Radiotherapy and Oncology xxx (2013) xxx–xxx

Contents lists available at ScienceDirect

Radiotherapy and Oncology

journal homepage: www.thegreenjournal .com

Original article

Long term rectal function after high-dose prostatecancer radiotherapy:Results from a prospective cohort study q

0167-8140/$ - see front matter � 2013 Elsevier Ireland Ltd. All rights reserved.http://dx.doi.org/10.1016/j.radonc.2013.09.028

q Results partially presented at: ESTRO 31, May 9–13, 2012, Barcelona, Spain (oralcommunication). ASTRO’s 54th Annual Meeting, October 28–31, 2012, Boston, USA(oral communication).⇑ Corresponding author. Address: Prostate Cancer Program, Fondazione IRCCS

Istituto Nazionale dei Tumori, Via Venezian 1, 20133 Milan, Italy.E-mail address: tiziana.rancati@istitutotumori.mi.it (T. Rancati).

1 Present address: Radiotherapy, Humanitas – Gavazzeni, Bergamo, Italy.2 Present address: Radiotherapy, A.O. Ordine Mauriziano, Torino.

Please cite this article in press as: Fellin G et al. Long term rectal function after high-dose prostatecancer radiotherapy: Results from a prospectivestudy. Radiother Oncol (2013), http://dx.doi.org/10.1016/j.radonc.2013.09.028

Gianni Fellin a, Tiziana Rancati b,⇑, Claudio Fiorino c, Vittorio Vavassori d,1, Paolo Antognoni d,Michela Baccolini e, Carla Bianchi f, Emanuela Cagna g, Valeria Casanova Borca h, Giuseppe Girelli i,Bruno Iacopino j, Giuseppe Maliverni k,2, Flora A. Mauro l, Loris Menegotti m, Angelo F. Monti n,Fabrizio Romani o, Michele Stasi p, Riccardo Valdagni b,q

a Department of Radiotherapy, Ospedale Santa Chiara, Trento; b Prostate Cancer Program, Ospedale San Raffaele, Milan; c Department of Medical Physics, Milan; d Department ofRadiotherapy, Ospedale di Circolo, Varese; e Department of Medical Physics, Ospedale Villa Maria Cecilia, Lugo di Romagna; f Department of Medical Physics, Ospedale di Circolo,Varese; g Department of Radiotherapy, Ospedale Sant’Anna, Como; h Department of Medical Physics, Ospedale ASL 9, Ivrea; i Department of Radiotherapy, Ospedale ASL 9, Ivrea;j Department of Radiotherapy, Policlinico Sant’Orsola Malpighi, Bologna; k Department of Radiotherapy, IRCC Candiolo; l Department of Radiotherapy, Ospedale Villa Maria Cecilia,Lugo di Romagna; m Department of Medical Physics, Ospedale Santa Chiara, Trento; n Department of Medical Physics, Ospedale Sant’Anna, Como; o Department of Medical Physics,Policlinico Sant’Orsola Malpighi Bologna; p Department of Medical Physics, IRCC Candiolo; and q Department of Radiation Oncology 1, Fondazione IRCCS Istituto Nazionale dei Tumori,Milan, Italy

a r t i c l e i n f o

Article history:Received 3 August 2012Received in revised form 5 July 2013Accepted 7 September 2013Available online xxxx

Keywords:Prostate cancerLate toxicityRadiotherapyDose–volume effectsNormal tissue complication probabilitymodeling

a b s t r a c t

Purpose: To prospectively evaluate long-term late rectal bleeding (lrb) and faecal incontinence (linc) afterhigh-dose radiotherapy (RT) for prostate cancer in the AIROPROS 0102 population, and to assess clinical/dosimetric risk factors.Materials and methods: Questionnaires of 515 patients with G0 baseline incontinence and bleeding scores(follow-up P6 years) were analysed. Correlations between lrb/linc and many clinical and dosimetricparameters were investigated by univariate and multivariate logistic analyses. The correlation betweenlrb/linc and symptoms during the first 3 years after RT was also investigated.Results: Of 515 patients lrb G1, G2 and G3 was found in 32 (6.1%), 2 (0.4%) and 3 (0.6%) patients while lincG1, G2 and G3 was detected in 50 (9.7%), 3 (0.6%) and 3 (0.6%), respectively. The prevalence of G2–G3 lrbevents was significantly reduced compared to the first 3-years (1% vs 2.7%, p = 0.016) PG1 lrb wassignificantly associated with V75Gy (OR = 1.07). In multivariate analysis, PG1 linc was associated withV40Gy (OR = 1.015), use of antihypertensive medication (OR = 0.38), abdominal surgery before RT(OR = 4.7), haemorrhoids (OR = 2.6), and G2–G3 acute faecal incontinence (OR = 4.4), a nomogram topredict the risk of long-term PG1 linc was proposed.

Results: Importantly, the prevalence of PG1 linc was significantly correlated with the mean inconti-nence score during the first 3 years after RT (OR = 16.3).Conclusions: Long-term (median: 7 years) rectal symptoms are prevalently mild and strongly correlatedwith moderate/severe events occurring in the first 3 years after RT. Linc was associated with several riskfactors.

� 2013 Elsevier Ireland Ltd. All rights reserved. Radiotherapy and Oncology xxx (2013) xxx–xxx

Late rectal injury is a major concern of radiotherapy (RT) forprostate cancer. An early retrospective study (AIROPROS 0101)conducted by the National Working Group on Prostate Radiother-apy of the Italian Association of Radiation Oncology (AIRO)revealed a strong correlation between rectal bleeding and rectal

dose–volume parameters [1]. However, symptoms of rectal toxic-ity other than bleeding are often exhibited by patients, and are dif-ficult to assess retrospectively [2,3].

Consequently, a prospective observational cohort study wasstarted in 2002 (AIROPROS 0102) using a self-administered ques-tionnaire based on the SOMA/LENT system [2,4,5]. The 3-year re-sults were previously reported [6–9]: several clinical andtreatment parameters were found to predict different adverseeffects.

The aim of the present study was to evaluate longer-term rectalfunction after high-dose RT in prostate cancer using the same self-reported questionnaire completed by patients followed up in theAIROPROS 0102 trial for P6 years.

cohort

2 Long-term rectal symptoms after prostate cancer radiotherapy

Materials and methods

The AIROPROS 0102 study

AIROPROS 0102 study was aimed to prospectively assesspredictors of late rectal toxicity after RT for prostate cancer.

In total, 1132 patients treated in 22 different institutions en-tered the study between July 2002 and March 2004. The selectioncriteria were: (a) histologically confirmed prostate adenocarci-noma; (b) participating centres using three-dimensional radical(no post-prostatectomy) radiotherapy with prescription dosesP70 Gy, at 1.8–2 Gy/fr (median dose 74 Gy, range:70–78).

Information concerning volume definition, planning and treat-ment modalities as well as the distribution of the main clinicalparameters of the whole population have previously been reportedin detail [6,9–11]. Of particular importance is the delineation of therectum: an anatomically based definition (from the anus to thepoint where it turns into the sigmoid) was used by the participat-ing centres and validated by a dummy-run investigation that wasmandatory before entering the study. The DVH of the solid rectumwas considered: as discussed elsewhere [6,9–11], as patients wereinstructed to empty their rectum at least at the simulation scan,the DVH of the solid organ may be considered sufficiently robustwith respect to systematic organ motion uncertainties.

The current analysis was limited to 515 patients with a follow-up of P6 years enrolled in 7 centres; only patients with G0 base-line incontinence/bleeding scores were included. Comorbidities,previous abdominal/pelvic surgery, use of antihypertensive/antico-agulant drugs, and previous/concomitant loco-regional diseases(disease to colon, rectum, anus and bladder) were recorded usinga questionnaire completed by the patient before undergoing RT.The type and duration of hormonal therapy, when prescribed, werealso recorded. The rectal DVH was recorded for all patients, and the% volumes receiving >20–75 Gy (termed V20Gy–V75Gy) werecalculated.

The characteristics of the patients included in this analysis arereported in Table 1, including dosimetry statistics.

Assessment of rectal toxicity

The self-administrated questionnaire used (the English versionis reported in [6]) was completed by patients before RT (baseline),

Table 1Patient characteristics and distribution of dosimetric parameters. V30Gy–V75Gy: per cent

Previous abdominal surgeryUse of anticoagulants/antiaggregantsUse of antihypertensivesCardiovascular disease (other than hypertension)DiabetesHaemorrhoidsHormonal therapyPelvic node irradiationSeminal vesicles irradiation

Minimum Maximum

Follow-up time (years) 6.4 8.7ICRU dose (Gy) 70.0 79.2Dose to the pelvis (Gy) 41.4 50.4Dose to seminal vesicles (Gy) 27.0 76.0Rectal V20Gy (%) 28.0 100.0Rectal V30Gy (%) 18.8 100.0Rectal V40Gy (%) 14.0 100.0Rectal V50Gy (%) 10.1 100.0Rectal V60Gy (%) 6.7 93.3Rectal V70Gy (%) 0.0 80.5Rectal V75Gy (%) 0.0 32.2Maximum rectal dose (Gy) 65.2 85.2Mean rectal dose (Gy) 15.4 69.0

Please cite this article in press as: Fellin G et al. Long term rectal function afterstudy. Radiother Oncol (2013), http://dx.doi.org/10.1016/j.radonc.2013.09.028

within 1 month after RT, and then every 6 months up to 3 yearsafter RT (AIROPROS 0102 protocol). The same questionnaire wasalso completed for the present analysis at P6 year follow-up(range 6–9 years). We focused on long term faecal incontinence(linc) and bleeding (lrb) (i.e. toxicity as registered at the time oflong term survey, regardless was happened in the first 3 years afterradiotherapy): Grade (G) 1 lrb was defined as lrb <2 times/week;G2 as lrb P2 times/week; G3 as daily lrb, or need of blood transfu-sion and/or laser coagulation. G1 linc was defined as unintentionalstool discharge ‘‘sometimes’’ experienced. G2 as unintentionalstool discharge ‘‘often’’ experienced or sporadic use of sanitarypads. G3 as daily unintentional stool discharge or sanitary pads>2 times/week.

Statistical analysis

The following parameters were considered: pre-treatment mor-bidities (hypertension, cardiovascular history, diabetes mellitus,and autoimmune diseases), hormonal therapy, use of anti-hyper-tensive and/or anticoagulant drugs, abdominal surgery before RT(rectum–sigmoid resection, kidney resection, cholecystectomyand appendectomy), presence of bowel disease (diverticulitis,chronic colitis, polyposis, colon adenoma and previous colon can-cer), presence of acute lower gastrointestinal toxicity (RTOG/EORTC grade P2), presence of grade P2 acute bleeding, presenceof grade P2 acute incontinence, presence of grade 3 acute inconti-nence, pelvic node/seminal vesicle irradiation, prescribed dose,mean rectal dose, and rectal DVH parameters (V20Gy–V75Gy).

We first conducted univariate logistic regression to identify fac-tors associated with lrb/linc. Next, multivariate stepwise logisticanalysis was performed by including all of the covariates that werepotentially associated with lrb/linc in univariate analyses withp 6 0.20. The odds ratio (OR) was used to express the strength ofassociation between each variable with lrb/linc. Discrimination ofthe models was expressed as the area under the receiver operatingcharacteristic (ROC) curve (AUC) and goodness of fit was testedusing the Hosmer–Lemeshow test. The grouping used for the Hos-mer–Lemeshow test was also used to draw the calibration graph.

Internal validation was carried out by bootstrapping (10,000resamples) and a 10-fold validation procedure to evaluate overfit-ting and optimism. Internal validation evaluates how much you are

volume of the rectum receiving more than 30–75Gy.

Yes (%) No (%)

6.6 93.417.4 82.646.1 53.93.1 96.96.2 93.819.0 81.089.4 10.65.5 94.579.0 21.0

25th percentile 50th percentile 75th percentile

6.8 7.3 7.672.0 74.0 76.045.0 45.0 45.060.0 66.6 72.088.6 97.7 100.079.1 93.4 100.060.1 78.1 92.339.7 52.1 62.127.6 37.6 52.210.3 18.2 26.6

0.0 0.2 6.572.5 74.6 76.445.7 52.8 58.4

high-dose prostatecancer radiotherapy: Results from a prospective cohort

G. Fellin et al. / Radiotherapy and Oncology xxx (2013) xxx–xxx 3

optimistic when you evaluate the performance of a model (in thiscase evaluation of discrimination capability through AUC) on thesame exact population on which the model was developed. Inter-nal validation generally has the effect of decreasing AUC and youcan expect that, in any external validation, the model will not per-form better than in internal validation.

The results of multivariate analyses were used to develop anomogram to predict long-term toxicity.

Overall, 347/515 patients completed the toxicity questionnaireat least 3 times within 3 years after RT (in addition to the baselinequestionnaire). Therefore, we determined the correlation betweenthe mean faecal incontinence score [9] in the first 3 years and long-term linc in this subgroup of patients using v2 tests and logisticregression.

Similarly the correlation between PG2 bleeding in the first3 years and long-term lrb was investigated.

Logistic regression and nomogram design were performedusing MedCalc v.12.1.4 (MedCalc Software bvba, Mariakerke,Belgium) and r-project (http://www.r-project.org) softwares.

Results

Late rectal bleeding

Median follow-up was 7 years (range: 6–9 years). Of 515 pa-tients, 32 (6.1%), 2 (0.4%), and 3 (0.6%) reported G1, G2, and G3lrb, respectively. In Fig. 1a incidence and prevalence PG2 lrb isshown: prevalence of PG2 events resulted to be 1% at 7 yearsagainst values around 3% in the first three years (p = 0.01). PG1lrb was only correlated with V75Gy (continuous variable;p = 0.02, OR = 1.07, AUC = 0.68, logistic plot in the supplementarymaterial, figure A4) and maximum dose to the rectum (continuousvariable; p = 0.02, OR = 1.2, AUC = 0.67). Maximum dose to the rec-tum and V75Gy are clearly correlated (Spearman’s rank correla-tion = 0.87, p = 0.001), for this reason they cannot be usedtogether in MVA. Notably, correlation between maximum rectaldose and V75Gy is weaker in the group of patients treated withdoses >75 Gy (Spearman’s rank correlation = 0.6), in addition max-imum rectal dose is not significantly correlated to lrb in thissubpopulation.

V75Gy = 6.0% was proposed as best-cutoff, based on ROC curve:prevalence of lrb was 5.1% below vs. 13.9% above the cutoff(p = 0.004, v2 test).

The prevalence of long-term PG1 lrb was significantly correlatedwith PG2 bleeding within 3 years after RT: PG1 lrb occurred in42.3% of patients with PG2 bleeding in the first 3 years vs. 5.6% inpatients with <G2 bleeding in the first 3 years (p < 0.0001, v2 test).

Late faecal incontinence

Of 515 patients, 50 (9.7%), 3 (0.6%), and 3(0.6%) reported G1, G2,and G3 linc, respectively. 15/56 (26.8%) long term linc were of newinsurgence (i.e. mean grade of late faecal incontinence in the firstthree years = 0).

Incidence and prevalence of grade P1 and grade P2 late faecalincontinence in the first three years of follow-up and at the time oflong term evaluation are reported in Fig. 2.

In multivariate analysis (global p < 0.0001), V40Gy (continuousvariable; p = 0.05, OR = 1.016, 95% confidence interval (CI): 1.00–1.03), use of antihypertensive (protective; p = 0.005, OR = 0.38,95% CI: 0.2–0.8), abdominal surgery before RT (p = 0.01, OR = 4.7,95% CI: 1.2–9.2), haemorrhoids (p = 0.008, OR = 2.6, 95% CI: 1.1–4.8), and the presence of G2–G3 acute incontinence (p = 0.007,OR = 4.4, 95% CI: 1.8–4.6) were significantly and independentlyassociated with long term linc. Resulting logistic curves arepresented in the supplementary material (Figs. A2 and A3).

Please cite this article in press as: Fellin G et al. Long term rectal function afterstudy. Radiother Oncol (2013), http://dx.doi.org/10.1016/j.radonc.2013.09.028

Using the results of the multivariate analyses, a nomogram wasdeveloped to predict long-term PG1 linc (Fig. 2). The AUC of themodel was 0.77 (p < 0.0001, 95% CI: 0.72–0.80; Hosmer–Leme-show: p = 0.50): Fig. 3 shows the calibration plot. Optimism, esti-mated through bootstrapping and 10-fold cross validation, wasapproximately 5%, leading to corrected AUCs of 0.71 and 0.72, forbootstrapping and cross validation, respectively.

Long-term linc was also significantly correlated with the meanincontinence score within 3 years after RT (p < 0.0001). The meanincontinence score at 3 years was 0.1, 0.5 and 0.78 in patients withG0, G1 and G2–G3 long-term linc, respectively (p < 0.0001; v2

test).The prevalence of long-term linc PG1 was significantly corre-

lated with the mean incontinence score (continuous variable) inthe first 3 years after RT (p < 0.0001, OR = 16.3, 95% CI: 6–43.8).When we dichotomised the mean incontinence score (cutoff:0.5), a significantly greater percentage of patients showed long-term linc in the group above cut-off (37.3% vs. 10%; p < 0.0001,v2 test).

Discussion

To our knowledge, this is the largest prospective longitudinalstudy designed to identify predictors of long-term rectal toxicityafter high-dose RT for prostate cancer. The prospective collectionof questionnaire-based information allowed us to carefully recordpatient-reported symptoms and their evolution over time and totest the possible impact of multiple cofactors on their generationand persistence.

The current study follows previous analyses revealing clear pre-dictors of rectal toxicity occurring within 3 years after RT [6–10].The first result is a sort of ‘‘down-staging’’ of rectal toxicity, withvery few grade P2 long term symptoms. For patients developinggrade P2 toxicity not performing Argon Plasma Coagulation(APC) a spontaneous improvement of lrb over time is described(which was already reported in some studies [12–14]), even if re-cover is not complete in a non-negligible proportion of patients.When APC is considered, in this subpopulation with long terminformation 6/26 patients with grade P2 lrb at 3 years performedAPC. 1/6 performing APC had grade 1 long term lrb, while 5/6 com-pletely recovered and were without symptoms.

On the other hand, for the first time, the correlations betweenthe longitudinal assessment of toxicities at 3 years and the long-term effects of RT could be investigated. The results are veryintriguing, showing that a large proportion of patients exhibitinglong-term rectal bleeding or faecal incontinence experienced clin-ically relevant symptoms in the first years after the treatment andthat the persisting symptoms are typically mild (i.e. grade 1).

Although several studies have reported the long-term outcomesof RT, including gastrointestinal symptoms, to our knowledge noprospective observational studies have attempted to identify pre-dictors of long term (>6 years) rectal toxicity, including longitudi-nal assessment of symptoms. Most of the studies involving largenumbers of patients were cross-sectional in design and comparedoutcomes of different therapeutic approaches for prostate orgynaecological cancer [15–19] or against an apparently healthypopulation in controlled studies [20,21].

Our findings were also strengthened by the exclusion of the pa-tients with rectal bleeding or faecal incontinence at baseline[17,18,22]. Long term toxicity in the subgroup of patients withpre-RT gastro-intestinal symptoms will be the topic of a dedicatedstudy.

Interestingly, the main predictors of lrb and linc werecompletely different. This is the first time such results have beenreported for long-term effects, although already shown for

high-dose prostatecancer radiotherapy: Results from a prospective cohort

Fig. 1. (a) Incidence and prevalence of grade P2 late rectal bleeding in the first three years of follow-up and at the time of long term evaluation; (b) incidence and prevalenceof grade P1 and grade P2 late faecal incontinence in the first three years of follow-up and at the time of long term evaluation.

Fig. 2. Nomogram for the prediction of PG1 long term faecal incontinence (i.e. presence of persistent mild symptoms at 6–9 years after radiotherapy).

4 Long-term rectal symptoms after prostate cancer radiotherapy

Please cite this article in press as: Fellin G et al. Long term rectal function after high-dose prostatecancer radiotherapy: Results from a prospective cohortstudy. Radiother Oncol (2013), http://dx.doi.org/10.1016/j.radonc.2013.09.028

Fig. 3. Calibration plot for the multivariate model of PG1 late faecal incontinence;the observed outcome in deciles are plotted to graphically illustrate the Hosmer–Lemeshow goodness-of-fit test. The equation was derived from interpolation of thecalibration points using the least-squares method.

G. Fellin et al. / Radiotherapy and Oncology xxx (2013) xxx–xxx 5

symptoms occurring within the first 3 years after RT [6–9,23–25].We found that lrb was only associated with V75Gy and maximumdose to the rectum; this confirmed the earlier findings showing apredominantly ‘‘serial’’ behaviour of the rectum for bleeding[6,7,23,24]. Not surprising, maximum rectal dose and V75Gy areweakly correlated (Spearman’s rank correlation = 0.6) when theprescribed dose exceeds 75Gy and maximum rectal dose is not apredictor of bleeding in this situation. This indicates that V75Gymight be of greater use when considering dose-escalation.

Interestingly, we did not confirm the previously reported im-pact of other clinical factors, such as previous abdominal surgeryand acute toxicity [6,26] that appear to have a greater effect inthe first few years after RT. On the other hand, other clinical factorsthat were previously reported as potential predictors of bleeding,such as cardiovascular disease or diabetes [24], were not confirmedin the present study. The clinical risk factors that were reported tobe correlated to moderate/severe lrb in the first 3 years continue tobe important predictors as reducing moderate/severe events is oneof the goals of toxicity modelling. It could be that clinical risk fac-tors increase the probability of moderate/severe events that com-pletely recover after some time, while V75Gy is more related toan enhanced local damage which partially, but not completely,recovers. The relatively low predictive power of the model thatonly included V75Gy (AUC = 0.68, 95% CI: 0.63–0.72, p = 0.0001)or maximum rectal dose (AUC = 0.67, 95% CI: 0.63–0.72,p = 0.001) is perhaps unsurprising, as other clinical and genetic[27] cofactors may also be important, but were not assessed in thisstudy.

For the first time, we found a clear relationship between PG2bleeding during the first 3 years and long-term PG1 lrb.

Incontinence symptoms are more frequent than bleeding symp-toms. When toxicity was prospectively evaluated, w10% of patientsexperienced any grade of faecal incontinence. Even grade 1 latefaecal incontinence has a key impact on quality of life. The main re-sults were consistent with the ones previously reported for PG2faecal incontinence at 3 years [9]; most patients with long-termlinc had persistent symptoms during the first 3 years.

The present results indicate that incontinence is a significantproblem after RT for prostate cancer [3,20,28] particularly for ‘‘sen-sitive’’ patients, (abdominal surgery, haemorrhoids). Several mech-anisms have been claimed to explain the impact of previoussurgery: it could predispose to inflammatory patterns that may in-crease tissue sensitivity, even at sites located some distance fromthe surgical site and/or may influence the regional (neuro)vascularsystem [6,9,23,24,28].

Please cite this article in press as: Fellin G et al. Long term rectal function afterstudy. Radiother Oncol (2013), http://dx.doi.org/10.1016/j.radonc.2013.09.028

Interestingly, the protective effects of antihypertensivemedication suggest that the vascular system [29] is involved inthe development and persistence of symptoms.

The risk of long-term incontinence may be reduced by limitingthe dose bath to the rectum (V40Gy), although this factor seems toexert a weaker influence compared with the other clinical factors.Other authors [30–32] recently found that doses to the anal canaland pelvic floor muscles were significantly related to late faecalincontinence, unfortunately we do not have detailed dosimetricinformation on the anal canal and pelvic floor muscles and cannotdeal with this topic.

The relatively high AUC of the logistic model (0.77), the con-firmed goodness-of-fit (Hosmer–Lemshow test), and the good cal-ibration for long-term incontinence allowed us to develop the firstnomogram to predict the risk of long-term faecal incontinence.

Our results on long term toxicity incidence and prevalence ofgrade2–3 events are quite consistent with those reported bySchmid and coworkers [33]: the majority of moderate/severe lategastro-intestinal effects after external beam radiotherapy for pros-tate cancer are transient. In addition, we found that a non-negligi-ble portion of patients still suffers of grade 1 long term bleedingafter a grade 2–3 bleeding event. On the other hand late faecalincontinence is a fairly persistent symptom, indicating that a lon-gitudinal score of faecal incontinence grade [9] is the best way totake into account duration and severity of this effect, which cancause important impairment of the patient quality of life (and thatcan also be due to causes other than radiation, first of all ageing). Inany case, even for incontinence, the prevalence of grade 2–3 eventsreduced from about 2–3% of the first 3 years to about 1% at 7 years.

In summary, this study shows the persistence of (mild) treat-ment-related rectal symptoms many years after high-dose RT.The main predictors of both endpoints showed greater predictivevalue for incontinence than for bleeding. An important finding ofthis study was a clear association between longitudinally assessedtoxicity occurring within 3 years after RT and long-term effects.

These results suggest that a non-negligible proportion of pa-tients with moderate/severe effects within the 3 years after RTare unlikely to fully recover from their symptoms by 6–9 yearsafter RT.

Disclaimer

No conflicts of interest to declare.

Acknowledgements

Fondazione Monzino is gratefully acknowledged for its supportto the AIRO Working Group on Prostate Radiotherapy. T. Magnaniis acknowledged for her technical assistance. All institutionsinvolved in the study are gratefully acknowledged for their enthu-siastic support.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.radonc.2013.09.028.

References

[1] Fiorino C, Sanguineti G, Cozzarini C, et al. Rectal dose–volume constraints inhigh dose conformal radiotherapy for localized prostate cancer. Int J RadiatOncol Biol Phys 2003;57:953–62.

[2] Denham JW, O’Brien PC, Dunstan RH, et al. Is there more than one lateradiation proctitis syndrome? Radiother Oncol 1999;51:43–53.

high-dose prostatecancer radiotherapy: Results from a prospective cohort

6 Long-term rectal symptoms after prostate cancer radiotherapy

[3] Maeda Y, Høyer M, Lundby L, Norton C. Faecal incontinence followingradiotherapy for prostate cancer: a systematic review. Radiother Oncol2011;98:145–53.

[4] LENT SOMA tables. Table of contents. Radiother Oncol 1995;35:17–60.[5] Davidson SE, Burns M, Routledge J, et al. Short report: a morbidity scoring

system for Clinical Oncology practice: questionnaires produced from the LENTSOMA scoring system. Clin Oncol (R Coll Radiol) 2002;14:68–9.

[6] Fellin G, Fiorino C, Rancati T, et al. Clinical and dosimetric predictors of laterectal toxicity after conformal radiation for localized prostate cancer: results ofa large multi-center observational study. Radiother Oncol 2009;93:197–202.

[7] Rancati T, Fiorino C, Fellin G, et al. Inclusion of clinical risk factors into NTCPmodelling of late rectal toxicity after high dose radiotherapy for prostatecancer. Radiother Oncol 2011;100:124–30.

[8] Valdagni R, Kattan MW, Rancati T, et al. Is it time to tailor the prediction ofradio-induced toxicity in prostate cancer patients? Building the first set ofnomograms for late rectal syndrome. Int J Radiat Oncol Biol Phys2012;82:1957–66.

[9] Fiorino C, Rancati T, Fellin G, et al. Late faecal incontinence after high-doseradiotherapy for prostate cancer: better prediction using longitudinaldefinitions. Int J Radiat Oncol Biol Phys 2012;83:38–45.

[10] Vavassori V, Fiorino C, Rancati T, et al. Predictors for rectal and intestinal acutetoxicities during prostate cancer high-dose 3D-CRT: results of a prospectivemulticenter study. Int J Radiat Oncol Biol Phys 2007;66:1401–10.

[11] Foppiano F, Fiorino C, Frezza G, et al. The impact of contouring uncertainty onrectal 3D dose-volume data: results of a dummy run in a multicenter trial(AIROPROS01-02). Int J Radiat Oncol Biol Phys 2003;57:573–9.

[12] Goldner G, Pötter R, Kranz A, et al. Healing of late endoscopic changes in therectum between 12 and 65 months after external beam radiotherapy.Strahlenther Onkol 2011;187:202–5.

[13] O’Brien PC, Hamilton CS, Denham JW, et al. Spontaneous improvement in laterectal mucosal changes after radiotherapy for prostate cancer. Int J RadiatOncol Biol Phys 2004;58:75–80.

[14] van Lin EN, Kristinsson J, Philippens ME, et al. Reduced late rectal mucosalchanges after prostate three-dimensional conformal radiotherapy withendorectal balloon as observed in repeated endoscopy. Int J Radiat OncolBiol Phys 2007;67:799–811.

[15] Fransson P, Lund JA, Damber JE, et al. Quality of life in patients with locallyadvanced prostate cancer given endocrine treatment with or withoutradiotherapy: 4-year follow-up of SPCG- 7/SFUO-3, an open-label,randomised, phase III trial. Lancet Oncol 2009;10:370–80.

[16] Wei JT, Dunn RL, Sandler HM, et al. Comprehensive comparison of health-related quality of life after contemporary therapies for localized prostatecancer. J Clin Oncol 2002;20:557–66.

[17] Pardo Y, Guedea F, Aguiló F. Quality-of-life impact of primary treatments forlocalized prostate cancer in patients without hormonal treatment. J Clin Oncol2010;28:4687–96. Erratum in: J Clin Oncol 2011: 29(779).

[18] Chen RC, Clark JA, Talcott JA. Individualizing quality-of-life outcomesreporting: how localized prostate cancer treatments affect patients withdifferent levels of baseline urinary, bowel, and sexual function. J Clin Oncol2009;27:3916–22.

Please cite this article in press as: Fellin G et al. Long term rectal function afterstudy. Radiother Oncol (2013), http://dx.doi.org/10.1016/j.radonc.2013.09.028

[19] Potosky AL, Davis WW, Hoffman RM, et al. Five-year outcomes afterprostatectomy or radiotherapy for prostate cancer: the prostate canceroutcomes study. J Natl Cancer Inst 2004;96:1358–67.

[20] Dunberger G, Lind H, Steineck G, et al. Self-reported symptoms of faecalincontinence among long-term gynaecological cancer survivors andpopulation-based controls. Eur J Cancer 2010;46:606–15.

[21] Mols F, van de Poll-Franse LV, Vingerhoets AJJM, et al. Long-term quality of lifeamong dutch prostate cancer survivors. Cancer 2006;107:2186–96.

[22] Coen JJ, Paly JJ, Niemierko A, et al. Long-term quality of life outcome afterproton beam monotherapy for localized prostate cancer. Int J Radiat Oncol BiolPhys 2012;82:e201–9.

[23] Peeters STH, Hoogeman M, Heemsbergen WD, et al. Rectal bleeding, fecalincontinence and high stool frequency after conformal radiotherapy forprostate cancer: normal tissue complication probability modeling. Int JRadiat Oncol Biol Phys 2006;66:11–9.

[24] Defraene G, Van den Bergh L, Al-Mamgani A, et al. The benefits of includingclinical factors in rectal normal tissue complication probability modeling afterradiotherapy for prostate cancer. Int J Radiat Oncol Biol Phys2012;82:1233–42.

[25] Barnett GC, De Meerleer G, Gulliford SL, et al. The impact of clinical factors onthe development of late radiation toxicity: results from the medical researchcouncil RT01 Trial (ISRCTN47772397). Clin Oncol 2011;23:613–24.

[26] Valdagni R, Vavassori V, Rancati T, et al. Increasing the risk of late rectalbleeding after high dose radiotherapy for prostate cancer: the case of previousabdominal surgery. Results from a prospective trial. Radiother Oncol2012;103:252–5.

[27] Valdagni R, Rancati T, Ghilotti M, et al. To bleed or not to bleed. A predictionbased on individual gene profiling combined with dose–volume histogramshapes in prostate cancer patients undergoing three-dimensional conformalradiation therapy. Int J Radiat Oncol Biol Phys 2009;74:1431–40.

[28] Fiorino C, Valdagni R, Rancati T, Sanguineti G. Dose–volume effects for normaltissues in external radiotherapy: pelvis. Radiother Oncol 2009;93:153–67.

[29] Agabiti-Rosei E, Rizzoni D. Regression of small resistance artery structuralalterations in hypertension by appropriate antihypertensive treatment. CurrHypertens Rep 2010;12:80–5.

[30] Smeenk RJ, Hoffmann AL, Hopman WP, et al. Dose–effect relationships forindividual pelvic floor muscles and anorectal complaints after prostateradiotherapy. Int J Radiat Oncol Biol Phys 2012;83:636–44.

[31] Smeenk RJ, Hopman WP, Hoffmann AL, et al. Differences in radiationdosimetry and anorectal function testing imply that anorectal symptomsmay arise from different anatomic substrates. Int J Radiat Oncol Biol Phys2012;82:145–52.

[32] Vordermark D, Schwab M, Ness-Dourdoumas R, et al. Association of anorectaldose–volume histograms and impaired fecal continence after 3D conformalradiotherapy for carcinoma of the prostate. Radiother Oncol 2003;69:209–14.

[33] Schmidt MP, Potter R, Bombosh V, et al. Late gastrointestinal and urogenitalside-effects after radiotherapy – incidence and prevalence. Subgroup-analysiswithin the prospective Austrian–German phase II multicentre trial forlocalized prostate cancer. Radiother Oncol 2012;104:114–8.

high-dose prostatecancer radiotherapy: Results from a prospective cohort