Second Primary Cancer Risk of Radiation Therapy After Radical Prostatectomy for Prostate Cancer: An Analysis of SEER Data

Oncology Second Primary Cancer Risk of Radiation Therapy After Radical Prostatectomy for Prostate Cancer: An Analysis of SEER Data May Abdel-Wahab, Isildinha M. Reis, Jiuhua Wu, and Robert Duncan OBJECTIVES METHODS

RESULTS

CONCLUSIONS

To determine the incidence of second primary cancer (SPC) and primary pelvic late SPC/ radiation-induced SPC after radical prostatectomy and radiation. A total of 228 235 prostate cancer patients in the 1973-2002 Surveillance, Epidemiology, and End Results database were studied. The age-adjusted estimates of SPCs was calculated. Competing risk multivariable Cox proportional hazards regression analysis was adjusted for age at diagnosis, race or ethnicity, and radiation and was used to evaluate the effect of treatment on SPC. The overall incidence of SPC was 8.4%. The most frequent pelvic SPCs were bladder (2303 cases), rectum or rectosigmoid junction (1006 cases). The most frequent nonpelvic SPCs were bronchus and lung (4131 cases), colon (2665 cases), and skin (1769 cases). The absolute risk of developing a second malignancy was 1747 cases per 100 000 in the “Radical surgery and x-ray treatment” group and 1581 in the “radical surgery” group. With regard to late primary pelvic SPC, a higher age-adjusted rate of 374 cases per 100 000 was seen in the radiated group. Radiation after radical surgery increased late primary pelvic SPC. No increases were seen in secondary pelvic or extrapelvic SPCs. UROLOGY 74: 866 – 872, 2009. © 2009 Elsevier Inc.

R

adiation has been used for salvage treatment for biochemical failure after radical prostatectomy.1 Recently, the results of 2 large randomized postoperative adjuvant trials have been reported.2,3 Bolla et al2 showed a significant improvement in biochemical progression-free survival (74.0% vs 52.6%, P ⬍ .0001), clinical progression-free survival, and locoregional failure in the group irradiated postoperatively. The Southwest Oncology Group study demonstrated that postprostatectomy adjuvant radiation significantly improves median prostate-specific antigen relapse-free and median recurrence-free survival.3 The advent of more effective treatments and the increase in long-term survivorship led to a realization of the high rate of second or multiple primary cancers in cancer survivors.4 Genetic and environmental factors play a major role in the development of second primary cancer (SPC) and most SPCs after treatment are not treatment induced. Radiation-induced SPC (RTSPC) within the radiation field is a rare side effect of radiation therapy,5 From the Department of Radiation Oncology, Miller School of Medicine, University of Miami, Miami, Florida; Division of Biostatistics, Sylvester Comprehensive Cancer Center, University of Miami, Miami, Florida; Department of Epidemiology and Public Health, Miller School of Medicine, University of Miami, University of Miami, Miami, Florida; and Biostatistics, RTP-PPD Inc, Morrisville, North Carolina Reprint requests: May Abdel-Wahab, M.D., Department of Radiation Oncology, Miller School of Medicine, University of Miami, 1475 NW 12th Ave (D-31), Miami, FL 33 136. E-mail: [email protected] Submitted: August 06, 2008, accepted (with revisions): February 27, 2009

866

© 2009 Elsevier Inc. All Rights Reserved

which is defined as tumor that arises ⱖ 5 years after radiation from tissue within the irradiated field and that have histopathologic features different from those of primary tumor.6-8 Although an increased incidence of rectal and bladder cancers after RTSPC has been reported by some investigators,9 others report no increase in rectal10 or in bladder cancer risk11 with radiation. Some even report a decrease in bladder cancer risk with radiation and an increase in sarcoma risk.12 To our knowledge, this is the first study to specifically address the risk of SPCs in patients receiving radiation in conjunction with radical surgery.

PATIENTS AND METHODS Eligible patients who fulfilled the study criteria were identified in the 1973-2002 Surveillance, Epidemiology, and End Results (SEER) database13 (228 235 patients). These were patients who were diagnosed between 1988 and 2002 with local/regional prostate adenocarcinoma as their first malignancy. Data before 1988 were excluded from this analysis because before 1988 there were no cases in which surgery information was equal to a definitive “no surgery”; 79% of the cases had undergone surgery and for the remaining 21% cases, surgery was coded as “unknown.” Other exclusions included cases diagnosed by autopsy or death certificate, cases with a follow-up of ⬍ 1 year, and those in which SPC diagnosis was within 1 year of primary prostate cancer (time to SPC ⬍ 1 year) to eliminate the possibility that prostate cancer was not the first malignancy. 0090-4295/09/$34.00 doi:10.1016/j.urology.2009.02.085

Table 1. Demographic and other characteristics by second primary cancer (SPC) status and for 2 treatment groups

Total Patients N % Age (y) at diagnosis ⬍ 65 65⫹ Range Mean (SD) Race/Ethnicity White Hispanic White non-Hispanic Black non-Hispanic Other non-Hispanic Year of diagnosis of 1st primary 1988-1992 1993-1996 1997-2000 Tumor grade of 1st primary 1 2 3/4 Histologic type (ICDO-3) 8140 Adenocarcinoma NOS Other* Number of primaries 1 2 ⱖ3

Diagnosed With Second Primary Cancer (SPC) Yes No N % N %

XRT ⫹ Radical Radical Surgery Surgery Only N % N %

228 235 100.0 19 182 100.0 209 053 100.0

5044

100.0 80 157 100.0

74 804 32.8 153 431 67.2 22-115 68.2 (8.9)

3838 20.0 15 354 80.0 38-99 70.4 (7.3)

70 976 33.9 138 077 66.1 22-115 68.0 (9.0)

2772 55.0 2272 45.0 39-87 62.9 (7.3)

44 092 55.0 36 065 45.0 32-100 62.9 (7.2)

13 175 25 12

948 719 770 798

6.1 77.0 11.3 5.6

782 15 559 1997 844

4.1 81.1 10.4 4.4

13 160 23 11

166 160 773 954

6.3 76.6 11.4 5.7

306 3992 444 302

6.1 79.1 8.8 6.0

5442 63 091 8085 3539

6.8 78.7 10.1 4.4

51 570 63 806 112 859

22.6 28.0 49.4

7640 6844 4698

39.8 35.7 24.5

43 930 56 962 108 161

21.0 27.2 51.7

1582 1459 2003

31.4 28.9 39.7

17 035 23 798 39 324

21.3 29.7 49.1

28 280 155 715 44 240

12.4 68.2 19.4

3416 12 235 3531

17.8 63.8 18.4

24 864 143 480 40 709

11.9 68.6 19.5

164 2839 2041

3.3 56.3 40.5

5665 60 025 14 467

7.1 74.9 18.0

222 488

97.5

18 810

98.1

203 678

97.4

4863

96.4

76 860

95.9

5747

2.5

372

1.9

5375

2.6

181

3.6

3297

4.1

209 053 17 685 1497

91.6 7.7 0.7

NA 17 685 1497

NA 92.2 7.8

209 053 100.0 NA NA NA NA

4613 392 39

91.5 7.8 0.8

74 470 5238 449

92.9 6.5 0.6

NA ⫽ not applicable. SPC and non-SPC groups differ regarding to each of these variables (P ⬍ .0001), given the large number of patients in both groups. * 8550 acinar cell carcinoma (5122 cases, 2.2%), 8141 scirrhous adenocarcinoma (13 cases), 8201 cribriform carcinoma NOS (107), 8260 papillary adenocarcinoma NOS (26), 8310 clear cell adenocarcinoma NOS (45), 8380 endometrioid adenocarcinoma NOS (24), 8480 mucinous adenocarcinoma (169), 8490 Signet ring cell carcinoma (37), 8481 mucin-producing adenocarcinoma (101), 8500 infiltrating duct carcinoma NOS (103).

Of the 228 235 patients analyzed, 13 834 patients received a combination of surgery and radiation (nonradical surgery in 8790 and radical surgery in 5044). A total of 105 949 patients had undergone surgery alone (80 157 patients had radical surgery alone while 25 792 patients had nonradical surgery alone). The remaining patients either received radiation alone (67 719) or no definitive treatment (40 733) (reference group, “no x-ray treatment [XRT], no surgery”) for their primary prostate cancer. Data retrieved from SEER database13 used in this analysis included several patient demographics and disease characteristics (Table 1), treatment modality and survival since prostate carcinoma diagnosis, as well as survival since SPC diagnosis for those with SPC. For SPC patients, time to SPC was given by the difference between survival since prostate cancer diagnosis and survival since SPC diagnosis. For non-SPC patients, censored observations, time to SPC equals to survival from primary prostate cancer. The development of SPCs was analyzed for the 6 groups including the reference group “no XRT, no surgery” (Table 2). We classified primary pelvic area SPCs as those arising from the bladder, rectum, anus or anal canal or anorectum, prostate, and other malignancies identified as involving the pelvis (soft tissue, bone and joints, and lymphoma). The primary pelvic area was UROLOGY 74 (4), 2009

designated as such because of the higher likelihood of these organs receiving some or all the radiation dose. Secondary pelvic area SPCs were those arising from the rectosigmoid region, penis, small intestine (ileum and jejunum), ureter, other urinary and male genital organs, testis and lymphoma within the pelvis. The secondary pelvic area may possibly have received radiation depending on the technique used although the probability is not as high as the primary pelvic area. Nonpelvic SPC tumors were those outside the pelvic area.

Statistical Analysis The focus of this analysis was to determine the effect, if any, of radiation after radical surgery on SPCs and RTSPCs. The original study group included eligible prostate cancer patients in the SEER database. Further analysis showed that the group comprised 6 groups, which included the following: no treatment, nonradical surgery, radical surgery, radiation alone, radiation plus nonradical surgery, and radiation plus radical surgery. A detailed analysis was done to compare patients treated with radiation plus radical surgery (XRT_R) and those treated with radical surgery only (R). For analysis of the incidence density of SPCs for each of the 2 treatment group, rates were calculated as the number of cases per 100 000 person-years of follow-up. 867

Table 2. Second primary cancer (SPC) specific site in pelvic and in nonpelvic area by treatment of primary prostate cancer Total Patients N % Total patients Patients diagnosed with SPC Primary pelvic area* Urinary bladder Rectum Other sites† Secondary pelvic area* Rectosigmoid junction Penis Small intestine Ureter Other urinary organs Other sites‡ Nonpelvic area Respiratory system Colon excluding rectum Skin excluding basal and squamous Lymphoma Leukemia Kidney and renal pelvis Pancreas Stomach Oral cavity and pharynx Myeloma Esophagus Brain and other nervous system Miscellaneous§ Other GI¶ Other sites㛳 Late SPC Primary pelvic area Secondary pelvic area Nonpelvic area

228 235 19 182 3101 2303 695 103 552 311 72 65 37 32 35 15 529 4131 2665 1769 1082 792 702 687 582 545 407 309 291 679 484 404 8010 1304 238 6468

100.0 16.2 12.0 3.6 0.5 2.9 1.6 0.4 0.3 0.2 0.2 0.2 81.0 21.5 13.9 9.2 5.6 4.1 3.7 3.6 3.0 2.8 2.1 1.6 1.5 3.5 2.5 2.1 100.0 16.3 3.0 80.7

XRT ⫹ Radical Surgery N % 5044 431 87 62 22 3 17 10 3 1 none 1 2 327 86 46 43 20 19 14 10 14 11 7 9 10 16 14 404 224 52 10 162

100.0 20.2 14.4 5.1 0.7 3.9 2.3 0.7 0.2 none 0.2 0.5 75.9 20.0 10.7 10.0 4.6 4.4 3.2 2.3 3.2 2.6 1.6 2.1 2.3 3.7 3.2 2.1 100.0 23.2 4.5 72.3

Radical Surgery Only N % 80 157 5687 802 563 214 25 163 88 22 26 10 4 13 4,722 1191 767 684 344 213 245 221 138 167 138 71 105 175 126 137 2849 378 86 2385

100.0 14.1 9.9 3.8 0.4 2.9 1.5 0.4 0.5 0.2 0.1 0.2 83.0 20.9 13.5 12.0 6.0 3.7 4.3 3.9 2.4 2.9 2.4 1.2 1.8 3.1 2.2 2.4 100.0 13.3 3.0 83.7

* Primary pelvic area includes sites that are absolutely irradiated while secondary pelvic area includes sites which are possibly irradiated when XRT is used for treatment of prostate cancer. † Other sites in secondary pelvic area includes the following: anus, anal canal, and anorectum (51 cases), soft tissue (30 cases), bones and joints (5 cases), lymphoma in bladder (7 cases), prostate (7 cases), and miscellaneous (pelvis, NOS; 3 cases). ‡ Other sites in secondary pelvic area include other male genital organs (19 cases), testis (8 cases), and lymphoma in testis (8 cases). § Miscellaneous includes hematopoietic system: C42.0 (blood, 15 cases), C42.1 (bone marrow, 167 cases), C42.4 (NOS, 1 case), C76.2 (other and ill-defined sites—abdomen, NOS; 2 cases), and 80.9 (unknown primary site; 488 cases). ¶ Other GI includes liver, bile duct, small intestine, and other digestive organs. 㛳 Other sites in nonpelvic area includes soft tissue including heart (123 cases), endocrine system (109 cases), breast (65 cases), eye and orbit (62 cases), retroperitoneum (20 cases), bones and joints (15 cases), and peritoneum, omentum, and mesentery (10 cases).

These incidences were adjusted for age. Five-year intervals were used in the adjustment for age using weights based on the age distribution of the entire population. The age-adjusted estimates and corresponding 95% confidence intervals of SPCs were computed as described in Rothman and Greenland.14 The data were then tabulated as occurring within 1-5 years (early) and ⱖ 5 years (late) for both SPCs and primary pelvic SPCs. Rates and 95% confidence intervals were computed as described earlier. Pairwise comparisons of incidence density were made using the z test for independent samples. These analyses were conducted using the SAS statistical software package (V 9.1 SAS Institute, Cary, NC). Cumulative incidence estimates based on a competing risk analyses as described by Gray15 were performed for the occurrence of primary pelvic SPCs, secondary pelvic SPCs, nonpelvic SPCs, with death as another competing event, using the “cuminc” procedure in the R statistical package “cmprsk.”16 Competing risk cumulative incidence analyses were then run univariately on the following variables: Treatment (XRT_R, 868

R), Age (⬍ 65, 65⫹), Ethnicity (white non-Hispanic [WNH], black non-Hispanic, white Hispanic, and other non-Hispanic), and Grade (1, 2, 3 plus 4). Variables significant in these analyses were then selected for the competing risk multivariable Cox proportional hazards regression analysis as described by Fine and Gray.17 Cox models were computed using the “crr” procedure in the R statistical package “cmprsk.”16 The significant univariate variables from the cumulative risk analyses were age, ethnicity, and treatment. Subsequently, ethnicity was recoded as WNH and other since white Hispanic, black nonHispanic, and other non-Hispanic behave similarly. The first Cox model tested whether there was any interaction between treatment and the other 2 variables. We found no significant interactions, so the final Cox model contained the variables treatment (XRT_R, R), age (⬍ 65, 65⫹), and ethnicity (WNH, other). Similar analyses were conducted for incidence of late SPC (after 5 years of primary prostate cancer) comparing radiation plus radical surgery and radical surgery alone treatments. The assumption of proportional hazards UROLOGY 74 (4), 2009

was tested using the interaction between time and treatment in the model.

RESULTS Table 1 presents demographics and disease characteristics by SPC status and for 2 treatment groups. There appears to be an aggregation of SPC occurrences among those aged ⬎ 65 years, a clear preponderance of SPCs among WNH. Table 2 presents SPC sites in the pelvic area and in the nonpelvic area by site and treatment. The frequency of SPC in the entire group was 8.4% (19 182 cases). These included 3101 (16.2%) cases in the primary pelvic area, 552 (2.9%) in the secondary pelvic area, and 15 529 (81%) cases in nonpelvic area. Thus, most SPCs (81%) were outside the pelvis. The most common sites in the pelvic area were cancers in bladder, rectum, and rectosigmoid junction. The vast majority of SPCs were carcinomas (13 999 cases or 73%), whereas sarcomas represented only 1.8% (340 cases) of all SPCs. Only 0.2% (46 cases) of sarcomas were in the primary pelvic area. Most sarcomas were in nonpelvic areas (279 cases representing 1.5% of SPCs). As expected, patients treated with radiation after radical surgery had worse tumor grade than those treated with radical surgery alone. Interestingly, there were 7 cases of prostate SPCs. Three were sarcomas in patients who had received radiation, 2 were transitional cell carcinomas in the “nonradical surgery only” group, and the remaining 2 included a squamous cell carcinoma in a patient who had been treated with radiation only, and a ductal carcinoma case in the “non-radical surgery only” group. The age-adjusted person-year estimates of incidence densities of all SPCs within 1-15 years was 1747 and 1581 per 100 000 in the radiation plus radical surgery group vs the radical surgery alone group, respectively. When only SPCs that occurred 5 or more years were analyzed, the age-adjusted rates were 1854 and 1745 per 100 000, respectively. There was also an increase in pelvic SPC incidence after ⱖ 5 years among patients treated with radical surgery and radiation compared with those treated with radical surgery alone (620 vs 246 per 100 000, respectively). Effect of Radiation After Radical Surgery Figure 1 shows unadjusted estimates of cumulative incidence of late SPC in primary pelvic and secondary pelvic as well as nonpelvic areas among patients who received either radiation plus radical surgery or radical surgery only. The difference reached statistical significance only for late SPCs occurring in the primary pelvic area (P ⬍ .0001); for instance the 7-, 9-, 11-, and 13-year cumulative incidence rates of late SPCs in the pelvic area were 476, 1338, 2237, and 2605 cases per 100 000, respectively, in patients treated with radiation plus radical surgery, and 385, 723, 1014, and 1455, respectively, for patients treated with radical surgery only. The correUROLOGY 74 (4), 2009

Figure 1. Unadjusted cumulative incidence of late second primary cancers (SPCs) (after 5 years) among patients treated with radiation and radical surgery or radical surgery alone and corresponding hazard ratios.

sponding estimated time-specific hazard ratios, reported in lower panel of Figure 1, indicate that the proportional hazard assumption is reasonable. The results of the competing risks multivariable Cox regression model focusing on the 2 treatment groups are shown in Table 3. Analysis of all SPCs among patients who received either radiation plus radical surgery or radical surgery only are shown in the upper panel. For primary pelvic SPCs, radiation after radical surgery, age, and ethnicity are all significant predictors, with hazards ratios of 1.53, 1.91, and 1.50, respectively. Thus, radiation, age ⬎ 65, and being WNH all significantly elevate the hazard of developing an SPC in the primary pelvic area. Only age is a significant factor in the development of secondary pelvic SPCs (hazard ratio ⫽ 1.53). Radiation after radical surgery was not a significant predictor of nonpelvic SPCs (hazard ratio 0.96, P ⫽ .510). In the lower panel of Table 3, we show results estimates of relative risk of occurrence of late SPC in the subset of patients who were alive and SPC-free after 5 years after treatment with either radiation plus radical surgery or radical surgery only for their first primary prostate cancer. For primary pelvic late SPCs, radiation after radical surgery, age, and ethnicity are all significant predictors, with hazards ratios of 1.82, 1.59, and 1.79, respectively. Thus, radiation after radical surgery, age ⬎ 65, and being WNH all significantly elevate the hazard of developing a late SPC in the primary pelvic area. Only age is a significant 869

Table 3. Estimated hazard ratios of occurrence of second primary cancer (SPC) and of late SPCs, based on competing risk* Cox regression analysis, in patients treated for first primary prostate cancer with radiation plus radical surgery or radical surgery only

Prognostic Factor XRT_R vs R Age: 65⫹ vs ⬍ 65 y Race: WNH vs other

Prognostic Factor XRT_R vs R Age: 65⫹ vs ⬍ 65 y Race: WNH vs other

HR 1.53 1.91 1.50

Primary Pelvic SPC 95% CI P 1.22-1.90 1.66-2.20 1.24-1.83

All SPCs Secondary Pelvic SPC HR 95% CI P

⬍ .001 ⬍ .0001 ⬍ .0001

1.45 1.53 0.97

0.88-2.39 1.13-2.07 0.66-1.41

.140 .006 .870

HR 0.96 1.65 1.09

Nonpelvic SPC 95% CI 0.86-1.08 1.56-1.75 1.01-1.17

P

.510 ⬍ .0001 .027

Late SPCs Among Patients who were Alive and SPC-Free after 5 y Primary Pelvic Late SPC Secondary Pelvic Late SPC Nonpelvic Late SPC HR 95% CI P HR 95% CI P HR 95% CI P 1.82 1.59 1.79

1.36-2.43 1.30-1.94 1.31-2.46

⬍ .0001 ⬍ .0001 .003

1.54 1.58 0.76

0.80-2.96 1.04-2.39 0.47-1.25

.190 .033 .290

0.89 1.40 1.10

0.76-1.04 1.30-1.52 0.98-1.22

.150 ⬍ .0001 .094

XRT_R ⫽ radiation plus radical surgery; R ⫽ radical surgery only; Age ⫽ age at diagnosis of primary prostate cancer; WNH ⫽ white non-Hispanic; other ⫽ white Hispanic, black non-Hispanic, and other non-Hispanic. * Death from any cause was also considered a competing event.

factor in the development of secondary pelvic late SPC (hazard ratio ⫽ 1.58) and in the development of nonpelvic late SPC (hazard ratio ⫽ 1.40). None of the models used in the analysis of late SPC showed a significant departure from the proportional hazards assumption.

COMMENT The risk of SPC after treatment for prostate cancer has been studied with conflicting results and without separate analyses of definitive radiation vs radiation used in conjunction with radical surgery.9-12,18-23 There are several caveats if one attempts to extrapolate the previous study results to the postprostatectomy setting. Patients treated definitely with radiation alone (whether brachytherapy or external beam radiation) and those treated with radiation as an adjunct to radical surgery may have different SPC patterns related to different doses, techniques, and volumes treated. Furthermore, patient selection bias may influence the results with healthier, younger patients generally undergoing surgery, with or without adjuvant radiation. The fact that 81% of SPCs were outside the pelvis demonstrates that SPCs are an issue in these patients regardless of radiation and that other predisposing factors such as lifestyle or genetic factors may be involved, particularly because the most common nonpelvic SPCs were lung (21.5%) and colon (13.9%) neoplasms commonly related to lifestyle and genetic factors. Another potential source of bias is the increase in the risk of developing SPCs with aging regardless of any intervention or treatment. Thus, even when we adjust for length of follow-up after the first cancer diagnosis, the actual age at which the primary cancer was diagnosed adds another variable. Kendal et al10 addressed this problem using age attained and found no increase in rectal cancer using this method. In our analysis, the SPC group 870

had a greater percentage of patients aged ⬎ 65 years than the group without SPC (80% vs 66%). Furthermore, the non-SPC group had a greater percentage of patients who were diagnosed in 1997 or later (51.7% vs 24.5%), suggesting a possible bias as to the length of follow-up and availability of modern techniques between the groups. Various studies have shown conflicting results with regard to SPC rate after radiation.9-12,18-22 In this study, an absolute increase was noted in 374 cases per 100 000 of pelvic SPC observed after 5 years (RTSPC) in patients who receive a combination of radiation and radical surgery, compared to radical surgery alone in this series. Of note is that the limitations of the current database preclude investigation of risk factors. The fact that the overall incidence of SPC was lower in the radical surgery group than the group that received no treatment suggests that there may be a selection bias at play. The importance of precise risk assessment becomes evident from the single institutional study by Liauw et al20 who showed that most patients who developed SPCs in their series had significant risk factors in their medical history that directly contributed to the development of those particular SPCs—these details are not available in larger population-based registries. Furthermore, the lack of detailed data concerning the radiation dose, volumes treated, and techniques used is also limiting because it is unlikely that the current available database included treatment by techniques such as intensity modulated or 3-dimensional treatment techniques that allow better target localization, conformality, sparing of normal tissue and possibly less treatmentrelated effects. This is due to the need to select patients who were treated a long time back to allow sufficient follow-up to detect possible RTSPC. We conclude that radiation increased late pelvic SPC if given in conjunction with radical surgery by an ageadjusted rate of 374 cases per 100 000, but that the results UROLOGY 74 (4), 2009

must be obtained with caution pending studies with databases that can accurately assess individual risk at baseline.

References 1. Stephenson AJ, Shariat SF, Zelefsky MJ, et al. Salvage radiotherapy for recurrent prostate cancer after radical prostatectomy. JAMA. 2004;291:1325-1332. 2. Bolla M, van Poppel H, Collette L, et al. Postoperative radiotherapy after radical prostatectomy: a randomised controlled trial (EORTC trial 22911). Lancet. 2005;366:572-578. 3. Thompson I, Tangen C, Paradelo J, et al. Adjuvant radiotherapy for pathologically advanced prostate cancer. JAMA. 2006;296: 2329-2335. 4. Alberts D. Second cancers are killing us! Cancer Epidemiol Biomarkers Prev. 2006;15:2019. 5. Hall EJ, Wuu CS. Radiation-induced second cancers: the impact of 3D-CRT and IMRT. Int J Radiat Oncol Biol Phys. 2003;56:83-88. 6. Amemiya K, Shibuya H, Yoshimura R, et al. The risk of radiationinduced cancer in patients with squamous cell carcinoma of the head and neck and its results of treatment. Br J Radiol. 2005;78: 1028-1033. 7. Warren S, Gates O. Multiple primary malignant tumors: a survey of the literature and a statistical study. Am J Cancer. 1932;16:13581414. 8. Cahan WG, Woodward HQ, Higinbotham NL, et al. Sarcoma arising in irradiated bone. Cancer. 1948;1:3-29. 9. Moon K, Stukenborg G, Keim J, et al. Cancer incidence after localized therapy for prostate cancer. Cancer. 2006;107:991-998. 10. Kendal WS, Eapen L, Macrae R, et al. Prostatic irradiation is not associated with any measurable increase in the risk of subsequent rectal cancer. Int J Radiat Oncol Biol Phys. 2006;65:661-668. 11. Chrouser K, Leibovich B, Bergstralh E, et al. Bladder cancer risk following primary and adjuvant external beam radiation for prostate cancer. J Urol. 2005;174:107-110. 12. Pickles T, Phillips N. The risk of second malignancy in men with prostate cancer treated with or without radiation in British Columbia, 1984-2000. Radiother Oncol. 2002;65:145-151. 13. National Cancer Institute. Surveillance, Epidemiology, and End Results (SEER) Program Public-Use Data (1973-2003). National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2006, based on the November 2005 submission. Available at: http://www.seer.cancer.gov. 14. Rothman K, Greenland S. Modern Epidemiology. 2nd ed. Philadelphia: Lippincott-Raven; 1998:260-264. 15. Gray RJ. A class of K-sample tests for comparing the cumulative incidence of a competing risk. Ann Stat. 1988;16:1141-1154. 16. The R Development Core Team. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Version 2.8.0 2008/10/20 ISBN: 3-9000-51-07-0. Available at: http://cran.r-project.org/doc/manuals/refman.pdf. 17. Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc. 1999;94:496-509. 18. Pawlish KS, Scottenfeld D, Severson R, et al. Risk of multiple primary cancers in prostate cancer patients in the Detroit metropolitan area: a retrospective cohort study. Prostate. 1997;33:75-86. 19. Neugat AI, Ahsan H, Robinson E, et al. Bladder carcinoma and other second malignancies after radiotherapy for prostate carcinoma. Cancer. 1997;79:1600-1604. 20. Liauw SL, Sylvester JE, Morris CG, et al. Second malignancies after prostate brachytherapy: incidence of bladder and colorectal cancers in patients with 15 years of potential follow-up. Int J Radiat Oncol Biol Phys. 2006;66:669-673. 21. Brenner DJ, Curtis RE, Hall EJ, et al. Second malignancies in prostate carcinoma patients after radiotherapy compared with surgery. Cancer. 2000;88:398-406.

UROLOGY 74 (4), 2009

22. Movsas B, Hanlon AL, Pinover W, et al. Is there an increased risk of second primaries following prostate irradiation? Int J Radiat Oncol Biol Phys. 1998;41:251-255. 23. Abdel-Wahab M, Reis I, Hamilton K. Second primary cancer after radiation therapy for prostate cancer—a SEER analysis of brachytherapy versus external beam radiation. Int J Radiat Oncol Biol Phys. 2007;72:58-68.

EDITORIAL COMMENT In my view, there is sufficient evidence that external beam radiation therapy (EBRT) in the doses used to treat adenocarcinoma of the prostate leads to a small but definite risk of in-field, and likely other, malignancies.1-3 There are at least 2 implications after the clinician acknowledges this. The first relates to the discussion with men with newly diagnosed clinically localized prostate cancer (PC). A urologist and possibly a radiation oncologist will review the treatment alternatives and in doing so will include the important discussion of side effects— both immediate and late. Typically, the treatment approaches will include a discussion of EBRT. I believe one should mention the risk, albeit small, of a radiation-induced second primary cancer. This same discussion should be invoked for the patient with a biochemical recurrence after total prostatectomy or for that matter for the patient being considered for adjuvant radiation because of a positive margin or other adverse pathology. Should the radiation oncologist point out the risk of a radiation-induced malignancy? In a recent discussion on this point, Anthony Zeitman, a noted radiation oncologist at the Massachusetts General Hospital, stated that the risk of a radiation-induced second primary cancer is stated on their consent form. The second consequence of this discussion is for clinicians who examine patients who had undergone EBRT for the treatment of PC. If they have hematuria, it might easily be dismissed as caused by radiation, that is, radiation cystitis. Although this is more likely than a radiation-induced bladder cancer (BC), we must not assume this as the case. Of course, men aged ⬎ 60 are at risk of BC unrelated to radiation and ideally should be worked up for BC. Just as women with hematuria are often treated for a urinary tract infection because it is more likely, a man who has been treated with radiation and has hematuria may be “assumed” to have bleeding caused by the effects of radiation. A delay in diagnosis of just a few months can adversely affect the prognosis of a high-grade BC. I have updated a previous report of personal experiences of men who developed BC after prior radiation therapy for the treatment of PC.3 The series now has 44 men who have BC after prior radiation for PC. Their mean age is 75 and the diagnosis of BC was made in an average of 6 years after the PC treatment. Thirty-seven had EBRT either as primary treatment for the PC or as adjuvant or salvage after prostatectomy. Seven men had only brachytherapy. Forty-two underwent a total cystoprostatectomy and 26 were pT2-pT4. One had a partial cystectomy and 1 a diversion only. Of interest during the same 15-year interval is that only 3 men had a cystectomy for BC and a history of a total prostatectomy for PC. This is not surprising as PC is not associated with a higher risk of second primary cancers. Mark S. Soloway, M.D., Department of Urology, University of Miami Miller School of Medicine, Miami, Florida 871