Machine Learning for the Prediction of Biochemical Recurrence in Patients Treated with Radical Prostatectomy

PURPOSE/OBJECTIVE(S)

Biochemical recurrence (BCR) occurs in about 40% of patients with prostate cancer following radical prostatectomy (RP). Our goal was to develop a machine learning model for the prediction of BCR five-years after RP, to improve patient prognostication.

MATERIALS/METHODS

Patients treated with RP at a tertiary care medical center between 1990 and 2017 were included. A gradient boosted decision trees-based machine learning model modified to handle survival data was trained on 80% of the dataset. The model's performance was evaluated on the remaining 20%. Input variables were age at surgery, prostate specific antigen (PSA) at diagnosis (in ng/mL), pathologic Gleason grade group (GG), pathologic T stage (organ confined disease vs. extracapsular extension (ECE) vs. seminal vesicle invasion (SVI)), lymph node involvement, and surgical margin status. Model performance was assessed using time-dependent area under curve of the receiver operator curve (AUC).

RESULTS

The full dataset included 11,139 patients, of whom 1,153 (10%) developed BCR. Median age at surgery was 59 and PSA at diagnosis was 5.4 ng/mL. Only 1,080 (9.7%) patients had GG 3, and 707 (6.3%) GG 4 and 5. 1,366 (12%) patients had positive surgical margins and 134 (1.2%) had lymph node involvement. Most patients had organ confined disease with EPE and SVI diagnosed in 2,759 (25%) and 392 (3.5%) patients, respectively. Median follow-up was 5 years and median time to BCR was 4 years. When validated on the hold-out set of 2,228 patients, the model shows a time-dependent AUC of 0.82 (95% CI 0.78 - 0.86) for BCR at t = 5 years.

CONCLUSION

Our machine learning model can be used to estimate risk of BCR following RP and shows exceptional performance, with implications on patient prognostication and follow-up. We are currently working on validating its performance on an external dataset.

Machine Learning for the Prediction of Adverse Pathological Outcomes in Patients Treated with Radical Prostatectomy

PURPOSE/OBJECTIVE(S)

Extracapsular extension (ECE) and seminal vesicle invasion (SVI) are associated with negative oncologic outcomes in patients with prostate cancer. We have developed and validated a machine learning model to more accurately identify patients at risk of these adverse surgical outcomes prior to radical prostatectomy (RP).

MATERIALS/METHODS

This study included a cohort of patients diagnosed with prostate cancer and treated with RP and lymph node dissection at a tertiary care medical center from 2010 to 2020. An ensemble model using a base gradient-boosted trees-based machine learning model and isotonic calibrators was trained on 80% of the cohort, with 20% held out for validation. The model uses age at surgery, prostate specific antigen level (PSA) at diagnosis, biopsy Gleason grade group, numbers of positive and negative cores on biopsy, and clinical T-stage (cT) as input variables. Model performance was assessed on the hold-out set using the area under the receiver operating curve (AUC).

RESULTS

The full dataset included 18,729 eligible patients. Median PSA at diagnosis was 7.3 ng/mL. Most patients had clinically organ confined disease (cT1 - cT2) with only 136 (0.7%) having cT3. The most common biopsy Gleason grade group was 2 (7,118 or 38% of patients), with Gleason grade 4 in 1,796 (9.6%), and 5 in 1,064 (5.7%) patients. After RP, 11,931 (64%) of patients had organ confined disease, 4,298 (23%) had ECE, and 2,500 (13%) had SVI. When validated on the hold-out set (n = 3,746), the model had AUCs of 0.79 (95%-CI 0.77 - 0.80), 0.67 (0.65 - 0.69), and 0.83 (0.81 - 0.85) for the prediction of organ confined disease, ECE, and SVI, respectively.

CONCLUSION

In conclusion, we have developed a machine learning model that predicts individual patient risk of pathologic T-stage. The model can be used to provide more accurate risk assessments and improve surgical treatment planning. We are currently working on externally validating our results on patients from different institutions.

Very-high-risk localized prostate cancer: definition and outcomes

BACKGROUND

Outcomes in men with National Comprehensive Cancer Network (NCCN) high-risk prostate cancer (PCa) can vary substantially-some will have excellent cancer-specific survival, whereas others will experience early metastasis even after aggressive local treatments. Current nomograms, which yield continuous risk probabilities, do not separate high-risk PCa into distinct sub-strata. Here, we derive a binary definition of very-high-risk (VHR) localized PCa to aid in risk stratification at diagnosis and selection of therapy.

METHODS

We queried the Johns Hopkins radical prostatectomy database to identify 753 men with NCCN high-risk localized PCa (Gleason sum 8-10, PSA >20 ng ml(-1), or clinical stage ≥T3). Twenty-eight alternate permutations of adverse grade, stage and cancer volume were compared by their hazard ratios for metastasis and cancer-specific mortality. VHR criteria with top-ranking hazard ratios were further evaluated by multivariable analyses and inclusion of a clinically meaningful proportion of the high-risk cohort.

RESULTS

The VHR cohort was best defined by primary pattern 5 present on biopsy, or ≥5 cores with Gleason sum 8-10, or multiple NCCN high-risk features. These criteria encompassed 15.1% of the NCCN high-risk cohort. Compared with other high-risk men, VHR men were at significantly higher risk for metastasis (hazard ratio 2.75) and cancer-specific mortality (hazard ratio 3.44) (P<0.001 for both). Among high-risk men, VHR men also had significantly worse 10-year metastasis-free survival (37% vs 78%) and cancer-specific survival (62% vs 90%).

CONCLUSIONS

Men who meet VHR criteria form a subgroup within the current NCCN high-risk classification who have particularly poor oncological outcomes. Use of these characteristics to distinguish VHR localized PCa may help in counseling and selection optimal candidates for multimodal treatments or clinical trials.

Metastasis-free survival is associated with overall survival in men with PSA-recurrent prostate cancer treated with deferred androgen deprivation therapy

BACKGROUND

Clinical trials in men with biochemically recurrent prostate cancer (BRPC) have been hampered by long survival times, making overall survival (OS) a difficult end point to reach. Intermediate end points are needed in order to conduct such trials within a more feasible time frame.

PATIENTS AND METHODS

This is a retrospective analysis of 450 men with BRPC following prostatectomy treated at a single institution between 1981 and 2010, of which 140 developed subsequent metastases. Androgen deprivation therapy (ADT) was deferred until after the development of metastases. Cox regression models were developed to investigate factors influencing OS.

RESULTS

Median metastasis-free survival (MFS) was 10.2 years [95% confidence interval (CI) 7.6-14.0 years]; median OS after metastasis was 6.6 years (95%CI 5.8-8.4 years). Multivariable Cox regressions identified four independently prognostic variables for OS: MFS (HR 0.77; 95% CI 0.63-0.94), number of metastases (≤3 versus ≥4; HR 0.50; 95% CI 0.29-0.85), pain (absent versus present; HR 0.43; 95% CI 0.25-0.72), and bisphosphonate use (yes versus no; HR 0.60; 95% CI 0.37-0.98).

CONCLUSIONS

MFS emerged as an independent predictor of OS in men with BRPC treated with deferred ADT after the development of metastases. MFS may be a reasonable intermediate end point in future clinical trials. This observation requires prospective validation.

Radical prostatectomy in men less than 50 years old

We studied the effect of age on tumor progression (defined as postoperative prostate specific antigen elevation) in 543 men who underwent radical prostatectomy for clinically localized prostate cancer. Patients were divided into two age groups: patients under 50 years (N = 85) and patients older than 50 years (N = 458). The mean follow-up for both groups was S.3 years. Clinical stage was similar in both groups, with only 3% in each group detected by screening techniques. By Kaplan-Meier analysis, men under 50 years showed slightly less progression than men older than 50 years (p = 0.04), especially during the first 5 years following surgery. The key differentiating feature was a lower incidence of positive margins in the younger age group (18.8%) than the older age group (42.6%; p < 0.0001). There was a higher incidence of lymph node metastasis in the younger age group (14.1%) than the older age group (6.1%; p = 0.01); this adverse feature was present in only a small fraction of the patients and did not play a major role in the difference in progression between age groups. There was no statistically significant difference between the two age groups in tumor grade, capsular penetration, or seminal vesicle invasion. Gland volume was significantly higher in the older age group. Within the younger age group, progression was not affected by a family history of prostate cancer. We found that, despite the tendency in younger men to preserve the neurovascular bundles and cut closer to the prostate, this age group still has a lower incidence of positive surgical margins possibly due to greater ease of surgical removal of small glands. Young men who are candidates for radical prostatectomy do not have a worse prognosis following surgery than older men, and even fare better during the first five postoperative years.

The natural history of metastatic progression in men with PSA-recurrent prostate cancer after radical prostatectomy: 25-year follow-up

5008

Background: In men with prostate specific antigen (PSA) recurrence following radical prostatectomy (RP) and no other therapy, the natural history of metastatic progression was previously described in 1999. We now report data reflecting up to 25 years of follow-up. Methods: We performed a retrospective analysis of 774 men treated with RP between 4/1982 and 7/2008 who developed PSA recurrence (>0.2 ng/ml) and never received adjuvant or salvage therapy. We investigated factors influencing the development of metastases. Results: Mean follow-up after RP was 8.5 y (median 8 y). Of 774 men with PSA recurrence, 295 (38%) developed metastases, and 433 had data on PSA doubling time (PSADT), forming our cohort. The mean time from RP to PSA recurrence in the entire cohort was 4.2 y (median 3 y). In those who developed metastases, the mean time from PSA recurrence to metastasis was 3.1 y (median 2 y). The mean PSA at the time of metastasis was 90.3 ng/ml (median 31.4 ng/ml). In Cox regression analysis: PSADT, Gleason score, and time to PSA progression were predictive of the development of metastases ( Table ). In Kaplan-Meier survival analysis, the median actuarial time from PSA recurrence to metastasis was 10 y (95% CI 9 - 15 y). Median actuarial metastasis-free survival from PSA recurrence for men with PSADT <3 mo, 3 - 8.9 mo, 9 - 14.9 mo, and >15 mo was 1 y (95% CI 0 - 1 y), 4 y (95% CI 2 - 6 y), 9 y (95% CI 7 - 13 y), and 15 y (95% CI 12 - 20 y), respectively. Conclusions: PSADT, Gleason score, and time to PSA progression are strong independent predictors of metastasis-free survival in men with PSA-recurrent prostate cancer. These data facilitate patient counseling and logical risk-based treatment planning. They also provide the background for appropriate selection of patients, treatments, and endpoints for clinical trials.

[Table: see text]

No significant financial relationships to disclose.

Predicting the long-term risk of prostate cancer-specific mortality after radical prostatectomy

5007

Background: Nomograms that predict prostate-specific antigen (PSA) defined biochemical recurrence (BCR) of prostate cancer after radical prostatectomy are the most widely used prediction tools in oncology for treatment decision making and counseling. While BCR universally antedates prostate cancer-specific mortality (PCSM), it is a limited surrogate endpoint due to its variable natural history. Nomograms that accurately predict the risk of PCSM are needed. Methods: Using Fine and Gray competing risk regression analysis, the clinical data and follow-up information of 11,521 patients treated with radical prostatectomy at four academic centers from 1987 to 2005 were modeled to predict PCSM. The model was externally validated on 12,893 patients treated at a separate institution during the same period. Results: The 15-year PCSM and all-cause mortality was 7% and 33%, respectively. The 15-year PCSM for patients with final pathological Gleason score 2–6, 3+4, 4+3, and 8–10 was 1%, 7%, 8%, and 49%, respectively. By pathologic stage, the risks were 2%, 7%, 29%, and 23% for organ-confined, extraprostatic extension, seminal vesicle invasion, and lymph node-positive prostate cancer. Of 3756 patients with organ-confined and Gleason 2–6 cancer, only 1 (0.03%) died from prostate cancer. Primary and secondary Gleason grade (p < 0.001 for both), seminal vesicle invasion (p < 0.001), and year of surgery (p = 0.002) were significant predictors of PCSM. A nomogram predicting 15-year PCSM based on pathologic parameters was accurate and discriminating with an externally-validated concordance index of 0.92. Conclusions: A nomogram has been constructed that predicts the long-term risk of PCSM after radical prostatectomy based on the pathologic grade and stage of the cancer. The presence of poorly differentiated cancer and seminal vesicle invasion are the prime determinants of PCSM. Our study suggests that biomarkers may have limited empiric prognostic utility as PCSM can be accurately predicted once the pathologic features of prostate cancer are known.

No significant financial relationships to disclose.

Survival following early hormone therapy for men with rapid PSA doubling time within 2 years following radical prostatectomy

5065

Background: Early hormonal therapy has been used in the salvage setting for men with biochemical recurrence following radical prostatectomy (RP), but no studies to date have been able to evaluate whether such treatment prolongs survival. We examined the impact of salvage hormonal therapy on overall survival (OS) in a cohort with long-term follow-up, and attempted to identify the subgroup most likely to benefit. Methods: Retrospective analysis of a cohort of 488 men undergoing RP at Johns Hopkins Hospital from 1982–2004, who experienced biochemical recurrence and received no salvage therapy (n = 386) or salvage hormonal therapy (n = 102); no one received adjuvant therapy. Survival was defined from biochemical recurrence to death from all causes, and analyzed with proportional hazards models with time-dependent covariates. Results: With median follow-up of 6 years after recurrence and 9 years after RP, there were 143 deaths (29%), including 105 from prostate cancer. After adjusting for PSA doubling time (PSADT), RP Gleason score, and year of surgery, hormonal therapy did not significantly improve OS for all men, compared to no salvage therapy: hazard ratio (HR) = 0.72 (95% confidence interval (CI): 0.45–1.17), p = 0.187. However, when restricted to men with early recurrence, i.e. within 2 years of RP, and with a rapid PSADT<6 months, hormonal therapy was associated with a large, significant improvement in OS: HR = 0.25 (95% CI: 0.08–0.71), p = 0.0095. This subgroup comprised 22% of the cohort. In contrast, there was no benefit of salvage hormonal therapy in men with early recurrence and PSADT>6 months: HR = 1.96 (95% CI: 0.89–4.31), p = 0.093, nor those who recurred more than 2 years after RP, regardless of PSADT. Conclusions: This study suggests that early salvage hormonal therapy may significantly and substantially prolong overall survival in the subgroup of men who experience an early biochemical recurrence with a rapid PSADT. These results are consistent with early recurrences being indicative of metastatic disease, while later recurrences are more likely to represent local recurrence. If validated, these results may provide useful stratification criteria for clinical trials.

No significant financial relationships to disclose.

A randomized, placebo-controlled trial of celecoxib in men prior to receiving prostatectomy for clinically localized adenocarcinoma of the prostate: Evaluation of drug-specific biomarker modulation

5001

Background: Cyclooxygenase-2 (COX-2) has been postulated as a pharmacological target for preventing a variety of epithelial malignancies including prostate cancer (PCa). We conducted a randomized, double-blind study evaluating celecoxib (C) on biomarkers in normal and PCa tissue at prostatectomy (RRP). Methods: Patients with Gleason sum ≥ 7, pre-study PSA ≥ 15 ng/ml, clinical stage T2b, T2c, or any combination of PSA, Stage, or two or more cores positive for PCa received either C at 400 mg po bid or placebo for 4–6 wks pre-RRP. The primary endpoint was to compare and correlate tissue PG levels with histologic and secondary endpoints performed on prostate tissue and serum from the two comparable groups. Outcomes included PG levels, quantitative RT-PCR for mRNA levels of COX-1 and COX-2, oxidized DNA bases; measurement of apoptosis, proliferation, angiogenic-potential assays and histologic comparison of treated/untreated tissue specimens; PSA levels; and tissue levels of C. Estimates of change in endpoints required 30 patients per arm. Results: Seventy three subjects consented with 64 randomized and included in the intent to treat analysis; 2 had missing data for primary endpoints. Age, baseline PSA, race and Gleason score were comparable across treatment groups. The regimen was well tolerated with no serious adverse events. There was no treatment effect observed in the PG, COX mRNA levels or oxidized DNA base levels in the RRP specimens. Tumor tissue contained significantly less COX-2 mRNA levels than benign tissue (p=<0.0001). Of the markers of apoptosis and proliferation assessed, Ki-67 was higher in tumor samples, and p21 was less in C treated samples. Celecoxib was present in tumor tissue demonstrating that it reached the target, but there were no observed effects in the study endpoints. There was no toxicity greater grade 1 except hepatic toxicity in a placebo group subject. Conclusions: Our results show a lack of effect of C on PCa despite demonstrating that C was present in tissue samples. At this time, we cannot recommend further studies of C as a PCa preventative agent when dosed at 400 mg PO BID. The study was supported by a grant from the NCI, DCP (#NO1-CN-95000–46) and Pfizer, Inc.

No significant financial relationships to disclose.

The impact of DNA methylation on the identification of recurrent prostate cancer

21086

Purpose: Biochemical (PSA) recurrence of prostate cancer following radical prostatectomy remains a major problem. Better biomarkers are needed to identify high and low-risk patients. DNA methylation of promoter regions leads to gene silencing in many cancers. In this study, we assessed the impact of changes in DNA methylation on biochemical recurrence in men with prostate cancer. Methods: We examined the methylation status of fifteen genes using MSP (Methylation Specific PCR) on tissue samples from 151 patients with clinically localized prostate cancer for whom at least five years of follow-up after prostatectomy was available. Results: In a multivariable logistic regression analysis, extra capsular penetration, high Gleason score, and involvement of the lymph nodes, seminal vesicles, or surgical margin were associated with an increased risk of recurrence. In addition, samples with methylation of 2 specific genes involved in cell-cell adhesion and apoptosis were associated with biochemical recurrence with an odds ratio of 5.64 (95% CI=1.47–21.7, p=0.012) compared to samples without methylation of both of these genes. The methylation status of these 2 genes had a higher sensitivity (72.3%; 95% CI=57–84.4%) for detecting recurrences than all the clinico-pathological variables (p<0.02) except extra-capsular penetration (p=0.346). The methylation status of these 2 genes had a similar negative predictive value (79.0%; 95% CI=66.8–88.3%) as the individual clinico-pathological variables examined. Conclusion: DNA Methylation of specific genes is independently associated with an increased risk of biochemical recurrence after radical prostatectomy even one considers the prognostic clinico-pathologic variables used in the clinic today. Our findings should be validated on another larger group of patients with prostate cancer who have undergone radical prosatetectomies.

No significant financial relationships to disclose.

Early Prostate Cancer Antigen Expression in Predicting Presence of Prostate Cancer in Men With Histologically Negative Biopsies

PURPOSE

Early prostate cancer antigen is a nuclear matrix protein that was recently shown to be expressed in prostate adenocarcinoma and adjacent benign tissue. Previous studies have demonstrated early prostate cancer antigen expression in benign prostate tissue up to 5 years before a diagnosis of prostate carcinoma, suggesting that early prostate cancer antigen could be used as a potential predictive marker.

MATERIALS AND METHODS

We evaluated early prostate cancer antigen expression by immunohistochemistry using a polyclonal antibody (Onconome Inc., Seattle, Washington) on benign biopsies from 98 patients. Biopsies were obtained from 4 groups that included 39 patients with first time negative biopsy (group 1), 24 patients with persistently negative biopsies (group 2), 8 patients with initially negative biopsies who were subsequently diagnosed with prostate carcinoma (group 3) and negative biopsies obtained from 27 cases where other concurrent biopsies contained prostate carcinoma (group 4). Early prostate cancer antigen staining was assessed by 2 of the authors who were blind to the group of the examined sections. Staining intensity (range 0 to 3) and extent (range 1 to 3) scores were assigned. The presence of intensity 3 staining in any of the blocks of a biopsy specimen was considered as positive for early prostate cancer antigen for the primary outcome in the statistical analysis. In addition, as secondary outcomes we evaluated the data using the proportion of blocks with intensity 3 early prostate cancer antigen staining, the mean of the product of staining intensity and staining extent of all blocks within a biopsy, and the mean of the product of intensity 3 staining and extent.

RESULTS

Primary outcome analysis revealed the proportion of early prostate cancer antigen positivity to be highest in group 3 (6 of 8, 75%) and lowest in group 2 (7 of 24, 29%, p=0.04 for differences among groups). A relatively higher than expected proportion of early prostate cancer antigen positivity was present in group 1 (23 of 39, 59%). Early prostate cancer antigen was negative in 41% of group 4 who were known to harbor prostate carcinoma. The proportion of early prostate cancer antigen positivity was statistically significantly lower in group 2 than in each of the other groups when compared pairwise. A lower proportion of early prostate cancer antigen positivity was encountered in older archival tissue blocks (p<0.0001) pointing to a potential confounding factor. Corrected for block age, group 3 was the only group to remain statistically significantly different in early prostate cancer antigen positivity compared to the reference group 2. Similar findings were obtained when adjustments for patient age were made and when analysis was based on secondary outcome measurements.

CONCLUSIONS

Our study showed a higher proportion of early prostate cancer antigen expression in initial negative prostate biopsy of patients who were diagnosed with prostate carcinoma on subsequent followup biopsies. We found a relatively high proportion of early prostate cancer antigen positivity (59%) in the group with first time negative biopsies and a potential 41% rate of false-negative early prostate cancer antigen staining in benign biopsies from cases with documented prostate carcinoma on concurrent cores. The lower early prostate cancer antigen positivity in cases with older blocks raises the question of a confounding effect of block age. Additional studies on the antigenic properties of early prostate cancer antigen in archival material are required to further delineate the usefulness of early prostate cancer antigen immunostaining on biopsy material.

Public health impact of PSA doubling time after radical prostatectomy on prostate cancer specific and overall survival

4568

Background: Among patients treated with radical prostatectomy (RP) with a PSA recurrence, we previously found men with a PSA doubling time (PSADT) <3 months were at increased risk of prostate cancer death, though these men constituted a small subset of patients. We sought to determine the actual and predicted number of prostate cancer deaths stratified by PSADT. Methods: We retrospectively studied 379 men treated with RP between 1982 and 2000 with a PSA recurrence. We calculated the actual and 15-year actuarial number of prostate cancer deaths in each of the following PSADT categories: <3, 3.0–8.9, 9.0–14.9, and ≥15.0 months. Results: Median follow-up after PSA recurrence was 7 years. During this time, there were 76 prostate cancer deaths; the majority (51%) were among men with a PSADT of 3.0–8.9 months. Though men with a PSADT <3 months were at the greatest risk of death, this group accounted for only 20% (n=15) of all prostate cancer deaths. Using actuarial 15-year estimates of prostate cancer specific survival, 50% of all prostate cancer deaths were among men with a PSADT of 3.0–8.9 months while men with a PSADT <3 months accounted for only 13% of prostate cancer deaths. Using actuarial 15-year estimates of all-cause and prostate cancer specific mortality, among men with a PSADT <15 months, prostate cancer was estimated to be the cause of death in 94% (145/155). Only among men with a PSADT >15 months was the risk of competing causes of mortality high enough such that the majority of deaths were not attributed to prostate cancer. Conclusions: Among a select cohort of men treated with RP who experienced a PSA recurrence, prostate cancer was estimated to account for 75% of all deaths. Though men with a PSADT <3 months were at the greatest risk, the majority of deaths occurred among men with a PSADT of 3.0–8.9 months. Efforts to reduce prostate cancer mortality should focus on men with intermediate PSADT times (3.0–15.0 months) as they represent the greatest public health concern among men with PSA recurrence following RP.

[Table: see text]

No significant financial relationships to disclose.

Factors influencing prostate cancer specific mortality in patients receiving delayed androgen deprivation therapy for metastasis after biochemical recurrence following radical prostatectomy

4571

Background: For men developing PSA failure after radical prostatectomy (RP), administration of immediate androgen deprivation therapy (ADT) has not been shown to improve survival compared to delaying ADT until evidence of metastatic disease. We evaluated factors influencing prostate cancer (PCa) specific mortality (PCSM) in a cohort of PSA era patients developing metastases after RP treated with delayed ADT. Methods: 3,658 men had RP by a single surgeon at Johns Hopkins Hospital from 4/82 until 6/05. 553 had PSA failure. 216 developed radiographically evident distant metastasis. Of these, 91 men formed the study cohort: initially treated during the PSA era (1987–2005), received ADT only after documented metastasis, and having complete data. 41 of these men died. Median failure times were estimated with the Kaplan-Meier method. Prognostic impact was estimated as the hazard ratio (HR) derived from the Cox proportional hazards model. Results: Median followup from RP was 10 yrs (range 2–18). Actuarial median failure times are: 1 yr from RP to PSA failure (range 1–11), 32 mos from PSA failure to metastasis (range 0–129), 79 mos from metastasis to death (range 7–181), and 13 yrs from RP to death (range 2–18). The following variables were significant prognostic factors for PCSM in univariate analyses: Pain at diagnosis of metastases (p < 0.01), time from RP to metastasis (p = 0.02), hematocrit at metastasis (p < 0.01) and PSADT <3 mos during the 2 years prior to metastasis (p < 0.01). A multivariable Cox proportional hazards model demonstrated the following independent predictors of PCSM: pain (HR = 10.5 p < 0.01), PSA at metastasis ≥100 ng/mL (HR = 5.3 p < 0.01) and PSADT < 3 months (HR = 7.1 p < 0.01). PSADT determined in the two years immediately after PSA failure (HR = 1.0 p = 0.37) and time from RP to bone metastasis (HR = 1.0 p = 0.80) were not independent predictors of PCSM. Conclusion: Men receiving delayed ADT for development of metastasis after RP may have a prolonged survival time (13 yrs post RP - range 2–18). Optimizing the time for ADT in these patients requires well-designed, prospective randomized studies. Our data may facilitate the selection of patients and thresholds for implementation of ADT.

No significant financial relationships to disclose.

Expectant management: an option for localized prostate cancer

Localized prostate cancer demonstrates tremendous heterogeneity in the natural history of the disease. To this end, although prostate cancer may be present histologically in nearly 30% of all men above the age of 50 y, the lifetime risk of developing clinically significant disease is 18% (one in six). Furthermore, the lifetime risk of dying from prostate cancer is less than 4%. Therefore, in order to avoid unnecessarily treating potentially insignificant prostate cancer, the concept of expectant management has been considered for this disease. In this brief review, we discuss the evolution of expectant management for men with localized prostate cancer.

The effect of hospital volume on cancer control after radical prostatectomy

PURPOSE

For complex oncological procedures, hospital volume affects short and long-term patient outcome. We examined the association of hospital volume and long-term cancer control after radical prostatectomy.

MATERIALS AND METHODS

With a cohort study design, we used the Surveillance, Epidemiology and End Results-Medicare linked files to identify a population based sample of men with newly diagnosed prostate cancer treated primarily with radical prostatectomy. Failure of cancer control was defined as the use of postoperative medical or surgical hormone ablation or treatment with radiation therapy more than 6 months after surgery.

RESULTS

A total of 12,635 men underwent radical prostatectomy for incident prostate cancer. After adjusting for age, comorbidity, histological grade and clinical stage, the risk of adjuvant therapy was greater among those treated at low (1 to 33 cases) and medium (34 to 61 cases) volume hospitals than at very high (more than 108 cases) volume hospitals (HR 1.25, p <0.001 and HR 1.11, p =0.023 respectively).

CONCLUSIONS

Patients treated at lower volume institutions are at increased risk of initiation of subsequent adjuvant therapy with radiation therapy, medical hormone ablation or orchiectomy. Noted differences in cancer control provide additional evidence regarding issues surrounding the debate over surgical volume standards for the surgical treatment of prostate cancer.

Stromal-epithelial measurements of prostate cancer in native Japanese and Japanese-American men

We measured the histologic stromal and epithelial tissue components of the benign (normal) and malignant tissue compartments of Japanese-Americans (J-A) and native Japanese (NJ) men living in Japan. The patient cohort included 25 NJ men undergoing radical prostatectomy (RP) in Nagoya, Japan and 25 J-A (second or third generation US born). We conducted tissue image quantitation (in-house image software) of the stromal and epithelial compartments in malignant and adjacent normal tissue areas from a tissue microarray (TMA) selected from radical prostatectomy (RP) blocks. Stromal-epithelial (S-E) areas were determined using immunohistochemical stains for CAM-5.2 epithelial cytokeratin marker and the Masson trichrome stain to measure the stroma component. We observed differences in the volumes of normal and cancer epithelium and stroma within both the J-A and NJ study populations (P<0.01). Only the individual average cancer epithelium (CE) volume (JA=24.1 vs NJ=29.9) differed significantly between the NJ and J-A study populations (P=0.03). Consequently, the S-E ratio in NJ group was significantly different from that of J-A population (P=0.05). The decrease in S-E ratio noted in the malignant tissues of NJ prostate tissue may provide a biological marker for differentiation of the two groups and suggests a need for further investigations into the molecular basis for these histologic differences.

Proteomic analysis of dunning prostate cancer cell lines with variable metastatic potential using SELDI-TOF

BACKGROUND

Surface enhanced laser desorption and ionization-time-of-flight (SELDI-TOF) is an evolving proteomic technology for improving biomarker discovery that allows for rapid and sensitive analysis of complex protein mixtures generated from body fluids, cells, and/or tissues. SELDI--based profiling identifies unique, differentially expressed proteins relating to specific cancer-related disease states. We utilized SELDI-TOF following pre-processing with molecular separation and chemical fractionation of cell membrane extracts from three Dunning rat prostate cancer cell lines of varying metastatic potential to search novel proteins that are differentially expressed.

METHODS

Dunning rat cell sublines of variable (%) metastatic potential; G (0%), AT-1 (20%), and Mat-Ly-Lu (100%) were cultured in two different laboratories. Cell lysis was performed in a homogenation buffer (320 mM sucrose/50 mM Tris/0.5 mM PSMF) using Dounce homogenation. After centrifugation, the membrane pellet was washed 2x and then solublized in 2% CHAPS/8 M urea. This sample was further processed using positive pressure molecular ultrafiltration at 30 kDa or precipitation with 50% ammonium sulfate. Next, each sample was applied to an IMAC3-Ni ProteinChip (Ciphergen Biosystems, Freemont, CA) and analyzed using Ciphergen's Protein Biology System with protein peak analysis software.

RESULTS

SELDI-TOF analysis differentiated the three Dunning rat cell sublines based upon protein concentration normalized profiles between 5,000 and 20,000 Da. The preparations from the three cells lines showed clear differences when the extracts from the metastatic sublines (AT-1 and MLL) were compared to the benign subline (G) for proteins with molecular weights of 9 kDa (decrease), 12 kDa (significant decrease), 14 kDa (decrease), and 17 kDa (significant gain). After pre-processing extracts with ammonium sulfate and molecular ultrafiltration, the molecular profile changes from one subline to the next became more apparent. Our results were reproducible using multiple runs including from Dunning cells cultured in a separate laboratory, and using different lots of SELDI ProteinChips.

CONCLUSIONS

The application of SELDI-TOF to a series of Dunning rat prostate cancer cell lines illustrated apparent changes in protein profiles among the three cell lines with known differences in metastatic biologic activity. SELDI-TOF identified four reproducible changes in protein expression in the AT1 and MLL metastatic cell sublines. Three of the expression changes were manifested as decreases, but one protein (17 kDa) was over-expressed in the AT1 and MLL cell lines. Emphasis will be placed on the isolation, purification, and characterization of the 17 kDa over-expressed protein and its potential role in PCa metastasis.

Contemporary identification of patients at high risk of early prostate cancer recurrence after radical retropubic prostatectomy

OBJECTIVES

To develop a model that will identify a contemporary cohort of patients at high risk of early prostate cancer recurrence (greater than 50% at 36 months) after radical retropubic prostatectomy for clinically localized disease. Data from this model will provide important information for patient selection and the design of prospective randomized trials of adjuvant therapies.

METHODS

Proportional hazards regression analysis was applied to two patient cohorts to develop and cross-validate a multifactorial predictive model to identify men with the highest risk of early prostate cancer recurrence. The model and validation cohorts contained 904 and 901 men, respectively, who underwent radical retropubic prostatectomy at Johns Hopkins Hospital. This model was then externally validated using a cohort of patients from the Mayo Clinic.

RESULTS

A model for weighted risk of recurrence was developed: R(W)'=lymph node involvement (0/1)x1.43+surgical margin status (0/1)x1.15+modified Gleason score (0 to 4)x0.71+seminal vesicle involvement (0/1)x0.51. Men with an R(W)' greater than 2.84 (9%) demonstrated a 50% biochemical recurrence rate (prostrate-specific antigen level greater than 0.2 ng/mL) at 3 years and thus were placed in the high-risk group. Kaplan-Meier analyses of biochemical recurrence-free survival demonstrated rapid deviation of the curves based on the R(W)'. This model was cross-validated in the second group of patients and performed with similar results. Furthermore, similar trends were apparent when the model was externally validated on patients treated at the Mayo Clinic.

CONCLUSIONS

We have developed a multivariate Cox proportional hazards model that successfully stratifies patients on the basis of their risk of early prostate cancer recurrence.

Contemporary update of prostate cancer staging nomograms (Partin Tables) for the new millennium

OBJECTIVES

We previously presented nomograms combining preoperative serum prostate-specific antigen (PSA), clinical (TNM) stage, and biopsy Gleason score to provide the likelihood of various final pathologic stages at radical retropubic prostatectomy. The data for the original nomograms were collected from men treated between 1982 and 1996. During the past 10 years, the stage at presentation has shifted, with more men presenting with Stage T1c, Gleason score 5 to 6, and serum PSA levels less than 10.0 ng/mL. In this work, we update the "Partin Tables" with a more contemporary cohort of men treated since 1994 and with revised PSA and Gleason categories.

METHODS

Multinomial log-linear regression analysis was used to estimate the likelihood of organ-confined disease, extraprostatic extension, seminal vesicle or lymph nodal status from the preoperative PSA stratified as 0 to 2.5, 2.6 to 4.0, 4.1 to 6.0, 6.1 to 10.0, and greater than 10 ng/mL, clinical (AJCC-TNM, 1992) stage (T1c, T2a, T2b, or T2c), and biopsy Gleason score stratified as 2 to 4, 5 to 6, 3 + 4 = 7, 4 + 3 = 7, or 8 to 10 among 5079 men treated with prostatectomy (without neoadjuvant therapy) between 1994 and 2000 at Johns Hopkins Hospital. The average age was 58 years.

RESULTS

In this cohort, more than 60% had T1c, more than 75% had Gleason score of 6, more than 70% had PSA greater than 2.5 and less than 10.0 ng/mL, and more than 60% had organ-confined disease. Nomograms of the robust estimated likelihoods and 95% confidence intervals were developed from 1000 bootstrap analyses. The probability of organ-confined disease improved across the groups, and further stratification of the Gleason score and PSA level allowed better differentiation of individual patients.

CONCLUSIONS

These updated "Partin Tables" were generated to reflect the trends in presentation and pathologic stage for men newly diagnosed with clinically localized prostate cancer at our institution. Clinicians can use these nomograms to counsel individual patients and help them make important decisions regarding their disease.

Does capsular incision at radical retropubic prostatectomy affect disease-free survival in otherwise organ-confined prostate cancer?

OBJECTIVES

To evaluate the influence of isolated, histologically identified capsular incision (CI) (exposure of benign or malignant glands to the inked surgical margin in the setting of organ-confined disease) on disease progression after anatomic radical retropubic prostatectomy (RRP) for clinically localized prostate cancer.

METHODS

Between March 1993 and September 1999, 4747 men underwent RRP at the Johns Hopkins Hospital; 107 men (2.3%) were diagnosed with CI in otherwise organ-confined disease; 92 (86%) had at least 6 months (mean 30) of follow-up. We matched these CI cases (based on surgeon, age, clinical stage, final pathologic Gleason grade, and preoperative serum prostate-specific antigen level) one-for-one with controls in three additional pathologically defined groups and compared the freedom from disease progression (prostate-specific antigen level greater than 0.2 ng/mL and/or local palpable recurrence) after RRP.

RESULTS

The actuarial 3-year likelihood of freedom from disease progression was 87.8% for the CI group, 96.4% for men with organ-confined disease (P = 0.10), 91.3% for men with extraprostatic extension and negative surgical margins (P = 0.99), and 73.9% for men with positive surgical margins resulting from extraprostatic extension (P <0.01). No statistically significant difference was found in the actuarial likelihood of freedom from disease progression between men with CI into benign glands (n = 22) and men with CI into tumor (n = 70) (P = 0.93).

CONCLUSIONS

No statistically significant difference was found in the likelihood of early recurrence between patients with isolated CI and other specimen-confined disease. Patients with isolated CI have a significantly lower likelihood of early recurrence than patients with positive surgical margins due to extraprostatic extension, regardless of whether the CI is into benign glands or tumor. Long-term follow-up is necessary to confirm these findings.

Safety and efficacy of exisulind for treatment of recurrent prostate cancer after radical prostatectomy

PURPOSE

We evaluated the safety and efficacy of exisulind for delaying disease progression in men with increasing prostate specific antigen (PSA) after radical prostatectomy.

MATERIALS AND METHODS

A total of 96 men with increasing PSA after radical prostatectomy were randomized to receive placebo (49) or 250 mg. exisulind twice daily (47) for 12 months. The primary efficacy parameter was the difference in change from baseline PSA between the placebo and exisulind groups. The PSA doubling time was also evaluated before and during study. A subgroup analysis classified patients based on the risk of developing metastatic disease.

RESULTS

Compared with placebo, exisulind significantly suppressed the increase in PSA in all patients (p = 0.017). The results were also statistically significant in men at high risk for metastasis (p = 0.0003) and those who could not be classified according to risk (p = 0.0009). In addition, median PSA doubling time was lengthened in high risk patients on exisulind (2.12 month increase) compared with those on placebo (3.37 month decrease, p = 0.048). Exisulind was well tolerated.

CONCLUSIONS

Exisulind inhibited the increase in PSA overall and prolonged PSA doubling time in high risk patients compared with placebo. These results suggest that Exisulind has the potential to extend the time from biochemical recurrence to the need for androgen deprivation therapy. Exisulind was well tolerated in this patient population. Our results support further study of Exisulind in the treatment of patients with prostate cancer.

Evaluation of artificial neural networks for the prediction of pathologic stage in prostate carcinoma

BACKGROUND

Currently, the standard for predicting pathologic stage from information available at the time of prostate biopsy is the "Partin nomograms" that were derived using logistic regression analysis. The authors retrospectively reviewed a large series of men with clinically localized prostate carcinoma who underwent staging pelvic lymphadenectomy and radical retropubic prostatectomy. They then utilized pathologic and clinical data at the time of prostate biopsy to develop and test an artificial neural network (ANN) to predict the final pathologic stage for this group of men. They then compared the results of ANN with the previous nomograms.

METHODS

Five thousand seven hundred forty-four men were treated at the authors' institution from 1985 to 1998. An ANN was developed using two randomly selected training and validation sets for predicting pathologic stage. Input variables included age, preoperative serum prostate specific antigen level, clinical TNM (tumor, lymph node, and metastasis) classification, and Gleason score from the biopsy specimen. Outcomes included organ confinement and lymph node involvement status.

RESULTS

The ANN was slightly superior to the nomograms in predicting pathologic stage, such as organ confinement and lymph node involvement status.

CONCLUSIONS

In predicting organ confinement and lymph node involvement status, ANN was more accurate and had a larger area under ROC than the nomograms based on the logistic regression method. Artificial neural network models can be developed and used to better predict final pathologic stage when preoperative pathologic and clinical features are known.

Quantitative nuclear grade (QNG): a new image analysis-based biomarker of clinically relevant nuclear structure alterations

This review addresses the potential clinical value of using quantitative nuclear morphometry information derived from computer-assisted image analysis for cancer detection and predicting outcomes such as tumor stage, recurrence, and progression. Today's imaging technology uses sophisticated hardware platforms coupled with powerful and user-friendly software packages that are commercially available as complete image analysis systems. There are many different mathematically derived nuclear morphometric descriptors (NMD's) (i.e. texture features) that can be calculated by these image analysis systems, but for the most part, these NMD's quantify nuclear size, shape, DNA content (ploidy), and chromatin organization (i.e. texture, both Markovian and non-Markovian) parameters. We have utilized commercially available image analysis systems and the NMD's calculated by these systems to create a mathematical solution, termed quantitative nuclear grade (QNG), for making clinical, diagnostic, and prognostic outcome predictions in both prostate and bladder cancer. A separate computational model is calculated for each outcome of interest using well-characterized and robust training, testing, and validation patient sample sets that adequately represent the selected population and clinical dilemma. A specific QNG solution may be calculated either by non-parametric statistical methods or non-linear mathematics employed by artificial neural networks (ANNs). The QNG solution, a measure of genomic instability, provides a unique independent variable to be used alone or to be included in an algorithm to assess a specific clinical outcome. This approach of customization of the nuclear morphometric descriptor (NMD) information through the calculation of a QNG solution mathematically adjusts for redundancy of features and reduces the complexity of the inputs used to create decision support tools for patient disease management. J. Cell. Biochem. Suppl. 35:151-157, 2000.

Long-term biochemical disease-free and cancer-specific survival following anatomic radical retropubic prostatectomy. The 15-year Johns Hopkins experience

In a large series of 2404 men with a mean follow-up of 6.3 plus or minus 4.2 years (range, 1-17) after anatomic RRP for clinically localized prostate cancer, 412 men (17%) have recurred. A detectable PSA was the only evidence of recurrence in 9.7%, whereas 1.7% and 5.8% had local recurrence and distant metastasis, respectively. The overall actuarial 5-, 10-, and 15-year recurrence-free survival rates for these men were 84%, 74%, and 66%, respectively. As demonstrated in the authors' previous reports, the actuarial likelihood of a postoperative recurrence increased with advancing clinical stage, Gleason-score, preoperative PSA level, and pathologic stage. Subdivision of men with Gleason 7 tumors resulted in better stratification. There was a similar actuarial likelihood of postoperative recurrence for men with Gleason 4 + 3 and Gleason score 8 to 10 disease. The actuarial rate of recurrence of tumor for men with Gleason 3 + 4 disease was statistically different from the rate for men with Gleason score 6 or Gleason 4 + 3 disease. The overall actuarial metastasis-free survival rates at 5, 10, and 15 years were 96%, 90%, and 82%, respectively. The overall actuarial cancer-specific survival rates at 5, 10, and 15 years were 99%, 96%, and 90%, respectively. This study provides long-term outcome of patients with clinically localized cancer who underwent RRP between 1982 and 1999. Recognizing that this long-term study includes many patients with more advanced disease diagnosed before the PSA era, caution must be exercised in comparing these results with the outcomes for cohorts of patients treated since 1989. Anatomic RRP is an effective way to manage clinically localized prostate cancer. Excellent long-term results can be obtained with RRP for early stage disease. The proportion of men with early stage prostate cancer will continue to increase with wide use of serum PSA testing and digital rectal examination.

Indicators of pathologic stage of prostate cancer and their use in clinical practice

Pathologic stage is the most reliable means of predicting the likelihood of curable prostate cancer at the time of definitive treatment. Its prediction is of the greatest importance to individuals with clinically localized disease, principally because of the therapeutic and prognostic implications. Multivariate models integrating variables that can be derived from clinical and pathologic assessment have been shown to be reliable and useful in urologic practice. Among these variables, the combination of clinical stage, serum PSA, and biopsy Gleason score provides reliable assessment of the risk for extraprostatic disease that can be used readily for counseling individual patients. Other biopsy-derived parameters may contribute additional information, but their value in multivariate analysis has not been validated in a multi-institutional setting. The development of new prognostic markers is a priority objective in current research to distinguish patients in whom cancer cannot be controlled by surgical treatment. For patients undergoing radical prostatectomy, definitive pathologic stage certainly will remain an important prognostic factor; therefore, clinical practice will continue to be determined by its accurate prediction.

Era specific biochemical recurrence-free survival following radical prostatectomy for clinically localized prostate cancer

PURPOSE

We retrospectively reviewed a large series of men with clinically localized prostate cancer who underwent surgery to define the extent of stage migration and its influence on biochemical recurrence in 3 different eras of prostate cancer management.

MATERIALS AND METHODS

A total of 2,370 men were treated with radical prostatectomy from 1982 to 1998. We analyzed the freedom from biochemical (prostate specific antigen) progression after radical prostatectomy. We compared the distribution of pathological stage by the year of surgery. We then compared the biochemical recurrence-free survival rate according to the different eras that reflect a change in prostate cancer management.

RESULTS

There was a significant downward stage migration of prostate cancer and an increasing proportion of men who presented with organ confined disease in recent years. The actuarial biochemical recurrence-free rate was significantly different for patients who underwent surgery between 1982 and 1988, compared with those between 1989 and 1998 (p <0.001). These changes may have reflected the benefits of early detection with prostate specific antigen and digital rectal examination, better preoperative selection of patients for surgery as well as the effect of lead time.

CONCLUSIONS

Widespread early detection programs for prostate cancer resulted in downward stage migration in men presenting with prostate cancer at our institution during the last 18 years. Also, we have demonstrated a biochemical recurrence-free survival advantage, probably secondary to an improved therapeutic outcome as well as lead time bias, in men who underwent surgery between 1989 and 1998, compared with those between 1982 and 1988. When trying to compare the efficacy of different treatment modalities for prostate cancer, the era in which patients underwent therapy is an important factor to be considered.

Prediction of prostate carcinoma stage by quantitative biopsy pathology

BACKGROUND

Considerable evidence has shown that the use of computational algorithms to combine pretreatment clinical and pathology results can enhance predictions of patient outcome. The aim of this study was to prove that the application of such methods to predict patient-specific likelihoods of organ-confined (OC) prostate carcinoma (PCA) may be helpful to patients and physicians when they are choosing an optimal treatment for carcinoma of the prostate.

METHODS

The authors used clinical and quantitative pathology results from the biopsy specimens of 817 PCA patients who had been evaluated at a large national pathology reference laboratory. The pathology parameters that were measured included the number of positive cores, Gleason grades and score, percentage of tumor involvement, and the tumor location. The pathologic stage of these cases, as determined by results from radical prostatectomy, lymphadenectomy, or bone scan, categorized the PCA as either OC, non-OC due to capsular penetration only (NOC-CP) or advanced disease with metastasis (NOC-Mets), i.e., seminal vesicle and/or lymph-node positive or bone-scan positive. There were a total of 481 OC cases, 185 NOC-CP cases, and 151 NOC-Mets cases. Patient-specific prediction models were trained by ordinal logistic regression (OLOGIT) and genetically engineered neural networks (GENNs), and the resulting trained models were validated by biopsy information from an independent set of 116 PCA patients.

RESULTS

When the authors applied a cutoff of >or= 35% for the n = 817 training set of OC, NOC-CP, and NOC-Mets predictive probabilities, the OLOGIT model predicted OC PCA with an accuracy of 91%, whereas the GENN model predicted the same with an accuracy of 95%. When the authors employed the n = 116 validation set (76 OCs, 31 NOC-CPs, and 9 NOC-Mets), the OLOGIT and GENN models correctly identified OC PCA with 91% and 97% accuracy, respectively.

CONCLUSIONS

The value of combining patient pretreatment diagnostic pathology parameters to make predictions concerning the postoperative extent of pathology was illustrated clearly in this study. This finding further confirms the need to pursue such approaches for PCA disease management in the future, especially with the increasing prevalence of clinical T1c (American Joint Committee on Cancer, 1977) disease.

Age, prostate-specific antigen, and digital rectal examination as determinants of the probability of having prostate cancer

OBJECTIVES

The decision to perform prostate biopsy has traditionally been based on an abnormal prostate-specific antigen (PSA) level or abnormal digital rectal examination (DRE) findings. For example, a 60-year-old man with a benign DRE and PSA level of 4.1 ng/mL would be counseled for biopsy, and the same man with a PSA level of 3.9 ng/mL might be counseled against biopsy. However, the difference in these PSA levels and in the likelihood of these two men having prostate cancer is not significant. We constructed a probability nomogram for the likelihood of detecting prostate cancer, thus aiding in the decision of whether to perform a prostate biopsy.

METHODS

Using multivariate logistic regression analysis and data from 2054 men (mean age 64 years) participating in the Tyrol Screening Project between January 9, 1993 and January 9, 1997, patient age, PSA level, and DRE findings were analyzed for their ability to determine the likelihood of finding prostate cancer on transrectal ultrasound-guided biopsy.

RESULTS

DRE was suspicious in 278 men (13.5%). Overall, 498 (24.5%) of 2054 men biopsied had prostate cancer. The probability of discovering cancer on biopsy was calculated using patient age, DRE findings, and PSA level.

CONCLUSIONS

DRE status had a large influence on the likelihood of positive biopsy across all PSA and age ranges. A combination of PSA, DRE result, and age better defined the probability of a positive biopsy than any factor alone. Using this nomogram, the decision to proceed with or defer prostate biopsy can be based on an actual probability of discovering prostate cancer rather than a single PSA-based cutpoint. These data may aid physicians and patients in decision-making.

Contemporary use of complexed PSA and calculated percent free PSA for early detection of prostate cancer: impact of changing disease demographics

OBJECTIVES

To assess the diagnostic performance of complexed prostate-specific antigen (cPSA), total PSA (tPSA), and calculated free/total PSA (f/t PSA) ratios in the differentiation of benign disease from prostate cancer (CaP) using a contemporary patient cohort.

METHODS

The cPSA, tPSA, and calculated fPSA values were determined using the Bayer Immuno-1 system. To validate our calculated f/t PSA ratio, we also retrospectively measured fPSA using the Abbott AxSYM immunoassay system in archival pretreatment sera obtained between 1990 and 1997 from 362 men with clinically and biopsy-confirmed benign prostatic hyperplasia (n = 179) or CaP (n = 183). The diagnostic utility of tPSA, cPSA, and the calculated f/t PSA ratio was assessed using a contemporary test population consisting of sera prospectively collected between June 1999 and June 2000 from 3006 men who had recently undergone a systematic biopsy by urologists in clinical practices throughout the United States. This contemporary patient sample had biopsy diagnoses of either no evidence of malignancy (n = 1857) or CaP (n = 1149). All serum samples had tPSA values between 2.0 and 20.0 ng/mL.

RESULTS

The measured versus calculated f/t PSA ratios had a Pearson's correlation coefficient of 0.9130 in the retrospectively studied population of 362 men. The areas under the receiver operating characteristic curves (ROC-AUCs) for the measured and calculated f/t PSA ratios were indistinguishable (69.6% versus 69.2%, respectively). In the contemporary population (n = 3006), the ROC-AUC for tPSA, cPSA, and the calculated f/t PSA ratio was 52.2%, 53.9%, and 58.4%, respectively. We also compared the diagnostic performance using published cutoffs for tPSA (greater than 4.0 ng/mL), cPSA (greater than 3.8 ng/mL), and the f/t PSA ratio (greater than 15% and greater than 25%) in tPSA reflex ranges of 2 to 20 ng/mL and 2 to 10 ng/mL. We found that both cPSA and the f/t PSA ratio (greater than 25% cutoff) outperformed tPSA and yielded similar results in terms of biopsies spared and cancers missed.

CONCLUSIONS

The calculated f/t PSA ratio and cPSA perform equally well in terms of the improvement of specificity in the discrimination of benign disease and CaP. The f/t PSA ratio and cPSA provide clinical benefits over the use of tPSA alone, such as an increased sparing of unnecessary biopsies performed with a manageable degree of risk of delayed cancer detection.

Evaluation of artificial neural networks for the prediction of pathologic stage in prostate carcinoma

BACKGROUND

Currently, the standard for predicting pathologic stage from information available at the time of prostate biopsy is the "Partin nomograms" that were derived using logistic regression analysis. The authors retrospectively reviewed a large series of men with clinically localized prostate carcinoma who underwent staging pelvic lymphadenectomy and radical retropubic prostatectomy. They then utilized pathologic and clinical data at the time of prostate biopsy to develop and test an artificial neural network (ANN) to predict the final pathologic stage for this group of men. They then compared the results of ANN with the previous nomograms.

METHODS

Five thousand seven hundred forty-four men were treated at the authors' institution from 1985 to 1998. An ANN was developed using two randomly selected training and validation sets for predicting pathologic stage. Input variables included age, preoperative serum prostate specific antigen level, clinical TNM (tumor, lymph node, and metastasis) classification, and Gleason score from the biopsy specimen. Outcomes included organ confinement and lymph node involvement status.

RESULTS

The ANN was slightly superior to the nomograms in predicting pathologic stage, such as organ confinement and lymph node involvement status.

CONCLUSIONS

In predicting organ confinement and lymph node involvement status, ANN was more accurate and had a larger area under ROC than the nomograms based on the logistic regression method. Artificial neural network models can be developed and used to better predict final pathologic stage when preoperative pathologic and clinical features are known.

The problem of cutoff levels in a screened population: appropriateness of informing screenees about their risk of having prostate carcinoma

BACKGROUND

Transrectal prostate biopsy decisions often have been based on absolute cutoff values for total and free prostate-specific antigen (PSA). The authors decided that it would be more appropriate to develop risk profiles for the individual patient to allow him to decide whether to undergo a prostate biopsy.

METHODS

To develop risk profiles, the authors first used multivariate logistic regression analysis to analyze 2054 males who were part of the Tyrol (Austria) PSA Screening Project. Second, artificial neural network (ANN) analyses were performed using data from 3474 males who also were part of the Tyrol PSA Screening Project and who had undergone prostate biopsy. These analyses were compared with standard cutoff levels of specificity for the detection of prostate carcinoma.

RESULTS

To the authors' knowledge, this was the first time that multivariate logistic regression analysis was used to decide whether to perform prostate biopsies based on risk profiles rather than on single cutoff levels. For the detection of prostate carcinoma, at sensitivity levels of 90--95%, the ANN was 150--200% more specific than the standard cutoff points. For screened volunteers with total PSA levels below 4 ng/mL, ANN showed a lower cancer predictive ability in comparison with volunteers with total PSA levels above 4 ng/mL. However, the ANN was approximately 150--200% more specific than the standard cutoff levels in both groups.

CONCLUSIONS

At high sensitivity levels, ANN increased the specificity for prostate carcinoma detection in a PSA-based screened population. The improvement in specificity between standard cutoff levels and ANN ranged between 150--200% and was not affected by the presence of benign prostatic hyperplasia or prostatitis.

Prostate-specific antigen and new serum biomarkers for evaluation of chemopreventive agents

A great deal of effort regarding the basic understanding and clinical relevance of the prevention of prostate cancer has emerged over the past decade. Chemoprevention or the administration of a drug or other agent in an attempt to prevent, inhibit, or delay the progression of localized prostate cancer has gained the most recent attention. Efforts have focused primarily in the identification of bioactive chemopreventive agents, risk factors identifying individuals with the highest likelihood of developing prostate cancer, pathologic identification of premalignant lesions, and epidemiologic studies to better understand the natural history of early prostate cancer. However, less work has been focused on identifying and characterizing our presently available biomarkers in an attempt to validate their use as surrogate endpoints or documenting their clinical utility in chemoprevention. This update will focus on a critical evaluation of prostate-specific antigen (PSA), percentage of free PSA, and human glandular kallikrein-2 (hK2) and how they may be used or misused for chemoprevention studies.

Histological markers of risk and the role of high-grade prostatic intraepithelial neoplasia

The marked discrepancy between the prevalence of preclinical prostate cancer and the incidence of clinically manifest disease indicates a long latency phase and significant heterogeneity in the progression potential of early neoplastic lesions. There are a variety of histologic changes within prostatic epithelium that have been termed atypical or dysplastic. The 2 most widely studied of these lesions are prostatic intraepithelial neoplasia (PIN) and atypical adenomatous hyperplasia (AAH). Although associations between AAH and adenocarcinoma are spurious, those linking high-grade PIN (HGPIN) to cancer are far more established. There is a significantly increased risk for patients with isolated HGPIN to have prostate cancer confirmed on subsequent biopsy, suggesting that HGPIN is a marker for prostate carcinoma in addition to its potential role as a premalignant lesion. Autopsy studies reveal that HGPIN is found in association with cancer in 63% to 94% of malignant and 25% to 43% of benign prostates. Data on age and race reveal that African American men develop more extensive HGPIN at a younger age than white men. A wide spectrum of molecular/genetic abnormalities appears to be common to both HGPIN and prostate cancer. Data loss of 8p, 10q, 16q, 18q, and gain of 7q31, 8q, multiple copies of the c-myc genes, along with changes in chromatin texture, telomerase activity, cell cycle status, and proliferative indices collectively suggest that HGPIN is intermediate between benign epithelium and prostatic carcinoma with respect to these markers. These data indicate that HGPIN is important in neoplastic progression, and may present an appropriate target/marker for chemoprevention.

The role of prostate needle biopsy in evaluation of chemopreventive agents

Needle biopsy of the prostate safely furnishes tissue that can be studied for the effects of promising chemopreventive agents on biomarkers. The modulation of biomarkers, such as those for apoptosis (TUNEL, Bcl-2, or nuclear morphometry), angiogenesis (factor 8), and cell proliferation (Ki-67), can indicate the potential of a new agent without waiting for the definitive evaluation of traditional endpoints, such as reduction in cancer mortality. A recent modification of prostate biopsy technique, including additional cores taken from the lateral peripheral zone, may improve the cancer yield by as much as 35% without increasing major complications, facilitating serial in vivo tests on cancer tissue. The serial biopsy approach may be especially valuable in "watchful waiting" cohorts.

Serum acid phosphatase level and biochemical recurrence following radical prostatectomy for men with clinically localized prostate cancer

OBJECTIVES

Serum acid phosphatase (ACP) was once used as the marker for advanced prostate cancer. However, with the development of assays for prostate-specific antigen (PSA), a more sensitive and specific tumor marker, the use of ACP has diminished. We investigated the prognostic value of preoperative serum ACP in predicting prognosis for men with localized prostate cancer following radical retropubic prostatectomy (RRP).

METHODS

Of 2293 men treated from 1982 to 1998, 1681 men had a preoperative ACP measurement using an enzymatic assay. We analyzed the actuarial freedom from biochemical (PSA) progression following RRP according to ACP levels. We used multivariate logistic regression and proportional hazards models to determine the independent prognostic value of ACP level with respect of pathologic stage and biochemical recurrence.

RESULTS

ACP was not an independent predictor of organ confinement or lymph node involvement in the multivariate logistic regression models using preoperative variables. However, in the proportional hazards model, ACP was an independent predictor of tumor recurrence following RRP, and there was a statistically significant improvement in biochemical recurrence-free survival for men with lower levels of ACP (P <0.001). Furthermore, the normalized hazard ratios of ACP and PSA for predicting biochemical recurrence were similar.

CONCLUSIONS

Stratification of men according to their preoperative ACP levels was predictive of patient outcome after RRP. Proportional hazards modeling using preoperative variables demonstrated that the serum ACP level is an independent predictor of tumor recurrence following RRP.

Executive Summary of the National Cancer Institute Workshop: Highlights and recommendations

Prostate cancer chemoprevention represents a relatively new and promising strategy for reducing the immense public health burden of this devastating cancer of men in the United States and Western societies. Chemoprevention is defined as the administration of agents (drugs, biologics, and natural products) that modulate (inhibit) one or more steps in the multistage carcinogenesis process culminating in invasive adenocarcinoma of the prostate. In 2000, there were an estimated 170,000 new cases of prostate cancer and 31,000 deaths in the United States. During the past decade, the National Cancer Institute (NCI) organized the chemoprevention research program and began testing the first generation of promising agents (eg, 4-(hydroxy)-fenretinide [4-HPR], difluoromethylornithine [DFMO], antiandrogens) in high-risk cohorts and launched the first-large scale US phase 3 primary prevention trial, known as Prostate Cancer Prevention Trial (PCPT-1), in 18,000 average-risk men (age more than 55 years and prostate-specific antigen [PSA] less than 3 ng/mL) treated for 7 years with finasteride or placebo. In the summer of 1998, the NCI Prostate Cancer Progress Review Group (PRG) Report to the director of NCI was published in response to the leadership of the prostate cancer advocacy community in conjunction with Congress. To further elucidate and address critical issues identified in this report and to develop a research agenda for the newly created Prostate and Urologic Cancer Research Group in the Division of Cancer Prevention at NCI, the NCI organized the workshop "New Clinical Trial Strategies for Prostate Cancer Chemoprevention." The major objectives were to promote understanding and cooperation among the NCI, US Food and Drug Administration (FDA), academia, pharmaceutical industry, and the public regarding new opportunities for clinical prevention trials for prostate cancer. The workshop was divided into three concurrent breakout panels and a fourth joint integrative panel. The workshop addressed multiple key areas identified in the PRG report in the following panels: (1) Molecular Targets and Promising Agents in Clinical Development; (2) Intermediate Endpoint Biomarkers for Prevention Trials; (3) High-Risk Study Populations for Prevention Trials, and (4) Preventive Clinical Trial Designs and Regulatory Issues. Expert panelists were drawn from leading academic, pharmaceutical, and government scientists in basic research and clinical investigation. Key pharmaceutical, biotechnology, academic, and National Institutes of Health scientists presented overviews of their new agents and products in clinical development (representing the next generation of promising agents). Senior FDA physicians from the Center for Drugs and Center for Biologics presented on current standards for new drug and biologic approval for chemoprevention efficacy. Some of the key topics included recent advances in the state of knowledge of promising agents in the clinic based on molecular targets as well as bottlenecks in drug development for pharmaceutical sponsors; strategic modulable biomarkers that can serve as primary endpoints in phase 1/2 trials to assess preventive efficacy; high-risk cohorts with precancer (high-grade prostatic intraepithelial neoplasia) and representative clinical trial designs that are ready for immediate translation into efficient prevention trials, such as Bayesian sequential monitoring for early assessment of biologic activity and factorial designs for assessment of multiagent combinations. Finally, each expert panel generated recommendations for areas of future research emphasizing opportunities and infrastructure needs.

Artificial neural network model for the assessment of lymph node spread in patients with clinically localized prostate cancer

OBJECTIVES

To develop an artificial neural network (ANN) model to predict lymph node (LN) spread in men with clinically localized prostate cancer and to describe a clinically useful method for interpreting the ANN's output scores.

METHODS

A simple, feed-forward ANN was trained and validated using clinical and pathologic data from two institutions (n = 6135 and n = 319). The clinical stage, biopsy Gleason sum, and prostate-specific antigen level were the input parameters and the presence or absence of LN spread was the output parameter. Patients with similar ANN outputs were grouped and assumed to be part of a cohort. The prevalence of LN spread for each of these patient cohorts was plotted against the range of ANN outputs to create a risk curve.

RESULTS

The area under the receiver operating characteristic curve for the first and second validation data sets was 0.81 and 0.77, respectively. At an ANN output cutoff of 0.3, the sensitivity achieved for each validation set was 63.8% and 44.4%; the specificity was 81.5% and 81.3%; the positive predictive value was 13.6% and 6.5%; and the negative predictive value was 98.0% and 98.1%, respectively. The risk curve showed a nearly linear increase (best fit R(2) = 0.972) in the prevalence of LN spread with increases in raw ANN output.

CONCLUSIONS

The ANN's performance on the two validation data sets suggests a role for ANNs in the accurate clinical staging of patients with prostate cancer. The risk curve provides a clinically useful tool that can be used to give patients a realistic assessment of their risk of LN spread.

Isolated local recurrence is rare after radical prostatectomy in men with Gleason 7 prostate cancer and positive surgical margins: therapeutic implications

PURPOSE

We determined whether the high biochemical failure rate in men with Gleason score 7 disease and positive surgical margins after radical retropubic prostatectomy is secondary to distant metastasis or to local tumor recurrence that could be eliminated by immediate adjuvant radiation therapy.

MATERIALS AND METHODS

Between 1982 and 1997, 112 men with Gleason score 7 disease and positive surgical margins but no seminal vesicle or lymph node involvement underwent radical retropubic prostatectomy without immediate adjuvant radiation or hormonal therapy. Median followup was 8 years (range 1 to 16) and 45 men (40%) were followed 10 years or more. Kaplan-Meier actuarial survival estimates were used to determine the actuarial 5 and 10-year post-prostatectomy, and 5-year post-radiation recurrence rates.

RESULTS

The actuarial 5 and 10-year post-prostatectomy biochemical, local and distant recurrence rates were 40% and 52%, 6% and 6%, and 7% and 16%, respectively. For 20 men who received radiation therapy for isolated prostate specific antigen elevation actuarial 5-year post-radiation biochemical recurrence-free rate was 34%. For 5 men who received radiation therapy for local recurrence actuarial 5-year post-radiation biochemical recurrence-free rate was 20%.

CONCLUSIONS

Isolated clinical local recurrence is rare during long-term followup of men with Gleason score 7 disease and positive surgical margins at radical prostatectomy. Radiation therapy given at prostate specific antigen elevation poorly controlled the disease. Because patients with biochemical failure rarely had local recurrence at long-term followup, they most likely harbored subclinical distant metastasis. These data suggest that immediate adjuvant radiation therapy will not have a major impact on outcome because most men with Gleason score 7 disease and positive surgical margins in whom treatment fails most likely had distant metastasis at surgery. To improve the outcome in cases of Gleason score 7 disease and positive surgical margins a systemic approach to adjuvant therapy is necessary.

Expression of metallothionein in seminal vesicles--an immunohistochemical study

OBJECTIVE

The seminal vesicles and prostate share the same blood supply and exposure to carcinogens. Despite these similarities, fewer than 60 adenocarcinomas of the seminal vesicles have been described, whereas prostate cancer is the most common cancer in men today. Metallothionein plays a significant role in the detoxification of heavy metals. Thus, this study investigated the expression of metallothionein in seminal vesicle tissue.

MATERIAL AND METHODS

Twenty individual tissue specimens each of normal seminal vesicle tissue and benign prostatic tissue underwent immunohistochemical staining with a monoclonal mouse anti-metallothionein antibody.

RESULTS

Positive immunostaining for metallothionein was found in 8 of 20 (40%) of the seminal vesicle tissues, but in 14 of 20 (70%) of the prostate specimens. Seminal vesicle tissue stained only with weak intensity.

CONCLUSION

Metallothionein expression is lower in seminal vesicles than in the prostate. The low cell turnover in seminal vesicle tissue may explain the lower staining activity of this tissue. These findings suggest that metallothionein expression cannot be regarded as the main reason for the vastly different cancer incidence in seminal vesicles and the prostate.

Current Clinical Applications of the In-capromab Pendetide Scan (ProstaScint(R) Scan, Cyt-356)

Prostate-specific antigen (PSA) has been extremely helpful in the detection of new or recurrent prostate cancer. However, localization of the recurrent tumor has been challenging with currently available radiographic modalities. The (111)In-capromab pendetide scan was developed to diagnose accurately and, more importantly, localize and stage a new or recurrent prostate cancer. Studies suggest that the (111)In-capromab pendetide scan can provide more accurate staging of clinically localized prostate cancer prior to staging lymphadenectomy or definitive therapy. It can also provide valuable information when local adjuvant radiation therapy is considered in men with biochemical cancer recurrence following radical prostatectomy.