Stage IV Prostate Cancer: Survival Differences in Clinical T4, Nodal and Metastatic Disease

Transcription

1 Stage IV Prostate Cancer: Survival Differences in Clinical T4, Nodal and Metastatic Disease Wayland Hsiao,* Kelvin A. Moses,* Michael Goodman, Ashesh B. Jani,* Peter J. Rossi* and Viraj A. Master*, From the Department of Urology (WH, KAM, VAM) and Radiation Oncology (ABJ, PJR), Emory University School of Medicine, Department of Epidemiology, Rollins School of Public Health (MG), and Winship Cancer Institute (ABJ, PJR, VAM), Emory University, Atlanta, Georgia Abbreviations and Acronyms AJCC American Joint Committee on Cancer OS overall survival PCa prostate cancer PCSS prostate cancer specific survival PSA prostate specific antigen SEER Surveillance, Epidemiology and End Results Submitted for publication December 20, * Nothing to disclose. Supported by the Atlanta Clinical and Translational Science Institute. Correspondence: Department of Urology, Emory University School of Medicine, 1365 Clifton Rd. NE, Building B, Atlanta, Georgia (telephone: ; vmaster@emory. edu). Editor s Note: This article is the third of 5 published in this issue for which category 1 CME credits can be earned. Instructions for obtaining credits are given with the questions on pages 800 and 811. Purpose: In the prostate specific antigen era most prostate cancer presents at an early stage. However, a significant number of patients have advanced disease, including those with stage IV disease. Assignment to stage IV prostate cancer may occur by different modes, namely as T4N0M0 vs N1 vs M1 disease. We hypothesize that patients with clinical T4 disease have better outcomes than those with N1 or M1 disease. Materials and Methods: A total of 17 SEER registries were queried from 1995 through Multivariate and univariate analyses examined overall survival and prostate cancer specific survival across subcategories of stage IV disease while controlling for various patient and disease related characteristics. Results: There were 615 patients with ct4n0m0 disease, 3,189 with TxN1M0 and 10,893 with TxNxM1 who met the study inclusion criteria. Survival differences were observed between ct4n0m0 and M1 cancer, between N1 and M1 disease, and were most pronounced in younger patients (age 50 years or younger), gradually narrowing with increasing patient age. Factors that demonstrated significant association with poor survival included higher tumor grade, unknown tumor grade and absence of a spouse. Conclusions: Staging systems based on American Joint Committee on Cancer/ TNM staging enables the grouping of patients into homogenous categories for treatment selection and prognostication. However, our data suggest that not all stage IV prostate cancers behave similarly. The difference in survival among locally advanced (T4), node positive and distantly metastatic stage IV prostate cancer appears to be dependent on patient age. Key Words: prostatic neoplasms, neoplasm staging, age groups, SEER program IN 2009, 192,280 patients were diagnosed with prostate cancer and 27,260 died of the disease, making it the most common noncutaneous cancer in American men according to the American Cancer Society. Significant stage migration of PCa has occurred in the PSA era. However, up to 15% of patients continue to have high risk disease and of these many will have advanced stage PCa (AJCC stage IV). 1 Distant spread at diagnosis has been found in 7% of patients in the SEER database 2 and data from the Cancer of the Prostate Strategic Urologic Research Endeavor disease registry reveal approximately 1% of patients have greater than clinical stage T3b tumors. 3 Stage IV PCa, like all solid tumors, is defined in the TNM /10/ /0 Vol. 184, , August 2010 THE JOURNAL OF UROLOGY Printed in U.S.A by AMERICAN UROLOGICAL ASSOCIATION EDUCATION AND RESEARCH, INC. DOI: /j.juro

2 SURVIVAL DIFFERENCES IN STAGE IV PROSTATE CANCER 513 classification by local invasion into adjacent structures, nodal metastasis or distant metastasis. Accurate staging aids in treatment planning and perhaps most importantly in determining prognosis. However, questions have been raised about the prognostic adequacy of AJCC staging in several cancers. 4 9 The AJCC system works well when there is homogeneous survival seen in every stage. However, the current grouping risks ignoring the marked survival differences in subcategories of various stages. Moreover there are additional limitations in not including pretreatment parameters such as PSA and Gleason score, or incorporating the extent of lymph node involvement. 10 Extirpative surgery is generally considered of limited usefulness in the treatment of stage IV cancers, which are usually treated systemically. 11,12 However, given the long natural history of prostate cancer and compelling recent reports of effective treatment of certain patients with stage III prostate cancer, we hypothesized that the different definitions of stage IV prostate cancer (ct4, N,M ) may have different biological outcomes and survival rates, and that stage IV prostate cancer may benefit from surgical treatment. 13,14 Additionally, we wanted to determine if there are any independent prognostic markers which predict improved survival in these patients. MATERIALS AND METHODS We extracted information on patients with prostate cancer reported to the SEER database. Eligibility criteria included 1) prostate cancer diagnosis, 2) AJCC (3rd edition) stage IV and 3) diagnosis between 1995 and Patients were excluded from study if prostate cancer was 1 of multiple malignancies, diagnosis was based on autopsy data or death certificate, prostate cancer was not microscopically confirmed and patient age was younger than 30 years. Followup and vital status information for this study was available through All eligible cases were divided into 3 subcategories of locally advanced disease (ct4n0m0), regional nodal disease (TxN1M0) and metastatic disease (TxNxM1). Patients with pathological T4 but without clinical stage T4 disease were excluded from the locally advanced group. The AJCC definition of T4 is tumor extension or fixation to any adjacent structures other than seminal vesicles. TxN1M0 cases are defined as having regional lymph node spread with no evidence of distant lymph node spread or distant metastasis. TxNxM1 cases are defined as having distant metastasis to bone, distant lymph node spread, soft tissue or other organ involvement. The 3 subcategories of stage IV cancer cases were compared with respect to age at diagnosis, race, marital status, grade, histology, year of diagnosis and treatment received. Treatment was categorized as surgery with or without radiation, radiation alone, or no definitive surgery or radiation. Surgery was defined as prostatectomy with or without node dissection, cystoprostatectomy, or pelvic exenteration. Radiation therapy was defined as brachytherapy and/or external beam radiation therapy. Kaplan-Meier survival curves were constructed for T4, N1 and M1 cases. Log rank tests were performed to compare survival across stage IV subcategories. Multivariate Cox regression analyses were performed to assess the association between stage IV subcategory and survival after controlling for age at diagnosis, race, grade, marital status, year of diagnosis and type of curative treatment received. All variables in the model were tested for proportional hazard assumptions, and all models were tested for the presence of interactions between the main independent variable (stage IV subcategories) and each covariate. The results of multivariable analyses were expressed as adjusted hazard ratios with corresponding 95% confidence intervals. In addition, multivariate adjusted survival plots were constructed to express Cox regression analysis results in a graphic form. All analyses were performed using SPSS 15.0 and SEER*Stat version (National Cancer Institute, Bethesda, Maryland). RESULTS The study cohort included 14,697 cases of stage IV prostate cancer, comprised of 615 ct4n0m0, 3,189 TxN1M0 and 10,893 with TxNxM1 disease at diagnosis. Of the 3,189 patients with TxN1M0 disease only 145 (4.5%) also had clinical T 4 tumors, whereas 10,893 with TxNxM1 disease included only 71 (0.6%) with ct4 tumors. When we examined the N-stage of TxNxM1 cases, for the majority (7,492 or 68%) the presence or extent of nodal involvement was unknown, whereas N0 status was confirmed in 1,884 (17%). Mean age of patients with ct4, N1 and M1 disease was 70.8, 63.6 and 71.3 years, respectively (table 1). In all 3 groups the population was predominantly white and most patients were unmarried. There were similar proportions of high grade disease across groups but there were more patients with unknown grade in the metastatic group. In the metastatic group 56% of patients died of prostate cancer, which is a higher percentage than in the locally advanced or nodal groups. Of patients with N1 disease 53% underwent surgery while a much smaller percentage of patients with ct4 and M1 had surgery (13.3% and 1.4%). Of patients in the metastatic group 78% did not receive radiation or surgical therapy. Given the large sample size of this study all cross tabulations produced statistically significant results (p 0.01). The 5 and 10-year overall survival for locally advanced (ct4n0m0) disease was 59% and 40%, respectively. The corresponding 5 and 10-year prostate cancer specific survival rates were 71% and 60%. For patients with nodal (TxN1M0) disease the 5-year OS and PCSS was 79% and 84%, respectively, and 10-year survival estimates were 55% and 68%.

3 514 SURVIVAL DIFFERENCES IN STAGE IV PROSTATE CANCER Table 1. Characteristics of patients with stage IV prostate cancer No. ct4n0m0 (%) No. TxN1M0 (%) No. TxNxM1 (%) Age: 50 or younger 18 (2.9) 187 (5.9) 347 (3.2) (43.3) 2,353 (73.8) 4,479 (41.1) (53.8) 649 (20.4) 6,067 (55.7) Race: White 443 (72.0) 2,572 (80.7) 7,967 (73.1) Black 113 (18.4) 410 (12.9) 1,979 (18.2) Other 59 (9.6) 207 (6.5) 947 (8.7) Marital status: Married 171 (27.8) 693 (21.7) 3,640 (33.4) Not married 418 (68.0) 2,385 (74.8) 6,575 (60.4) Unknown 26 (4.2) 111 (3.5) 678 (6.2) Grade: I II 236 (38.4) 1,296 (40.6) 2,937 (27.0) III IV 356 (57.9) 1,812 (56.8) 6,210 (57.0) Unknown 23 (3.7) 81 (2.5) 1,746 (16.0) Vital status: Alive 359 (58.4) 2,367 (74.2) 2,738 (25.1) Dead of PCa 147 (23.9) 549 (17.2) 6,129 (56.3) Dead of other causes 109 (17.7) 273 (8.6) 2,026 (18.6) Yr of diagnosis: (21.6) 872 (27.3) 3,019 (27.7) (31.7) 1,017 (31.9) 3,273 (30.0) (46.7) 1,300 (40.8) 4,601 (42.2) Treatment: Prostatectomy 82 (13.3) 1,692 (53.1) 157 (1.4) Radiation 259 (42.1) 517 (16.2) 2,181 (20.0) No surgery or radiation 274 (44.6) 980 (30.7) 8,555 (78.5) The metastatic group had 22% OS and 35% PCSS at 5 years, and 11% OS and 21% PCSS at 10 years. On univariate Kaplan-Meier analysis there was evidence of decreased survival among patients with TxNxM1 disease. The difference in survival between ct4n0m0 and TxN1M0 was not as pronounced, although patients with nodal disease appeared to have better survival compared to those with locally advanced disease (fig. 1). Using Cox regression analyses the effects of stage IV subcategory (ct4 vs N1 vs M1) and treatment (no curative therapy vs prostatectomy vs radiation) on survival were modified by age at diagnosis (p for interaction 0.05). Therefore, all further analyses were stratified by age and all results are reported separately for the 3 age categories of 50 years or younger, 51 to 70 years and older than 70 years. The multivariate adjusted PCSS plots are shown in figure 2. There is improved survival for patients with locally advanced disease compared to those with nodal or metastatic disease and 50 years old or younger. This survival advantage was not extended to men 51 to 70 years old. Patients with locally advanced and nodal diseased fared similarly, while those with metastatic disease still had worse survival. Of men older than 70 years those with nodal disease had the best survival followed by those with locally advanced disease and metastatic disease, respectively. Table 2 shows that using ctxnxm1 subcategory as reference, the adjusted hazard ratio for N1 disease was 0.18 (95% CI ) among patients younger than 50 years, 0.33 (95% CI ) among those 51 to 70 and 0.36 (95% CI ) in those older than 70 years. The survival difference between ct4 and M1 disease was greatest in the youngest patients (HR 0.05, 95% CI ), but appeared to weaken gradually with increasing age with a HR of 0.54 (95% CI ) in the oldest group. Other factors that demonstrated significant association with poor OS (irrespective of the age group) included higher tumor grade, unknown tumor grade and the absence of a spouse. The different treatment approaches had different frequencies in different age groups. In patients younger than 50 years radiation was the only treatment in 29%, surgery (with or without radiation) was used in 23% and the remaining 48% received no curative treatment. The oldest age group included 78% of patients without curative treatment, 18% with radiation only and only 4% with surgery (with or without radiation). In the middle age group the percentage of patients receiving radiation and surgery (with or Figure 1. Overall survival (A) and prostate cancer specific survival (B) of patients with stage IV prostate cancer stratified by mode of inclusion into stage IV category. All values p

4 SURVIVAL DIFFERENCES IN STAGE IV PROSTATE CANCER 515 Figure 2. Prostate cancer specific survival stratified by patient age 50 years or younger (A) 51 to 70 years (B) and older than 70 years (C). All values p without radiation) was 21% and 22%, respectively, while the remaining 57% of patients had no curative treatment. Compared to patients who received no curative treatment surgery demonstrated evidence of a moderate improvement in survival with HR ranging from 0.38 to 0.47 depending on the age group. HRs for radiation were 1.42 (95% CI ) in patients younger than 50 years, 0.99 (95% CI ) in those 51 to 70 years old and 1.02 (95% CI ) in those older than 70 years (table 2). The corresponding results using PCSS as the outcome of interest demonstrated similar patterns for stage IV subcategories, marital status, tumor grade and treatment received (table 3). To avoid cross contamination we reexamined the data by limiting our analyses to nonoverlapping stage IV subcategories, ie T4N0M0 (615), T1 3N1M0 (3,044) and T1 3N0M1 (1,884). Although the sample size decreased, the results remained essentially the same. There was still a significant interaction (p 0.001) between age and stage IV subcategory and, thus, the results still had to be stratified by age. DISCUSSION Stage IV prostate cancer comprises a multitude of differing presentations. Traditionally it has been thought that all patients with AJCC stage IV prostate cancer would fare poorly. However, these data from SEER suggest that this may not be true in all cases. We have shown that differing presentations of stage IV prostate cancer, whether locally advanced disease, regional lymph node spread, or distant metastatic disease, may confer different prognoses. The 5 and 10-year prostate cancer survival rates are comparable to other published series While we have demonstrated a significant difference between T4 vs M1 disease and N1 vs M1 disease, it is likely that the difference between T4 and N1 disease may not be significant. Our observations that younger patients with locally advanced prostate cancer did well after surgery were confirmed in our stratified analyses of the SEER database, which demonstrated a particularly strong survival advantage for patients with ct4 vs M1 disease in the youngest age group. In contrast, in the most advanced age group (older than 70 years)

5 516 SURVIVAL DIFFERENCES IN STAGE IV PROSTATE CANCER Table 2. Multivariate Cox survival analyses of the association between all cause mortality and various patient disease and treatment related characteristics Younger Than 50 yrs yrs Older Than 70 yrs HR (95% CI) p Value HR (95% CI) p Value HR (95% CI) p Value Race: White Black 1.08 ( ) ( ) ( ) 0.50 Other 1.07 ( ) ( ) ( ) Grade: I II III IV 1.76 ( ) ( ) ( ) Unknown 2.49 ( ) ( ) ( ) Marital status: Married Not married 1.16 ( ) ( ) ( ) Unknown 0.98 ( ) ( ) ( ) 0.27 Yr of diagnosis: ( ) ( ) ( ) ( ) ( ) ( ) 0.09 Stage IV subcategory: TxNxM TxN1M ( ) ( ) ( ) T4N0M ( ) ( ) ( ) Curative treatment: None Prostatectomy 0.43 ( ) ( ) ( ) Radiation 1.42 ( ) ( ) ( ) 0.62 Results stratified by age group due to statistically significant interaction (p 0.05) between age and stage IV subcategory, and between age and treatment. the difference between M1 and other subcategories was far less pronounced. Importantly the association between stage IV subcategory and survival were independent of disease grade, patient demographics or treatment received as those factors were included in the multivariate models. This beneficial effect of young age at diagnosis disappears in patients with nodal or metastatic disease. This suggests that young patients with locally advanced disease should be offered prostatectomy as part of a multimodal treatment approach. Interestingly the N1 group was significantly younger than the ct4 or M1 groups and had the best overall survival when all age groups were considered. The N1 group is likely the most heterogeneous in our study because inclusion in this group may be according to radiological or pathological findings, either after pelvic lymphadenectomy with or without radical prostatectomy. Indeed approximately half of the patients in this group were diagnosed in conjunction with a radical prostatectomy. This high percentage of patients having undergone curative therapy in conjunction with their younger age may explain these results. This is also the group that has the lowest percentage of patients (16.2%) undergoing radiation therapy alone. In our age specific multivariate model the most favorable survival was observed in patients with surgical therapy (with or without radiation) compared to patients who had either no radiation or surgery, or those with radiation therapy only. This does not reflect the superiority of any treatment modality (specifically it does not reflect the superiority of surgery compared to radiotherapy), but may reflect the fact that SEER does not give information about comorbidity, patient/physician preference or practice setting. Patients with greater comorbidity may in fact be more likely to undergo radiation therapy. In addition, variation in radiation technique may manifest as variable outcomes after radiation, especially given the advancements in radiation technique with time. Therefore, whether this observed survival difference is due to effectiveness of treatment or is simply selection bias cannot be answered with these data. There are several confounders which make it difficult to answer the critical question of treatment superiority outside of a randomized clinical trial. There are several limitations that are inherent to this type of study. With SEER capturing approximately 26% of the population, our results are not necessarily referable to the entire United States population. 20 Also SEER data do not include PSA, patient comorbidities or the use of androgen deprivation therapy, which may be significant confounders significantly affecting survival. We are also limited by the SEER grading classification which does

6 SURVIVAL DIFFERENCES IN STAGE IV PROSTATE CANCER 517 Table 3. Multivariate Cox survival analyses of the association between prostate cancer specific mortality, and various patient disease and treatment-related characteristics Younger Than 50 Yrs Yrs Older Than 70 Yrs HR (95% CI) p Value HR (95% CI) p Value HR (95% CI) p Value Race: White Black 1.09 ( ) ( ) ( ) 0.71 Other 0.78 ( ) ( ) ( ) Grade: I II III IV 1.96 ( ) ( ) ( ) Unknown 2.81 ( ) ( ) ( ) Marital status: Married Not married 1.20 ( ) ( ) ( ) Unknown 1.02 ( ) ( ) ( ) Yr of diagnosis: ( ) ( ) ( ) ( ) ( ) ( ) 0.15 Stage IV subcategory: TxNxM TxN1M ( ) ( ) ( ) T4N0M ( ) ( ) ( ) Curative treatment: None Prostatectomy 0.36 ( ) ( ) ( ) Radiation 1.40 ( ) ( ) ( ) Results stratified by age group due to statistically significant interaction (p 0.05) between age and stage IV subcategory, and between age and treatment. not allow for comparison of specific Gleason patterns. The reason for the finding that radiotherapy correlates adversely with outcome is difficult to ascertain in the SEER database because of limitations of the system of radiation coding in SEER. Radiation coding in SEER does not capture the intent of radiation, radiotherapy dose, or the site of administration. Thus, patients receiving palliative radiation for bony metastases may be coded as receiving radiation therapy. If this use of radiotherapy was included for even a small sample of patients, the radiotherapy group could have worse outcomes than those seen for other modalities. In our database a total of 3,233 patients underwent radiation therapy. Of these patients 276 received radiation in conjunction with surgery. Of those 276 there were 11 patients in the ct4 group, 244 in the N1 group and 21 in the M1 group which comprised 1.8% (11 of 615), 7.7% (244 of 3,189) and 0.2% (21 of 10,893) of their respective groups. Since the number of patients having undergone both treatments was so small, we reported on these patients along with those who underwent surgery only. While separate categorization of radiotherapy delivered as external beam radiotherapy, brachytherapy or adjuvant treatment might cause 1 of these subsets to be different, this subcategorization may have its own set of limitations. Thus, for purposes of our report, which surveys the relative importance of major patient demographic, disease and treatment factors, it is appropriate to include radiation therapy as a dichotomous variable. CONCLUSIONS Our results indicate that the survival of patients with stage IV prostate cancer depends on the specific type of tumor progression which can be subcategorized as locally advanced (ct4), regional nodal (N1) or metastatic (M1) disease. Moreover in locally advanced prostate cancer patient age at diagnosis appears to affect survival. Surgery (with or without radiation) was associated with better survival in younger patients with T4N0M0 prostate cancer. Patients with metastatic disease fared poorly across all age groups. The particular groupings that we have formed allow for relative comparison of subsets within the AJCC classification of stage IV disease. Thus, the prognosis of patients with stage IV prostate cancer should take into consideration the specific subcategory of disease within the defined age groups in addition to traditional prognostic factors such as grade, comorbidities and patient social/demographic characteristics. ACKNOWLEDGMENTS SEER Citation: Surveillance, Epidemiology, and End Results (SEER) Program (

7 518 SURVIVAL DIFFERENCES IN STAGE IV PROSTATE CANCER gov) SEER*Stat Database: Incidence - SEER 17 Regs Limited-Use Hurricane Katrina Impacted Louisiana Cases, Nov 2007 Sub ( varying) - Linked To County Attributes - Total U.S., Counties, National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2008, based on the November REFERENCES 1. Cooperberg MR, Park S and Carroll PR: Prostate cancer 2004: insights from national disease registries. Oncology (Williston Park) 2004; 18: Ahaghotu C, Baffoe-Bonnie A, Kittles R et al: Clinical characteristics of African-American men with hereditary prostate cancer: the AAHPC Study. Prostate Cancer Prostatic Dis 2004; 7: Scales CD Jr, Moul JW, Curtis LH et al: Prostate cancer in the Baby Boomer generation: results from CaPSURE. Urology 2007; 70: Leijte JA, Gallee M, Antonini N et al: Evaluation of current TNM classification of penile carcinoma. J Urol 2008; 180: Chun FK, Briganti A, Lebeau T et al: The 2002 AJCC pt2 substages confer no prognostic information on the rate of biochemical recurrence after radical prostatectomy. Eur Urol 2006; 49: Hutterer GC and Karakiewicz PI: Can the AJCC staging system risk-stratify patients with localized prostate cancer? Nat Clin Pract Urol 2007; 4: May F, Hartung R and Breul J: The ability of the American Joint Committee on Cancer staging system to predict progression-free survival after radical prostatectomy. BJU Int 2001; 88: Roach M 3rd, Weinberg V, Sandler H et al: Staging for prostate cancer: time to incorporate pretreatment prostate-specific antigen and Gleason score? Cancer 2007; 109: Taylor SH, Merriman KW, Spiess PE et al: Inadequacies of the current American Joint Committee on Cancer staging system for prostate cancer. Cancer 2006; 106: Briganti A, Karnes JR, Da Pozzo LF et al: Two positive nodes represent a significant cut-off value for cancer specific survival in patients with node positive prostate cancer. A new proposal based on a two-institution experience on 703 consecutive N patients treated with radical prostatectomy, extended pelvic lymph node dissection and adjuvant therapy. Eur Urol 2009; 55: Guinan P, Stewart AK, Fremgen AM et al: Patterns of care for metastatic carcinoma of the prostate gland: results of the American College of Surgeons patient care evaluation study. Prostate Cancer Prostatic Dis 1998; 1: Resnick MI and Grayhack JT: Treatment of stage IV carcinoma of the prostate. Urol Clin North Am 1975; 2: Johnstone PA, Ward KC, Goodman M et al: Radical prostatectomy for clinical T4 prostate cancer. Cancer 2006; 106: Van Poppel H and Joniau S: An analysis of radical prostatectomy in advanced stage and high-grade prostate cancer. Eur Urol 2008; 53: Boorjian SA, Thompson RH, Siddiqui S et al: Long-term outcome after radical prostatectomy for patients with lymph node positive prostate cancer in the prostate specific antigen era. J Urol 2007; 178: Cheng CW, Bergstralh EJ and Zincke H: Stage D1 prostate cancer. A nonrandomized comparison of conservative treatment options versus radical prostatectomy. Cancer 1993; 71: Gjertson CK, Asher KP, Sclar JD et al: Local control and long-term disease-free survival for stage D1 (T2-T4N1-N2M0) prostate cancer after radical prostatectomy in the PSA era. Urology 2007; 70: Kroepfl D, Loewen H, Roggenbuck U et al: Disease progression and survival in patients with prostate carcinoma and positive lymph nodes after radical retropubic prostatectomy. BJU Int 2006; 97: Powell IJ, Tangen CM, Miller GJ et al: Neoadjuvant therapy before radical prostatectomy for clinical T3/T4 carcinoma of the prostate: 5-year followup, Phase II Southwest Oncology Group Study J Urol 2002; 168: Shuch B, Mikhail M, Satagopan J et al: Racial disparity of epidermal growth factor receptor expression in prostate cancer. J Clin Oncol 2004; 22: 4725.