductal adenocarcinoma often is under staged before surgery (Table 2). Although all of their patients with prostate ductal adenocarcinoma appeared to have resectable tumors before surgery, Christensen and colleagues5 reported that 93% had extracapsular extension, 47% had positive margins, 40% had seminal vesicle involvement, and 27% had pelvic lymph node metastasis. Similarly, Brinker et al12 reported that 65% of their patients with ductal tumors had T3 disease. These observations could account for the shorter time to local progression after prostatectomy among patients with pure ductal adenocarcinoma. Indeed, our results indicated that a higher clinical stage (ie, T3 or T4) conferred a worse prognosis to prostate ductal adenocarcinoma. We expect that the incidence of local recurrence would be much diminished for a completely resected tumor. We propose that neoadjuvant approaches to downstage pure ductal prostate adenocarcinomas may further improve the clinical outcome of definitive local therapies, especially radical prostatectomy.
It remains unknown whether definitive radiotherapy also would improve local control of prostate ductal adenocarcinoma.7 Although more patients with prostate ductal adenocarcinoma from our nonsurgical group (56%) had developed a local recurrence compared with those from the surgical group (17%), we emphasize that the 2 patient populations were not equivalent and, thus, are not comparable (Tables 1 and 5). The subsequent receipt of salvage radiotherapy by 20 patients (27%) and of adjuvant radiotherapy by 11 patients (15%) may account for the lower incidence of local recurrence in the surgical group (Table 1). It is conceivable that, without salvage or adjuvant radiation therapy, the local recurrence rate could have been higher; however, this would only enhance the main message of this study that local control might be important for the optimal management of prostate ductal adenocarcinomas. Inadequate radiation dosages without concurrent hormone-ablative therapy also could account for the greater incidence of local recurrence in our nonsurgical group. We did not determine the role of pure or mixed ductal prostate cancer in the nonsurgical group, because the biopsy specimens (in contrast to prostatectomy specimens) provided only limited tissue and may have produced sampling bias. Therefore, determining whether primary, salvage, or adjuvant radiotherapy improves the local control of certain subtypes of prostate cancer (eg, pure ductal prostate adenocarcinoma) will require further study.
Until recently, surgical resection of the primary tumor generally was not recommended for metastatic cancer. The rationale for this sentiment seems obvious: If it is the systemic cancer that ultimately hurts or kills a patient, then why worry about the local tumor, which may not even be symptomatic or affect their eventual clinical outcome? However, accumulating data suggest that extirpation of the primary tumor, even in the setting of metastatic disease, does provide a superior clinical outcome in some cancers. For example, patients with metastatic renal cell carcinoma who underwent radical nephrectomy survived longer than those who did not undergo surgery.15, 16 Complete resection of the primary tumor in metastatic breast cancer also may enhance patient survival.17 There is evidence suggesting the potential advantages of treating primary prostate cancer despite the presence of metastatic disease.18-20 The identification of a particular subtype of prostate cancer with its unique biologic and clinical characteristics (ie, indolent clinical course and predilection for local recurrence) may improve the selection of appropriate patients and thereby enhance the therapeutic efficacy of surgery.
We hypothesize that various malignant entities have unique and distinct cellular origins.21 Consequently, they pursue separate and different clinical courses. This theory explains why some tumors are intrinsically more indolent, (eg, pure ductal adenocarcinoma), whereas others are inherently more lethal, (eg, mixed ductal adenocarcinoma). Tumors derived from a later progenitor stem cell in a stem-cell hierarchy tend to be homogeneous and exhibit a more pure phenotype, whereas tumors derived from an earlier (more pluripotent) progenitor stem cell tend to be heterogeneous and express a more mixed phenotype. It is noteworthy that this notion (contrary to the traditional model of multistep carcinogenesis) views an aggressive tumor as a distinct and separate entity rather than a product of rapid transformation from an indolent tumor, ie, pure ductal does not evolve to become mixed ductal adenocarcinoma. It predicts that a mixed tumor may express a pure tumor under the right conditions, but not vice versa. A corollary to this theory assumes that certain prostate cancers contain unique cancer stem cells, which tend to remain confined in the primary site for a prolonged period and may differentiate to form drug-resistant “teratomatous” progenies. Therefore, an effective way to treat such cancer types is to remove the cancer stem cells and the differentiated “teratomatous” components within a primary tumor by surgery.
In conclusion, the current results suggest that patients with pure ductal adenocarcinoma of the prostate have a better clinical outcome after radical prostatectomy than patients with mixed ductal adenocarcinoma. Pure ductal adenocarcinoma may be a separate and distinct biologic and clinical entity from mixed ductal adenocarcinoma of the prostate. Further studies need to be performed to improve the selection of patients with prostatic ductal adenocarcinoma (or other subtypes of prostate cancer) who may benefit from localized therapeutic strategies in the future. Additional studies need to be conducted in a prospective setting to confirm our findings.
We thank Angela C. Chinweze for compiling the patient list from the GURU database for this study, Karen F. Phillips for editorial assistance, and Linda Johnson for preparation of this article.