Long term experience with 3D image guided brachytherapy and clinical outcome in cervical cancer patients

Radiotherapy and Oncology xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Radiotherapy and Oncology journal homepage: www.thegreenjournal.com

Original article

Long term experience with 3D image guided brachytherapy and clinical outcome in cervical cancer patients Ivone Ribeiro a,⇑, Hilde Janssen a, Marisol De Brabandere a, An Nulens a, Dominique De Bal c, Ignace Vergote b, Erik Van Limbergen a a

Department of Radiation Oncology; b Department of Gynecologic Oncology, Leuven Cancer Institute; and c Department of Radiation Oncology, AZ. St. Maarten, Belgium

a r t i c l e

i n f o

Article history: Received 16 September 2015 Received in revised form 10 April 2016 Accepted 10 April 2016 Available online xxxx Keywords: Cervical cancer Image-guided adaptive brachytherapy (IGABT) DVH parameters Local control and late rectal morbidity

a b s t r a c t Background and purpose: To report our 10 years’ experience and learning curve of the treatment of cervical cancer patients with chemo radiotherapy and MRI (or CT in 9 selected patients) guided brachytherapy using pulsed dose rate (PDR) brachytherapy (BT). Methods and materials: Hundred and seventy consecutive patients with cervical cancer FIGO stage IB–IVB (without metastases beyond the para-aortic nodal region) were treated in our institute between 2002 and 2012. Patients received external beam radiotherapy (nodal boost to the lymph nodes positive at diagnosis) ± chemotherapy followed by a pulsed or low dose rate brachytherapy boost. MRI (or CT) images were taken with the applicator in situ. The first 16 patients were treated according to X-ray-based plans, optimized on MRI. High-risk CTV, intermediate-risk CTV, bladder, rectum and sigmoid were retrospectively contoured according to the GEC-ESTRO recommendations. In all other patients, treatment plans were optimized after delineation of the target volumes and organs at risk at MRI (or CT). Doses were converted to the equivalent dose in 2 Gy (EQD2) by applying the linear quadratic model. The median age of the patients was 55 years (range 16–88). 41% had stage III or IV disease. Of the 170 patients, 91 patients had on imaging metastatic lymph nodes at diagnosis (62 patients pelvic lymph node involvement and 29 para-aortic). In 27 (16%) patients the intracavitary technique was combined with interstitial brachytherapy. Results: The mean D90 and D100 for the high-risk CTV were 84.8 ± 8.36 Gy and 67.5 ± 6.29 Gy for the entire patient group. Mean D90 and D100 values for the IR CTV were 68.7 ± 5.5 Gy and 56.5 ± 6.25 Gy. There was an important learning curve between both patient groups, with an increase in mean D90 of 75.8 Gy for the first 16 patients compared to 85.8 Gy for the second group. At the same time, the mean dose to 2 cm3 of bladder and sigmoid decreased from 86.1 Gy to 82.7 Gy and from 70 Gy to 61.7 Gy, respectively. At a median follow-up of 37 months (range 2–136 months), local control rate for all patients was 96%, the regional control (pelvic and para-aortic) rate 81% and crude disease free survival rate 55%. The overall survival at 5 years is 65%. The higher dose to the target volume resulted in an increase in local control from 88% in the first 16 patients compared to 97% in the second patient group. Regarding late toxicity, 21 patients (12%) presented grade 3–4 late morbidity. Rectal, urinary, sigmoid and vaginal morbidity was 5%, 6%, 2% and 5%, respectively. A correlation between rectal D2 cm3 >65 Gy and grade >3 late morbidity was found (p = 0.006). Conclusion: Although the majority of the patients presented with locally advanced carcinoma, excellent local and regional control rates were achieved. Rectal, urinary, sigmoid and vaginal grade 3–4 morbidity was 5%, 6%, 2% and 5%, respectively. A correlation between rectal D2 cm3 >65 Gy and grade >3 late morbidity was found (p = 0.006). Ó 2016 Elsevier Ireland Ltd. All rights reserved. Radiotherapy and Oncology xxx (2016) xxx–xxx

The current standard treatment for locally advanced cervical cancer is a combination of external beam radiotherapy (EBRT) and chemotherapy (ChT) followed by a boost of the primary tumor ⇑ Corresponding author at: Department of Radiation Oncology, University Hospitals Leuven, Herestraat 49, 3000 Leuven, Belgium. E-mail address: [email protected] (I. Ribeiro).

with brachytherapy (BT) [1,2]. BT plays a crucial role in the treatment of cervical cancer [3,4] allowing a high dose to the tumor while relatively sparing the organs at risk. In the last decade, 3dimensional image based treatment planning was introduced for the treatment of BT for cervical cancer. Using Magnetic Resonance imaging (MRI) in BT has allowed several improvements. First of all, using MRI target volumes and organs at risk can be delineated

http://dx.doi.org/10.1016/j.radonc.2016.04.016 0167-8140/Ó 2016 Elsevier Ireland Ltd. All rights reserved.

Please cite this article in press as: Ribeiro I et al. Long term experience with 3D image guided brachytherapy and clinical outcome in cervical cancer patients. Radiother Oncol (2016), http://dx.doi.org/10.1016/j.radonc.2016.04.016

2

3D image guided brachytherapy and clinical outcome

more accurately. Secondly the dose can be optimized to increase the dose to the target and minimize the dose to the organs at risk. In 2005, the GEC-ESTRO published guidelines for MRI-based planning of intracavitary BT in cervical cancer [5,6]. Several studies have shown that MRI-based optimization resulted in an improvement of dosimetric parameters [7–12] and also clinical outcome data became available supporting the benefit of this technique. Using this image-guided conformal technique, Pötter et al. [13–15] reported 3-year pelvic control rates of 90% with a rate of serious bowel and urinary toxicity of only 2.8%. In 2011, these data were published with longer follow-up, reporting a 3-year overall local control of 95%, with a 3-year grade 3 + 4 late morbidity of 2%, 4% and 1% (bladder, gastrointestinal and vagina, respectively) [16]. We started to treat patients with MRI guided brachytherapy in our institute in 2002 using a tandem ovoid applicator and pulsed dose rate (PDR) or low dose rate (LDR) brachytherapy. In this paper we present outcome and toxicity data of 170 consecutively treated patients, related to their dose–volume histogram (DVH) parameters.

Material and methods Patient and tumor characteristics Between 2002 and 2012 a total of 170 patients with primary cervical carcinoma were treated at our institution with image guided adaptive brachytherapy (IGABT) after initial (chemo) radiotherapy. The median age of the patients was 55 years (range 16–88). The initial loco-regional staging examination included clinical examination under general anesthesia, whole body Positron Emission Tomography (PET) computed tomography (CT) and pelvic MRI. For tumors larger than 4 cm or with involved pelvic lymph nodes on PET-CT, a laparoscopic or robotic para-aortic lymph node staging was performed from the bifurcation of the aorta up to the inferior mesenteric artery. The disease was staged

according to the International Federation of Gynecology and Obstetrics (FIGO) classification and UICC TNM classification. Patients with distant metastases (except para-aortic lymph nodes) were excluded from this treatment. All included patients were treated with curative intent. The clinico-pathological features of the patients are given in Table 1. External beam radiotherapy Prior to BT, patients were treated with 45 Gy (25 fractions of 1.8 Gy) EBRT using a 3D conformal four-field box technique with 18 MV, 15 MV or 10 MV photons. 16 (9%) patients received 50 Gy (25 times 2 Gy without CT) and 8 (5%) patients with advanced bulky disease received 50.4 Gy (28 times 1.8 Gy). Concomitantly, a weekly dose of 40 mg/m2 cisplatin was administered in 143 patients (84%) aiming a total of 6 courses. 10 patients (6%) received a combination of weekly taxol and cisplatin. In 17 patients (10%) chemotherapy was not given due to poor renal function, age or other comorbidity. 18 patients (11%) with para-aortic lymph nodes were treated with neoadjuvant chemotherapy (12 patients with taxol and carboplatin and 6 with taxol, ifosfamide and cisplatinum). In 32 patients (19%) with pathological para-aortic lymph nodes, the radiotherapy field was extended to the level of the upper border of the D12 vertebra to include all the para-aortic lymph nodes. Enlarged pelvic lymph nodes were boosted to a total dose of 65 Gy in 78 patients (46%); enlarged para-aortic lymph nodes were boosted to a dose of 60 Gy in 20 patients (12%). Brachytherapy application On completion of EBRT, all patients proceeded to BT. The application was performed under general anesthesia. During clinical examination, the tumor response to EBRT was assessed, as well as the tumor topography with regard to the position of the cervical os. After hysterometry and dilatation, the MRI-compatible Nucletron Standard tandem-ovoid applicator was inserted, together with

Table 1 Patient, tumor and treatment characteristics. Age (y) (median; range)

55

(16–88)

Follow-up (months) (median; range)

132

(2.3–132.2)

FIGO stage(n° pts; % pts)

IB1 IB2 IIA IIB IIIA IIIB IVA IVB

3 17 11 70 4 26 10 29

1.8% 10.0% 6.5% 41.2% 2.4% 15.3% 5.9% 17.1%

Hystopathology (n° pts; % pts)

Squamous cell carcinoma Adenocarcinoma Adenosquamous carcinoma Other

139 19 5 7

81.8% 11.2% 2.9% 4.1%

Nodal involvement (n° pts; % pts)

N0 Pelvic nodes Pelvic and para-aortic nodes

79 62 29

46.5% 36.5% 17.1%

84

49.4%

Chemotherapy (n° pts; % pts)

Tumor width >5 cm (n° pts; % pts) Neoadjuvant Concomitant with EBRT No ChT

18 153 17

10.6% 90.0% 10.0%

Boost to lymph nodes (n° pts; % pts)

No boost Boost to pelvic LN Boost to para-aortic LN Boost to pelvic and para-aortic nodes

88 61 3 17

51.8% 35.9% 1.8% 10.0%

27

15.9%

Interstitial needles (n° pts; % pts) pts – Patients; EBRT – external beam radiotherapy; LN – lymph nodes; ChT – chemotherapy.

Please cite this article in press as: Ribeiro I et al. Long term experience with 3D image guided brachytherapy and clinical outcome in cervical cancer patients. Radiother Oncol (2016), http://dx.doi.org/10.1016/j.radonc.2016.04.016

3

I. Ribeiro et al. / Radiotherapy and Oncology xxx (2016) xxx–xxx

a rectal retractor (Nucletron, Veenendaal, the Netherlands). The vagina was packed with gauze soaked in gadolinium to increase the distance of the source positions to the bladder and rectum and to fix the applicator. In patients with large tumors and poor response to EBRT, interstitial plastic needles were placed, transperineally and/or through Utrecht applicator ovoids. After the insertion, MRI or CT images were obtained with the applicator in situ; in the first 16 patients, also X-ray images were acquired with variable gantry angles, respecting a difference in angle of at least 60° between the two images for reconstruction of the catheters. Details of the imaging and reconstruction process have been described previously by De Brabandere et al. [10]. CT scanning was performed in 9 patients and treated with LDR for logistical reasons. We included these patients in this series because they all had no locally advanced disease and a complete remission on clinical examination at the time of BT. Seven of these patients had FIGO stage IIb disease and 2 patients FIGO stage Ib. The protocol of brachytherapy delivery is described in supplements. MRI (or CT) guided 3D treatment planning Delineation procedure Three target volumes, i.e. gross tumor volume (GTV), high-risk clinical target volume (HR CTV) and intermedium risk clinical target volume (IR CTV), were delineated following the recommendations of the GYN GEC-ESTRO working group [5,6]. For the rectum, bladder and sigmoid colon the outer contour was delineated. The rectum was contoured from the anal–rectal junction to the recto sigmoid flexure. Delineation of the sigmoid colon started at the recto sigmoid and ended 2 cm above the uterus, unless the descending colon started below this level (then the sigmoid colon was delineated cranially unto the transition into descending colon). When appropriate, also small bowel loops lying in the vicinity of the uterus were delineated. Dose and volume parameters All the absorbed doses delivered to CTV and organs at risk (OAR) were converted into a radio biologically equi-effective dose of 2 Gy per fraction (EQD2). The linear quadratic model was applied, using the following tissue parameter values: a/b = 10 Gy for the tumor, a/b = 3 Gy for critical organs, with T½ (half time for sub-lethal damage repair) 1.5 h for both of them. For the three target volumes, the D90 and D100 were derived from the cumulative DVH. The dose delivered to the organs at risk was evaluated using the DVH parameters D0.1 cm3 and D2 cm3 representing the minimum doses calculated at the most irradiated 0.1 and 2 cc volumes, respectively. All doses reported below are the summed EQD2 doses of EBRT and BT. It was assumed that the BT target volumes and the high dose regions of the OAR received 100% of the prescribed EBRT dose. MRI-based optimization The first 16 patients (2002–2005) were treated using traditional X-ray-based plans according to the ICRU 38 guidelines [17] using Plato BPS version 14 (Nucletron, Veenendaal, The Netherlands) and with dosimetry plan optimization based on MRI; MRI information was used to do small modifications in case of clear underdosage of target region or severe overdosage of a critical organ. This procedure has been previously reported in detail by De Brabandere et al. [10]. Target volumes were retrospectively contoured and DVH parameters calculated. In all other patients, treatment plans were 3D-optimized [10]. Dose optimization was done manually by adjusting dwell positions and dwell times in a trial and error procedure, continuously checking the effect on the dose distribution. A strict optimization protocol was followed. In order of priority, the objectives of this protocol were: (1) to reduce

the dose to the critical organs until the dose constraints were met and (2) to increase the target dose. For the bladder a total EQD2 dose of 85–90 Gya in the 2 cm3 volume was allowed. For the rectum and sigmoid a dose of 70–75 Gy was allowed, all with a maximum pulse size of 60 cGy/h. For the HR-CTV, we aimed to achieve D90 of 85 Gy while for the IR-CTV the aim was to obtain 100% target coverage with the 60 Gy isodose, i.e. V(60 Gy) = 100%. In some cases additional dose improvement was obtained by changing the number of pulses (radiobiological optimization). Lowering the pulse size by increasing the number of pulses, results in a dose reduction that is more pronounced in the organs at risk (lower a/b ratio) than in the target volume, enabling an additional sparing of OAR. Analysis of treatment outcome Treatment failures were classified according to the site(s) of first tumoral relapse and were defined as local (cervix, uterine corpus, vagina, parametria), pelvic node, para-aortic node or distant metastases. Time intervals for disease free survival and local and regional control rates were calculated from the date of biopsy to the date of event or last follow-up appointment. Late toxicity was graded according to site and severity using the National Cancer Institute CTCAE v4.03 (Common Terminology Criteria for Adverse Events version 4.03) guidelines. Late toxicity was defined as complications present at or after 6 months from the completion of radiotherapy. Cox regression models were used to analyze the association between predictors and time-to-event outcomes. Fisher’s exact and Mann–Whitney U tests were used to evaluate patients without and with grade 3–4 toxicity. Kaplan–Meier test was used to calculate survival curves. All analyses were performed using SAS software, version 9.4 of the SAS System for Windows. Results 161 patients were treated with MRI based PDR BT and 9 patients with LDR based BT. For PDR patients, the mean number of pulses was 70 ± 12. Radiobiological optimization using an increased number of pulses was more frequently done in the second group with a mean of 70 ± 17 pulses compared to 57 ± 12 in the first group. 27 patients were treated with combined intracavitary and interstitial MRI guided PDR BT. The overall treatment time (EBRT, BT and boost to lymph nodes included) was 53.6 days (±11.6 days). Table 2 shows the histogram values obtained for all patients and differentiated for groups, for target volumes and OAR: limited optimization (first 16 patients treated between 2002 and 2005) and optimized (patients treated after 2005 with complete MRI optimization). Comparing the values of the optimized and patients with limited optimization, it is possible to

Table 2 Histogram parameters (target volumes and organs at risk). All patients (170 pts)

Limited optimization (16 pts)

Optimized (154 pts)

35.7 ± 21.0 84.8 ± 8.4 67.5 ± 6.3

48.1 ± 19.1 75.8 ± 9.0 61.5 ± 5.7

34.4 ± 20.9 85.8 ± 7.7 68.1 ± 6.0

HR CTV

Volume (cc) D90 (Gy) D100 (Gy)

IR CTV

D90 (Gy) D100 (Gy)

68.7 ± 5.5 56.5 ± 6.2

65.3 ± 4.7 55.4 ± 3.0

69.0 ± 5.5 56.6 ± 6.5

D2 cm3 (Gy) D2 cm3 (Gy) D2 cm3 (Gy)

61.7 ± 7.8 83.0 ± 8.6 62.5 ± 9.2

59.3 ± 2.8 86.1 ± 8.5 70.1 ± 12.4

62.0 ± 8.2 82.7 ± 8.5 61.7 ± 8.4

OAR

Rectum Bladder Sigmoid

HR CTV – high-risk clinical target volume; IR CTV – intermedium risk clinical target volume; OAR – organs at risk; pts – patients.

Please cite this article in press as: Ribeiro I et al. Long term experience with 3D image guided brachytherapy and clinical outcome in cervical cancer patients. Radiother Oncol (2016), http://dx.doi.org/10.1016/j.radonc.2016.04.016

4

3D image guided brachytherapy and clinical outcome

Table 3 Pattern of relapses. Relapses ? 55 patients 1st local of relapse (n° pts; % pts)

Local (only) Regional⁄ (only) Systemic (only) Regional + systemic Local + regional + systemic

3 11 24 13 4

1.8% 6.5% 14.1% 7.6% 2.4%

Total (n° pts; % pts)

Local Regional

7 19 17 41

4.1% 11.2% 10.0% 24.1%

Pelvic Para-aortic

Systemic Regional⁄ – pelvis and para-aortic lymph nodes; pts – patients.

see a better coverage of HR CTV and IR CTV with optimization of the dosimetric plans. Local Control and survival The median follow-up for surviving patients was 37 months (range 2–136 months). A total of 55 failures were observed. In Table 3 patterns of failure are shown. Local failure occurred in 7 patients: three vaginal relapses, three in the parametrium and one in the uterus. There were 2 local recurrences (13%) in the early treatment group compared to 5 (3%) in the second group. Five were simultaneously diagnosed with lymph nodes (pelvic or para-aortic) and distant metastasis. One patient with local failure was treated by an exenteration. However, she subsequently developed pelvic lymph node recurrence. Of the 7 patients who experienced local failure, one had a T3b lesion, one T3a, two had T2b and the 3 others had T2a lesion. Histological analysis revealed that one of the 7 patients with local failure had adenocarcinoma; another had a squamo-papillary carcinoma and the others (5 patients) had a squamous cell carcinoma. The median D100 (HR CTV) for the 7 patients was 66.6 Gy (range 44.3– 78.9 Gy); the median D90 (HR CTV) was 83.5 Gy(range 66.3– 98.9 Gy). The estimated local control at 3 and 5 years is 95% (Fig. 1). Pelvic lymph node recurrence occurred in 19 patients (11%) and para-aortic lymph node recurrence in 17 (10%) patients. Twentyfour patients (14%) had distant failures without local or regional

failures. The estimated regional control at 3 and 5 years is 80% and 77% (Fig. 2), respectively. At the last follow-up, 39 patients had died of their cancer and 108 patients were alive, with 94 disease free and 14 with progressive disease. Sixteen patients died of unrelated medical disease. The estimated systemic control at 3 and 5 years is 76% and 70% (Fig. 3). The local, regional and systemic control was 96%, 81% and 73%, respectively. The estimated overall survival rate, for all stages, at 3 and 5 years is 73% and 65%, respectively (Fig. 4a). The estimated overall survival for all the stages, excluding IVb (para-aortic LN) patients, at 3 and 5 years is 76% and 66%, respectively (Fig. 4b). Late toxicity 21 patients (12%) developed grade 3–4 morbidity, 15 patients (9%) toxicity only in one organ and 6 (4%) in more than one. Rectal morbidity grade 3 was observed in 9 patients (5%): 1 patient with diarrhea with necessity of colostomy, 4 patients with active bleeding and necessity of blood transfusion and 4 patients with fistulas. The mean dose to D2 cm3 to the rectum on these patients was 67.7 Gy (range of 56.7–74.9 Gy). 10 patients (6%) had urinary late toxicity grade 3, 4 with ureteral stenosis and consequent hydronephrosis with necessity of nephrostomy or cystectomy, 4 patients with fistulization and 2 patients presenting with urinary incontinence. The mean D2 cm3 to the bladder for these patients was 88.4 Gy (range of 78.3– 100.3 Gy). Of the patients with rectal and vesical fistulas (4), 2 had tumors FIGO IVA. None of the patients presented with G4 rectal or bladder toxicity. Three patients had sigmoid morbidity grade 3–4 (2%): 2 patients grade 4 (sigmoid perforation with necessity of urgent surgery for acute abdomen) and 1 patient grade 3 (fistulization). The D2 cm3 was, respectively, 70.3 Gy, 48.6 and 70.8 Gy. Two of these patients had pelvic lymph node boost (extra 16 Gy) that on posterior dosimetry plan analysis was perceived having contributed to the dose on the sigmoid. Vaginal morbidity grade 3 was present in 8 patients (5%). 5 patients had stenosis, 2 dyspareunia and one vaginal bleeding.

Fig. 1. Local control curve.

Please cite this article in press as: Ribeiro I et al. Long term experience with 3D image guided brachytherapy and clinical outcome in cervical cancer patients. Radiother Oncol (2016), http://dx.doi.org/10.1016/j.radonc.2016.04.016

I. Ribeiro et al. / Radiotherapy and Oncology xxx (2016) xxx–xxx

5

Fig. 2. Regional control curve.

Fig. 3. Systemic control curve.

Analyzing patients with or without grade 3–4 morbidity, we found a correlation between rectal morbidity and the HR CTV volume at time of BT application – the higher the HR CTV volume, the higher the probability of grade 3 rectal toxicity (p = 0.009) (Supplement material). We also found a relation between rectal toxicity Pgrade 3 and D2 cm3 P65 Gy (p = 0.006). It was not possible to find a relation between urinary toxicity and the bladder D2 cm3.

Discussion This retrospective study assessed dose–volume parameters, outcome and toxicity data in patients with cervical cancer treated by EBRT with concurrent CT followed by MRI guided PDR or LDR BT. The use of 3-dimensional (3D) imaging in the treatment planning of BT for cervical cancer has several obvious advantages over the conventional 2-dimensional (2D) approach. It allows detailed

Please cite this article in press as: Ribeiro I et al. Long term experience with 3D image guided brachytherapy and clinical outcome in cervical cancer patients. Radiother Oncol (2016), http://dx.doi.org/10.1016/j.radonc.2016.04.016

6

3D image guided brachytherapy and clinical outcome

Fig. 4. Overall survival curve.

delineation of target volumes, taking into account regression of the tumor after EBRT and organs at risk. MRI has been shown to be superior in the delineation of these volumes compared to CT [18]. In addition, the dose can be prescribed to a 3D target volume rather than to point A. As De Brabandere el al showed [10], there is no correlation between dose to point A and dose range to HR CTV, mainly because the point A dose reporting method does not take into account the tumor size of each patient individually. Furthermore, dose optimization can be performed by modifying standard loading patterns and dwell times to achieve individually shaped dose distributions. Several studies have compared dose distributions from conventional X-ray-based planning with three-dimensional treatment planning using CT or MRI. All have shown that three-dimensional optimizations can improve tumor target dose coverage while reducing the dose to OAR [7–12]. BT in our institute is delivered by PDR or LDR (in selected patients) using a tandem ovoid applicator. Since 2002 MRI and/or CT imaging is systematically performed after the application. We report here our experience in 170 patients. To assess our learning curve, we analyzed and report two different treatment periods separately: first 16 patients with treatment planning 2D and minor MRI optimization (limited optimized patients) and others with full 3D dose optimization (optimized group). Comparing to previously published data [7–12], we found that individual MRI (or CT) optimization allowed us to increase the dose to the target volumes (mean HR CTV D90: 75.8 Gy in the limited optimized patients group compared to 85.8 Gy in the optimized group) and decrease the dose in the bladder and sigmoid (Table 2). The largest benefit was seen at the mean dose to 2 cm3 of the sigmoid which decreased from 70.1 Gy to 61.7 Gy. This can be explained by the fact that in our initial phase no constraint for the sigmoid was used. The mean dose to 2 cm3 of the rectum increased from 59.3 ± 2.8 Gy to 62 ± 8.2 Gy, due to accepting 75 Gy as the dose constraint in this organ in the 3D treatment plans. Our optimized DVH parameters are comparable to published values in the literature. The mean D90 of the HR CTV reported by Pötter et al. [13] was 86 ± 16 Gy in a series of 145 patients treated with image guided HDR. In the update of these data, in 2011 [16], the mean D90 was 93 ± 13 Gy. Lindegaard et al. [19] reported for the optimized HR CTV D90 a mean of 91 Gy (range of 69–107) using PDR BT in 140 consecutive patients. Mazeron et al. [20]

reported a mean D90 of 78.1 ± 9.6 in a series 163 patients treated with PDR. Nomden et al. [21] reported a mean HR CTV D90 of 84 ± 9 Gy in a series of 46 patients treated with PDR. Our mean D90 value for the HR CTV (85.8 Gy) is lower than those reported from Vienna and Aarhus. One possible explanation is that they used interstitial needles more frequently. Interstitial needles were placed in 27 patients (16%) in our patient group compared to 44% in the Vienna series [16] and 43% in the patient group from Aarhus [19]. The number of BT applications with interstitial needles has been increasing in our department over time: of the 27 procedures, 74% of them were performed in the last 3 years (2010–2012). In 2013 and 2014, in 41% of the procedures interstitial needles were used (data not included on this paper). The average D2 cm3 values for bladder, rectum and sigmoid are comparable with the ones described by other groups [9,11,13,16,19,20]. Evaluation of treatment outcome showed an excellent local control of disease, with local treatment failure occurring in only 7 patients (4%) at a median follow-up of 37 months (range 2–132). There was an improvement in local control rate between the two patient cohorts, from 88% in the first group to 97% in the second group, showing that the improvement in the DVH parameters also resulted in a better clinical outcome. Overall local control, regional control (pelvic and para-aortic) and disease-free survival rates were 96%, 81% and 55.3% resp. The overall survival rate at 3 and 5 years was 73% and 65%, respectively. The estimated local control at 5 years is 95%. Our local control data are comparable with previously published results. In 2007, Pötter et al. [13–15] from Vienna reported the first clinical outcome data from a single institution experience of 145 patients treated with MRI-based image-guided HDR BT using a tandem-ring applicator. They found a local control rate for the true pelvis of 88% at 3 years and a cancer specific survival rate of 68%. They compared a cohort of patients treated between 1998 and 2000 with a cohort treated between 2001 and 2003 after the introduction of systematic MRI-based planning and found a major improvement in the local control of tumors >5 cm in diameter from 71% in the earlier cohort to 90% in the second cohort. They stated that a local control rate greater than 95% can be achieved if the D90 (EQD2) for HR CTV is 87 Gy or greater. At the same time, the rate for serious bowel and urinary toxicity was only 2% in the second cohort compared with 10% in the earlier cohort.

Please cite this article in press as: Ribeiro I et al. Long term experience with 3D image guided brachytherapy and clinical outcome in cervical cancer patients. Radiother Oncol (2016), http://dx.doi.org/10.1016/j.radonc.2016.04.016

I. Ribeiro et al. / Radiotherapy and Oncology xxx (2016) xxx–xxx

This group published an update of these data in 2011 [16], in a series of 156 patients. They found an overall local control at 3-years of 95%, a cancer specific survival at 3-years of 74% and an overall survival at 3 years of 68%. Mazeron et al. [20] from Institut Gustave Roussy in Paris reported their experience with PDR image-guided adaptive BT (IGABT) in 163 patients with a median follow-up of 36 months. The 3-year overall survival rate was 76%, with a local and pelvic control rate of 92% and 86%, respectively. In their series, 61 patients underwent total hysterectomy after the radiotherapy; of these, only 13 patients had macroscopic residual disease. The grade 3 or 4 morbidity reported was 7.4% Nomden et al. [21] from Utrecht reported the clinical outcome of 46 patients treated with PDR image-guided adaptive BT. With a median follow-up of 41 months, they showed a 3-year local control, progression free survival and overall survival rates of 93%, 71% and 65%, respectively. Late grade 3–4 morbidity was 9.5%. Lindegaard et al. [19] from Aarhus published their results for 140 patients treated with MRI-based IGABT. They presented an actuarial 3-year local control, pelvic control and overall survival rate of 91%, 85% and 79%, respectively. The amount of very advanced disease stage (III or higher) was 27.6% in the study of Pötter [16], 14.7% in the study of Mazeron [20], 21.7% in the study of Nomden [21], 29.3% in Lindegaard study [19] and 40.6% in our series. Despite the higher rate of advanced disease, we report an excellent local control (95.2% at 5y). The group of Paris [22] also reported their experience on LDR IGABT on a series of 84 patients. They showed a local control rate of 88% for true pelvis and only 4.7% of grade 3 complications. Of interest is that the local relapse rate in their LDR study was higher than in their PDR series [23], which might suggest that PDR for BT is superior to LDR due to its better optimization possibilities. Concerning to treatment complications, the urinary toxicity found was higher than reported by other centers [13,21]. Seven of the ten patients with genito-urinary toxicity Pgrade 3 respected the dose constraints (D2 cm3 < 90 Gy) and all of our patients followed the protocol: during all treatments, the bladder was catheterized in order to avoid volume changes. In fact, Morgia et al. [24] showed that in PDR treatments, if the bladder is catheterized throughout treatment, there are no changes in the mean bladder volume and, consequently, no differences in bladder D2 cm3. They described however small fluctuations in bladder volume in some patients, due to differences in the position of the urine collection bag relative to the bladder during the imaging. De Leeuw et al. [27] found that applicator shifts during PDR treatments, which occur mainly in ventral and cranial directions, may result in an increase in bladder dose. The data presented for the applicator shift were measured after one night. In our institution the treatment lasts on average 70 pulses (±12), which means three days of treatment, and hence applicator shifts might likely have occurred. Of the 3 patients with sigmoid toxicity grade 3–4, only 2 of them received a pelvic lymph node boost. It was not possible to find a statistical correlation between sigmoid morbidity and the boost dose to the lymph nodes. Analyzing a posteriori the EBRT boost dosimetry plans, it became clear that the boost had contributed to the sigmoid dose. The boosted lymph nodes were located at right external iliac artery and the planned CTV was close to the sigmoid, probably contributing with an extra dose to the sigmoid. The vaginal toxicity in our patients (5%) was similar to the one reported by the EMBRACE study after 2 years of follow-up [26], i.e. 3.6% probability of vaginal morbidity grade 3–4. Rectal Toxicity In this study we found 21 patients (12%) with late morbidity PG3.

7

Our rectal late morbidity (5%) is slightly higher than that reported for Vienna, Aarhus and Paris (3.2%, 3% and 5%, respectively) [16,19,20], but is lower than the one reported by the Utrecht group (9.5%) [21]. The dose constraint used was D2 cm3 <75 Gy for all the centers, except Utrecht and Vienna, that lowered the constraint for rectum and sigmoid to 70 Gy in the last years. We found a correlation between probability of late rectal morbidity and rectal D2 cm3 >65 Gy (Fisher’s exact test p = 0.006; Supplemental material). Nomden et al. [21] suggested as a possible explanation for the high rectal morbidity, even respecting the rectal dose constraints, the unpredictability of rectal movements and volume changes that may occur during PDR BT even after a regular bowel preparation. In 2013, Morgia et al. [24] presented data that supported this theory: they studied changes in HR CTV and OAR (bladder and rectum) anatomy and respective dose implications during PDR cervix treatment, performing MRI in day 1, 2 and 3 of treatment and applying the day 1 plan to the re-contoured volumes. They found no significant changes in bladder volume during the treatment and an increase in mean rectal volume. This higher volume was converted into an increase in the mean rectal dose. Thus, the rectal dosimetry met target constraints (D2 cm3 < 75 Gy) in 77% of the patients on day 1, 61% of the patients on day 2 and 67% of the patients on day 3, meaning 33% of patients on day 3 with rectal D2 cm3 >75 Gy. These data were recently supported by Mazeron et al. [25]. They analyzed the intra-fractional movements of OAR and their dosimetric impact during the delivery of PDR BT in cervical cancer and also found a significant variation on the rectum dose – an increase of 6% of the planned D2cm3 in EQD2. Thus, we can expect that, with PDR, an additional dose may be given to the rectum, suggesting that a lower dose constraint should be applied (probably D2cc <65 Gy, once it was the dose we could correlate with a bigger probability of rectal morbidity G3). So, we suggest that more studies might be needed to determine this value. Conclusion The presented data show our long-term experience with 3D IGABT in cervical cancer. A good local and regional control was observed in our series with a considerable proportion of FIGO stage III and IV tumors (41%), showing that chemo-radiation and MRIIGABT results in high local control and encouraging survival rates. The toxicity was higher than reported by other centers. We found a correlation between rectal D2 cm3 >65 Gy and risk of grade P3 late rectal morbidity. More studies are needed to determine the rectal dose constrain, and to assure if there is a difference between this value in PDR and HDR treatments. Conflict of interest statement The authors have nothing to declare. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.radonc.2016.04. 016. References [1] Vale C, Tierney JF, Stewart LA, et alChemotherapy for Cervical Cancer MetaAnalysis Collaboration. Reducing uncertainties about the effects of chemoradiotherapy for cervical cancer: a systematic review and metaanalysis of individual patient data from 18 randomized trials. J Clin Oncol 2008;26:5802–12. [2] Logsdon MD, Eifel PJ. Figo IIIB squamous cell carcinoma of the cervix: an analysis of prognostic factors emphasizing the balance between external beam

Please cite this article in press as: Ribeiro I et al. Long term experience with 3D image guided brachytherapy and clinical outcome in cervical cancer patients. Radiother Oncol (2016), http://dx.doi.org/10.1016/j.radonc.2016.04.016

8

[3]

[4]

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

3D image guided brachytherapy and clinical outcome and intracavitary radiation therapy. Int J Radiat Oncol Biol Phys 1999;43:763–75. Han K, Milosevic M, Fyles A, Pintilie M, Viswanathan AN. Trends in the utilization of brachytherapy in cervical cancer in the United States. Int J Radiation Oncol Biol Phys 2013;87:111–9. Gill BS, Lin JF, Krivak TC, Sukumvanich P, et al. national cancer data base analysis of radiation therapy consolidation modality for cervical cancer: the impact of new technological advancements. Int J Radiation Oncol Biol Phys 2014;90:1083–90. Haie-Meder C, Pötter R, Van Limbergen E, et al. 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. Radiother Oncol 2005;74:235–45. Pötter R, Haie-Meder C, Van Limbergen E, et al. 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. Radiother Oncol 2006;78:67–77. Kirisits C, Pötter R, Lang S, Dimopoulos J, Wachter-Gerstner N, Georg D. Dose and volume parameters for MRI-based treatment planning in intracavitary brachytherapy for cervical cancer. Int J Radiat Oncol Biol Phys 2005;62:901–11. Kirisits C, Lang S, Dimopoulos J, Berger D, Georg D, Pötter R. The Vienna applicator for combined intracavitary and interstitial brachytherapy of cervical cancer: Design, application, treatment planning, and dosimetric results. Int J Radiat Oncol Biol Phys 2006;65:624–30. Lindegaard JC, Tanderup K, Nielsen SK, Haack S, Gelineck J. MRI-guided 3D optimization significantly improves DVH parameters of pulsed-dose-rate brachytherapy in locally advanced cervical cancer. Int J Radiat Oncol Biol Phys 2008;71:756–64. De Brabandere M, Mousa AG, Nulens A, Swinnen A, Van Limbergen E. Potential of dose optimisation in MRI-based PDR brachytherapy of cervix carcinoma. Radiother Oncol 2008;88:217–26. Jürgenliemk-Schulz IM, Tersteeg RJ, Roesink JM, et al. MRI-guided treatmentplanning optimisation in intracavitary or combined intracavitary/interstitial PDR brachytherapy using tandem ovoid applicators in locally advanced cervical cancer. Radiother Oncol 2009;93:322–30. Tanderup K, Nielsen SK, Nyvang GB, et al. From point A to the sculpted pear: MR image guidance significantly improves tumour dose and sparing of organs at risk in brachytherapy of cervical cancer. Radiotherapy and Oncol 2010;94:173–80. Pötter R, Dimopoulos J, Georg P, et al. Clinical impact of MRI assisted dose– volume adaptation and dose escalation in brachytherapy of locally advanced cervix cancer. Radiother Oncol 2007;83:148–55. Dimopoulos JC, Lang S, Kirisits C, et al. Dose-volume histogram parameters and local tumor control in magnetic resonance image-guided cervical cancer brachytherapy. Int J Radiat Oncol Biol Phys 2009;75:56–63.

[15] Dimopoulos JC, Pötter R, Lang S, et al. Dose-effect relationship for local control of cervical cancer by magnetic resonance image-guided brachytherapy. Radiother Oncol 2009;93:311–2. [16] Pötter R, Georg P, Dimopoulos JC, et al. 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. Radiother Oncol 2011;100:116–23. [17] International Commission on Radiation Units and Measurements. ICRU report 38: dose and volume specification for reporting intracavitary therapy in gynaecology. Bethesda, MD: ICRU; 1985. [18] Viswanathan AN, Dimopoulos J, Kirisits C, Berger D, Pötter R. Computed tomography versus magnetic resonance imaging-based contouring in cervical cancer brachytherapy: results of a prospective trial and preliminary guidelines for standardized contours. Int J Radiat Oncol Biol Phys 2007;68:491–8. [19] Lindegaard JC, Fokdal LU, Nielsen SK, Juul-Christensen J, Tanderup K. MRIguided adaptive radiotherapy in locally advanced cervical cancer from a Nordic perspective. Acta Oncol 2013;52:1510–9. [20] Mazeron R, Gilmore J, Dumas I, et al. Adaptive 3D image-guided brachytherapy: a strong argument in the debate on systematic radical hysterectomy for locally advanced cervical cancer. Oncologist 2013;18:415–22. [21] Nomden CN, de Leeuw AA, Roesink JM, et al. 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. Radiother Oncol 2013;107:69–74. [22] Haie-Meder C, Chargari C, Rey A, Dumas I, Morice P, Magné N. MRI-based low dose-rate brachytherapy experience in locally advanced cervical cancer patients initially treated by concomitant chemoradiotherapy. Radiother Oncol 2010;96:161–5. [23] Chargari C, Magné N, Dumas I, et al. Physics contributions and clinical outcome with 3D-MRI-based pulsed-dose-rate intracavitary brachytherapy in cervical cancer patients. Int J Radiat Oncol Biol Phys 2009;74:133–9. [24] Morgia M, Cuartero J, Walsh L, et al. Tumor and normal tissue dosimetry changes during MR-guided pulsed-dose-rate (PDR) brachytherapy for cervical cancer. Radiother Oncol 2013;107:46–51. [25] Mazeron R, Champoudry J, Gilmore J, et al. Intrafractional organs movement in three-dimensional image-guided adaptive pulsed-dose-rate cervical cancer brachytherapy: assessment and dosimetric impact. Brachytherapy 2015;14:260–6. [26] Kirchheiner K, Nout RA, Tanderup K, et al. Manifestation pattern of early-late vaginal morbidity after definitive radiation (chemo)therapy and image-guided adaptive brachytherapy for locally advanced cervical cancer: an analysis from the EMBRACE study. Int J Radiat Oncol Biol Phys 2014;89:88–95. [27] De Leeuw AA, Moerland MA, Nomden C, Tersteeg RH, Roesink JM, JürgenliemkSchulz IM. Applicator reconstruction and applicator shifts in 3D MR-based PDR brachytherapy of cervical cancer. Radiother Oncol 2009;93:341–6.

Please cite this article in press as: Ribeiro I et al. Long term experience with 3D image guided brachytherapy and clinical outcome in cervical cancer patients. Radiother Oncol (2016), http://dx.doi.org/10.1016/j.radonc.2016.04.016