- Email: [email protected]
Characterization of implant displacement and deformation in gynecologic interstitial brachytherapy
Brachytherapy 13 (2014) 100e109
Characterization of implant displacement and deformation in gynecologic interstitial brachytherapy Antonio L. Damato*, Robert A. Cormack, Akila N. Viswanathan Department of Radiation Oncology, Brigham and Women’s Hospital and Dana-Farber Cancer Institute, Boston, MA
ABSTRACT
PURPOSE: To determine the uncertainties in implant position during multifraction gynecologic interstitial brachytherapy, we analyzed the interfraction displacements and deformations of gynecologic interstitial implants. METHODS AND MATERIALS: Fourteen gynecologic patients treated with multifraction highdose-rate interstitial brachytherapy received two CT scans each at the time of implantation and 48e72 h later. Rigid fusions on the pubic symphysis were performed. This analysis included catheter shifts in the cranial (CR), caudal (CA), anterior, posterior, left, and right directions; template shifts; the change in the catheter length measured along the path from catheter tip to catheter connector (offset); the change in relative distances between catheters (deformations); and changes in rectum and bladder D2cc and tumor D90. RESULTS: Of the 198 catheters analyzed, the number of catheter shifts (%) and mean standard deviation were 43% CA (5.0 2.0 mm), 22% CR (7.9 4.0 mm), 14% anterior (6.3 2.1 mm), 48% posterior (8.7 3.1 mm), 7% left (4.8 0.4 mm), and 9% right (5.4 0.9 mm). Catheter offsets were 3% CA (7.2 6.3 mm) and 11% CR (6.1 2.6 mm). Template shifts were 43% CA (5.2 1.6 mm) and 14% CR (6.6 4.0 mm). Deformations were 10 shrinkages (4.7 0.9 mm) and 32 expansions (4.7 0.5 mm). Dosimetric changes were 5.2% 10.8% for rectum D2cc, 1.1% 18.5% for bladder D2cc, and 5.1% 6.7% for tumor D90. CONCLUSIONS: On average, less than 1 cm displacements and deformations of the implant occurred over the course of treatment. Proper quality assurance methodologies should be in place to detect shifts that can potentially result in inadvertent insertion into normal tissue. Ó 2014 American Brachytherapy Society. Published by Elsevier Inc. All rights reserved.
Keywords:
Brachytherapy; Gynecologic; Interstitial; Displacement; Uncertainty; Catheter
Introduction The use of interstitial brachytherapy for the treatment of advanced and recurrent gynecologic malignancies with significant vaginal and/or sidewall extensions increases local control with an acceptable risk of morbidity compared with external beam alone (1). Multiple implantation methodologies for gynecologic interstitial brachytherapy have been described in the literature (2e6). Templates with preset hole patterns are used for needle placement and to help with guidance and quality assurance (QA). The corners of the Received 16 May 2013; received in revised form 30 August 2013; accepted 25 September 2013. * Corresponding author. Department of Radiation Oncology, Brigham and Women’s Hospital, Dana-Farber Cancer Institute, 75 Francis St, ASB1-L2, Boston, MA 02115. Tel.: þ1-617-525-7242; fax: þ1-617732-7347. E-mail address: [email protected] (A.L. Damato).
template are sutured to the perineal skin for stability. The two most common template systems are the Syed-Neblett template (SNT) (2) and the Martinez Universal Perineal Interstitial Template (MUPIT) (3). Plastic catheters and hollow titanium needles (4) have been used. Custom-made templates and catheters inserted ‘‘free hand (FH)’’ into the patient may also be used. Catheters can be secured in place with a variety of anchoring techniques, such as buttons or liquid adhesive. Most multifractionated interstitial brachytherapy treatments require the hospitalization of the patient for the duration of the treatment; some ambulatory techniques have been proposed and used for gynecologic treatments (5). A single implant can be used for a treatment that extends from 1 to 5 days. Significant uncertainties may exist related to the catheter/needle anchorage to the template or the patient’s skin, the stability of the template, and deformations of the patient anatomy.
1538-4721/$ - see front matter Ó 2014 American Brachytherapy Society. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.brachy.2013.09.010
A.L. Damato et al. / Brachytherapy 13 (2014) 100e109
An assessment of the interfraction variability of catheter location is essential to understand the uncertainties in multifraction gynecologic interstitial brachytherapy and can be used to inform planning and QA policies. A substantial literature exists (7e14) on implant displacements for interstitial prostate brachytherapy. However, insights that prostate literature may provide for gynecologic applications may be limited. Differences in tumor invasion into organ structures, imaging methodology, and applicators used in prostate and gynecologic interstitial brachytherapy procedures may result in significant differences in implant geometry, stabilization, and deformation. Implant displacement has been studied for gynecologic ambulatory procedures (15) and for cervical cancer patients receiving interstitial brachytherapy through a MUPIT (16). In this work, we analyze the variability of gynecologic interstitial implants recently performed according to the standard practice in our clinic. To our knowledge, this is the first detailed investigation of implant displacements and deformations in gynecologic interstitial brachytherapy using SNT.
Methods and materials A data set of 28 CT scans from 14 patients treated from March 2011 to December 2011 with high-dose-rate (HDR) interstitial gynecologic brachytherapy in our institution was retrospectively analyzed with institutional review board approval. All patients received a CT scan immediately after implantation and again 2 (n 5 11 patients) or 3 (n 5 3) days after implantation. All patients had vaginal extension of gynecologic malignancy necessitating interstitial treatments. One patient did not have a rectum because of a history of rectal cancer.
101
Implants All patients were implanted under general anesthesia. Three implantations were performed using real-time MRI guidance from a 3 T wide-bore unit, with a CT scan acquired immediately after implantation. All the other implantations were performed in a CT brachytherapy suite under CT guidance. A vaginal obturator (VO) was inserted through the center of an SNT (Best Medical International Inc., Springfield, VA) (Fig. 1). Plastic ProGuide catheters (Nucletron, an Elekta company, Elekta AB, Stockholm, Sweden) were implanted with a central stainless steel removable insert to ensure stability during the insertion. Of the 198 total catheters implanted in all patients, 7 were inserted FH, 75 through grooves in the central VO, and 112 through holes in the SNTs. In 4 patients with an intact uterus, a central tandem was inserted. After the implant, the central VO and the VO and SNT catheters were secured to the template using a sterile liquid adhesive, and the template was sutured in four corners to the perineal skin. FH catheters were secured to the patient’s skin through buttons when possible. The extent of each catheter’s entry into the template or the patient body was marked on the catheter for pretreatment visual inspection. All patients received one fraction on the day of implantation and an additional four to eight fractions twice daily. The doses from the brachytherapy and prior radiation were summed via the EQD2 (equivalent dose 2 Gy) formalism. Dose optimization and fractionation scheme selection had the goal of achieving a combined clinical target volume (CTV) D90 between 70 and 80 Gy, while respecting limits on the combined D2cc metrics for bladder (!90e95 Gy), rectum (!70e75 Gy), and sigmoid (!70e75 Gy). The fractionation schemes were prescribed by the physician on a patient-by-patient basis based on tumor, normal tissue, and patient factors and are reported in Table 4.
Fig. 1. Graphical representation of a Syed-Neblett template with holes for catheter insertion, vaginal obturator with grooves for catheter insertion, and catheter inserted into a tumor.
102
A.L. Damato et al. / Brachytherapy 13 (2014) 100e109
Imaging CT scans for all the patients were obtained in a dedicated brachytherapy suite equipped with a CT scanner. The slice thicknesses were 1.25 mm (20 scans), 2 mm (2 scans), and 2.5 mm (6 scans). Copper dummy markers were inserted in all catheters for every scan for easy tip identification. Scans were performed with the legs down. This permitted accurate catheter localization and reconstruction. All patients received a first scan immediately after implantation, before being extubated and transferred to the postoperative recovery area. The first treatment occurred on the day of implantation, subsequent treatments occurred twice a day with 6 h between fractions. All other imaging was performed immediately before a treatment: 11 scans before fraction 4 and 3 scans before fraction 6. Visual inspection of catheter entry level into the template and eventual corrections were performed after imaging. Fusion For each patient, the two CT scans were fused using the Eclipse Treatment Planning System (Varian Medical Systems, Inc., Palo Alto, CA) registration tool (Varian Rigid Registration Algorithm 10.0.28). The first scan was taken as the reference coordinate system, and each subsequent scan was registered on the reference coordinate system based on a rigid fusion of the pubic symphysis. Fusions were automatically performed on pixel data prioritized to the pubic symphysis only, in a volume of interest excluding the template, femoral heads, and the superior portion of the bladder. Visual inspection confirmed excellent fusion of the pubic symphysis, with uncertainty associated with the rigid registration of the pubic symphysis assumed to be !1mm. The quality of fusion of other organs varied depending on organ filling and organ motion that may have occurred between scans. Fused sets were used to compare implant position across CT scans in a common coordinate system. Catheter reconstruction Catheters were reconstructed in Oncentra Brachy (Nucletron, an Elekta company) by an experienced brachytherapy planner. The first digitization point (connector end) was positioned on the most superior slice of the template. Catheter location was highlighted in the scan by the presence of a copper dummy marker. Implant displacement Catheter displacement Catheter displacement refers to the displacement of the tip of a catheter between scans. Catheter displacement is subdivided into shifts in the cranial (CR), caudal (CA), anterior (A), posterior (P), left (L), and right (R) directions. The A direction is defined as toward the bladder and the P direction as toward the rectum. L and R are defined as
toward the left and right of the patient. Displacements can arise from a variety of causes including malfunction of the catheter attachment to the template, displacement of the template, and deformation of tissues. To gain better insight into the cause of displacement, we measured two primary parameters: catheter offset and template shift (Fig. 2). Catheter offset Catheter offset is defined as the change in the catheter length measured along the path from catheter tip to catheter connector (i.e., within the patient), likely resulting from inadequate anchorage of the catheter to the template. A shorter catheter path is scored as a CA offset and a longer catheter path as a CR offset. Template shift Template shift is the average CR/CA shift of the catheters’ connector ends between scans. Template shift represents the overall displacement of the template/implant in relation to the pubic symphysis because of perineal tissue swelling, handling of the patient during hospitalization, and pressure on the implant due to patient movement or other anatomic changes. The slice thickness of the CT scan imposes a lower limit on the size of observable CR/CA displacements. In this work, we consider a CR/CA displacement to be observable when it is greater than twice the slice thickness of the underlying CT scan. For A/P and L/R displacements, the limiting factor is the accuracy of the manual point digitization. We assume an accuracy of 2 mm, taking into account the size of the dummy marker on CT and rare but possible confounding factors like calcifications or nearby reconstruction artifacts from possible implanted clips; a shift greater than 4 mm is therefore considered observable. Throughout this manuscript, mean and standard deviation values are reported as mean standard deviation. Implant deformation Implant deformation is defined as the change in relative distances between catheters. More specifically, implant deformation is the change in relative distance between the points where any two given catheters intersect the most superior plane perpendicular to the template that contains all the catheters of the patient (i.e., just before the tip of the most shallow catheter of the implant). This definition is insensitive to catheter offset and is designed to evaluate deformations due to implant expansion (X) from edema or implant shrinkage (S) from tissue compression and/or tumor response. Similarly to A/P and L/R displacements, we consider a deformation observable if it is greater than 4 mm. Dosimetry The rectum and bladder were contoured on both CTs. The dosimetry from the clinical treatment plan was calculated based on catheter locations on both scans, and D2cc metrics as a percentage of the prescription dose are reported.
A.L. Damato et al. / Brachytherapy 13 (2014) 100e109
103
Fig. 2. Graphical representation of the catheter displacement, template shift, and catheter offset. On the top half (Shift A), we show a net caudal catheter shift displacement (solid arrow), resulting from a caudal template shift (dashed line). On the bottom half (Shift B), we show a net cranial catheter displacement (solid arrow), resulting from the superposition of the same caudal template shift (dashed line) and a cranial catheter offset (dotted line). The net effect of offsets and template shifts may introduce both cranial and caudal shifts in the same implant.
Because of uncertainties in contouring of CTV on CT, and the unavailability of magnetic resonance (MR) images at the time of the secondary scan for the 3 patients who were implanted under MR guidance, CTV recontouring was not performed. Changes in CTV D90 due to catheter shift were evaluated based on the difference between the dosimetry at time of planning and the dosimetry resulting from a redigitization of the catheters, reflecting the CR/ CA component of their shift. A/P and L/R shifts were not considered under the assumption that the CTV and the catheters will translate and deform equally in those directions. A limit of this approach is that possible shift and deformation of the CTV in the same CR/CA direction of the implant shift is not considered, and the changes in CTV D90 thus calculated may overestimate the actual change in CTV D90.
Statistics A two-tailed paired Student t test was performed to assess statistical significance of the dosimetric changes. A p-value of !0.05 was associated with statistical significance.
Results A total of 198 catheters in 14 implants were analyzed. The median number of catheters per implant was 15 (range, 9e23).
Implant displacement CR/CA implant displacement Table 1 lists the summary of the observable CR/CA catheter shifts and offsets on a patient-by-patient basis and as an average of the mean value per patient among those who experienced an observable shift in that direction. The number of patients experiencing observable shifts was 8 for template shifts, 13 for catheter shifts, and 8 for offsets. Table 2 lists the summary of the CR/CA catheter shifts and offsets per catheter type. A total of 129 CR/CA catheter shifts (65%) were observed, of which 43 were CR (22%) (9 shifts! 5 mm, 20 shifts between 5 and 10 mm, and 14 shiftsO10 mm) and 86 were CA (43%) (51 shifts!5 mm, 33 shifts between 5 and 10 mm, and 2 shifts O10 mm). Twenty-seven catheter offsets (14%) were observed, of which 21 were CR (11%) (10 offsets !5 mm, 9 offsets between 5 and 10 mm, and 2 offsets O10 mm) and 6 were CA (3%) (4 offsets !5 mm, 1 offset between 5 and 10 mm, and 1 offset O10 mm). A total of 20 of the 75 catheters in the VO group (27%) experienced an observable offset in any direction compared with 2 of the 112 catheters in the SNT group (2%). A total of four of the seven catheters in the FH group (57%) experienced an observable offset in any direction. Figure 3 summarizes the components of the CR/CA implant displacements. The strong correlation between template shift and mean catheter shift per patient (1.05 linearity, R2 5 0.946) suggests that rigid template shift was the main contributor to the overall catheter shift. In addition to the rigid displacement driven by the template shift, the offset contributed to isolated large shifts of catheters.
104
A.L. Damato et al. / Brachytherapy 13 (2014) 100e109
Table 1 Summary of CR and CA implant displacements per patient Patient
Template shift (mm)
Mean catheter shift CR/CA (mm)
CR shifts (%)
CA shifts (%)
A B C D E F G H I J K L M Na
CA CA CA CA CA CA Not Not Not Not Not Not CR CR
d/6.7 d/5.5 d/5.1 d/4.8 d/3.7 5.6/3.7 d/3.3 12.5/16.3 d/d 3.6/d 6/d 5.6/3.6 5.5/d 11.9/d
d d d d d 5/12 (42) d 1/13 (8) d 2/10 (20) 1/9 (11) 8/16 (50) 11/11 (100) 15/15 (100)
9/15 21/23 17/17 11/16 13/15 5/12 8/16 1/13
Average
5.9 6.8 7.0 3.7 4.5 3.4 observable observable observable observable observable observable 3.7 9.4
(60) (91) (100) (69) (87) (42) (50) (8) d d d 1/16 (6) d d
Mean catheter offsets CR/CA (mm)
CR offsets (%)
d/d 4.0/6.7 d/d 3.3/4.5 d/d 7.7/d d/2.9 10.5/19.4 d/d 3.7/d d 5.4/d d/d 7.0/d
2/23 1/16 6/12 1/13 2/10 7/16 2/15
d (9) d (6) d (50) d (8) d (20) d (44) d (13)
CA offsets (%) d 2/23 (9) d 1/16 (6) d d 2/16 (13) 1/13 (8) d d d d d d
Template shift (mm)
Mean template shift (mm)
Mean catheter shift (mm)
Mean offset (mm)
CR CA
6.6 4.0 5.2 1.6
7.2 3.5 5.9 4.1
6.0 2.6 8.4 7.5
CR 5 cranial; CA 5 caudal. CR (CA) average values are across patients who have experienced at least one CR (CA) shift. Averages are expressed as mean standard deviation. a Patient without a rectum.
CR implant displacement subset analysis Of the 14 CR shifts exceeding 10 mm, 13 occurred in the 1 patient who did not have a rectum. In subset analysis, omitting this case, the remaining 13 patients had the following CR displacements: observable CR catheter shifts occurred in 6 patients (46%) (6.5 3.1 mm; range, 3.6e12.5 mm),
observable CR catheter offsets occurred in 6 patients (46%) (5.8 2.8 mm; range, 3.3e10.5 mm), and 1 CR template shift of 3.7 mm was observed. Among the 183 catheters in this subset, 28 CR catheter shifts (15%) were observed (5.7 2.2 mm; range, 2.7e12.5 mm; 9 shifts !5 mm; 18 shifts between 5 and
Table 2 Summary of CR and CA implant displacements per catheter insertion site Catheter shift Site Total Number of displacements Mean (mm) Range (mm) Subset Number of events Mean (mm) Range (mm) VO Number of displacements Mean (mm) Range (mm) SNT Number of displacements Mean (mm) Range (mm) FH Number of displacements Mean (mm) Range (mm) Tandem Number of displacements Mean (mm) Range (mm)
Catheter offset
CR (%)
CA (%)
CR (%)
CA (%)
43/198 (22) 7.9 4.0 2.7e20.0
86/198 (43) 5.0 2.0 2.6e16.3
21/198 (11) 6.1 2.6 3.0e12.4
6/198 (3) 7.2 6.3 2.7e19.4
28/183 (15) 5.7 2.2 2.7e12.5
86/183 (47) 5.0 2.0 2.6e16.3
19/183 (11) 6.0 2.7 3.0e12.4
6/183 (3) 7.2 6.3 2.7e19.4
24/75 (32) 6.6 3.1 2.7e12.5
22/75 (29) 5.5 2.8 3.5e16.3
17/75 (23) 5.7 2.4 3.0e10.5
3/75 (4) 9.0 9.0 3.2e19.4
16/112 (14) 8.7 2.9 5.3e16.2
59/112 (53) 4.6 1.4 3.0e8.0
1/112 (1) 12.4 12.4
1/112 (1) 2.7 2.7
2/7 (29) 10.9 12.9 1.8e20.0
4/7 (57) 7.2 3.0 4.4e10.3
2/7 (29) 7.0 1.1 6.2e7.8
2/7 (29) 6.8 2.9 4.7e8.8
1/4 (25) 3 3
1/4 (25) 5.8 5.8
1/4 (25) 4.7 4.7
d d d
Subset 5 excluding patient without a rectum; CR 5 cranial; CA 5 caudal; VO 5 vaginal obturator; SNT 5 Syed-Neblett template; FH 5 free hand. Average values are expressed as mean standard deviation.
A.L. Damato et al. / Brachytherapy 13 (2014) 100e109
105
Fig. 3. Relationship between template shifts and mean catheter shifts for each patient. Range of observable catheter offset is marked with a slashed line. Only patients with an observable template shift are plotted and contributed to the trendline. Nonobservable region is marked in gray. Cranial shifts have the positive sign, and caudal shifts have the negative sign.
10 mm; and 1 shift O10 mm). Nineteen CR offsets (10%) were observed (6.0 2.7 mm; range, 3.0e12.4 mm; 10 offsets !5 mm; 7 offsets between 5 and 10 mm; and 2 offsets O10 mm). A/P catheter shifts A total of 123 A/P catheter shifts (62%) were observed. All 14 patients experienced at least one A/P catheter displacement. Twenty-seven A shifts (14%) (6.3 2.1 mm; range, 4.3e12.0 mm) were observed in 6 patients. Ninety-six P shifts (48%) (8.7 3.1 mm; range, 4.1e20.8 mm) were observed in 9 patients.
L/R catheter shifts Twenty-nine L/R catheter shifts (15%) were observed in 10 patients. Twelve L shifts (6%) (4.8 0.4 mm; range, 4.2e5.8 mm) were observed in 4 patients. Seventeen R shifts (9%) (5.4 0.9 mm; range, 4.1e7.1 mm) were observed in 6 patients. Implant deformation Table 3 summarizes the observable deformations seen in each patient. Because deformation is measured at different depths, we report the depth used for each patient. Forty-two
Table 3 Summary of implant deformations S deformations
X deformations
Patient
Measurement depth (mm)
Number of deformations/number of possible deformations (%)
Mean (mm)
Range (mm)
Number of deformations/number of possible deformations (%)
Mean (mm)
Range (mm)
A B C D E F G H I J K L M Na
80 75 95 135 110 80 120 80 75 60 110 70 75 70
4/105 d 4/136 d 1/105 d d d d d 1/36 d d d
4.4 0.3 d 5.2 1.3 d 4 d d d d d 4.2 d d d
4.2e4.8 d 4.0e6.5 d 4 d d d d d 4.2 d d d
d d 7/136 16/120 d d 1/120 d d d d d d 8/105
d d 4.7 0.5 4.7 0.5 d d 4.2 d d d d d d 4.9 0.7
d d 4.1e5.4 4.1e5.5 d d 4.2 d d d d d d 4.0e6.0
(4) (3) (1)
(3)
(5) (13)
(1)
(8)
S 5 shrinkage deformation (tumor response/tissue compression); X 5 expansion deformation (edema). a Patient without a rectum.
106
A.L. Damato et al. / Brachytherapy 13 (2014) 100e109
observable implant deformations were observed. Ten S deformations (4.7 0.9 mm; range, 4.0e6.5 mm) were observed in 4 patients. No patient presented with a number of S deformations exceeding 5% of the number of possible permutations (N (N 1)/2, where N is the number of catheters in a patient). Thirty-two X deformations (4.7 0.5 mm; range, 4.0e6.0 mm) were observed in 4 patients. Three patients presented with a number of X deformations exceeding 5% of the number of possible permutations (one exceeding 10%). Dosimetry Table 4 summarizes the dosimetric changes that have occurred between scans. The average change in CTV D90 between scans was 5.1 6.7% of the prescription dose, from an average of 100 26% to an average of 95 25%. In three cases, a difference in D90 exceeding 10% of the prescription dose was observed. The change was statistically significant ( p ! 0.02). Changes in rectum and bladder D2cc between scans have been observed. The average change in D2cc between scans was þ5.2 10.8% of the prescription dose for rectum ( p 5 0.09) and 1.1% 18.5% of the prescription dose for bladder ( p 5 0.82). These differences are likely because of anatomic differences between scans due to a change in organ filling. Clinical results No perforations or other adverse events occurred for the 14 patients considered in this study. In three instances, a verification of the dosimetry on the QA CT scan was deemed
advisable because of implant displacement. In one instance, a replan was deemed appropriate because of an increase in rectal dose. This work did not investigate the origin of this dosimetric change, although it seems likely that the main contributing factor was a change in rectal filling. Discussion We observed that a majority (129/198, 65%) of catheters experienced an observable CR or CA shift between the planning scan and the second day of treatment. Most shifts were small (86 CA shifts with a mean of 5.0 mm and 43 CR shifts with a mean of 7.9 mm), but shifts as large as 16.3 mm in the CA direction and 20.0 mm in the CR direction were observed. These results show that template shifts and shifts of single catheters relative to the template may occur, and a QA methodology to monitor such shifts should be implemented. Possible methodologies include visual inspection of catheters before each fraction, aided by the marking on the catheter of the insertion length at the time of the planning scan. Moreover, visual inspection of the position of the template relative to the perineal area should be performed, aided by pictures of the area taken at the time of the planning scan. When clinically meaningful shifts are suspected, corrective actions or confirmation through CT scan may be considered. Routine CT scans may not be appropriate as patient movement to the CT table may possibly contribute to the implant shifts. Moreover, when a CT scan is used to evaluate implant shifts, care should be taken in evaluating the measuring uncertainties that depend on clinic-specific imaging protocols and to allocate resources for scanning and the timely catheter
Table 4 Summary of dosimetric results
Patient A B C D E F G H I J K L M Na Average
Prescription dose (cGy) number of fractions 350 500 400 750 325 450 450 300 425 500 550 450 350 450
9 6 7 5 7 5 9 9 7 5 5 7 6 6
CTV D90 (% of prescription dose)
R/B D2cc (% of prescription dose)
Planning
Re-scan
Difference
Planning
Re-scan
Difference
74.7 129.6 114.3 124.9 98.4 54.1 122.5 64.2 106.1 74.5 115.7 126.6 75.3 119.3 100.0 26.1
76.3 109.9 108.2 125.4 86.9 52.1 119.5 57.0 106.6 76.8 117.1 121.6 67.0 104.7 94.9 25.0
þ1.6 19.7 6.1 þ0.4 11.5 2.0 3.0 7.1 þ0.4 þ2.3 þ1.4 5.0 8.3 14.6 5.1 6.7 p ! 0.02
52.1/114 43.6/50.3 78.3/102 68.9/81.1 68.5/79.9 63.2/62.2 75.9/74.7 67.6/78.5 41.3/70.2 87.7/83.5 85.3/117 47.9/55.1 68.4/84.5 N/A/48.8 R: 60.6 22.7 B: 78.2 21.4
65.0/101 47.2/52.3 75.2/88.3 95.7/87.1 81.3/96.0 53.9/61.7 67.9/115 64.9/57.6 40.8/63.2 93.3/90.3 98.0/107 48.4/56.6 90.0/48.7 N/A/53.9 R: 65.8 26.8 B: 77.1 23.0
þ12.9/12.7 þ3.6/þ2.1 3.1/13.7 þ26.8/6.0 þ12.8/þ22.1 9.3/0.4 8.0/þ40.7 2.7/20.9 0.4/7.0 þ5.6/þ6.8 þ12.7/9.3 þ0.5/þ1.5 þ21.6/35.8 N/A/þ5.1 R: þ5.2 10.8 p 5 0.09 B: 1.1 18.5 p 5 0.82
CTV 5 clinical target volume; R 5 rectum; B 5 bladder; N/A 5 not applicable. Average values are expressed as mean standard deviation. a Patient without a rectum.
A.L. Damato et al. / Brachytherapy 13 (2014) 100e109
redigitization, plan generation, and evaluation, which may also require recontouring. The use of interstitial brachytherapy for the treatment of gynecologic malignancy was first reported in 1913 by Abbe (6). More recently, the use of three-dimensional image guidance to aid with implantation and treatment planning was proposed using CT (17) and MR (18, 19), in which localization of the catheters in relation to the tumor and the organs at risk (OARs, defined as the bladder, rectum, sigmoid, and small bowel) is used to achieve optimal dosimetry. In this analysis, we demonstrate that interfraction displacement of the catheters occurs preferentially in the CR, CA, and P directions. The displacements may introduce uncertainty in the delivered dosimetry. In the CR/CA axis, we observed the largest number of shifts in the CA direction (86/198, 43%): Most (51) were smaller than 5 mm, with a single shift as large as 16.3 mm occurring. The average mean CA shift observed (5.9 mm) in this study is lower than what is observed in prostate brachytherapy (from 7.6 to 20 mm) (7e12) and gynecologic brachytherapy using the MUPIT (17.4 mm) (16). A total of 43/198 CR shifts (22%) were observed in this study, with a mean shift of 7.9 mm and a single maximum shift of 20.0 mm. The CR shifts observed were larger and more frequent than what has been observed in prostate brachytherapy, in which most authors report not observing any CR shifts (7e9). The CR shifts observed in this study were also higher than what is reported (15) in ambulatory uterine cancer treatment and in MUPIT implantations (mean, 2.5 mm; range, 0e7.4 mm) (16). Differences in measurement methodology (referenced to the fusion of the pubic symphysis in this study, to a relocatable bony point as in ref. (16), and to the center of mass of implanted markers as in ref. (15)) and differences in implantation technique (using SNT in this study, MUPIT as in ref. (16), and a thin vinyl plate as in ref. (15)) are possible explanations of the variability of the results. Notably, 15/43 CR shifts (35%) occurred in the 1 patient without a rectum. This patient also accounted for 13 of the 14 CR shifts exceeding 1 cm and for the maximum CR shift of 20.0 mm. A subset analysis was done omitting this patient: The mean CR shift was 5.7 mm, with a single maximum shift of 12.5 mm, suggesting that the rectum may act as a stabilizer in the perineal tissue of patients with a vaginal interstitial implant. Patients requiring vaginal interstitial brachytherapy who do not have a rectum may benefit from daily imaging, given the potential for a large CR shift. The observed CR/CA catheter shifts may result in significant dosimetric changes, as shown by the observed changes in CTV D90 in this study. Similarly, the literature on prostate brachytherapy suggests that significant loss of coverage occurs for CA displacements of the order seen in this study. A study on change in CTV dosimetry (19), in which the CTV was delineated on three MRs taken daily in the first 3 days of treatment, shows an average decrease of CTV D90 of 5.7%. Despite the different methodology
107
(redelineation of the CTV on repeated MR (19) vs. the calculation of the effect of CR/CA shift on CTV as done in this study), this result is in line with the results of our work. Further studies are needed to assess the clinical efficacy of possible corrective actions, such as adjusting planning methodologies to minimize the effect of shifts, correcting the shifts when they occur, or replanning during treatment. Current data on clinical end points such as OAR toxicity and local control probability for multifraction gynecologic interstitial brachytherapy already account for average changes in dosimetry due to catheter shift. QA methodologies for the detection of large dosimetric changes, which in this study were up to 19.7% of the CTV D90, need to be developed and implemented. Swelling of the peritoneal area and handling of the patient during hospitalization are possible causes of the rigid CR/CA implant shifts. In the analysis of the offsets, that is the relative motion between the catheter and the template, contrary to some of the prostate brachytherapy literature (10) in which no offset in any direction has been reported, we observed infrequent (27/198, 14%) but large (up to 12.4 mm CR and up to 19.4 mm CA) offsets. Offsets were more likely to occur in the VO than in the template, possibly because of a smoother interface between the catheters and the obturator than between the catheters and the template. The lack of an equivalent to the VO in the prostate brachytherapy literature is a likely explanation for the uniqueness of our results. Large CR shifts may result, depending on the associated OAR motion, in an inadvertent close proximity of an activated catheter to an OAR. We observed 96 P catheter shifts (48%) with mean of 8.7 mm, with three shifts exceeding 20.0 mm; these results do not conform to the prostate brachytherapy literature (7). This difference can arise from a variety of causes, such as bladder catheterization, diet affecting rectal filling, and difference in tissue elasticity. Fusion methodologies based on registration of fiducials implanted in the tumor are expected to minimize the observed A/P shift because of the likely rigid motion of implant and tumor in the A/P direction. Dosimetric changes in the OAR were observed in this study. It is likely that the main contributing factor to the increase in rectal dose was a change in rectal filling, suggesting that a control of rectal filling may be appropriate in patients with rectal doses close to tolerance and treatments extending over many days when constipation is suspected. Changes in bladder dose were also observed, likely due to changes in bladder filling. Because of the high tolerance to radiation of the bladder, these changes were not clinically meaningful. Edema of the implanted tissue has been demonstrated to adversely affect the dosimetry of permanent prostate implants (20). Although it has been suggested that edema of tissue might be a cause for CA displacement in prostate HDR brachytherapy (14), we are not aware of studies correlating tumor edema with loss of coverage for HDR
108
A.L. Damato et al. / Brachytherapy 13 (2014) 100e109
treatments. In this work, we demonstrate that tissue edema and possible shrinkage because of tumor response and tissue compression may induce an X/S deformation in some of the implants. The effect is small in size (mean, 4.7 mm), and only 1 patient experienced deformations involving more than 10% of possible X/S deformations. Our data suggest that deformations are a source of minor uncertainty in gynecologic interstitial brachytherapy compared with implant displacements. Changes in tumor size over 1 day ranging from a shrinkage of 11 cm3 to an expansion of 21 cm3 (with a nonsignificant average increase of 3.1 cm3) have been reported (21) in gynecologic brachytherapy. In that study, changes were assessed through contouring of the HR-CTV (High Risk CTV) for 13 cervical cancer patients receiving intracavitary applications with an interstitial component on two CTs taken 1 day apart. A limitation of this study is the analysis as one single group of patients with secondary scans performed at different time intervals from the planning scan. Previous reports on gynecologic (16) and prostate (14) brachytherapy have shown that the shift between Days 2 and 3 is less frequent and smaller than the shift between Days 1 and 2. We therefore assumed that this limitation does not detract from the validity of our data. A second limitation of this study is that the results are dependent on clinic practices ranging from insertion technique, nursing care of the hospitalized patient, and materials used for the implant. The CT scans used in this work were acquired with 2.5-, 2-, and 1.25-mm slice thickness. This is because of a change in scanning policy of interstitial implants, from 2.5- to 1.25-mm slice thickness, which occurred during the time this study took place to improve the accuracy of the catheter tip localization. Although a 1.25-mm scan provides better digitization accuracy, and therefore more accurate dosimetry, we do still consider 2.5 mm a clinically acceptable choice. The different accuracy is accounted for in this work through a different uncertainty associated with the CR and CA displacements, as discussed in the Methods and Materials section. In summary, we demonstrated that displacements and deformations of the implant occur in gynecologic interstitial brachytherapy. The cause of these shifts may be because of patient movement, changes in tissue, or lack of reinforced attachments of the catheters to the template. CR, CA, and P displacements have been shown to be the largest and most frequent. Our data suggest that CR and CA implant shifts can have an adverse dosimetric effect on tumor coverage. Large A and P shifts may indicate that a large change in organ filling has occurred, potentially resulting in changes in OAR dosimetry. These data suggest that possible displacements should be considered during treatment planning and QA. Strategies involving physician contouring, planning margins, pretreatment displacement correction, and replanning have been used to address the loss of coverage because of displacement in HDR prostate brachytherapy; similar approaches can be considered for gynecologic applications. Proper QA methodologies should be in place to detect CR
shifts that can potentially result in piercing of OARs. Visual inspection of the implant to detect catheter offset and template shift before each fraction is advised.
Conclusions Mean displacements averaging less than 1 cm in any direction occur in gynecologic interstitial brachytherapy. QA of implant displacement before each fraction is necessary to detect CR and CA shifts, which are greater than 1 cm in some cases, and may result in a decrease in CTV D90. Routine visual inspection of the implant before each fraction is recommended, aided by markings of the catheter insertion depth made at the time of the planning scan and pictures of the template location. A midtreatment CT scan may be considered if visual inspection cannot be performed or is unsatisfactory.
Acknowledgments The authors thank Sam Song, Ph.D., for assistance with the figures and Barbara Silver for proofreading the text. References [1] Erickson B, Gillin MT. Interstitial implantation of gynecologic malignancies. J Surg Oncol 1997;66:285e295. [2] Fleming P, Nisar Syed AM, Neblett D, et al. Description of an afterloading 192Ir interstitial-intracavitary technique in the treatment of carcinoma of the vagina. Obstet Gynecol 1980;55:525e530. [3] Martinez A, Cox RS, Edmundson GK. A multiple-site perineal applicator (MUPIT) for treatment of prostatic, anorectal, and gynecologic malignancies. Int J Radiat Oncol Biol Phys 1984;10:297e305. [4] Popowski Y, Hiltbrand E, Joliat D, et al. Open magnetic resonance imaging using titanium-zirconium needles: Improved accuracy for interstitial brachytherapy implants? Int J Radiat Oncol Biol Phys 2000;47:759e765. [5] Yoshida K, Nose T, Shiomi H, et al. New ambulatory implant technique of high-dose-rate interstitial brachytherapy for prostate cancer. Radiat Med 2006;24:595e599. [6] Abbe R. The use of radium in malignant disease. Lancet 1913;2: 524e527. [7] Damore SJ, Syed AM, Puthawala AA, et al. Needle displacement during HDR brachytherapy in the treatment of prostate cancer. Int J Radiat Oncol Biol Phys 2000;46:1205e1211. [8] Martinez AA, Pataki I, Edmundson G, et al. Phase II prospective study of the use of conformal high-dose-rate brachytherapy as monotherapy for the treatment of favorable stage prostate cancer: A feasibility report. Int J Radiat Oncol Biol Phys 2001;49:61e69. [9] Hoskin PJ, Bownes PJ, Ostler P, et al. High dose rate afterloading brachytherapy for prostate cancer: Catheter and gland movement between fractions. Radiother Oncol 2003;68:285e288. [10] Mullokandov E, Gejerman G. Analysis of serial CT scans to assess template and catheter movement in prostate HDR brachytherapy. Int J Radiat Oncol Biol Phys 2004;58:1063e1071. [11] Yoshida K, Yamazaki H, Nose T, et al. Needle applicator displacement during high-dose-rate interstitial brachytherapy for prostate cancer. Brachytherapy 2010;9:36e41. [12] Foster W, Cunha JA, Hsu IC, et al. Dosimetric impact of interfraction catheter movement in high-dose rate prostate brachytherapy. Int J Radiat Oncol Biol Phys 2011;80:85e90.
A.L. Damato et al. / Brachytherapy 13 (2014) 100e109 [13] Kolkman-Deurloo IK, Roos MA, Aluwini S. HDR monotherapy for prostate cancer: A simulation study to determine the effect of catheter displacement on target coverage and normal tissue irradiation. Radiother Oncol 2011;98:192e197. [14] Simnor T, Li S, Lowe G, et al. Justification for inter-fraction correction of catheter movement in fractionated high dose-rate brachytherapy treatment of prostate cancer. Radiother Oncol 2009;93: 253e258. [15] Mikami M, Yoshida K, Takenaka T, et al. Daily computed tomography measurement of needle applicator displacement during high-dose-rate interstitial brachytherapy for previously untreated uterine cervical cancer. Brachytherapy 2011;10: 318e324. [16] Shukla P, Chopra S, Engineer R, et al. Quality assurance of multifractionated pelvic interstitial brachytherapy for postoperative recurrences of cervical cancers: A prospective study. Int J Radiat Oncol Biol Phys 2012;82:e617ee622.
109
[17] Erickson B, Albano K, Gillin M. CT-guided interstitial implantation of gynecologic malignancies. Int J Radiat Oncol Biol Phys 1996;36: 699e709. [18] Viswanathan AN, Cormack R, Holloway CL, et al. Magnetic resonance-guided interstitial therapy for vaginal recurrence of endometrial cancer. Int J Radiat Oncol Biol Phys 2006;66:91e99. [19] Viswanathan AN, Szymonifka J, Tempany-Afdhal CM, et al. A prospective trial of real-time magnetic resonance-guided catheter placement in interstitial gynecologic brachytherapy. Brachytherapy 2013;12:240e247. [20] Wang JZ, Mayr NA, Nag S, et al. Effect of edema, relative biological effectiveness, and dose heterogeneity on prostate brachytherapy. Med Phys 2006;33:1025e1032. [21] Pinnaduwage DS, Cunha JA, Weinberg V, et al. A dosimetric evaluation of using a single treatment plan for multiple treatment fractions within a given applicator insertion in gynecologic brachytherapy. Brachytherapy 2013;12:487e494.
Characterization of implant displacement and deformation in gynecologic interstitial brachytherapy Antonio L. Damato*, Robert A. Cormack, Akila N. Viswanathan Department of Radiation Oncology, Brigham and Women’s Hospital and Dana-Farber Cancer Institute, Boston, MA
ABSTRACT
PURPOSE: To determine the uncertainties in implant position during multifraction gynecologic interstitial brachytherapy, we analyzed the interfraction displacements and deformations of gynecologic interstitial implants. METHODS AND MATERIALS: Fourteen gynecologic patients treated with multifraction highdose-rate interstitial brachytherapy received two CT scans each at the time of implantation and 48e72 h later. Rigid fusions on the pubic symphysis were performed. This analysis included catheter shifts in the cranial (CR), caudal (CA), anterior, posterior, left, and right directions; template shifts; the change in the catheter length measured along the path from catheter tip to catheter connector (offset); the change in relative distances between catheters (deformations); and changes in rectum and bladder D2cc and tumor D90. RESULTS: Of the 198 catheters analyzed, the number of catheter shifts (%) and mean standard deviation were 43% CA (5.0 2.0 mm), 22% CR (7.9 4.0 mm), 14% anterior (6.3 2.1 mm), 48% posterior (8.7 3.1 mm), 7% left (4.8 0.4 mm), and 9% right (5.4 0.9 mm). Catheter offsets were 3% CA (7.2 6.3 mm) and 11% CR (6.1 2.6 mm). Template shifts were 43% CA (5.2 1.6 mm) and 14% CR (6.6 4.0 mm). Deformations were 10 shrinkages (4.7 0.9 mm) and 32 expansions (4.7 0.5 mm). Dosimetric changes were 5.2% 10.8% for rectum D2cc, 1.1% 18.5% for bladder D2cc, and 5.1% 6.7% for tumor D90. CONCLUSIONS: On average, less than 1 cm displacements and deformations of the implant occurred over the course of treatment. Proper quality assurance methodologies should be in place to detect shifts that can potentially result in inadvertent insertion into normal tissue. Ó 2014 American Brachytherapy Society. Published by Elsevier Inc. All rights reserved.
Keywords:
Brachytherapy; Gynecologic; Interstitial; Displacement; Uncertainty; Catheter
Introduction The use of interstitial brachytherapy for the treatment of advanced and recurrent gynecologic malignancies with significant vaginal and/or sidewall extensions increases local control with an acceptable risk of morbidity compared with external beam alone (1). Multiple implantation methodologies for gynecologic interstitial brachytherapy have been described in the literature (2e6). Templates with preset hole patterns are used for needle placement and to help with guidance and quality assurance (QA). The corners of the Received 16 May 2013; received in revised form 30 August 2013; accepted 25 September 2013. * Corresponding author. Department of Radiation Oncology, Brigham and Women’s Hospital, Dana-Farber Cancer Institute, 75 Francis St, ASB1-L2, Boston, MA 02115. Tel.: þ1-617-525-7242; fax: þ1-617732-7347. E-mail address: [email protected] (A.L. Damato).
template are sutured to the perineal skin for stability. The two most common template systems are the Syed-Neblett template (SNT) (2) and the Martinez Universal Perineal Interstitial Template (MUPIT) (3). Plastic catheters and hollow titanium needles (4) have been used. Custom-made templates and catheters inserted ‘‘free hand (FH)’’ into the patient may also be used. Catheters can be secured in place with a variety of anchoring techniques, such as buttons or liquid adhesive. Most multifractionated interstitial brachytherapy treatments require the hospitalization of the patient for the duration of the treatment; some ambulatory techniques have been proposed and used for gynecologic treatments (5). A single implant can be used for a treatment that extends from 1 to 5 days. Significant uncertainties may exist related to the catheter/needle anchorage to the template or the patient’s skin, the stability of the template, and deformations of the patient anatomy.
1538-4721/$ - see front matter Ó 2014 American Brachytherapy Society. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.brachy.2013.09.010
A.L. Damato et al. / Brachytherapy 13 (2014) 100e109
An assessment of the interfraction variability of catheter location is essential to understand the uncertainties in multifraction gynecologic interstitial brachytherapy and can be used to inform planning and QA policies. A substantial literature exists (7e14) on implant displacements for interstitial prostate brachytherapy. However, insights that prostate literature may provide for gynecologic applications may be limited. Differences in tumor invasion into organ structures, imaging methodology, and applicators used in prostate and gynecologic interstitial brachytherapy procedures may result in significant differences in implant geometry, stabilization, and deformation. Implant displacement has been studied for gynecologic ambulatory procedures (15) and for cervical cancer patients receiving interstitial brachytherapy through a MUPIT (16). In this work, we analyze the variability of gynecologic interstitial implants recently performed according to the standard practice in our clinic. To our knowledge, this is the first detailed investigation of implant displacements and deformations in gynecologic interstitial brachytherapy using SNT.
Methods and materials A data set of 28 CT scans from 14 patients treated from March 2011 to December 2011 with high-dose-rate (HDR) interstitial gynecologic brachytherapy in our institution was retrospectively analyzed with institutional review board approval. All patients received a CT scan immediately after implantation and again 2 (n 5 11 patients) or 3 (n 5 3) days after implantation. All patients had vaginal extension of gynecologic malignancy necessitating interstitial treatments. One patient did not have a rectum because of a history of rectal cancer.
101
Implants All patients were implanted under general anesthesia. Three implantations were performed using real-time MRI guidance from a 3 T wide-bore unit, with a CT scan acquired immediately after implantation. All the other implantations were performed in a CT brachytherapy suite under CT guidance. A vaginal obturator (VO) was inserted through the center of an SNT (Best Medical International Inc., Springfield, VA) (Fig. 1). Plastic ProGuide catheters (Nucletron, an Elekta company, Elekta AB, Stockholm, Sweden) were implanted with a central stainless steel removable insert to ensure stability during the insertion. Of the 198 total catheters implanted in all patients, 7 were inserted FH, 75 through grooves in the central VO, and 112 through holes in the SNTs. In 4 patients with an intact uterus, a central tandem was inserted. After the implant, the central VO and the VO and SNT catheters were secured to the template using a sterile liquid adhesive, and the template was sutured in four corners to the perineal skin. FH catheters were secured to the patient’s skin through buttons when possible. The extent of each catheter’s entry into the template or the patient body was marked on the catheter for pretreatment visual inspection. All patients received one fraction on the day of implantation and an additional four to eight fractions twice daily. The doses from the brachytherapy and prior radiation were summed via the EQD2 (equivalent dose 2 Gy) formalism. Dose optimization and fractionation scheme selection had the goal of achieving a combined clinical target volume (CTV) D90 between 70 and 80 Gy, while respecting limits on the combined D2cc metrics for bladder (!90e95 Gy), rectum (!70e75 Gy), and sigmoid (!70e75 Gy). The fractionation schemes were prescribed by the physician on a patient-by-patient basis based on tumor, normal tissue, and patient factors and are reported in Table 4.
Fig. 1. Graphical representation of a Syed-Neblett template with holes for catheter insertion, vaginal obturator with grooves for catheter insertion, and catheter inserted into a tumor.
102
A.L. Damato et al. / Brachytherapy 13 (2014) 100e109
Imaging CT scans for all the patients were obtained in a dedicated brachytherapy suite equipped with a CT scanner. The slice thicknesses were 1.25 mm (20 scans), 2 mm (2 scans), and 2.5 mm (6 scans). Copper dummy markers were inserted in all catheters for every scan for easy tip identification. Scans were performed with the legs down. This permitted accurate catheter localization and reconstruction. All patients received a first scan immediately after implantation, before being extubated and transferred to the postoperative recovery area. The first treatment occurred on the day of implantation, subsequent treatments occurred twice a day with 6 h between fractions. All other imaging was performed immediately before a treatment: 11 scans before fraction 4 and 3 scans before fraction 6. Visual inspection of catheter entry level into the template and eventual corrections were performed after imaging. Fusion For each patient, the two CT scans were fused using the Eclipse Treatment Planning System (Varian Medical Systems, Inc., Palo Alto, CA) registration tool (Varian Rigid Registration Algorithm 10.0.28). The first scan was taken as the reference coordinate system, and each subsequent scan was registered on the reference coordinate system based on a rigid fusion of the pubic symphysis. Fusions were automatically performed on pixel data prioritized to the pubic symphysis only, in a volume of interest excluding the template, femoral heads, and the superior portion of the bladder. Visual inspection confirmed excellent fusion of the pubic symphysis, with uncertainty associated with the rigid registration of the pubic symphysis assumed to be !1mm. The quality of fusion of other organs varied depending on organ filling and organ motion that may have occurred between scans. Fused sets were used to compare implant position across CT scans in a common coordinate system. Catheter reconstruction Catheters were reconstructed in Oncentra Brachy (Nucletron, an Elekta company) by an experienced brachytherapy planner. The first digitization point (connector end) was positioned on the most superior slice of the template. Catheter location was highlighted in the scan by the presence of a copper dummy marker. Implant displacement Catheter displacement Catheter displacement refers to the displacement of the tip of a catheter between scans. Catheter displacement is subdivided into shifts in the cranial (CR), caudal (CA), anterior (A), posterior (P), left (L), and right (R) directions. The A direction is defined as toward the bladder and the P direction as toward the rectum. L and R are defined as
toward the left and right of the patient. Displacements can arise from a variety of causes including malfunction of the catheter attachment to the template, displacement of the template, and deformation of tissues. To gain better insight into the cause of displacement, we measured two primary parameters: catheter offset and template shift (Fig. 2). Catheter offset Catheter offset is defined as the change in the catheter length measured along the path from catheter tip to catheter connector (i.e., within the patient), likely resulting from inadequate anchorage of the catheter to the template. A shorter catheter path is scored as a CA offset and a longer catheter path as a CR offset. Template shift Template shift is the average CR/CA shift of the catheters’ connector ends between scans. Template shift represents the overall displacement of the template/implant in relation to the pubic symphysis because of perineal tissue swelling, handling of the patient during hospitalization, and pressure on the implant due to patient movement or other anatomic changes. The slice thickness of the CT scan imposes a lower limit on the size of observable CR/CA displacements. In this work, we consider a CR/CA displacement to be observable when it is greater than twice the slice thickness of the underlying CT scan. For A/P and L/R displacements, the limiting factor is the accuracy of the manual point digitization. We assume an accuracy of 2 mm, taking into account the size of the dummy marker on CT and rare but possible confounding factors like calcifications or nearby reconstruction artifacts from possible implanted clips; a shift greater than 4 mm is therefore considered observable. Throughout this manuscript, mean and standard deviation values are reported as mean standard deviation. Implant deformation Implant deformation is defined as the change in relative distances between catheters. More specifically, implant deformation is the change in relative distance between the points where any two given catheters intersect the most superior plane perpendicular to the template that contains all the catheters of the patient (i.e., just before the tip of the most shallow catheter of the implant). This definition is insensitive to catheter offset and is designed to evaluate deformations due to implant expansion (X) from edema or implant shrinkage (S) from tissue compression and/or tumor response. Similarly to A/P and L/R displacements, we consider a deformation observable if it is greater than 4 mm. Dosimetry The rectum and bladder were contoured on both CTs. The dosimetry from the clinical treatment plan was calculated based on catheter locations on both scans, and D2cc metrics as a percentage of the prescription dose are reported.
A.L. Damato et al. / Brachytherapy 13 (2014) 100e109
103
Fig. 2. Graphical representation of the catheter displacement, template shift, and catheter offset. On the top half (Shift A), we show a net caudal catheter shift displacement (solid arrow), resulting from a caudal template shift (dashed line). On the bottom half (Shift B), we show a net cranial catheter displacement (solid arrow), resulting from the superposition of the same caudal template shift (dashed line) and a cranial catheter offset (dotted line). The net effect of offsets and template shifts may introduce both cranial and caudal shifts in the same implant.
Because of uncertainties in contouring of CTV on CT, and the unavailability of magnetic resonance (MR) images at the time of the secondary scan for the 3 patients who were implanted under MR guidance, CTV recontouring was not performed. Changes in CTV D90 due to catheter shift were evaluated based on the difference between the dosimetry at time of planning and the dosimetry resulting from a redigitization of the catheters, reflecting the CR/ CA component of their shift. A/P and L/R shifts were not considered under the assumption that the CTV and the catheters will translate and deform equally in those directions. A limit of this approach is that possible shift and deformation of the CTV in the same CR/CA direction of the implant shift is not considered, and the changes in CTV D90 thus calculated may overestimate the actual change in CTV D90.
Statistics A two-tailed paired Student t test was performed to assess statistical significance of the dosimetric changes. A p-value of !0.05 was associated with statistical significance.
Results A total of 198 catheters in 14 implants were analyzed. The median number of catheters per implant was 15 (range, 9e23).
Implant displacement CR/CA implant displacement Table 1 lists the summary of the observable CR/CA catheter shifts and offsets on a patient-by-patient basis and as an average of the mean value per patient among those who experienced an observable shift in that direction. The number of patients experiencing observable shifts was 8 for template shifts, 13 for catheter shifts, and 8 for offsets. Table 2 lists the summary of the CR/CA catheter shifts and offsets per catheter type. A total of 129 CR/CA catheter shifts (65%) were observed, of which 43 were CR (22%) (9 shifts! 5 mm, 20 shifts between 5 and 10 mm, and 14 shiftsO10 mm) and 86 were CA (43%) (51 shifts!5 mm, 33 shifts between 5 and 10 mm, and 2 shifts O10 mm). Twenty-seven catheter offsets (14%) were observed, of which 21 were CR (11%) (10 offsets !5 mm, 9 offsets between 5 and 10 mm, and 2 offsets O10 mm) and 6 were CA (3%) (4 offsets !5 mm, 1 offset between 5 and 10 mm, and 1 offset O10 mm). A total of 20 of the 75 catheters in the VO group (27%) experienced an observable offset in any direction compared with 2 of the 112 catheters in the SNT group (2%). A total of four of the seven catheters in the FH group (57%) experienced an observable offset in any direction. Figure 3 summarizes the components of the CR/CA implant displacements. The strong correlation between template shift and mean catheter shift per patient (1.05 linearity, R2 5 0.946) suggests that rigid template shift was the main contributor to the overall catheter shift. In addition to the rigid displacement driven by the template shift, the offset contributed to isolated large shifts of catheters.
104
A.L. Damato et al. / Brachytherapy 13 (2014) 100e109
Table 1 Summary of CR and CA implant displacements per patient Patient
Template shift (mm)
Mean catheter shift CR/CA (mm)
CR shifts (%)
CA shifts (%)
A B C D E F G H I J K L M Na
CA CA CA CA CA CA Not Not Not Not Not Not CR CR
d/6.7 d/5.5 d/5.1 d/4.8 d/3.7 5.6/3.7 d/3.3 12.5/16.3 d/d 3.6/d 6/d 5.6/3.6 5.5/d 11.9/d
d d d d d 5/12 (42) d 1/13 (8) d 2/10 (20) 1/9 (11) 8/16 (50) 11/11 (100) 15/15 (100)
9/15 21/23 17/17 11/16 13/15 5/12 8/16 1/13
Average
5.9 6.8 7.0 3.7 4.5 3.4 observable observable observable observable observable observable 3.7 9.4
(60) (91) (100) (69) (87) (42) (50) (8) d d d 1/16 (6) d d
Mean catheter offsets CR/CA (mm)
CR offsets (%)
d/d 4.0/6.7 d/d 3.3/4.5 d/d 7.7/d d/2.9 10.5/19.4 d/d 3.7/d d 5.4/d d/d 7.0/d
2/23 1/16 6/12 1/13 2/10 7/16 2/15
d (9) d (6) d (50) d (8) d (20) d (44) d (13)
CA offsets (%) d 2/23 (9) d 1/16 (6) d d 2/16 (13) 1/13 (8) d d d d d d
Template shift (mm)
Mean template shift (mm)
Mean catheter shift (mm)
Mean offset (mm)
CR CA
6.6 4.0 5.2 1.6
7.2 3.5 5.9 4.1
6.0 2.6 8.4 7.5
CR 5 cranial; CA 5 caudal. CR (CA) average values are across patients who have experienced at least one CR (CA) shift. Averages are expressed as mean standard deviation. a Patient without a rectum.
CR implant displacement subset analysis Of the 14 CR shifts exceeding 10 mm, 13 occurred in the 1 patient who did not have a rectum. In subset analysis, omitting this case, the remaining 13 patients had the following CR displacements: observable CR catheter shifts occurred in 6 patients (46%) (6.5 3.1 mm; range, 3.6e12.5 mm),
observable CR catheter offsets occurred in 6 patients (46%) (5.8 2.8 mm; range, 3.3e10.5 mm), and 1 CR template shift of 3.7 mm was observed. Among the 183 catheters in this subset, 28 CR catheter shifts (15%) were observed (5.7 2.2 mm; range, 2.7e12.5 mm; 9 shifts !5 mm; 18 shifts between 5 and
Table 2 Summary of CR and CA implant displacements per catheter insertion site Catheter shift Site Total Number of displacements Mean (mm) Range (mm) Subset Number of events Mean (mm) Range (mm) VO Number of displacements Mean (mm) Range (mm) SNT Number of displacements Mean (mm) Range (mm) FH Number of displacements Mean (mm) Range (mm) Tandem Number of displacements Mean (mm) Range (mm)
Catheter offset
CR (%)
CA (%)
CR (%)
CA (%)
43/198 (22) 7.9 4.0 2.7e20.0
86/198 (43) 5.0 2.0 2.6e16.3
21/198 (11) 6.1 2.6 3.0e12.4
6/198 (3) 7.2 6.3 2.7e19.4
28/183 (15) 5.7 2.2 2.7e12.5
86/183 (47) 5.0 2.0 2.6e16.3
19/183 (11) 6.0 2.7 3.0e12.4
6/183 (3) 7.2 6.3 2.7e19.4
24/75 (32) 6.6 3.1 2.7e12.5
22/75 (29) 5.5 2.8 3.5e16.3
17/75 (23) 5.7 2.4 3.0e10.5
3/75 (4) 9.0 9.0 3.2e19.4
16/112 (14) 8.7 2.9 5.3e16.2
59/112 (53) 4.6 1.4 3.0e8.0
1/112 (1) 12.4 12.4
1/112 (1) 2.7 2.7
2/7 (29) 10.9 12.9 1.8e20.0
4/7 (57) 7.2 3.0 4.4e10.3
2/7 (29) 7.0 1.1 6.2e7.8
2/7 (29) 6.8 2.9 4.7e8.8
1/4 (25) 3 3
1/4 (25) 5.8 5.8
1/4 (25) 4.7 4.7
d d d
Subset 5 excluding patient without a rectum; CR 5 cranial; CA 5 caudal; VO 5 vaginal obturator; SNT 5 Syed-Neblett template; FH 5 free hand. Average values are expressed as mean standard deviation.
A.L. Damato et al. / Brachytherapy 13 (2014) 100e109
105
Fig. 3. Relationship between template shifts and mean catheter shifts for each patient. Range of observable catheter offset is marked with a slashed line. Only patients with an observable template shift are plotted and contributed to the trendline. Nonobservable region is marked in gray. Cranial shifts have the positive sign, and caudal shifts have the negative sign.
10 mm; and 1 shift O10 mm). Nineteen CR offsets (10%) were observed (6.0 2.7 mm; range, 3.0e12.4 mm; 10 offsets !5 mm; 7 offsets between 5 and 10 mm; and 2 offsets O10 mm). A/P catheter shifts A total of 123 A/P catheter shifts (62%) were observed. All 14 patients experienced at least one A/P catheter displacement. Twenty-seven A shifts (14%) (6.3 2.1 mm; range, 4.3e12.0 mm) were observed in 6 patients. Ninety-six P shifts (48%) (8.7 3.1 mm; range, 4.1e20.8 mm) were observed in 9 patients.
L/R catheter shifts Twenty-nine L/R catheter shifts (15%) were observed in 10 patients. Twelve L shifts (6%) (4.8 0.4 mm; range, 4.2e5.8 mm) were observed in 4 patients. Seventeen R shifts (9%) (5.4 0.9 mm; range, 4.1e7.1 mm) were observed in 6 patients. Implant deformation Table 3 summarizes the observable deformations seen in each patient. Because deformation is measured at different depths, we report the depth used for each patient. Forty-two
Table 3 Summary of implant deformations S deformations
X deformations
Patient
Measurement depth (mm)
Number of deformations/number of possible deformations (%)
Mean (mm)
Range (mm)
Number of deformations/number of possible deformations (%)
Mean (mm)
Range (mm)
A B C D E F G H I J K L M Na
80 75 95 135 110 80 120 80 75 60 110 70 75 70
4/105 d 4/136 d 1/105 d d d d d 1/36 d d d
4.4 0.3 d 5.2 1.3 d 4 d d d d d 4.2 d d d
4.2e4.8 d 4.0e6.5 d 4 d d d d d 4.2 d d d
d d 7/136 16/120 d d 1/120 d d d d d d 8/105
d d 4.7 0.5 4.7 0.5 d d 4.2 d d d d d d 4.9 0.7
d d 4.1e5.4 4.1e5.5 d d 4.2 d d d d d d 4.0e6.0
(4) (3) (1)
(3)
(5) (13)
(1)
(8)
S 5 shrinkage deformation (tumor response/tissue compression); X 5 expansion deformation (edema). a Patient without a rectum.
106
A.L. Damato et al. / Brachytherapy 13 (2014) 100e109
observable implant deformations were observed. Ten S deformations (4.7 0.9 mm; range, 4.0e6.5 mm) were observed in 4 patients. No patient presented with a number of S deformations exceeding 5% of the number of possible permutations (N (N 1)/2, where N is the number of catheters in a patient). Thirty-two X deformations (4.7 0.5 mm; range, 4.0e6.0 mm) were observed in 4 patients. Three patients presented with a number of X deformations exceeding 5% of the number of possible permutations (one exceeding 10%). Dosimetry Table 4 summarizes the dosimetric changes that have occurred between scans. The average change in CTV D90 between scans was 5.1 6.7% of the prescription dose, from an average of 100 26% to an average of 95 25%. In three cases, a difference in D90 exceeding 10% of the prescription dose was observed. The change was statistically significant ( p ! 0.02). Changes in rectum and bladder D2cc between scans have been observed. The average change in D2cc between scans was þ5.2 10.8% of the prescription dose for rectum ( p 5 0.09) and 1.1% 18.5% of the prescription dose for bladder ( p 5 0.82). These differences are likely because of anatomic differences between scans due to a change in organ filling. Clinical results No perforations or other adverse events occurred for the 14 patients considered in this study. In three instances, a verification of the dosimetry on the QA CT scan was deemed
advisable because of implant displacement. In one instance, a replan was deemed appropriate because of an increase in rectal dose. This work did not investigate the origin of this dosimetric change, although it seems likely that the main contributing factor was a change in rectal filling. Discussion We observed that a majority (129/198, 65%) of catheters experienced an observable CR or CA shift between the planning scan and the second day of treatment. Most shifts were small (86 CA shifts with a mean of 5.0 mm and 43 CR shifts with a mean of 7.9 mm), but shifts as large as 16.3 mm in the CA direction and 20.0 mm in the CR direction were observed. These results show that template shifts and shifts of single catheters relative to the template may occur, and a QA methodology to monitor such shifts should be implemented. Possible methodologies include visual inspection of catheters before each fraction, aided by the marking on the catheter of the insertion length at the time of the planning scan. Moreover, visual inspection of the position of the template relative to the perineal area should be performed, aided by pictures of the area taken at the time of the planning scan. When clinically meaningful shifts are suspected, corrective actions or confirmation through CT scan may be considered. Routine CT scans may not be appropriate as patient movement to the CT table may possibly contribute to the implant shifts. Moreover, when a CT scan is used to evaluate implant shifts, care should be taken in evaluating the measuring uncertainties that depend on clinic-specific imaging protocols and to allocate resources for scanning and the timely catheter
Table 4 Summary of dosimetric results
Patient A B C D E F G H I J K L M Na Average
Prescription dose (cGy) number of fractions 350 500 400 750 325 450 450 300 425 500 550 450 350 450
9 6 7 5 7 5 9 9 7 5 5 7 6 6
CTV D90 (% of prescription dose)
R/B D2cc (% of prescription dose)
Planning
Re-scan
Difference
Planning
Re-scan
Difference
74.7 129.6 114.3 124.9 98.4 54.1 122.5 64.2 106.1 74.5 115.7 126.6 75.3 119.3 100.0 26.1
76.3 109.9 108.2 125.4 86.9 52.1 119.5 57.0 106.6 76.8 117.1 121.6 67.0 104.7 94.9 25.0
þ1.6 19.7 6.1 þ0.4 11.5 2.0 3.0 7.1 þ0.4 þ2.3 þ1.4 5.0 8.3 14.6 5.1 6.7 p ! 0.02
52.1/114 43.6/50.3 78.3/102 68.9/81.1 68.5/79.9 63.2/62.2 75.9/74.7 67.6/78.5 41.3/70.2 87.7/83.5 85.3/117 47.9/55.1 68.4/84.5 N/A/48.8 R: 60.6 22.7 B: 78.2 21.4
65.0/101 47.2/52.3 75.2/88.3 95.7/87.1 81.3/96.0 53.9/61.7 67.9/115 64.9/57.6 40.8/63.2 93.3/90.3 98.0/107 48.4/56.6 90.0/48.7 N/A/53.9 R: 65.8 26.8 B: 77.1 23.0
þ12.9/12.7 þ3.6/þ2.1 3.1/13.7 þ26.8/6.0 þ12.8/þ22.1 9.3/0.4 8.0/þ40.7 2.7/20.9 0.4/7.0 þ5.6/þ6.8 þ12.7/9.3 þ0.5/þ1.5 þ21.6/35.8 N/A/þ5.1 R: þ5.2 10.8 p 5 0.09 B: 1.1 18.5 p 5 0.82
CTV 5 clinical target volume; R 5 rectum; B 5 bladder; N/A 5 not applicable. Average values are expressed as mean standard deviation. a Patient without a rectum.
A.L. Damato et al. / Brachytherapy 13 (2014) 100e109
redigitization, plan generation, and evaluation, which may also require recontouring. The use of interstitial brachytherapy for the treatment of gynecologic malignancy was first reported in 1913 by Abbe (6). More recently, the use of three-dimensional image guidance to aid with implantation and treatment planning was proposed using CT (17) and MR (18, 19), in which localization of the catheters in relation to the tumor and the organs at risk (OARs, defined as the bladder, rectum, sigmoid, and small bowel) is used to achieve optimal dosimetry. In this analysis, we demonstrate that interfraction displacement of the catheters occurs preferentially in the CR, CA, and P directions. The displacements may introduce uncertainty in the delivered dosimetry. In the CR/CA axis, we observed the largest number of shifts in the CA direction (86/198, 43%): Most (51) were smaller than 5 mm, with a single shift as large as 16.3 mm occurring. The average mean CA shift observed (5.9 mm) in this study is lower than what is observed in prostate brachytherapy (from 7.6 to 20 mm) (7e12) and gynecologic brachytherapy using the MUPIT (17.4 mm) (16). A total of 43/198 CR shifts (22%) were observed in this study, with a mean shift of 7.9 mm and a single maximum shift of 20.0 mm. The CR shifts observed were larger and more frequent than what has been observed in prostate brachytherapy, in which most authors report not observing any CR shifts (7e9). The CR shifts observed in this study were also higher than what is reported (15) in ambulatory uterine cancer treatment and in MUPIT implantations (mean, 2.5 mm; range, 0e7.4 mm) (16). Differences in measurement methodology (referenced to the fusion of the pubic symphysis in this study, to a relocatable bony point as in ref. (16), and to the center of mass of implanted markers as in ref. (15)) and differences in implantation technique (using SNT in this study, MUPIT as in ref. (16), and a thin vinyl plate as in ref. (15)) are possible explanations of the variability of the results. Notably, 15/43 CR shifts (35%) occurred in the 1 patient without a rectum. This patient also accounted for 13 of the 14 CR shifts exceeding 1 cm and for the maximum CR shift of 20.0 mm. A subset analysis was done omitting this patient: The mean CR shift was 5.7 mm, with a single maximum shift of 12.5 mm, suggesting that the rectum may act as a stabilizer in the perineal tissue of patients with a vaginal interstitial implant. Patients requiring vaginal interstitial brachytherapy who do not have a rectum may benefit from daily imaging, given the potential for a large CR shift. The observed CR/CA catheter shifts may result in significant dosimetric changes, as shown by the observed changes in CTV D90 in this study. Similarly, the literature on prostate brachytherapy suggests that significant loss of coverage occurs for CA displacements of the order seen in this study. A study on change in CTV dosimetry (19), in which the CTV was delineated on three MRs taken daily in the first 3 days of treatment, shows an average decrease of CTV D90 of 5.7%. Despite the different methodology
107
(redelineation of the CTV on repeated MR (19) vs. the calculation of the effect of CR/CA shift on CTV as done in this study), this result is in line with the results of our work. Further studies are needed to assess the clinical efficacy of possible corrective actions, such as adjusting planning methodologies to minimize the effect of shifts, correcting the shifts when they occur, or replanning during treatment. Current data on clinical end points such as OAR toxicity and local control probability for multifraction gynecologic interstitial brachytherapy already account for average changes in dosimetry due to catheter shift. QA methodologies for the detection of large dosimetric changes, which in this study were up to 19.7% of the CTV D90, need to be developed and implemented. Swelling of the peritoneal area and handling of the patient during hospitalization are possible causes of the rigid CR/CA implant shifts. In the analysis of the offsets, that is the relative motion between the catheter and the template, contrary to some of the prostate brachytherapy literature (10) in which no offset in any direction has been reported, we observed infrequent (27/198, 14%) but large (up to 12.4 mm CR and up to 19.4 mm CA) offsets. Offsets were more likely to occur in the VO than in the template, possibly because of a smoother interface between the catheters and the obturator than between the catheters and the template. The lack of an equivalent to the VO in the prostate brachytherapy literature is a likely explanation for the uniqueness of our results. Large CR shifts may result, depending on the associated OAR motion, in an inadvertent close proximity of an activated catheter to an OAR. We observed 96 P catheter shifts (48%) with mean of 8.7 mm, with three shifts exceeding 20.0 mm; these results do not conform to the prostate brachytherapy literature (7). This difference can arise from a variety of causes, such as bladder catheterization, diet affecting rectal filling, and difference in tissue elasticity. Fusion methodologies based on registration of fiducials implanted in the tumor are expected to minimize the observed A/P shift because of the likely rigid motion of implant and tumor in the A/P direction. Dosimetric changes in the OAR were observed in this study. It is likely that the main contributing factor to the increase in rectal dose was a change in rectal filling, suggesting that a control of rectal filling may be appropriate in patients with rectal doses close to tolerance and treatments extending over many days when constipation is suspected. Changes in bladder dose were also observed, likely due to changes in bladder filling. Because of the high tolerance to radiation of the bladder, these changes were not clinically meaningful. Edema of the implanted tissue has been demonstrated to adversely affect the dosimetry of permanent prostate implants (20). Although it has been suggested that edema of tissue might be a cause for CA displacement in prostate HDR brachytherapy (14), we are not aware of studies correlating tumor edema with loss of coverage for HDR
108
A.L. Damato et al. / Brachytherapy 13 (2014) 100e109
treatments. In this work, we demonstrate that tissue edema and possible shrinkage because of tumor response and tissue compression may induce an X/S deformation in some of the implants. The effect is small in size (mean, 4.7 mm), and only 1 patient experienced deformations involving more than 10% of possible X/S deformations. Our data suggest that deformations are a source of minor uncertainty in gynecologic interstitial brachytherapy compared with implant displacements. Changes in tumor size over 1 day ranging from a shrinkage of 11 cm3 to an expansion of 21 cm3 (with a nonsignificant average increase of 3.1 cm3) have been reported (21) in gynecologic brachytherapy. In that study, changes were assessed through contouring of the HR-CTV (High Risk CTV) for 13 cervical cancer patients receiving intracavitary applications with an interstitial component on two CTs taken 1 day apart. A limitation of this study is the analysis as one single group of patients with secondary scans performed at different time intervals from the planning scan. Previous reports on gynecologic (16) and prostate (14) brachytherapy have shown that the shift between Days 2 and 3 is less frequent and smaller than the shift between Days 1 and 2. We therefore assumed that this limitation does not detract from the validity of our data. A second limitation of this study is that the results are dependent on clinic practices ranging from insertion technique, nursing care of the hospitalized patient, and materials used for the implant. The CT scans used in this work were acquired with 2.5-, 2-, and 1.25-mm slice thickness. This is because of a change in scanning policy of interstitial implants, from 2.5- to 1.25-mm slice thickness, which occurred during the time this study took place to improve the accuracy of the catheter tip localization. Although a 1.25-mm scan provides better digitization accuracy, and therefore more accurate dosimetry, we do still consider 2.5 mm a clinically acceptable choice. The different accuracy is accounted for in this work through a different uncertainty associated with the CR and CA displacements, as discussed in the Methods and Materials section. In summary, we demonstrated that displacements and deformations of the implant occur in gynecologic interstitial brachytherapy. The cause of these shifts may be because of patient movement, changes in tissue, or lack of reinforced attachments of the catheters to the template. CR, CA, and P displacements have been shown to be the largest and most frequent. Our data suggest that CR and CA implant shifts can have an adverse dosimetric effect on tumor coverage. Large A and P shifts may indicate that a large change in organ filling has occurred, potentially resulting in changes in OAR dosimetry. These data suggest that possible displacements should be considered during treatment planning and QA. Strategies involving physician contouring, planning margins, pretreatment displacement correction, and replanning have been used to address the loss of coverage because of displacement in HDR prostate brachytherapy; similar approaches can be considered for gynecologic applications. Proper QA methodologies should be in place to detect CR
shifts that can potentially result in piercing of OARs. Visual inspection of the implant to detect catheter offset and template shift before each fraction is advised.
Conclusions Mean displacements averaging less than 1 cm in any direction occur in gynecologic interstitial brachytherapy. QA of implant displacement before each fraction is necessary to detect CR and CA shifts, which are greater than 1 cm in some cases, and may result in a decrease in CTV D90. Routine visual inspection of the implant before each fraction is recommended, aided by markings of the catheter insertion depth made at the time of the planning scan and pictures of the template location. A midtreatment CT scan may be considered if visual inspection cannot be performed or is unsatisfactory.
Acknowledgments The authors thank Sam Song, Ph.D., for assistance with the figures and Barbara Silver for proofreading the text. References [1] Erickson B, Gillin MT. Interstitial implantation of gynecologic malignancies. J Surg Oncol 1997;66:285e295. [2] Fleming P, Nisar Syed AM, Neblett D, et al. Description of an afterloading 192Ir interstitial-intracavitary technique in the treatment of carcinoma of the vagina. Obstet Gynecol 1980;55:525e530. [3] Martinez A, Cox RS, Edmundson GK. A multiple-site perineal applicator (MUPIT) for treatment of prostatic, anorectal, and gynecologic malignancies. Int J Radiat Oncol Biol Phys 1984;10:297e305. [4] Popowski Y, Hiltbrand E, Joliat D, et al. Open magnetic resonance imaging using titanium-zirconium needles: Improved accuracy for interstitial brachytherapy implants? Int J Radiat Oncol Biol Phys 2000;47:759e765. [5] Yoshida K, Nose T, Shiomi H, et al. New ambulatory implant technique of high-dose-rate interstitial brachytherapy for prostate cancer. Radiat Med 2006;24:595e599. [6] Abbe R. The use of radium in malignant disease. Lancet 1913;2: 524e527. [7] Damore SJ, Syed AM, Puthawala AA, et al. Needle displacement during HDR brachytherapy in the treatment of prostate cancer. Int J Radiat Oncol Biol Phys 2000;46:1205e1211. [8] Martinez AA, Pataki I, Edmundson G, et al. Phase II prospective study of the use of conformal high-dose-rate brachytherapy as monotherapy for the treatment of favorable stage prostate cancer: A feasibility report. Int J Radiat Oncol Biol Phys 2001;49:61e69. [9] Hoskin PJ, Bownes PJ, Ostler P, et al. High dose rate afterloading brachytherapy for prostate cancer: Catheter and gland movement between fractions. Radiother Oncol 2003;68:285e288. [10] Mullokandov E, Gejerman G. Analysis of serial CT scans to assess template and catheter movement in prostate HDR brachytherapy. Int J Radiat Oncol Biol Phys 2004;58:1063e1071. [11] Yoshida K, Yamazaki H, Nose T, et al. Needle applicator displacement during high-dose-rate interstitial brachytherapy for prostate cancer. Brachytherapy 2010;9:36e41. [12] Foster W, Cunha JA, Hsu IC, et al. Dosimetric impact of interfraction catheter movement in high-dose rate prostate brachytherapy. Int J Radiat Oncol Biol Phys 2011;80:85e90.
A.L. Damato et al. / Brachytherapy 13 (2014) 100e109 [13] Kolkman-Deurloo IK, Roos MA, Aluwini S. HDR monotherapy for prostate cancer: A simulation study to determine the effect of catheter displacement on target coverage and normal tissue irradiation. Radiother Oncol 2011;98:192e197. [14] Simnor T, Li S, Lowe G, et al. Justification for inter-fraction correction of catheter movement in fractionated high dose-rate brachytherapy treatment of prostate cancer. Radiother Oncol 2009;93: 253e258. [15] Mikami M, Yoshida K, Takenaka T, et al. Daily computed tomography measurement of needle applicator displacement during high-dose-rate interstitial brachytherapy for previously untreated uterine cervical cancer. Brachytherapy 2011;10: 318e324. [16] Shukla P, Chopra S, Engineer R, et al. Quality assurance of multifractionated pelvic interstitial brachytherapy for postoperative recurrences of cervical cancers: A prospective study. Int J Radiat Oncol Biol Phys 2012;82:e617ee622.
109
[17] Erickson B, Albano K, Gillin M. CT-guided interstitial implantation of gynecologic malignancies. Int J Radiat Oncol Biol Phys 1996;36: 699e709. [18] Viswanathan AN, Cormack R, Holloway CL, et al. Magnetic resonance-guided interstitial therapy for vaginal recurrence of endometrial cancer. Int J Radiat Oncol Biol Phys 2006;66:91e99. [19] Viswanathan AN, Szymonifka J, Tempany-Afdhal CM, et al. A prospective trial of real-time magnetic resonance-guided catheter placement in interstitial gynecologic brachytherapy. Brachytherapy 2013;12:240e247. [20] Wang JZ, Mayr NA, Nag S, et al. Effect of edema, relative biological effectiveness, and dose heterogeneity on prostate brachytherapy. Med Phys 2006;33:1025e1032. [21] Pinnaduwage DS, Cunha JA, Weinberg V, et al. A dosimetric evaluation of using a single treatment plan for multiple treatment fractions within a given applicator insertion in gynecologic brachytherapy. Brachytherapy 2013;12:487e494.