If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Dose–volume effect relationships for late rectal morbidity in patients treated with chemoradiation and MRI-guided adaptive brachytherapy for locally advanced cervical cancer: Results from the prospective multicenter EMBRACE study
To establish dose volume–effect relationships predicting late rectal morbidity in cervix cancer patients treated with concomitant chemoradiation and MRI-guided adaptive brachytherapy (IBABT) within the prospective EMBRACE study.
Material and method
All patients were treated with curative intent according to institutional protocols with chemoradiation and IGABT. Reporting followed the GEC-ESTRO recommendations (, ), applying bioeffect modeling (linear quadratic model) with equieffective doses (EQD23). Morbidity was scored according to the CTC-AE 3.0. Dose–effect relationships were assessed using comparisons of mean doses, the probit model and log rank tests on event-free periods.
Results
960 patients were included. The median follow-up was 25.4 months. Twenty point one percent of the patients had grade 1 events, 6.0% grade 2, 1.6% grade 3 and 0.1%, grade 4. The mean DICRU, , and were respectively: 66.2 ± 9.1 Gy, 72.9 ± 11.9 Gy, and 62.8 ± 7.6 Gy. Increase of dose was associated with increase in severity of single endpoints and overall rectal morbidity (grade 1–4) (p < 0.001–0.026), except for stenosis (p = 0.24–0.31). The probit model showed significant relationships between the , , and DICRU and the probability of grade 1–4, 2–4, and 3–4 rectal events. The equieffective for a 10% probability for overall rectal grade ⩾ 2 morbidity was 69.5 Gy (p < 0.0001). After sorting patients according to 6 levels, less favorable outcome was observed in the high dose subgroups, for bleeding, proctitis, fistula, and overall rectal morbidity. A ⩾ 75 Gy was associated with a 12.5% risk of fistula at 3 years versus 0–2.7% for lower doses (p > 0.001). A < 65 Gy was associated with a two times lower risk of proctitis than ⩾ 65 Gy.
Conclusions
Significant correlations were established between late rectal morbidity, overall and single endpoints, and dose–volume (, ) and dose-point (DICRU) parameters. A ⩽ 65 Gy is associated with more minor and less frequent rectal morbidity, whereas a ⩾ 75 Gy is associated with more major and more frequent rectal morbidity.
]. During the last 20 years, image guided adaptive brachytherapy (IGABT), a high precision radiation technique has been developed through progress in afterloaders, applicators and computer software which enable the integration of 3D images such as MRI into treatment planning. This approach includes accurate delineation of tumor and targets and the organs at risk (OAR), as well as optimized 3D treatment planning based on dose–volume histograms. Recent monocentric series showed high local control rates with promising results and only limited to moderate morbidity in regard to classical outcomes from historical cohorts [
Clinical outcomes of definitive chemoradiation followed by intracavitary pulsed-dose rate image-guided adaptive brachytherapy in locally advanced cervical cancer.
Clinical efficacy and toxicity of radio-chemotherapy and magnetic resonance imaging-guided brachytherapy for locally advanced cervical cancer patients: a mono-institutional experience.
Reporting and validation of gynaecological Groupe Euopeen de Curietherapie European Society for Therapeutic Radiology and Oncology (ESTRO) brachytherapy recommendations for MR image-based dose volume parameters and clinical outcome with high dose-rate brachytherapy in cervical cancers: a single-institution initial experience.
Clinical outcome and dosimetric parameters of chemo-radiation including MRI guided adaptive brachytherapy with tandem-ovoid applicators for cervical cancer patients: a single institution experience.
Clinical outcome of protocol based image (MRI) guided adaptive brachytherapy combined with 3D conformal radiotherapy with or without chemotherapy in patients with locally advanced cervical cancer.
In 2000, in a will of harmonizing contouring and reporting in image guided cervix cancer brachytherapy, the GEC-ESTRO (Groupe Européen de Curiethérapie – European Society for Radiation Oncology) launched a Gynaecologic working group which finally led to publish recommendations to define clinical target volumes (low, intermediate and high risk CTV) for response-adaptive brachytherapy and consensual dosimetric parameters (D100 and D90 for the CTV, , and for OAR) in 2005–06 [
Recommendations from Gynaecological (GYN) GEC-ESTRO Working Group (I): concepts and terms in 3D image based 3D treatment planning in cervix cancer brachytherapy with emphasis on MRI assessment of GTV and CTV.
Recommendations from gynaecological (GYN) GEC ESTRO working group (II): concepts and terms in 3D image-based treatment planning in cervix cancer brachytherapy-3D dose volume parameters and aspects of 3D image-based anatomy, radiation physics, radiobiology.
]. In the aftermath, the GEC-ESTRO launched a prospective multicentric study, EMBRACE (an intErnational study on MRI-guided BRachytherapy in locally Advanced CErvical cancer) in 2008, aiming to evaluate the outcome and morbidity of MRI-guided brachytherapy [
]. At that time, little evidence was available on dose constraints for organs at risk and most clinicians were used to refer to data based on 2D experience and monoinstitutional reports [
Correlation of dose–volume parameters, endoscopic and clinical rectal side effects in cervix cancer patients treated with definitive radiotherapy including MRI-based brachytherapy.
Computed tomography-based high-dose-rate intracavitary brachytherapy for uterine cervical cancer: preliminary demonstration of correlation between dose–volume parameters and rectal mucosal changes observed by flexible sigmoidoscopy.
]. In EMBRACE, the participants were free to apply their institutional protocols, but had to perform contouring and to report their data following the GEC-ESTRO recommendations. Among secondary study aims, aim #6 was to establish dose–volume effects correlations between morbidity and dose–volume parameters. The objective of this work was to establish dose–volume correlations between volumetric dosimetric parameters ( and , minimal dose delivered to a maximally exposed volume of an OAR) and the occurrence of rectal morbidity.
Material and methods
Patients
The EMBRACE study included prospectively patients with histologically proven cervical carcinomas from 24 centers located in Europe, Asia, and North America. Centers went through a dummy run before being accepted in the study [
]. Briefly, to be eligible, patients should have no history of cancer except carcinoma in situ of the cervix or basal cell carcinoma of the skin, and no total or partial hysterectomy. An MRI at diagnosis was required. Patients were included prior to treatment initiation (EBRT). Their treatment had to combine pelvic or extended-fields external beam radiotherapy with concomitant chemotherapy when not contra-indicated and MRI guided brachytherapy, and to be led with curative intent. Patients with involvement of the para-aortic lymph nodes could be included provided pathological nodes were located below the level of L1–L2. The study was approved by local ethic committees and patients consent was obtained if required according to institutional rules.
Treatment
The study was prospective and observational. However, the EBRT technique was harmonized with standard fractionation (1.8–2 Gy, once a day, 5 times a week), prescribed doses ranging from 45 to 50 Gy, and delivery of 4 field conformal EBRT or IMRT/VMAT. Midline block during EBRT was not allowed. Nodal boosts could be sequential or performed with simultaneous integrated boost technique. Parametrial boosts were also allowed. Concomitant chemotherapy was foreseen for all patients, except for those with major comorbidity. Neo-adjuvant or adjuvant treatment was not allowed. Patients could be treated either with high dose rate (HDR) or pulsed-dose rate (PDR) brachytherapy. No planning aims were imposed, but the reporting had to follow the GEC-ESTRO recommendations including the and for OAR [
Recommendations from Gynaecological (GYN) GEC-ESTRO Working Group (I): concepts and terms in 3D image based 3D treatment planning in cervix cancer brachytherapy with emphasis on MRI assessment of GTV and CTV.
Recommendations from gynaecological (GYN) GEC ESTRO working group (II): concepts and terms in 3D image-based treatment planning in cervix cancer brachytherapy-3D dose volume parameters and aspects of 3D image-based anatomy, radiation physics, radiobiology.
Outcomes and morbidity were assessed at 3, 6, 9, 12, 18, 24, 30, 36, 48, and 60 months after treatment completion. Morbidity was evaluated using the Common Toxicity Criteria for adverse events (CTC-AE), version 3.0. Late morbidity was defined as any toxicity event occurring or lasting over 90 days after treatment initiation (first fraction of EBRT). Therefore, patients with follow-up less than 90 days were not eligible for this subgroup analysis. Four types of rectal events were reported in EMBRACE: rectal bleeding, proctitis, stenosis, and fistula (Additional material: Table 1). They were analyzed independently, and together as overall rectal morbidity. Patients with persistent disease after completion of the treatment were not eligible for morbidity assessment. Patients who experienced relapses were excluded from morbidity analysis from the date of their recurrence. Morbidity was censored if present at the same time of any local and/or regional and/or systemic evidence of disease.
Dose–volume parameters
Dose volume parameters ( and ) and rectal DICRU (dose to the rectal International Commission for Radiation measurements and Units point, ICRU, report 38 were reported and converted in 2 Gy equivalent (EQD2) using the linear quadratic model with α/β = 3 Gy and a half-time repair of 1.5 h [
]. In cases of multiple brachytherapy fractions, doses were summed by adding the dose from each fraction, assuming that the most exposed area of the rectum remained stable in succeeding brachytherapy fractions. Brachytherapy doses were summed with the EBRT prescribed dose converted into EQD2, using the same model, assuming that studied volumes (points) were located in the 100% isodose of the EBRT prescribed dose.
Data collection
Data were extracted from the study database in August 2015. For the purpose of this analysis, patients who completed their treatment at least 12 months before (August 2014) were retained.
Statistics
Prevalence and cumulative incidence were calculated. Times to onset were defined from treatment initiation to the date of event occurrence. For crude incidence, as well as for dose effect correlations, the maximal graded event was considered for analyses. In case of equally graded events in the same patients, the earliest was considered for analyses. For patients with rectal symptoms at baseline, treatment related morbidity was defined as an increase in the score. As DICRU, , and did not follow normal distributions (positive Shapiro Wilk and Kolmogorov–Smirnov tests), mean doses were compared using Kruskal–Wallis tests while comparing three or more variables and Mann–Whitney–U tests for analyses limited to 2 variables. Dose–volume effects were analyzed using two methods. First, dose effect correlations were tested using the logistic regression analysis with the probit model, without assumption of a threshold dose. Second, log rank tests were performed on Kaplan–Meier event-free period curves sorting patients according levels, by steps of 5 Gy. Statistical significance was considered for p ⩽ 0.05. All statistics were performed using XLSTAT 2014 (Addinsoft SARL, Paris, France).
Results
Patients
From the 1 129 patients who completed their treatment at least 1 year before the data extraction, 113 were excluded for missing rectal values, 8 for inconsistencies in the reported dosimetric parameters ( ⩾ ), and 48 who were not evaluable for late morbidity (follow-up < 90 days, progressive disease or incomplete response before first assessment of morbidity). Finally, 960 patients were eligible for analyses.
Patients’ mean age was 50.5 ± 13.1 years. Stage was classified as IA in 0.1% (n = 1) IB in 19.2% (n = 184), IIA in 5.5% (n = 53), IIB in 53.1% (n = 510), IIIA in 0.6% (n = 6), IIIB in 17.1% (n = 165), IVA in 3.0% (n = 29), and unknown in 1.3% (n = 12, missing value or inconstancies in the reported data), according to the FIGO (Fédération Internationale de Gynécologie Obstétrique) staging system. Rectal involvement was noticed on MRI in 4 patients, confirmed by endoscopy in 1 patient (2 negative endoscopies and 1 not performed). Median follow-up of the whole cohort was 25.4 months (3.0–75.6). IMRT was applied in 28.2% of the patients and the remaining received 3D conformal radiotherapy. Concomitant chemotherapy was delivered in 92.8% of the patients. High dose rate brachytherapy was performed in 57.1% of the cases. Techniques combining intracavitary application with an interstitial component were used in 34.4%. The mean rectal DICRU, , and were 66 ± 9.1 Gy, 72.9 ± 11.9 Gy, and 63.8 ± 7.6 Gy, respectively.
Descriptive findings and analyses
Rectal symptoms were noticed in 15 patients at diagnosis (1.5%): 1 patient with grade 1 rectal bleeding, 7 patients with proctitis, grade 1 in all cases except 1 with grade 2, 2 patients with rectum stenosis, all grade 2, and finally 5 patients with mixed symptoms (proctitis and bleeding in 4 patients, proctitis and stenosis in one patient).
A total of 265 patients (265 out of 960, 27.6%) reported rectal treatment-related events during the observation period. Proctitis and bleeding were the most frequent events with 178 patients (18.5%) and 155 patients (16.3%), respectively. Stenosis and fistula occurred in 11 (1.1%) and 9 patients (0.9%), respectively. Among the 9 patients who developed a fistula, none was treated for a IVA lesion, nor had been suspicious of rectal involvement on initial pelvic MRI. Proctitis and rectal bleeding occurred both in 74 patients (7.7%). Rectal morbidity grade 1 was reported in 20.1% of the patients, grade 2 in 6.0%, grade 3 in 2.0%, and grade 4 in 0.1%. No rectal event was reported in 72.3% of the patients. Details on frequency of rectal events and grading are provided in Table 1.
The prevalence of rectal bleeding increased from 3 to 24 months (1.2% to 9.8%), followed by a gradual decrease: 5.9% at 3 years and 4.8% at 5 years. The prevalence of proctitis followed a similar profile with a peak of 9.2% at 2 years followed by a decrease to 3.4% at 4 years. The prevalence of fistula and stenosis increased continuously with time (0.1, 0.4, and 0.5% at 1, 2, 4 years for fistula, and 0.4, 1.0, 1.2% at 1, 2, 5 years for stenosis (Additional Fig. 1, in Supplementary material). Overall, the prevalence of rectal morbidity was 6.2% at 3 months, 15.7% at 2 years, and 7.8% at 4 years (Fig. 1).
Fig. 1Incidence and prevalence of rectal morbidity. BL: baseline. m: months. Prevalence is represented using histograms, with legends on the left, and incidence using curves, legends on the right. Prevalence rates according to grade and time are presented on the histograms.
The actuarial estimates of cumulative incidence of grade 1–4 overall rectal morbidity was 34.6% at 3 years, whereas it was 10.3% for grade 2–4, and 2.3% for grade 3–4 (Fig. 1). For proctitis, the 3-year actuarial estimate was 23.1% for grade 1–4, 5.8% for grade 2–4, and 0.6% for grade 3–4 (Fig. 2B). The corresponding rates for bleeding were 20.7%, 5.4%, and 1.4% (Fig. 2A. Concerning stenosis, grade 1–4 and 2–4 actuarial estimates at 3 years were 1.7% and 0.9% (Fig. 2C). For fistula, actuarial estimates were 1.4% for grade 2–4 and 0.6% for grade 3–4, at 3 years (Fig. 2D).
Fig. 2Cumulative actuarial incidence of rectal bleeding, proctitis, fistula, and stenosis. A: bleeding, B: proctitis, C: stenosis, D: fistula. Number of patients at risk, number of events and 3-year event probability are reported in tables below the plots.
Median times to onset were 12.6, 13.7, and 12.9 months for grade 1, 2 and 3–4 events, respectively. There was no major difference among the four types of endpoints analyzed: grade 1 ranging from 13.1 to 14.5 months, grade 2 from 11.7 to 14.5 months, and grade 3–4, from 10.8 to 15.9 months.
Dose–volume effect analyses
Overall rectal morbidity grade progressively increased with increase in mean , and DICRU progressively (Additional Table 2, p < 0.0001 for the 3 dosimetric parameters). Significant increase in grades in parallel to an increase in mean doses was also observed for proctitis, bleeding, and fistula, but not for stenosis (p ranging from <0.0001 to 0.026). Dividing patients into two groups, grade 0–1 versus 2–4, the mean doses were significantly higher in the latter group for proctitis, bleeding, fistula and the overall rectal morbidity (Additional Table 2, p ranging <0.0001–0.003).
The probit model showed significant relationships between the DICRU, and and the probability of grade 1–4, 2–4 and 3–4 fistula and rectal bleeding (p ranging from <0.0001 to 0.014, Table 3 in Supplementary material) except for grade 3–4 rectal bleeding DICRU (p = 0.42). Significant relationships were also observed for and and the probability of occurrence of grade 2–4 proctitis, as well as for grade 1–4/2–4 and DICRU (p ranging from 0.004 to 0.013). No relationship was observed between stenosis and any dosimetric parameter. , and DICRU were significantly related to the probability of overall rectal morbidity. According to this model the delivery of 60, 65, 70, and 75 Gy corresponded to 6.0% (4.5–7.8), 7.9% (6.4–9.7), 10.2% (8.0–13.0), and 13.1% (9.3–15.6) of grade 2–4 rectal morbidity (Fig. 3A and Additional Fig. 2 in Supplementary material). The ED10 for rectal morbidity ( corresponding to a 10% probability of grade 2–4 events) was 69.5 Gy (65.2–76.9). The effect of the follow-up was analyzed in this model, by sorting patients according to their follow-up in 4 groups: <12 months (n = 199, 20.7% of the patients), 12–24 months (n = 256, 26.7%), 24–36 months (n = 181, 18.9%), ⩾36 months (n = 324, 33.8%). According to the probit model, the ED10 for all endpoints at the different follow-up times were 90.6 Gy, 73.2 Gy, 60.7 Gy and 63.6 Gy, respectively (Fig. 3B, p < 0.0001).
Fig. 3Dose–volume effects for grade 2–4 overall rectum morbidity (A) and impact of follow-up on the relationships (B). Figure A focuses on the common prescription range of . The full plot is provided as an additional material (Additional Fig. 2). Gray dashes: 95% confidence interval.
Actuarial estimates of cumulative incidence were generated according to 6 levels: <55 Gy (n = 129, 13.4%), 55–60 (n = 242, 25.2%), 60–65 (n = 243, 25.3%), 65–70 (n = 196, 20.4%), 70–75 (n = 115, 12.0%), and ⩾75 Gy (n = 35, 3.6%). Log rank test showed significant differences for grade 1–4, 2–4, and 3–4 bleeding, fistula, and all rectal events (Fig. 4A, B, D from <0.0001 to 0.032). The probability of grade 2–4 rectal morbidity was 3.5%, 6.5%, 8.6%, 15.1%, 18.0%, and 26.0% for <55 Gy, 55–60, 60–65, 65–70, 70–75, and ⩾75 Gy respectively (Fig. 4A, p < 0.0001). The cumulative incidence of proctitis at 3 years ranged between 8.3 and 14.3% in patients with ⩾65 Gy, whereas it was between 3.4 and 4.6% in patients with <65 Gy (Fig. 4C). The risk of rectal bleeding increased progressively with : 0.9%, 3.1%, 5.2%, 6.3% in the groups with <55 Gy, 55–60 Gy, 60–65, and 65–70 Gy, respectively. The rates were 12.8 and 11.8% in the subgroups with ranging between 70 and 75, and ⩾75 Gy, respectively (Fig. 4D). For fistula, the risk was 12.5% at 3 years in the subgroup of patients who received ⩾75 Gy, whereas it was ranging between 0 and 2.7% in all other subgroups with lower (p < 0.0001, Fig. 4B). Finally, no significant difference was observed for stenosis. The cumulative incidences at 3 years of each event according to subgroups are reported in Fig. 4. Results for grade 1–4 and 3–4 event-free periods are provided in the Supplementary Table 4.
Fig. 4Grade 2–4 rectal morbidity probability according to levels. A: overall rectal morbidity, B: Fistula, C: proctitis, D: bleeding. Number of patients at risk, number of events and 3-year event probability are reported in tables below the plots.
Significant correlations were established between late rectal morbidity, overall and single endpoints (proctitis, bleeding, fistula) and dose–volume (, ) and dose point (DICRU) parameters. These correlations were found within a large multicenter cohort (N = 960) with prospective data collection (n = 265 rectal events, 27.6%, for four items) using different methods for evaluation, such as comparison of mean doses, logistic regression analyses (probit model) and log rank tests on actuarial estimates. The actuarial estimate evaluation showed that overpassing the so far discussed constraint of 75 Gy for the is associated in the EMBRACE cohort with a risk of approximately 30% of grade 2–4 overall rectal morbidity at 3 years, whereas ⩽65 Gy has an actuarial risk of <10% (Fig. 4) [
Dose effect relationship for late side effects of the rectum and urinary bladder in magnetic resonance image-guided adaptive cervix cancer brachytherapy.
]. If we compare the crude incidence of rectal events reported in the EMBRACE cohort with a median follow-up of 25 months (see above) with those in the so far published large institutional cohorts from Gustave Roussy (52 out of 225 patients: 23.1% with 68 events for 5 items and a median follow-up of 35 months) and Vienna (11 out of 143 patients: 7.7%, with a median follow-up of 51 months), there seems to be considerable underreporting in the large institutional cohorts [
Dose effect relationship for late side effects of the rectum and urinary bladder in magnetic resonance image-guided adaptive cervix cancer brachytherapy.
]. This is in particularly true for the retrospective Vienna cohort, which results also differed from those from a previous prospective Vienna cohort undergoing an endoscopy study (11 out of 35 patients, 31%, with rectal bleeding events) [
Correlation of dose–volume parameters, endoscopic and clinical rectal side effects in cervix cancer patients treated with definitive radiotherapy including MRI-based brachytherapy.
]. The crude data from this large multicenter EMBRACE dataset with prospective morbidity assessment and with continuous data quality assessment may therefore be regarded as the most valid and reliable data and therefore the EMBRACE based evaluations for dose and volume effects should have a prominent place for further research development and clinical practice.
The large number of patients accrued in the EMBRACE cohort enabled also the analysis of different types of rectal events individually, in particular proctitis, bleeding, stenosis and fistula as defined in CTC-AE v3.0. This allows weighting the results and potential dose constraints depending on the nature of the studied event and its severity (grade). For instance a transient rectal bleeding is probably less impacting on quality of life than a fistula that may have lifetime consequences, and may require a stoma, even with an equivalent severity (score). As rectal events are predicted using the same dosimetric parameter, it is crucial to estimate the risk of each event individually. Our data showed that a ⩾75 Gy exposes to a risk of 12.5% of rectal fistula. Conversely, maintaining the below this threshold decreases this risk substantially (<3%). These data suggest to use 75 Gy as a hard dose-constraint for the rectal in clinical practice. In addition, a ⩽65 Gy seems to be valid as a cut-off to minimize the risk of grade 2–4 proctitis and bleeding with a 2-fold decrease below this value (see above). These values have been chosen e.g. as planning aims and dose constraints for the rectal in the prospective interventional study EMBRACE II (www.embracestudy.dk). The overall grade 2–4 rectal morbidity in the subgroup of patients with between 60 and 65 Gy was 8.6% at 3 years; 3 times lower than the rate corresponding to ⩾75 Gy (26.0%).
The EMBRACE cohort also confirms the predictive value of the DICRU. Probit analyses and mean DICRU comparisons were significant within a comparable range as the and . DICRU has been successfully used in 2D-based brachytherapy reporting and prescribing for long and specific constraints using such dose points have been widely used [
Cancer of the uterine cervix: dosimetric guidelines for prevention of late rectal and rectosigmoid complications as a result of radiotherapeutic treatment.
]. In a multivariate analysis within the EMBRACE cohort, the rectal DICRU has been shown to be an independent predictor of vaginal stenosis, an area where quantitative dose–effect associations have been lacking so far [
]. Therefore, the ICRU recto-vaginal point is recommended for future clinical use serving to predict rectal and vaginal morbidity and clinical evidence for constraints has been provided [
Dose-effect relationship and risk factors for vaginal stenosis after definitive radio(chemo)therapy with image-guided brachytherapy for locally advanced cervical cancer in the EMBRACE study.
This EMBRACE analysis using the probit model also allows for assessing the ED10 with 69.5 Gy (65.2–76.9 Gy). Several values based on mono-center and retrospective series are available in the literature (Table 2) [
Dose effect relationship for late side effects of the rectum and urinary bladder in magnetic resonance image-guided adaptive cervix cancer brachytherapy.
]. The series are heterogeneous, with the use of different morbidity scoring scales, different endpoints, and different frequencies of rectal symptom reporting and are based on the evaluation of limited numbers of patients. Three series evaluated HDR brachytherapy and one PDR. Two focused on the definitive treatment of locally advanced cervical cancer by concomitant chemoradiation and IGABT (University of Vienna and Gustave-Roussy) whereas two investigated patients in a post-operative setting, or in a situation of relapses (Tata Mumbai, and BWG, Brigham and Women’s hospital). The BWH series is on patients with different gynecological primary tumors. The reported ED10 varied from 55 Gy to 78 Gy. The two series reporting on patients treated exclusively with interstitial components reported the lowest ED10 (55 and 61.8 Gy). In the retro-EMBRACE multicenter cohort, using definitive radiochemotherapy and IGABT it has been shown that the combined intracavitary and interstitial brachytherapy was not associated with an increase of the radiation-induced morbidity [
Image guided adaptive brachytherapy with combined intracavitary and interstitial technique improves the therapeutic ratio in locally advanced cervical cancer: an analysis from the retroEMBRACE study.
]. In the large EMBRACE cohort analysis the ED10 had a small confidence interval (borders: −4.3 to +7.4 Gy, Fig. 3A) which underlines its reliability and reproducibility. Subgroup analyses in patients with a longer follow-up suggest that the ED10 may evolve toward lower values (e.g. 60–65 Gy, Fig. 3B).
Table 2Effective for a 10% probability (ED10) in literature.
Dose effect relationship for late side effects of the rectum and urinary bladder in magnetic resonance image-guided adaptive cervix cancer brachytherapy.
N: number of patients, FU: follow-up, excl: exclusively; LENT-SOMA: late effects normal tissue – subjective objective management analytic, CTC-AE: common toxicity criteria for adverse event, ED10: corresponding to a 10% probability of rectal morbidity. BWH: Bragham Whomen’s hospital, IGR: Gustave-Roussy.
It is also to be noted that the was as effective as the as predictor of late rectal morbidity. The dose level of the is closer to that of the DICRU which has been used as a reference for decades (mean difference of −3.4 ± 7.1 Gy in this cohort) and less close to that of the (mean difference of 9.9 ± 6.7 Gy) [
Correlation of dose–volume parameters, endoscopic and clinical rectal side effects in cervix cancer patients treated with definitive radiotherapy including MRI-based brachytherapy.
Computed tomography-based high-dose-rate intracavitary brachytherapy for uterine cervical cancer: preliminary demonstration of correlation between dose–volume parameters and rectal mucosal changes observed by flexible sigmoidoscopy.
Dose effect relationship for late side effects of the rectum and urinary bladder in magnetic resonance image-guided adaptive cervix cancer brachytherapy.
] Therefore, Planning aims for rectal dose volume constraints are usually presented in terms of . So far evidence and therefore dose constraints for the use of is limited. Initially, it was suggested that high doses (>70–80 Gy) represented by the might be associated with local effects such as ulceration, necrosis, fistula, whereas intermediate doses (60–70 Gy), evaluated with the would be associated with fibrosis, telangiectasia, or inflammation [
Recommendations from gynaecological (GYN) GEC ESTRO working group (II): concepts and terms in 3D image-based treatment planning in cervix cancer brachytherapy-3D dose volume parameters and aspects of 3D image-based anatomy, radiation physics, radiobiology.
]. However, the findings from this EMBRACE analysis indicate that both parameters are predictive of all these rectal events. They appear highly correlated with a R2 value of 0.90, and linked by the following equation: = 0.574 x + 20.937 (Supplementary material Fig. 3). However, at an individual patient level, a major difference between the and can be caused by the use of interstitial needles in the vicinity of the rectum. Such high value of may be regarded as a rectal wall hot spot which may indicate an increased risk of a relevant clinical effect [
The findings and the analyses based on this large EMBRACE cohort allow a more accurate evaluation of dose–effect relationships for rectal morbidity and individual endpoints compared to so far published series with heterogeneous factors as indicated by larger confidence intervals. However, even if based on a large amount of patients our study has some limitations. First, only small volume dose parameters have been investigated, assumed as typical for brachytherapy related morbidity [
Recommendations from gynaecological (GYN) GEC ESTRO working group (II): concepts and terms in 3D image-based treatment planning in cervix cancer brachytherapy-3D dose volume parameters and aspects of 3D image-based anatomy, radiation physics, radiobiology.
]. However, intermediate and large volume dose parameters which also reflect the EBRT dose distribution may also be of relevance for morbidity (e.g. stenosis), and will be investigated prospectively in the new EMBRACE II protocol. Secondly, uncertainties exist in assessing the delivered dose in our study comparable to previously published studies [
Image guided adaptive brachytherapy with combined intracavitary and interstitial technique improves the therapeutic ratio in locally advanced cervical cancer: an analysis from the retroEMBRACE study.
Uncertainty analysis for 3D image-based cervix cancer brachytherapy by repetitive MR imaging: assessment of DVH-variations between two HDR fractions within one applicator insertion and their clinical relevance.
]. The prescribed doses were used to assess the dose–effect relationships, assuming that the prescribed doses are equivalent to delivered doses. It has been demonstrated, however, that in PDR brachytherapy the prescribed rectal may underestimate the delivered dose due to OAR movements and deformations during treatment delivery, whereas no significant difference has been reported in HDR series [
Uncertainty analysis for 3D image-based cervix cancer brachytherapy by repetitive MR imaging: assessment of DVH-variations between two HDR fractions within one applicator insertion and their clinical relevance.
Intrafractional organs movement in three-dimensional image-guided adaptive pulsed-dose-rate cervical cancer brachytherapy: assessment and dosimetric impact.
Intrafractional organs movement in three-dimensional image-guided adaptive pulsed-dose-rate cervical cancer brachytherapy: assessment and dosimetric impact.
]. In EMBRACE, nearly half of the patients were treated with PDR. A third limitation is that our analyses focus on only one factor: radiation dose. It seems logical that potential co-factors may significantly impact dose–effect relationships. Comorbidities (diabetes, inflammatory bowel diseases…), tobacco use, past history of abdomino-pelvic surgery, radiation techniques, concomitant chemotherapy, could for example have significant impact on this correlation. Further multivariate analyses are therefore required for a more comprehensive understanding and to develop nomograms taking into account all pertinent parameters, allowing for individualization of dose prescription consequently. These analyses have been limited so far and require a further maturation of data.
Conclusions
Significant dose–volume effect correlations between , and DICRU and the probability of late overall rectal morbidity and for single events were established based on the large prospective EMBRACE study. A ⩽ 65 Gy is associated with more minor and less frequent rectal morbidity, in particular for bleeding and proctitis, whereas a ⩾ 75 Gy is associated with more major and more frequent rectal morbidity, also with an increased risk of fistula. Multi-variate analyses taking into account cofactors of these dose–volume effects are necessary. Furthermore, prospective validation of application of these dose constraints as well as the ability to achieve them within clinical practice is necessary.
Conflict of interest statement
The authors declare no conflict of interest
Acknowledgements
EMBRACE was supported by Nucletron, an Elekta company and Varian Medical System, through grants and study sponsoring through Vienna Medical University and research grants from CIRRO – The Lundbeck Foundation Center for Investigational Research in Radiation Oncology and the Danish Cancer Society.
Appendix A. EMBRACE collaborative group
–
Department of Oncology, Aarhus University Hospital, Denmark: Jacob Lindegaard, Kari Tanderup, Lars Fokdal.
–
Department of Radiotherapy, Arnhem, The Netherlands: Elzbieta Van Der Steen Banasik.
–
Department of Radiotherapy, Gustave-Roussy, Villejuif, France: Christine Haie-Meder, Isabelle Dumas, Cyrus Chargari.
–
Department of Radiation Oncology, Leiden University Medical Center, The Netherlands: Remi A. Nout.
–
Department of Radiation Oncology, UZ Leuven, Belgium: Erik Van Limbergen.
–
Department of Radiotherapy, Institute of Oncology Ljubljana, Slovenia: Barbara Segedin, Robert Hudej.
–
Department of Radiation Oncology, Medical College of Wisconsin, Milwaukee, USA: Beth Erickson.
–
Cancer Centre, Mount Vernon Hospital, London, UK: Peter Hoskin, Gerry Lowe.
–
Department of Radiation Oncology, Tata Memorial Hospital, Mumbai, India: Umesh Mahantshetty, Jamema Swamidas, Shyam Kishore Shrivastava.
–
Department of Radiation Oncology, University Medical Centre Utrecht, The Netherlands: Ina M. Jürgenliemk-Schulz, Astrid De Leeuw.
–
Department of Radiation Oncology, Comprehensive Cancer Center, Medical University of Vienna/General Hospital of Vienna, Vienna, Austria: Christian Kirisits, Alina Sturdza, Maximilian Schmid.
Department of Oncoradiology, University of Kaposvár, Healthsciences Center, Kaposvar, Hungary: Janaki Hadjiev.
–
Leeds Cancer Centre, St James’s University Hospital, Leeds, UK: Rachel Cooper, Peter Bownes.
–
Clinic of Oncology and Women’s Clinic, St. Olavs Hospital, Trondheim, Norway: Marit Sundset.
–
Department of Oncology, The Norwegian Radium Hospital, Oslo University Hospital, Oslo, Norway: Kjersti Bruheim, Taran Paulsen Hellebust.
–
Department of Oncology, Cross Cancer Institute and University of Alberta, Edmonton, Canada: Fleur Huang, Geetha Menon.
–
Department of Radiation Oncology, Hospital of Navarra, Pamplona, Spain: Elena Villafranca.
–
Department of Radiotherapy and Oncology, Postgraduate Institute of Medical Education and Research, Chandigarh, India: Bhavana Rai, Arun S. Oinam.
–
Oncology Centre, Cambridge University Hospitals NHS Foundation Trust, Addenbrooke’s Hospital, Cambridge, UK: Li Tee Tan.
–
Department of Radiation Oncology, British Columbia Cancer Agency, British Columbia, Canada: Francois Bachand.
–
Department of Radiation Oncology, University of Iowa, Iowa, USA: Geraldine Jacobson.
–
Department of Obstetrics and Gynecology, Kuopio University Hospital, Finland: Maarit Anttila.
–
Department of Radiation Oncology Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands: Bradley Pieters.
Appendix B. Supplementary data
Supplementary Figure 1Prevalence of rectal events, reported individually. BL: baseline, m: months. Numbers of patents evaluated at each follow-up is presented in abscises, with proportion of the whole cohort. Rates according to grade and time are superimposed on the corresponding histogram.
Supplementary Figure 2Relationship between D2cm3 and grade 2–4 rectal morbidity probability based on the probit model (entire curve, compare figure 2A). Grey dashes: 95% confidence interval.
Supplementary Figure 3Relationships between D0.1cm3 and D2cm3. Each point represents a patient. Red dashes: linear trend curve. All doses are presented in 2 Gy-equivalent.
Clinical outcomes of definitive chemoradiation followed by intracavitary pulsed-dose rate image-guided adaptive brachytherapy in locally advanced cervical cancer.
Clinical efficacy and toxicity of radio-chemotherapy and magnetic resonance imaging-guided brachytherapy for locally advanced cervical cancer patients: a mono-institutional experience.
Reporting and validation of gynaecological Groupe Euopeen de Curietherapie European Society for Therapeutic Radiology and Oncology (ESTRO) brachytherapy recommendations for MR image-based dose volume parameters and clinical outcome with high dose-rate brachytherapy in cervical cancers: a single-institution initial experience.
Clinical outcome and dosimetric parameters of chemo-radiation including MRI guided adaptive brachytherapy with tandem-ovoid applicators for cervical cancer patients: a single institution experience.
Clinical outcome of protocol based image (MRI) guided adaptive brachytherapy combined with 3D conformal radiotherapy with or without chemotherapy in patients with locally advanced cervical cancer.
Recommendations from Gynaecological (GYN) GEC-ESTRO Working Group (I): concepts and terms in 3D image based 3D treatment planning in cervix cancer brachytherapy with emphasis on MRI assessment of GTV and CTV.
Recommendations from gynaecological (GYN) GEC ESTRO working group (II): concepts and terms in 3D image-based treatment planning in cervix cancer brachytherapy-3D dose volume parameters and aspects of 3D image-based anatomy, radiation physics, radiobiology.
Correlation of dose–volume parameters, endoscopic and clinical rectal side effects in cervix cancer patients treated with definitive radiotherapy including MRI-based brachytherapy.
Computed tomography-based high-dose-rate intracavitary brachytherapy for uterine cervical cancer: preliminary demonstration of correlation between dose–volume parameters and rectal mucosal changes observed by flexible sigmoidoscopy.
Dose effect relationship for late side effects of the rectum and urinary bladder in magnetic resonance image-guided adaptive cervix cancer brachytherapy.
Cancer of the uterine cervix: dosimetric guidelines for prevention of late rectal and rectosigmoid complications as a result of radiotherapeutic treatment.
Dose-effect relationship and risk factors for vaginal stenosis after definitive radio(chemo)therapy with image-guided brachytherapy for locally advanced cervical cancer in the EMBRACE study.
Image guided adaptive brachytherapy with combined intracavitary and interstitial technique improves the therapeutic ratio in locally advanced cervical cancer: an analysis from the retroEMBRACE study.
Uncertainty analysis for 3D image-based cervix cancer brachytherapy by repetitive MR imaging: assessment of DVH-variations between two HDR fractions within one applicator insertion and their clinical relevance.
Intrafractional organs movement in three-dimensional image-guided adaptive pulsed-dose-rate cervical cancer brachytherapy: assessment and dosimetric impact.
☆These results were partly presented as an oral communication at last ESTRO forum, held in Barcelona April 2015. The current study has been presented at the World Congress of Brachytherapy, held in San Francisco, June 2016.
31154-9&id=gr1.jpg){kind=link}
31154-9&id=gr2.jpg){kind=link}
31154-9&id=gr3.jpg){kind=link}
31154-9&id=gr4.jpg){kind=link}
31154-9&id=fx1.jpg){kind=link}
31154-9&id=fx2.jpg){kind=link}
31154-9&id=fx3.jpg){kind=link}