Symptomatic local recurrence of prostate carcinoma (SLRPC) after radiation therapy (RT) is associated with morbidity and debilitating symptoms that have a substantial impact on the patient's quality of life. Most reports on the results of RT for localized prostate carcinoma (PC) do not address this endpoint. The objective of this study was to determine the incidence of SLRPC and to identify the risk factors for this endpoint.
The medical charts of 1006 patients who received RT for localized PC at the University of Texas M. D. Anderson Cancer Center between 1987 and 1997 were reviewed. Local symptoms were defined as hematuria, voiding symptoms, urinary obstruction, and pelvic pain. Progressive symptoms accompanied by either confirmatory histology or cystoscopic findings were attributed to PC. Univariate and multivariate analyses using Cox proportional hazards models were applied to identify risk predictors.
Among 964 patients for whom follow-up data were available, 277 patients had prostate-specific antigen (PSA) progression, and 45 patients died of PC during a median follow-up of 9.4 years. In total, 33 patients (3.4%) developed SLRPC. In patients who experienced biochemical progression, the actuarial 5-year incidence of SLRPC was 8.3%. Among the patients who had developed SLRPC, 23 patients (69.7%) died of PC at a median of 25.3 months from the onset of local symptoms. Adverse histologic tumor subtypes (ductal, small cell, and sarcomatoid) were associated significantly with SLRPC (hazard ratio, 8.4; 95% confidence interval, 2.99–23.63). Clinical T classification at diagnosis, Gleason score, and initial PSA level showed a trend toward an increased hazard ratio.
Overall, the long-term risk for clinically evident prostate carcinoma (PC) recurrence after patients receive radiation therapy is 30%, depending on initial tumor stage and grade.1 PC may recur at the primary site and may involve the prostate gland or become disseminated, metastatic disease. Although disease progression to the metastatic stage is the ultimate factor that determines cause-specific survival, local recurrence of PC can affect the clinical course of the disease significantly by becoming the source of subsequent metastasis and because the symptoms associated with local recurrence can have an impact on the patient's quality of life. Symptoms arising from local recurrence of PC are caused by direct invasion of the tumor into adjacent organs, such as the bladder, rectum, or pelvic sidewall. Affected patients may present with hematuria, voiding symptoms, pelvic pain, or incontinence. Ureteral or bladder outlet obstruction are common and predispose these patients to the potentially life-threatening morbidity of renal failure and infection. Preservation of renal function may require persistent tube drainage in the form of nephrostomies, ureteral stents, and catheters that may be needed indefinitely, and additional invasive procedures occasionally will be required. Because some patients with local recurrence of PC without metastasis may survive for > 5 years,2, 3 the endpoint of symptomatic local recurrence deserves attention independent of disease-specific survival (DSS).
Most published reports documenting the outcome of radiation therapy for localized PC have focused on biochemical progression-free and DSS rates.4–7 Consequently, the incidence of symptomatic local recurrence of PC (SLRPC) is not well documented. Moreover, specific risk factors for the development of SLRPC have not been identified. To address these issues, we retrospectively reviewed the records of a cohort of patients who received radiation therapy as the initial therapy for localized PC at the University of Texas M. D. Anderson Cancer Center.
MATERIALS AND METHODS
Between January 1987 and October 1997, a total of 1006 patients who were diagnosed with Stage T1c–T3b PC received definitive radiation therapy at our institution. For all patients, staging was determined by computerized tomography scans of the abdomen and pelvis and by bone scan. Androgen ablation was not used prior to radiation therapy.
Radiation was administered to the isocenter at doses ranging from 64 grays (Gy) in the earlier years to 78 Gy more recently using techniques that corresponded to the prescribed dose: a four-field approach followed later by four fields with a conformal six-field boost. After the completion of radiation therapy, patients were followed at 3-month intervals for the first 2 years and at 6-month intervals thereafter, unless symptoms mandated more frequent visits. Patients were censored from the study at last follow-up visit or at the time of death. Patients were followed for a median of 9.4 years after radiation therapy, and 75% of the patients were followed for > 7 years (range, from 3 months to 15.8 years). Each patient's living status was determined at the time of data collection using the Social Security Death Index update for January, 2004. From the patient's medical records, age, ethnicity, clinical stage of disease, and Gleason score at the time of diagnosis were recorded along with the presence of potentially more aggressive histologic subtypes of PC, including ductal, sarcomatoid, and small cell carcinomas. Because many of our patients were treated before prostate-specific antigen (PSA) testing became widespread, we included baseline PSA levels for all patients who were treated after January, 1990.
Using the American Society for Therapeutic Radiology and Oncology definition of biochemical progression,8 we defined the PSA progression date as the time midway between the first and second of three consecutive PSA elevations. The occurrence of SLRPC was assessed in patients who experienced PSA progression after radiation therapy. The following were defined as local symptoms or signs: pelvic pain, hematuria, severe urinary frequency and urgency (> 10 episodes of voiding daily or urgency incontinence), and obstructive voiding symptoms (e.g., weak or intermittent urinary stream, sensation of incomplete bladder emptying, need to strain to urinate, and obstructive renal failure). Due to their poor specificity, nocturia and isolated mild urinary frequency that occurred within 6 months after radiation therapy were not considered local symptoms. Local symptoms were considered to be associated with locally recurrent PC when all of the following criteria were met: 1) aggravation of symptoms over time; 2) histologic evidence of PC or distortion of the prostatic urethra by irregular ingrowth of tissue, as observed at cystoscopy; and 3) absence of cystoscopic findings characteristic of radiation-induced cystitis. For patients in whom the origin of local symptoms was not clear, the cause of symptoms was classified as undetermined.
Thirty-six of 334 patients who experienced biochemical progression underwent salvage surgery. Surgical procedures included cryosurgery (27 patients), salvage radical prostatectomy (7 patients), cystoprostatectomy (1 patient), and total pelvic exenteration (1 patient). The analysis of local symptoms in these patients referred to the period between radiation therapy and salvage surgery; and, to avoid introducing a confounding bias with regard to the origin of symptoms that occurred after salvage surgery, these patients were censored out of the analysis of local symptoms on the day of salvage surgery. Patients in whom androgen ablation was initiated at the time of biochemical progression were not censored from the symptom analysis. Data on the need for and duration of tube drainage of the obstructed urinary tract and on transurethral channeling procedures also were recorded.
Overall survival duration was calculated from the date radiation therapy was completed to the date of death. The cause of death was attributed to PC when there was evidence of symptomatic, metastatic disease at the time of death. Patients in whom the only sign of disease progression was an elevated PSA level at the time of death were considered to have died of unrelated causes. Patients who died without SLRPC were censored. Worst and best scenario calculations of the proportion of patients with SLRPC were applied to the patients in whom the etiology of local symptoms was unclear. The survival distribution was estimated using the Kaplan–Meier product-limit method. The two-sided log-rank test was used to test the difference between local symptom-free survival for each of the risk factors in univariate analysis. The Cox proportional hazard model was used to determine the association of the risk factor of interest with SLRPC-free survival after adjusting for the role of other risk factors. All P values are two-sided. Statistical analysis was carried out using the S-plus 6.0 software package.
From an initial cohort of 1006 patients, 42 patients were excluded due to incomplete follow-up. Thus, in total, 964 patients were available for the current analysis (Table 1). During the follow-up period, PSA progression occurred in 334 patients; however, complete data on clinical symptoms were available in only 277 patients (82.9%). Among 277 patients, 78 patients experienced local symptoms. On the basis of the criteria described above, local symptoms were attributed to local recurrence of PC (33 patients) or to radiation therapy or other conditions (36 patients). In nine patients who had local symptoms, the precise etiology was unclear. In all patients in whom the origin of local symptoms was attributed to PC, there was persistence and progression of symptoms with time; in 23 patients, local recurrence of PC was confirmed by biopsy; and, in the other 10 patients, cystoscopy demonstrated obstruction of the prostatic urethra by bulky, irregular, friable tumor tissue. In 32 patients with SLRPC, there were no concomitant cystoscopic findings compatible with radiation cystitis; however, in another patient who had a bulky tumor in the prostatic urethra with a Gleason score of 10, there also were typical findings of radiation cystitis. This patient manifested with hematuria and bladder outlet obstruction and thus, was considered to have developed SLRPC, although the hematuria was related in part to radiation therapy. The risk calculations for SLRPC are shown in Table 2. Among the patients who experienced PSA progression, in total, 18 patients developed SLRPC during the first 5 years after radiation therapy. Among the patients who experienced biochemical progression, the actuarial incidence of SLRPC at 5 years (defined as the ratio of the number of events to the number of patients at risk who were followed for 5 years) was 8.3% (95% confidence interval, 4.6–12.0%), and the median period between radiation therapy and the development of SLRPC was 4.5 years (range, from 3.6 months to 12 years). Among 36 patients in whom local symptoms were attributed to another etiology, in 27 patients, the symptoms were related to the radiation treatment; in the other 9 patients, the etiology of symptoms was a combination of radiation therapy-related and benign prostate hyperplasia or bladder hyperreflexia after a cerebrovascular accident.
Table 1. Clinical Characteristics of the Study Group
No. of patients (%)
PSA progression group
SLRPC: symptomatic local recurrence of prostate carcinoma; PSA: prostate-specific antigen; XRT: radiation therapy.
No. of patients
Age at study entry (yrs)
Clinical T stage
Table 2. Risk Calculations for Acquiring Symptomatic Local Recurrence of Prostate Carcinoma After Radiation Therapy
A best-case scenario assumes that none of the nine patients with undetermined etiology had symptoms related to recurrent prostate carcinoma. A worst-case scenario assumes that all nine patients with undetermined etiology of symptoms had a symptomatic local recurrence of prostate carcinoma.
Patients with PSA progression
The most frequent presenting symptoms in patients who developed SLRPC were those of bladder outlet obstruction. Obstructive symptoms were reported by 26 patients with SLRPC (78.8%), including 9 patients (27.2%) who experienced urinary retention. Obstructive renal failure, hydronephrosis, or both occurred in 6 patients (18.1%), and a total of 12 patients (36.3%) required insertion of a urethral catheter (n = 9 patients), suprapubic cystostomy (n = 1 patient), nephrostomy (n = 10 patients), or some combination of these to alleviate urinary tract obstruction. The average duration of indwelling tubes in these patients was 5.2 months (range, from 1 week to 48 months). Transurethral resection of prostatic tumor that was obstructing the urethra was required 17 times in 13 patients (39.4%) who developed SLRPC. Macroscopic recurrent hematuria occurred in 14 patients (42.4%) with SLRPC, including 2 patients with clot retention and another patient with urethral bleeding. This usually accompanied other local symptoms but was the single problem in three patients. Pelvic pain was reported by three patients. At the time of onset of SLRPC, 26 patients (78.8%) had been treated with androgen ablation, and 21 of those patients had progressed to androgen-independent disease, as defined by a consistent PSA elevation despite therapy. Systemic chemotherapy was used in 11 patients who had androgen-independent PC.
At the time of onset of SLRPC, 16 patients (48.5%) had progressed to metastatic disease. Of the 33 patients who developed SLRPC, 26 patients (78.7%) died during the follow-up period, including 23 patients (69.7%) who died of disease at a median of 25.3 months after the onset of local symptoms (range, 8.8–72.2 months). Of the 10 patients with SLRPC who were alive at the end of this study or who died of an unrelated cause, only 1 patient had androgen-independent PC, and metastatic disease was found in another patient. Among 36 patients who experienced biochemical progression and who underwent salvage surgical procedures, SLRPC was present in 4 of 9 patients who underwent open salvage surgery and in 0 of 27 patients who underwent salvage cryosurgery.
In a univariate analysis of potential risk factors for development of SLRPC, only the adverse histologic subtypes of PC (ductal carcinoma, sarcomatoid tumor, and small cell carcinoma) were significant. (Tables 3, 4; Fig. 1). However, there was a trend toward a shorter SLRPC-free survival duration in patients with higher clinical T classification (Fig. 2), higher Gleason scores (Fig. 3), and higher pretreatment PSA levels (Fig. 4).
Table 3. Univariate Analysis of Predictors of Symptomatic Local Recurrence of Prostate Carcinoma
Pretreatment baseline PSA level and patient age were cut at the median values. The analysis was based on a total population of 276 patients and 33 occurrences of symptomatic local recurrence of prostate carcinoma.
Among 277 patients who experienced PSA progression after radiation therapy, 98 patients (35.4%) died during the follow-up period, including 45 patients (16.2%) who died of PC. Forty-three patients died of unrelated disease. In another group of 10 patients with limited recent follow-up, the cause of death could not be determined. The overall survival rates at 5 years and 10 years after the completion of radiation therapy were 94% and 69%, respectively; and the DSS rates at 5 years and 10 years after treatment were 97% and 84%, respectively.
Our study demonstrated a low risk for developing SLRPC after radiation therapy for patients with localized PC. Even among patients who had biochemical progression after treatment and who were expected to develop subsequent clinical disease progression, the overall likelihood that they would be free of SLRPC was 85–88%. Furthermore, SLRPC may occur many years after radiation therapy; and, in our cohort, the likelihood that patients would be free of SLRPC at 5 years and 10 years after treatment was 93% and 87%, respectively. Because most patients who develop disease progression will not have SLRPC, it seems that radiation therapy provides effective clinical local control in the majority of patients.
Various series have reported diverse rates of local failure after radiation therapy, ranging between 6% and up to 61% of patients.5, 9–11 This apparently wide distribution may be related to whether the patients had biochemical failure at the time of biopsy, the dose of radiation administered, and the timing of biopsy after radiation therapy. Our results showed a 12–15% rate of SLRPC among patients who had experienced biochemical progression. This relatively lower rate reflects the fact that, in our study, prostate biopsy was performed to confirm the etiology of symptoms and not necessarily due to the rising PSA level. Furthermore, to cause symptoms, a locally recurrent tumor possibly may need to acquire sufficient size and aggressiveness that would facilitate invasion to neighboring organs. Therefore, SLRPC is a phenomenon that occurs in a subset of the patients who experience local failure after radiation therapy.
SLRPC was a late event in the course of the disease: 78% of patients had androgen-independent disease, and 48% of patients had distant metastasis. Whether the use of androgen-deprivation therapy delayed the onset of SLRPC in this cohort is unclear. Despite the occurrence of SLRPC late in the course of the disease, the median survival duration with SLRPC was 2 years. Although it was a rare event, SLRPC had a considerably detrimental impact on the patients' quality of life and was associated with significant morbidity, with persistent and debilitating symptoms, and, in some patients, with the necessity of further interventions. Treatment modalities that achieved effective and durable palliation for these patients still are lacking. Early identification of predictors that would define the population at risk for developing SLRPC could provide the opportunity to study new alternative or combination therapies that may prevent the occurrence of SLRPC or that may offer more effective palliation. Despite our relatively large cohort, due to the small number of SLRPC events, our analysis could not identify significant predictors for SLRPC other than aggressive histologic subtypes. Nevertheless, a trend toward an increased risk for SLRPC was observed in patients who presented with high Gleason scores, high T classification, and high pretreatment PSA levels (Gleason scores of 8–10 vs. scores of 2–5; T3 tumor status vs. T1–T2 tumor status; and PSA > 11.4 ng/mL vs. PSA ≤ 11.4 ng/mL). In addition, the presence of more aggressive subtypes of PC (i.e., ductal carcinoma, small cell carcinoma, and sarcomatoid) was statistically more frequent among patients who subsequently developed SLRPC than among patients with PSA progression who did not have SLRPC.
In our series, although the radiation dose did not affect the risk for SLRPC, current evidence suggests that higher doses > 70 Gy are associated with a decreased risk for disease progression.4, 6 A multiinstitutional study addressing the occurrence of SLRPC could generate enough statistical power to establish or rule out these predictors.
Others have reported that the baseline Gleason score and clinical tumor stage are reliable predictors of outcome in patients who are treated with radiation therapy.12 In addition to Gleason score and tumor stage, we reported that a pretreatment PSA level < 10 ng/mL and a PSA nadir < 0.99 ng/mL were associated with a decreased hazard ratio for biochemical progression.13 These clinical and pathologic features have been accepted widely as predictors of tumor biology and outcome. In our series, there was a nonsignificant trend toward an increased hazard risk for the development of SLRPC in patients who presented with Gleason scores of 8–10 and locally advanced (T3) tumors.
Most published series on radiation therapy for patients with localized PC have focused on more general endpoints to assess outcome, such as biochemical progression, overall survival, and DSS rates. Most studies addressed to the occurrence of local symptoms as part of the assessment of treatment-related toxicity, with no attempt made to distinguish between side effects of radiation therapy and SLRPC. Zietman et al.14 reported on a cohort of 205 men who were treated with radiation therapy. In that group, the 10-year DSS was 89%, and the local recurrence rate, as determined by digital rectal examination, was 18%. However, the risk for SLRPC was not reported. In a series from Europe, the 10-year DSS ranged between 38% and 76%, depending on tumor stage and grade, and 25% of the patients had local recurrence 10 years after treatment; however, the rate of symptomatic local recurrence was not reported.15 Treatment-induced urinary side effects occurred in 5.5% of the patients in that series. Other studies addressing the toxicity of radiation therapy have not attempted to distinguish between treatment-induced and disease-related local symptoms.16, 17 One group reported that local symptoms that developed after radiation therapy were tolerated well and did not have a significant impact on the patients' quality of life.4 Again, there was no distinction between disease-related and treatment-related symptoms.
A significant potential source of bias in our report is the confounding between radiation-induced symptoms and symptoms related to local disease progression, as in patients with SLRPC, especially when the two symptom sources coexist. We attempted to corroborate the accurate origin of symptoms based on rigid definitions and to support our judgment with objective evidence of local disease recurrence either in the form of histology analysis or by cystoscopy. Radiation-induced symptoms include hematuria secondary to cystitis, urethral stricture, and irritable bladder symptoms. Diagnosis of the first two problems is straightforward and can be established by cystoscopy. When the only prevailing symptoms were related to irritable bladder, we did not consider those as symptoms that originated from local disease recurrence.
In the current study, we focused on local symptoms that clearly were related to local recurrence of PC. To date, this clinical endpoint has not been reported thoroughly. Our results demonstrated that, although SLRPC is an uncommon event after patients receive radiation therapy, it is clinically important. Further studies will be necessary to define better the specific risk factors associated with SLRPC.