Clinical and dosimetric factors associated with acute rectal toxicity in patients treated with 131Cs brachytherapy for prostate cancer

Results One hundred six patients were treated with 131 Cs from September 2006 to May 2008. Thirty-eight percent of patients met our criteria for patient-appreciated acute bowel symptoms. On multivariate analysis, the volume of rectum receiving 50% of the prescribed dose (R- V 50 ; 4.1 vs. 2.6 cc, p = 0.01), R- V 75 (1.3 vs. 0.62 cc, p = 0.01), the percentage of the prescribed dose received by 1 cc of the rectum (R-D-1cc; 75% vs. 64%, p = 0.02), and R-D-2cc (63% vs. 54%, p = 0.003) were found to be factors associated with a greater risk of severe acute bowel toxicity. At 3-month followup, 28% of patients had persistent acute bowel toxicity. On multivariate analysis, no factors were identified that correlated with persistent acute bowel toxicity. Conclusions This study identifies R- V 50 , R- V 75 , R-D-1cc, and R-D-2cc as factors associated with patient-appreciated acute rectal toxicity. We are performing dosimetric analysis to determine the optimal distance for the posterior needles from the prostate–rectal interface to decrease rectal dose while still maintaining adequate coverage of prostate. Keywords 131 Cs Quality of life Prostate brachytherapy Rectal toxicity Introduction Prostate brachytherapy (PB) has emerged as a viable and an efficacious treatment modality for patients with prostate adenocarcinoma, either as monotherapy for low-risk disease or in conjunction with external beam radiation therapy (EBRT) and/or hormonal therapy for higher risk disease. In comparison with other treatment approaches for prostate cancer, notably radical prostatectomy and definitive EBRT, PB carries the advantage of a minimally invasive one-time procedure, allowing for a rapid return to normal activity. PB does, however, come with a constellation of known toxicities. Most men experience acute irritative and obstructive voiding systems. Acute and late bowel toxicities occur but are less common. The most common rectal symptom is mild self-limiting proctitis, and incidence rates of Grade 2 rectal toxicity have been reported to range from 4% to 13% (1–7) . Incidence rates of <1% for more severe toxicities, such as ulceration and/or fistula, have been reported (1, 2, 8–11) . Recently, 131 Cs (IsoRay Medical, Richland, WA) has been introduced as a new isotope for use in PB. 131 Cs, with its average energy of 30.4 keV and half-life of 9.7 days, carries the potential advantage of limiting the duration of acute symptoms (12, 13) . Because most of the data with regard to bowel toxicity and PB have been collected in the 125 I and 103 Pd eras, little is known about the bowel toxicity profile of 131 Cs. In this study, we sought to identify clinical and dosimetric factors associated with acute bowel toxicity in patients treated with 131 Cs PB for prostate adenocarcinoma. Bowel toxicity was assessed using the expanded prostate cancer index composite (EPIC), a validated prostate cancer–specific patient quality of life (QOL) self-assessment tool (14) . Methods and materials One hundred six consecutive patients with clinically localized prostate cancer underwent PB. Patient characteristics are shown in Table 1 . Thirty-three patients with intermediate- or high-risk disease defined by National Comprehensive Cancer Network (NCCN) criteria were treated with EBRT before the PB procedure (15) . All patients treated with EBRT received intensity-modulated radiation therapy (IMRT). Patients with high-risk disease treated with IMRT received 45 Gy to the pelvic lymph nodes and prostate and seminal vesicles, whereas those with intermediate-risk disease treated with IMRT received 45 Gy to the prostate and seminal vesicles with a 1-cm margin alone. PB was performed using an afterloading technique with a Mick applicator. Needles were placed throughout the prostate by a single urologist (RB). Peripheral needles were placed first, approximately 0.8–1.0 cm apart, with two to eight (depending on prostate volume) central needles then placed to ensure adequate dosing of the central portion of the gland. The prescribed dose of 131 Cs was 85 Gy for those who received prior EBRT and 115 Gy for those treated with PB alone. Implant characteristics are shown in Table 2 . Thirteen to 27 (median, 18) needles were placed during the procedure through which the seeds were delivered. Fifty to 145 (median, 85) 131 Cs seeds were implanted. Prostate volumes obtained by ultrasound in the operating room ranged from 20.2 to 85.2 cc (median, 39.3). Dosimetric goals included a prostate D 90 = 100%, V 100 ≥ 90%, V 150 ≤ 50%, and V 200 ≤ 15%. After the implantation, patients underwent a CT scan within 24 h. Using 3-mm cuts, the prostate, as well as the anterior rectal wall and urethra, was outlined on every CT slice to obtain the volumes of these structures. Isodose curves were generated for each individual slice and dose volume histograms were calculated from these. Patients were instructed to use loperamide hydrochloride or Anusol HC after therapy as needed. Patients were asked to complete EPIC surveys at baseline and at 2 weeks, 4 weeks, and 3 months after seed placement. For patients who received EBRT before PB, the baseline survey was completed before the start of EBRT, whereas for all others, the baseline study was done before PB. EPIC surveys are designed to evaluate patient function and bother before and after treatment of prostate cancer with assessments of urinary, bowel, sexual, and hormonal function. With regard to bowel function, the EPIC survey includes seven questions that assess a patient's frequency of bowel movements, rectal urgency, rectal incontinence, loose stools, hematochezia, pain with defecation, and crampy abdominal pain over the prior 4 weeks, respectively, on a scale from 1 (never) to 5 (always). Additionally, patients record the impact each of the seven bowel symptoms has on their lives on a scale from 0 (not a problem) to 4 (big problem) and record their assessment of their overall bowel health on the same scale. An established scoring system converts patient-recorded answers to a value between 0 and 100. For example, questions assessing the frequency of specific bowel symptoms on the scale of 1 (never), 2 (rarely), 3 (about half the time), 4 (usually), and 5 (always) converted to 100, 75, 50, 25, and 0, respectively. A final bowel summary score was calculated by taking the average of the scores from each question. Based on the EPIC surveys, we identified two cohorts of patients. The first cohort consisted of patients who experienced acute rectal/bowel toxicity as defined in this study by a difference from baseline EPIC bowel summary score at 2- to 4-week followup that exceeded 2 standard deviations (SDs) of the baseline EPIC bowel summary values of all patients. The second cohort consisted of patients whose acute bowel/rectal toxicity persisted 3 months after seed placement as defined in this study by a difference from baseline EPIC bowel summary score at 3-month followup that exceeded half an SD of the baseline value (16, 17) . We then analyzed for clinical and dosimetric factors that were associated with either situation. The factors analyzed included age, use of hormones, use of preimplant EBRT, prostate volume, number of seeds used, number of needles used, the dose received by 90% of the prostate ( D 90 ), the volume of prostate receiving 100% of the prescribed dose ( V 100 ), V 150 , V 200 , the volume (cc) of rectum receiving 150% of the prescribed dose (R- V 150 ), R- V 100 , R- V 75 , R- V 50 , R- V 25 , the percentage of the prescribed dose received by 1 cc of the rectum (R-D-1cc), and R-D-2cc. Statistical analysis was performed using SPSS version 16.0 software (SPSS Inc., Chicago, IL). Student's t tests and χ 2 tests were performed for univariate analyses. Multivariate analysis was performed via logistic regression using factors found to be significant on univariate analyses. All analyses were two-sided, and significance was set to the level of p < 0.05. Results Rectal dosimetry and EPIC bowel summary scores at various time points are shown in Tables 3 and 4 . The SD of all patient EPIC bowel summary scores at baseline was 12.9, and this value was used to define the cohort of patients who experienced acute rectal symptoms (≥2 SD difference or ≥25.8-point drop in bowel summary score from baseline at 2–4 weeks) or persistent acute rectal symptoms (≥0.5 SD difference or ≥6.5-point drop in bowel summary score from baseline at 3 months). Of the 106 patients in our study, 102 patients had completed EPIC surveys at both baseline and at 2–4 weeks postimplantation. Thirty-eight patients (37%) experienced a change in EPIC bowel summary scores from baseline to 2–4 weeks that met our criteria for patient-appreciated acute rectal toxicity. Clinical and dosimetric variables compared between the group of patient who did and did not experience acute rectal toxicity are displayed in Table 5 . On univariate analysis, statistically significant differences were noted between those who did and did not experience acute rectal toxicity, respectively, with regard to number of needles (19 vs. 17, p = 0.005), number of seeds (97 vs. 86, p = 0.03), preimplant EBRT (18% vs. 39%, p = 0.05), R- V 100 (0.37 vs. 0.16 cc, p = 0.02), R- V 75 (1.3 vs. 0.62 cc, p = 0.001), R- V 50 (4.1 vs. 2.6 cc, p = 0.003), R- V 25 (11.3 vs. 9.1, p = 0.02), R-D-1cc (75% vs. 64%, p = 0.004), and R-D-2cc (63% vs. 54%, p = 0.003). On multivariate analysis, R- V 75 ( p = 0.01), R- V 50 ( p = 0.01), R-D-1cc ( p = 0.02), and R-D-2cc ( p = 0.02) remained significant. At the time of this analysis, 87 of the 106 patients had EPIC bowel summary score data available at 3 months. Of these 87 patients, 24 (28%) had EPIC bowel summary scores at 3 months that met our criteria for persistent acute rectal toxicity. Clinical and dosimetric variables compared between those who did and did not have persistent acute rectal toxicity are shown in Table 6 . Statistically significant differences between the groups were noted on univariate analysis with respect to the number of needles used (19 vs. 18, p = 0.05), number of seeds (98 vs. 86, p = 0.02), prostate volume measured in the operating room by ultrasound (47.7 vs. 39 cc, p = 0.02), prostate volume as measured on CT scan (57.6 vs. 46.5 cc, p = 0.01), and R- V 25 (11.6 vs. 9.1, p = 0.01). On multivariate analysis, no factor was found to be significant, though there seemed to a trend toward the significance of R- V 25 ( p = 0.08). A separate analysis was performed among patients treated with PB without EBRT. Of the 106 patients, 70 were treated without EBRT. In terms of acute rectal toxicity at 2–4 weeks, 31 patients (44%) met our criteria for acute rectal toxicity. Factors found to be significant for acute rectal toxicity were similar to those seen for the entire group. These factors included number of needles (19 vs. 18, p = 0.03), R- V 75 (1.4 vs. 0.73, p = 0.02), R- V 50 (4.2 vs. 2.7, p = 0.02), R- V 25 (11.4 vs. 9.0, p = 0.03), R-D-1cc (76% vs. 64%, p = 0.02), and R-D-2cc (64% vs. 54%, p = 0.02). On multivariate analysis, R- V 75 , R- V 50 , R-D-1cc, and R-D-2cc remained statistically significant ( p < 0.05) as had been also seen in the total group. With regard to persistent acute rectal toxicity at 3 months, 56 of 70 patients had EPIC bowel summary data available at 3 months. Sixteen patients (29%) experienced persistent acute toxicity. Significant factors were similar to those found for the total group. These included number of needles (21 vs. 18, p = 0.02), number of seeds (112 vs. 94, p = 0.004), prostate height (35.3 vs. 31.1 mm) and width (54.0 vs. 48.9 mm, p = 0.02) measured in the operating room by ultrasound, prostate volume as measured on CT (68.0 vs. 49.5 cc, p = 0.001), and R- V 25 (0.04). Similar to our findings for the group as a whole, no factors were found to be significant for persistent acute rectal symptoms on multivariate analysis. Discussion Compared with the other commonly used isotopes for PB, 131 Cs sources emit a similar mean photon energy to that of 125 I (30.4 vs. 28.5 keV, respectively) but higher than that of 103 Pd (21 keV). The sharper difference emerges with respect to half-life, with 131 Cs having the lowest (9.7 days) compared with 125 I (60 days) or 103 Pd (17 days) (12) . This shorter half-life, along with equivalent to slightly higher mean energy, carries the theoretical advantage of delivering the same therapeutic dose to the prostate while limiting the duration of acute toxicities. For most men treated with 131 Cs, acute urinary and rectal/bowel toxicities should reach their peak 2–4 weeks after therapy and subsequently subside by 3 months (13, 18, 19) . Rectal toxicity related to PB is generally mild. In the immediate postimplant period, transient rectal discomfort and bleeding are associated with the trauma of the brachytherapy procedure, notably the insertion and removal of the ultrasound probe. In the weeks after the implant, some experience acute changes in bowel habits, which present as diarrhea, constipation, tenesmus, or pressure (20) . The incidence of acute rectal toxicity after PB with 125 I or 103 Pd has been reported to range from 5% to 40%, with the majority experiencing Grade 1 toxicity based on Radiation Therapy Oncology Group (RTOG) criteria (7, 21, 22) . Sanda et al . performed a detailed analysis of QOL among prostate cancer survivors treated with radical prostatectomy, definitive EBRT, or PB based on EPIC survey scores. They reported that at 2 months after PB ( 131 Cs not used), 19% experienced rectal urgency, 17% experienced more frequent bowel movements, 6% experienced incontinence, 1% experienced bloody stools, 10% experienced rectal pain, and 12% noted some bowel issue (16) . In our cohort, 37% of patients experienced acute rectal symptoms at 2–4 weeks after PB implantation and 28% had persistent acute symptoms at 3 months. Our rate of acute rectal toxicity is similar to that reported by Shah and Ennis (22) and Zelefsky et al . (21) who reported acute rectal toxicity rates of 38% and 39%, respectively. The mechanism of acute rectal injury in PB has not been clearly elucidated, but mucositis likely plays a key role. Acute rectal symptoms can be quite distressing for patients and can significantly impair their QOL. More importantly, the presence of acute symptoms may portend a higher risk of late toxicities. Late rectal toxicities include proctitis, rectal ulceration, fistula formation, and incontinence. Radiation proctitis is the most common of these late effects occurring in 4–12% of patients (1–7, 20) . Patients with proctitis typically experience painless self-limiting hemorrhoidal-like bleeding. The mechanism of radiation proctitis is thought to involve edema and fibrosis of the arterioles in the luminal crypts of colonic mucosa. With time and increasing fibrosis, the mucosal lining becomes friable and bleeding can ensue (20) . In this study, we assessed acute rectal toxicity associated with 131 Cs using EPIC surveys. It is known that patient self-assessments of QOL are more accurate and sensitive than physician-recorded measurements of patient QOL, and the EPIC survey has been validated as a robust self-assessment tool for patients undergoing prostate cancer treatment (14, 23) . The EPIC survey addresses patient bowel function by assessing the frequency of bowel movements, rectal urgency, rectal incontinence, loose stools, hematochezia, pain with defecation, and crampy abdominal pain over the prior 4 weeks, respectively, and determines the impact each of the seven bowel symptoms, as well the impact of general bowel function, has on a patient's life. We attempted to correlate clinical and dosimetric factors associated with two EPIC-derived measures of acute rectal toxicity using EPIC bowel summary scores. Determination of clinically relevant changes in QOL scores is difficult; however, the use of a difference from baseline to followup score exceeding half an SD of the baseline value as marker of “minimum important change” seems to be a reasonable approximation (16, 17) . We used this 0.5 SD difference as a marker of persistent acute rectal symptoms at 3 months. To define our threshold for patient-appreciated acute rectal symptoms at 2–4 weeks postimplantation, we used a larger change, that is, a 2 SD difference. We felt that a 2 SD change would be clinically meaningful and would be a marker of severity of acute symptoms. On multivariate analysis, we found that R- V 50 , R- V 75 , R-D-1cc, and R-D-2cc were found to be significant factors for patient-appreciated acute rectal toxicity. In terms of persistent acute rectal toxicity, no clinical or dosimetric factors seemed to be significant on multivariate analysis. R- V 25 was closest ( p = 0.08) but did not meet statistical significance. Although a number of studies have identified dosimetric variables associated with late rectal toxicities, there is a paucity of data regarding dosimetric correlates with acute rectal toxicity. Much of the data regarding late toxicities have identified high maximal rectal doses to limited volumes of rectal mucosa playing a major role in severe late toxicities. In this study, R- V 50 , R- V 75 , R-D-1cc, and R-D-2cc were found to be predictors of patient-appreciated acute rectal toxicity. Interestingly, a factor that was found to be significant on univariate analysis of factors associated with persistent acute rectal toxicity and the one most nearing statistical significance on multivariate analysis was R- V 25 . With more patients and added statistical power, if R- V 25 emerges in the future as a significant predictor of persistent acute rectal toxicity, then this might suggest a possible differential mechanism of injury where lower radiation doses distributed over a larger volume of rectal mucosa may increase the risk of acute toxicity, possibly by inducing mucositis, whereas severe late toxicities, such as ulceration, may occur as a result of high doses of radiation delivered to relatively smaller volumes. Though being treated with EBRT was not found to be a significant factor on multivariate analysis for either acute or persistent acute rectal symptoms, we performed an analysis examining only patients treated with PB without EBRT to remove EBRT as a factor. In this analysis, the same factors were found on multivariate analysis to be significant for acute rectal toxicity as had been seen for the entire group, and no factors were noted to be significant on multivariate analysis for persistent acute rectal symptoms, the same finding seen for the group as a whole. We did not perform a separate analysis of rectal toxicity in patients treated with both EBRT and PB because there were only 36 such patients in the study. Based on our finding of the R- V 75 , R- V 50 , R-D-1cc, and R-D-2cc as possible predictive factors in determining those patients at highest risk of having severe acute toxicity after 131 Cs PB, there might be measures to reduce this risk. Waterman and Dicker indicated that the major contributor of dose received by the anterior rectal wall comes from the most posteriorly implanted seeds. In their study of 9 patients treated with either 125 I or 103 Pd in which patients had CT scans performed on Day 0 and then rescanned approximately 1, 3, 7, and 15 weeks later, the authors found that rectal dose increases with time after implantation as a result of resolution of PB-induced prostate edema. They found a direct relationship between an increase in rectal dose and a decrease in the distance between the most posterior implanted seeds and the anterior rectal surface with time after implantation (24) . Therefore, determination of the optimal distance of the posterior needles from the prostate–rectal interface that can provide adequate coverage of the prostate while minimizing the R- V 25 , R- V 50 , and R- V 75 would provide valuable information with regard to potentially limiting rectal toxicities. Ongoing analysis of this optimal distance is currently under way at our institution. In addition, with longer term followup, we hope to determine whether patients in our cohort who were identified to have acute rectal toxicity are at increased risk of experiencing late rectal toxicity and also whether clinical and/or dosimetric variables can be identified that correlate with late rectal toxicity when 131 Cs is used. Certainly, there are a number of limitations of this study. First, the sample size was relatively small (73 patients treated with PB alone and 33 patients treated with EBRT + PB). This small sample size is one possible reason why EBRT did not emerge as a factor associated with an increased rate of acute bowel/rectal toxicity, which is contrary to evidence from other studies of rectal toxicity in patients treated with both PB and EBRT (7, 8, 23) . With a larger sample size, other clinical or dosimetric factors, including having been treated with pre-PB EBRT, may have proved to be significant. Second, we did not quantify or record in our analysis the use of aftercare interventions for rectal toxicity, such as antidiarrheals or rectal suppositories, that may have played a factor in patients' assessment of their bowel symptoms. Third, comorbid conditions such as diabetes, heart disease, history of hemorrhoids, or prior rectal/anal surgeries were not included in the analysis. Though the rate of hemorrhoids or prior rectal/anal surgeries in our patient population was rare, inclusion of these factors and other comorbid conditions may have proven to be significant. Conclusion With a half-life of 9.7 days and average energy of 30.4 keV, the use of 131 Cs in PB carries the potential of limiting the duration of acute symptoms. This study identifies R- V 75 , R- V 50 , R-D-1cc, and R-D-2cc as predictive factors in identifying patients at highest risk of developing significant acute bowel symptoms after 131 Cs PB. Our institution is currently performing dosimetric analysis to reduce dose to the rectum while maintaining adequate coverage of the prostate by determining the optimal distance between posterior needles and the prostate–rectal interface. References [1] G.S. Merrick W.M. Butler K.E. Wallner Long-term rectal function after permanent prostate brachytherapy Cancer J 13 2007 95 104 [2] D.Y. Gelblum L. Potters Rectal complications associated with transperineal interstitial brachytherapy for prostate cancer Int J Radiat Oncol Biol Phys 48 2000 119 124 [3] K.M. Snyder R.G. Stock S.M. Hong Defining the risk of developing grade 2 proctitis following 125I prostate brachytherapy using a rectal dose-volume histogram analysis Int J Radiat Oncol Biol Phys 50 2001 335 341 [4] S.K. Kang R.H. Chou R.K. Dodge Gastrointestinal toxicity of transperineal interstitial prostate brachytherapy Int J Radiat Oncol Biol Phys 53 2002 99 103 [5] F.M. Waterman A.P. Dicker Probability of late rectal morbidity in 125I prostate brachytherapy Int J Radiat Oncol Biol Phys 55 2003 342 353 [6] K. Wallner J. Roy L. Harrison Dosimetry guidelines to minimize urethral and rectal morbidity following transperineal I-125 prostate brachytherapy Int J Radiat Oncol Biol Phys 41 1998 263 265 [7] R.N. Lesperance R.J. Kjorstadt J.B. Halligan Colorectal complications of external beam radiation versus brachytherapy for prostate cancer Am J Surg 195 2008 616 620 [8] G.S. Merrick W.M. Butler K.E. Wallner Late rectal function after prostate brachytherapy Int J Radiat Oncol Biol Phys 57 2003 42 48 [9] A. Howard K. Wallner B. Han Clinical course and dosimetry of rectal fistulas after prostate brachytherapy J Brachyther Int 17 2001 37 42 [10] D. Theodorescu J. Gillenwater P. Koutrovelis Prostatourethral-rectal fistula after prostate brachytherapy: Incidence and risk factors Cancer 89 2000 2085 2091 [11] A. Tran K.E. Wallner G.S. Merrick Rectal fistulas after prostate brachytherapy Int J Radiat Oncol Biol Phys 62 2005 150 154 [12] Z.C. Chen J. Deng K. Roberts On the need to compensate for edema-induced dose reductions in preplanned 131Cs prostate brachytherapy Int J Radiat Oncol Biol Phys 70 2008 303 310 [13] R.P. Smith S. Beriwal K. Komanduri Planning based on postneedle volume with early dosimetric assessment is beneficial for Cesium-131 permanent prostate seed implantation Brachytherapy 7 2008 237 241 [14] J.T. Wei R.L. Dunn M.S. Litwin Development and validation of the expanded prostate cancer index composite (EPIC) for comprehensive assessment of health related quality of life in men with prostate cancer Urology 56 2000 899 905 [15] The NCCN Clinical Practice Guidelines in Oncology™ Prostate Cancer (Version 2.2.2009). National Comprehensive Cancer Network, Inc. Available at: NCCN.org . Accessed August 22, 2009. [16] M.G. Sanda R.L. Dunn J. Michalski Quality of life and satisfaction with outcome among prostate-cancer survivors N Engl J Med 358 2008 1250 1261 [17] G.R. Norman J.A. Sloan K.W. Wyrwich Interpretation of changes in health-related quality of life: The remarkable universality of half a standard deviation Med Care 41 2003 582 592 [18] Prestidge B, Bice W. Cesium-131—The isotope of choice in permanent prostate brachytherapy [Abstract]. Presented at American Brachytherapy Society. Chicago, IL; 2007;6:80. [19] Moran B, Braccioforte M. Cesium 131—Prostate brachytherapy: An early experience [Abstract]. Presented at American Brachytherapy Society. Chicago, IL; 2007;6:80. [20] L. Potters Rectal complications following permanent seed implants A.P. Dicker G. Merrick L. Gomella R.K. Valicenti F. Waterman Basic and advanced techniques in prostate brachytherapy 2005 Taylor & Francis, Inc Philadelphia, PA 419 424 [21] M.J. Zelefsky T. Hollister A. Raben Five-year biochemical outcome and toxicity with transperineal CT-planned permanent I-125 prostate implantation for patients with localized prostate cancer Int J Radiat Oncol Biol Phys 47 2000 1261 1266 [22] J.N. Shah R.D. Ennis Permanent prostate brachytherapy: Use of comprehensive toxicity scoring system and identification of rectal dosimetric toxicity predictors Int J Radiat Oncol Biol Phys 64 2006 817 824 [23] G.S. Merrick W.M. Butler A.T. Dorsey Rectal function following prostate brachytherapy Int J Radiat Oncol Biol Phys 48 2000 667 674 [24] F.M. Waterman A.P. Dicker Effect of post-implant edema on the rectal dose in prostate brachytherapy Int J Radiat Oncol Biol Phys 45 1999 571 576