A surrogate urethra for real-time planning of high-dose-rate prostate brachytherapy

Brachytherapy 18 (2019) 675e682

Physics

A surrogate urethra for real-time planning of high-dose-rate prostate brachytherapy Heather Halperin1, Michelle Hilts1,2,*, Juanita Crook1, Deidre Batchelar1,2, Steven Tisseverasinghe1, Audrey Tetreault-LaFlamme3, Francois Bachand1 1 BC Cancer - Kelowna, Kelowna, BC Department of Physics, Irving K Barber School of Arts and Sciences, University of British Columbia, Kelowna, BC 3 Universite de Sherbrooke, Sherbrooke, QC

2

ABSTRACT

PURPOSE: This study characterizes prostatic urethra cross-section to develop a surrogate urethra for accurate prediction of urethral dose during real-time high-dose-rate prostate brachytherapy. MATERIALS AND METHODS: Archived preoperative transrectal ultrasound images from 100 patients receiving low-dose-rate prostate brachytherapy were used to characterize the prostatic urethra, contoured on ultrasound using aerated gel. Consensus contours, defined using majority vote, described commonalities in cross-sectional shape across patients. Potential simplified surrogates were defined and evaluated against the true urethra. The best performing surrogate, a circle of varying size (CS) was retrospectively contoured on 85 high-dose-rate prostate brachytherapy treatment plans. Dose to this recommended surrogate was compared with urethral doses estimated by the standard 6 mm circle surrogate. RESULTS: Clear variation in urethral cross-sectional shape was observed along its length and between patients. The standard circle surrogate had low predictive sensitivity (61.1%) compared with true urethra because of underrepresentation of the verumontanum midgland. The CS best represented the true urethra across all validation metrics (dice: 0.73, precision: 67.0%, sensitivity: 83.2%, conformity: 0.78). Retrospective evaluation of planned doses using the CS surrogate resulted in significant differences in all reported urethral dose parameters compared with the standard circle, with the exception of D100%. The urethral dose limit (115%) was exceeded in 40% of patients for the CS surrogate. CONCLUSIONS: The proposed CS surrogate, consisting of circles of varying diameter, is simple yet better represents the true urethra compared with the standard 6 mm circle. Higher urethral doses were predicted using CS, and the improved accuracy of CS may offer increased predictive power for urethral toxicity, a subject of future work. Crown Copyright Ó 2019 Published by Elsevier Inc. on behalf of American Brachytherapy Society. All rights reserved.

Keywords:

HDR brachytherapy; Prostate brachytherapy; Dosimetry; Urethral dose

Introduction High-dose-rate prostate brachytherapy (HDR-PB) is a preferred treatment for prostate cancer, associated with low toxicity and excellent tumor control for patients with

Received 20 March 2019; received in revised form 23 May 2019; accepted 28 May 2019. Conflict of interest: The authors certify that they have no affiliations with or involvement in any organization or entity with any financial interest or nonfinancial interest in the subject matter or materials discussed in this manuscript. * Corresponding author. Medical Physics, BC Cancer - Kelowna, 399 Royal Ave, Kelowna, BC, Canada V1Y5L3. Tel.: þ1 250 712 3966; fax: þ1 250 712 3911. E-mail address: [email protected] (M. Hilts).

intermediate- and high-risk disease (1). Real-time planning and steep HDR dose gradients allow HDR-PB to deliver a high dose to the prostate while better sparing organs at risk (the rectum, bladder, and urethra) (2, 3) compared with external beam radiation (4, 5, 6) and low-dose-rate prostate brachytherapy (LDR-PB) (7). Furthermore, HDR-PB enables focal dose escalation to the dominant intraprostatic lesion (8, 9). This individualized, focused dose escalation has led to excellent outcomes, including improved causespecific survival and distant metastasis-free survival for intermediate- and high-risk patients (10). Although severe toxicity from HDR-PB is rare, with Grade 3 toxicity typically occurring in fewer than 5% of patients and Grade 4 or 5 extremely rare (11, 12), larger urethral high-dose volumes correlate with both the incidence

1538-4721/$ - see front matter Crown Copyright Ó 2019 Published by Elsevier Inc. on behalf of American Brachytherapy Society. All rights reserved. https://doi.org/10.1016/j.brachy.2019.05.009

676

H. Halperin et al. / Brachytherapy 18 (2019) 675e682

and severity of acute and late genitourinary (GU) toxicities (4, 6, 13, 14, 15, 16). The urethra is a primary dose-limiting structure for HDR-PB and thus accurate urethral definition is important for treatment plan optimization. For dose escalation studies, the proximity of the urethra to targeted dominant intraprostatic lesions can result in urethral dose constraints limiting dose escalation, rendering accurate urethral representation particularly critical. Current prostate brachytherapy recommendations use the 6 mm diameter of the Foley catheter as a landmark to define the urethral location from the base to apex of the prostate (17). Aerated gel within the catheter can improve catheter visualization on ultrasound (17, 1). Although the urethra is circular in cross-section at both the base and apex, it expands in the midgland around the paired ejaculatory ducts and verumontanum (18, 19). Therefore, the 6 mm diameter catheter underrepresents the true urethral shape and volume at midgland. In this study, we characterize urethral cross-sectional shape throughout the prostate. From this characterization, urethra commonalities are observed and used to define possible surrogates with improved representation of urethral cross-section along the length of the prostate over the customary 6 mm circular surrogate. Possible surrogates are compared by evaluating their ability to represent true urethral contours and a high performing, simple to implement surrogate is defined as the preferred new surrogate. The impact of this improved surrogate on presumed HDR-PB urethral dose is evaluated by retrospectively comparing the surrogate urethral dose to that calculated using the standard 6 mm circle (Circle) on an independent group of patients. Methods Urethral characterization Urethral cross-sectional shape was defined retrospectively on preoperative transrectal ultrasound images (TRUS) used for planning for 100 patients who received LDR-PB between September 2013 and September 2017

at our institution. At the time of prostate mapping, a retrograde urethrogram is performed using aerated gel for increased urethral visualization. The urethra as visualized on these ultrasound images was defined as the true urethra. Using VariSeed 8.0 (Varian Medical Systems, Palo Alto CA), the true urethra was contoured on each 5 mm crosssectional slice from 5 mm above the prostate base to 5 mm below the apex by one of two trained observers (H.H. and S.T.). Interobserver variability in contouring was minimized by a single expert (J.C.) training both observers; frequent peer review between the two observers; and review of all challenging contours by the primary expert (J.C.). 17 patients treated within this time period were excluded because of inadequate urethral visualization. For all cases, urethral length, volume, and cross-sectional shape along urethral length were recorded. After this characterization of individual urethras, crosssectional urethral shapes were examined at five locations along the prostate length: the prostate base, one-quarter gland (Q1), midgland, three-quarter gland (Q3), and apex (Fig. 1a). With the aim of determining commonalities in cross-section across patients, contours at each location were shifted to a common point on the template grid (Figs. 1b and 1c). This procedure effectively standardized patient-specific urethral trajectories and thus allowed for comparison of urethral cross-sections at each location across all patients. At each of the defined locations along the urethra (Fig. 1a) shifted urethral contours for all patients were superimposed as shown in Fig. 2a for the midgland. To quantify observed cross-section commonalities consensus surrogates of superimposed contours were created at each location using a majority vote (MV) algorithm in MIM Maestro 6.5.5 (MIM Software, Cleveland, Ohio) (Fig. 2b). This MV algorithm selects voxels where predefined threshold fractions of the contours agree or overlap. (20). Consensus contours for 50% (MV50) and 75% overlap (MV75) were created at each of the five defined locations shown in Fig. 1a. As an example, Fig. 2b shows the MV consensus contours at urethral midgland.

Fig. 1. Characterization of urethral cross-section. (a) A sagittal view of urethral cross-sections defined at five locations along urethral length from the prostate base to apex. (b) Contours were shifted within each cross-sectional plane to a standard position on the template grid. (c) The original (blue) and shifted (pink) urethra contours that allow comparison of urethral cross-section across patients irrespective of urethral trajectory. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

H. Halperin et al. / Brachytherapy 18 (2019) 675e682

677

Fig. 2. Defining urethral commonalities. (a) Urethral contours at the midgland shifted to common grid location and superimposed and (b) commonalities between these contours quantified by majority vote urethral consensus contours, threshold 50% (orange) and 75% (green). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Surrogate definition Urethral surrogates were designed to represent the cross-sectional urethral shape at the five locations defined in Fig. 1a. The MV contours, MV50 and MV75, used to characterize urethra commonalities, were tested as surrogates. Other defined surrogates were simplified approaches designed by examination of commonalities characterized by the consensus contours. The simplified surrogates were based on combinations of circles and triangles as these contour shapes are easily accessible within the HDR-PB planning software, Vitesse (Varian Medical Systems, Palo Alto CA) and thus easy to implement clinically. Two strong options emerged: surrogates consisting of circles of varying size (CS) and surrogates consisting of a triangle at the midgland, defined by mean dimensions of midgland urethral contours, and circles at the base and apex (circletriangle-circle [CTC]). Optimal CS and CTC surrogate definitions were defined by iterating circle diameters and triangle size and comparing to consensus MV75 urethra cross-sections. The results presented here are for the top performing CS and CTC surrogates. For the CS, these were circles of the following diameters: base: 5 mm, Q1: 7 mm; midgland: 9 mm; Q3: 7 mm, apex: 4 mm. For CTC, the optimal definition was 6 mm diameter circle at base and apex, and a triangle (11.8 mm width, 9.4 mm height) at the midgland. Q1 and Q3 were linear interpolations between these structures. All four potential surrogates (MV50, MV75, CS, and CTC) are illustrated in Fig. 3.

urethra length. An example at the midgland is shown in Fig. 4. After positioning on the visualized urethra, the five urethral contours were then linearly interpolated along the entire urethral length to define a full surrogate urethra. Surrogates were then compared with the true urethra using dice similarity coefficient, positive predictive value, true positivity rate, and conformity coefficient (MIM Maestro 6.5.5). The following definitions of these validation metrics were used. Given each patient’s contoured cross-sectional urethra, defined as the true structure (U), and the specific surrogate (X), if X correctly identified U, this defined a true positive (TP). The regions where there was no overlap were defined as false positives (FP) (where X extended beyond U) and false negatives (FN) (where U extended beyond X). Dice similarity coefficient The similarity of two samples, where 0 is no intersection, and 1 is perfect overlap (21e23). Dice 5

2 ðTPÞ 2TP þ FP þ FN

Positive predictive value (precision, %) The probability that X predicts U or the measure of compromise between true and false positive (24). Precision 5

TP  100 TP þ FP

Surrogate evaluation To evaluate surrogate performance, all potential surrogates (MV50, MV75, CS, and CTC) and the standard Circle were compared with the visualized true urethra contours. To facilitate this, the surrogate cross-sections as defined in Fig. 3 were repositioned at the visualized urethra at each of their respective positions along the

True positive rate (sensitivity, %) Represents the ability of X to correctly represent U as a proportion of total volume (25). Sensitivity 5

TP  100 TP þ FN

678

H. Halperin et al. / Brachytherapy 18 (2019) 675e682

Fig. 3. Defining potential urethral surrogates. Surrogate urethral contours at the base (red), Q1 (orange), midgland (yellow), Q3 (green), and apex (blue) for potential urethral surrogates (a) MV50, (b) MV75, (c) CS, and (d) CTC. Q1 5 one-quarter gland; Q3 5 three-quarter gland; CS 5 varying circle size; CTC 5 circle-triangle-circle; MV50 5 majority vote, 50%; MV75 5 majority vote, 75%. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Conformity coefficient Measures the ratio between missegmented and correctly segmented voxels. A result of 0 indicates no missegmented voxels (26). Conformity 5

FP þ FN TP

The effectiveness of the surrogates at representing the true urethra was compared using paired Students t-tests of the above validation metrics. From this evaluation, a preferred surrogate was defined. Both the preferred and Circle surrogate were then independently validated using a separate sample of 30 LDR-PB image sets from patients treated at our institution from August 2012 to September 2013. In all image sets, a urethrogram with aerated gel was used for urethral visualization. Dice coefficients were used to determine if a statistical difference was observed between the initial 100 LDR-PB patients (September 2013 to September 2017) and external population’s 30 LDR-PB patients (August 2012 to September 2013) using a Student’s t-test. Surrogate impact on urethral dose The impact of using the newly defined preferred urethral surrogate as compared with the Circle surrogate was retrospectively assessed for 85 HDR-PB patients treated between December 2013 and January 2016 using Vitesse 3.0. All cases included use of a Foley catheter and aerated gel for urethral visualization. Needle position, source dwell positions, and treatment doses were already defined as utilized for treatment. Prostate coverage objectives were: V100% $ 98%, V125% 55e62%, and D90% $ 100% with urethral dose constraint of V115% 5 0 cc. The Circle surrogate was previously defined during the procedure. The new surrogate was defined retrospectively for all patients as described previously. Urethral doses for both surrogates were calculated (Vitesse 3.0) and compared using paired Student’s t-test. Mean (SD) doses are also reported. The following urethral dosimetric parameters were considered: D10%, D20%, D30%, D50%, D80%, D100%, D31.0cm, Dmax, V115%, V125%, and V150%. Analysis of outliers was completed by Grubb’s test for single outlier (27).

Results Urethral characterization The mean prostate length of the 100 prostates observed via preoperative TRUS images was 38  6 mm, and the average prostatic urethral volume, 1.4  0.5 mL. Urethral trajectory varied significantly, as is exemplified in Fig. 5a. Urethral cross-sectional shape at the five defined locations along the prostate (Fig. 1) also varied among patients as is illustrated in Fig. 5. However, commonalities can be observed, including a smaller circular apex and base and a larger triangular structure at the midgland (Fig. 5d). Evaluation of surrogates The results of the validation metrics for the five surrogates tested are reported in Table 1. With the exception of the CTC surrogate’s precision, all surrogates were an improvement over the standard Circle across all validation metrics. No statistical difference was observed between the MV75 and CS ( p 5 0.3). Based on these results, the CS surrogate was selected as the preferred alternate surrogate because of its good representation of the true urethra and its simplicity relative to the other high performing surrogate, MV75.

Fig. 4. To allow for evaluation, surrogates are repositioned on the visualized Foley catheter. Shown here is the standard 6 mm circle, circle (yellow) and one of the potential alternate surrogates, CS (orange) at the midgland. CS 5 varying circle size. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

H. Halperin et al. / Brachytherapy 18 (2019) 675e682

679

Fig. 5. Characterization of 100 prostatic urethras. (a) Sagittal overlay of five representative prostatic urethra contours before showing observed differences in trajectory. (b-f): Overlay of 100 urethral contours shifted to a common position on template grid (Fig. 1) at defined locations of the prostate (b) base, (c) Q1, (d) midgland, (e) Q3, and (f) apex show variation but also commonalities. The blue contour represents the MV75 contour at each location. Q1 5 one-quarter gland; Q3 5 three-quarter gland; MV75 5 majority vote, 75%. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Application of the standard Circle and the CS surrogates to an independent population indicated that the aforementioned results are robust, with no statistical difference between the two populations (treated September 2013 to September 2017 vs. August 2012 to September 2013) for both CS and Circle, p 5 0.6 and p 5 0.1, respectively. Surrogate impact on urethral dose The urethral doses calculated for Circle and CS are summarized in Table 2. One patient from the 85 HDR-PB patients was excluded as an extreme outlier because of the close proximity of a needle to the urethra in the base of the prostate ( p ! 0.001). The remaining 84 patients are included here. Urethral V125% and V150% were equal to 0 cc for all patients regardless of the surrogate type and thus are not included here. All reported dose parameters, with the exception of D100%, showed an increase in mean urethral dose with CS. This was expected due to CS having a larger diameter midgland than the standard surrogate. Looking at individual cases, CS resulted in an increase reported dose in more than 85% of cases for all dose parameters, again with the exception of D100%. This increase had measurable impact on meeting the planned constraint for urethral dose (!115%); the Circle urethra had V115% of 0 cc for all patients, whereas the V115% clinic constraint was exceeded for 34/84 patients (40% of cases) using the CS surrogate. Furthermore, the increase in dose to 1 cc of urethra was also substantial for several cases, increasing by as much as 6.7 Gy. Overall, these results demonstrate that urethral dose is frequently underestimated by the standard surrogate, as is clearly illustrated in the example in Fig. 6.

Discussion This study defines, through evaluation of 100 prostate urethras, a surrogate urethra that better reflects prostatic urethral cross-sections than standard approaches, yet remains sufficiently simple to implement in clinical settings. Although visualization of the true urethra for each individual patient would provide the most accurate representation of urethral dose, the width of the urethra at midgland when opacified hinders ultrasound visualization of paramedian treatment needles at the anterior base. The use of a surrogate not only simplifies urethral definition but provides a standardized urethra that if used consistently would facilitate comparison of urethral doses across sites and ultimately help establish a relationship between dose and observed toxicity. The characterization of 100 prostatic urethras found urethral length, volume, trajectory, and cross-sectional shape to vary substantially among patients. Cross-sectional shape is not consistent along urethra length, as it enlarges at midgland because of the verumontanum. At our institution

Table 1 Validation metrics of surrogate methods Validation

Circle

MV50

MV75

CTC

CS

Dice Precision (%) Sensitivity (%) Conformity

0.65 62.1 61.1 1.61

0.75 80.3 71.2 0.74

0.74 68.2 85.5 0.71

0.69 58.8 80.0 1.11

0.73 67.0 83.2 0.78

Circle 5 standard 6 mm diameter circle; MV50 5 majority vote, 50%; MV75 5 majority vote, 75%; CTC 5 circle-triangle-circle; CS 5 circles of varying size.

680

H. Halperin et al. / Brachytherapy 18 (2019) 675e682

Table 2 Retrospective comparison of urethral dose for the standard Circle and proposed CS surrogates for 84 HDR-PB patients treated to a prescription dose of 15 Gy Parameter

Circle mean  SD

D10% (Gy) D20% (Gy) D30% (Gy) D50% (Gy) D80% (Gy) D100% (Gy) D31.0cm (Gy) Dmax (Gy) V115% (cc)

16.4 16.3 16.2 15.9 14.0 8.0 15.0 17.1 0.00

        

0.2 0.2 0.2 0.3 1.0 1.0 1.0 0.2 0

CS mean  SD 16.6 16.4 16.3 16.1 15.4 8.0 15.9 17.5 0.02

        

0.3 0.2 0.2 0.2 0.7 2.0 0.3 0.6 0.03

p-value

Diff CS e Circle Median (range)

% With CS $ circle

!0.001 !0.001 !0.001 !0.001 !0.001 0.279 !0.001 !0.001 !0.001

0.16 0.14 0.14 0.14 1.61 0.02 0.86 0.22 0.01

86% 89% 88% 88% 86% 66% 96% 85% 100%

(0.7 to 0.63) (0.34 to 0.69) (0.33 to 0.75) (0.33 to 1.10) (0.40 to 4.93) (4.07 to 4.17) (0.32 to 6.72) (1.01 to 2.39) (0.00 to 0.14)

Circle 5 standard 6 mm diameter circle; CS 5 circles of varying size. Dosimetry is reported as mean  SD with a Paired-T test for significance. The median and range of difference between the CS and Circle reported values also given, as well as the percentage of the 84 treated cases that showed elevated dosimetry when CS was used.

for HDR-PB, we currently use a Circle placed at the visualized catheter to represent the urethra along the whole urethral trajectory. Although Circle is not adequate anatomically, it is a commonly used standard approach (used, e.g., in RTOG 0321), and use of standards is important for comparison of results between sites. However, the low sensitivity we observed for this standard surrogate illustrates clearly the deficiency in this approach and thus motivated our search for a preferred surrogate that could be easily used across all sites as a new standard. Defining urethral surrogates based on commonalities observed in urethral cross-section should lead to improvement in surrogate representation. Consensus contour structures (MV50 and MV75) provided significantly improved representation of the true urethra over the standard approach, showing enhancements in all validation metrics. However, because these structures cannot be readily produced in current HDR planning systems, they are not easily implemented clinically. Our approach here is focused on developing clinically relevant tools, where the desired surrogate is a fast, simple, and standardized approximation that can be easily applied during real-time HDR-PB planning. The CTC, an attempt to mimic the longitudinal mucosal fold of the prostatic urethra, provided improved sensitivity but showed decreased precision over the standard Circle as it did not adequately include the anterior region of most midgland cross-sectional contours. The surrogate based on circles of varying diameter (CS) performed better; in fact, no statistical difference was observed between the MV75 and CS ( p 5 0.3). Thus, the CS surrogate was selected as the preferred surrogate as it provides better representation of the true urethra compared with the Circle and it meets the criteria for a clinically useful surrogate. CS can be adapted for catheterized urethra by defining a minimum diameter appropriate for the catheter size. Furthermore, CS could be applied to prostate brachytherapy broadly, beyond the ultrasound-based planning technique described here, for example, to CT-based HDR techniques. Retrospective analysis of the impact of using this proposed surrogate on reported HDR-PB urethral dose showed

that the doses observed using the new CS surrogate showed statistically significant increases compared with the 6 mm Circle surrogate for all urethral dosimetric parameters evaluated, with the exception of D100% (Table 2). Given that CS is shown here to be an improved representation of urethral cross-sectional shape, this suggests actual delivered urethral dose may differ from what was calculated using the Circle surrogate at the time of planning. At our institution, the urethral constraint of V115% 5 0 cc was exceeded in 40% of patients when the evaluation was completed with the CS surrogate. This indicates it is likely that the true urethra is receiving a higher dose than intended, although the dose difference is not extreme. The corollary is that use of the improved CS surrogate to better represent the true urethra may impact the ability to dose escalate safely within the prostate, particularly when dominant lesions are in close proximity to the midgland urethra. The relationship between urethral dose and the severity of late GU toxicity is well established (6). Hsu et al. found evidence to support that higher urethral dose, specifically increased high dose volumes, was associated with greater acute and late GU toxicity (13). Ishiyama et al. reported values of D70%, D80%, V12Gy, and V13Gy were significantly higher in patients with late Grade 3 GU toxicity than in those with Grade 0e2 toxicity (6). However, due to

Fig. 6. An example illustrating the difference in V115% urethra dose (115% isodose shown in blue) observed using the standard circle (yellow) and proposed CS (orange) defined urethral surrogates. CS 5 varying circle size. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

H. Halperin et al. / Brachytherapy 18 (2019) 675e682

heterogeneity in dose fractionation schedules and generally low toxicity, it is difficult to establish absolute dose guidelines for normal tissues including the urethra (1). Other clinical and patient-related factors play a role in the occurrence of acute and late GU toxicity, complicating the establishment of guidelines (14, 28e31). Further research into the association between urethral dose and GU toxicity is necessary. In such work, the establishment and use of a well-defined surrogate is important as accurate estimation of urethral dose is critical if we are to understand the relationship between urethral dose and toxicity. Future work will focus on whether the use of a standardized, improved urethral representation, the CS surrogate, improves the predictive power for GU toxicity in HDR-PB.

Conclusion This study characterized the prostatic urethra along its length and through detailed examination of commonalities in cross-sectional shape proposes an improved surrogate urethral description over the standard circle. This recommended surrogate, which consists of circles of varying diameter throughout the prostate length, provides a simple and standardized yet more accurate representation of the urethra for use in HDR-PB real-time planning. A comparison of urethral dose reported using these two surrogates suggests higher urethral dose predicted by the proposed surrogate. Future work will evaluate the impact of this improved surrogate on the relation between urethral dose and observed GU toxicity.

Acknowledgments The authors would like to acknowledge Varian Medical Systems for support for Vitesse 4.0 and the BC CancerKelowna KARO fund for support of H.H. MIM Software is also acknowledged for their assistance with implementation of majority vote calculations. References [1] Yamada Y, Rogers L, Demanes DJ, et al. American Brachytherapy Society consensus guidelines for high-dose-rate. Brachytherapy 2012;11:20e32. [2] Cunha JA, Pouliot J, Weinberg V, et al. Urethra low-dose tunnels: validation of and class solution for generating urethra-sparing dose plans using inverse planning simulated annealing for prostate highdose-rate brachytherapy. Brachytherapy 2012;11:348e353. [3] Keyes M, Crook J, Morris WJ, et al. Canadian prostate brachytherapy in 2012. Can Urol Assoc J 2013;7:51e58. [4] Akimoto T, Kato H, Kitamoto Y, et al. Anatomy-based inverse optimization in high-dose-rate brachytherapy combined with hypofractionated external beam radiotherapy for localized prostate cancer. Int J Radiat Oncol Biol Phys 2006;64:1360e1366. [5] Bose S, Bahri S, Lalonde R. Treatment planning comparison between high dose rate and intensity-modulated radiation therapy for prostate cancer as a means of boost dose. J Radiother Pract 2014;13:332e339.

681

[6] Ishiyama H, Kitano M, Satoh T, et al. Genitourinary toxicity after high-dose-rate (HDR) brachytherapy combined with hypofractionated external beam radiotherapy for localized prostate cancer. Int J Radiat Oncol Biol Phys 2009;75:23e28. [7] King CR. LDR vs. HDR brachytherapy for localized prostate cancer: the view from radiobiological models. Brachytherapy 2002;1:219e226. [8] Pouliot J, Kim Y, Lessard E, et al. Inverse planning for HDR prostate brachytherapy used to boost dominant intraprostatic lesions defined by magnetic resonance spectroscopy imaging. Int J Radiat Oncol Biol Phys 2004;59:1196e1207. [9] Crook J, Ots A, Gazta~naga M, et al. Ultrasound-planned high-doserate prostate brachytherapy: dose painting to the dominant intraprostatic lesion. Brachytherapy 2014;13:433e441. [10] Kovacs G, M€uller K, Soror T, et al. Results of multiparametric transrectal ultrasound-based focal high-dose-rate dose escalation combined with supplementary external beam irradiation in intermediate- and high-risk localized prostate cancer patients. Brachytherapy 2017;16:277e281. [11] Zaorsky NG, Davis BJ, Nguyen PL, et al. The evolution of brachytherapy for prostate cancer. Nat Rev Urol 2017;14:415e439. [12] Demanes DJ, Rodriquez RR, Schour L. High-dose-rate intensitymodulated brachytherapy with external beam radiotherapy for prostate cancer: California endocurie therapy’s 10-year results. Int J Radiat Oncol Biol Phys 2005;61:1306e1316. [13] Hsu IC, Hunt D, Straube W, et al. Dosimetric analysis of radiation therapy oncology group 0321: the importance of urethral dose. Pract Radiat Oncol 2014;4:27e34. [14] Akimoto T, Ito K, Saitoh J, et al. Acute genitourinary toxicity after high-dose-rate (HDR) brachytherapy combined with hypofractionated external-beam radiation therapy for localized prostate cancer: correlation between the urethral dose in HDR brachytherapy and the severity of acute genitourinary toxicity. Int J Radiat Oncol Biol Phys 2005;63:463e471. [15] Akimoto T, Katoh H, Noda SE, et al. Genitourinary toxicity after high dose rate (HDR) brachytherapy combined with hypofractionated external-beam radiation therapy for localized prostate cancer: second analysis to determine the correlation between the urethral dose in HDR brachytherapy. Int J Radiat Oncol Biol Phys 2005;63:472e478. [16] Duchesne GM, Das R, Toye W, et al. Dose distribution and morbidity after high dose rate brachytherapy for prostate cancer: influence of V150 and V200 parameters. Australas Radiol 2002;46:384e389. [17] Hoskin PJ, Colombo A, Henry A, et al. GEC/ESTRO recommendations on high dose rate afterloading brachytherapy for localised prostate cancer: an update. Radiother Oncol 2013;107:325e332. [18] Dicker AP, Merrick G, Gomella L, et al. Basic and advanced techniques in prostate brachytherapy. Oxfordshire, ENG: Taylor & Francis Group; 2005. [19] Gonzales GF. Function of seminal vesicles and their role on male fertility. Asian J Androl 2001;3:251e258. [20] Bhavsar A, Verma S. Anatomic imaging of the prostate. Biomed Res Int 2014;2014:728539. [ePub]. [21] Penrose LS. The elementary statistics of majority voting. J R Stat Soc 1946;1:53e57. [22] Dice L. Measures of the amount of ecologic association between species. Ecology 1945;26:297e302. [23] Fotina I, Lutgendorf-Caucig C, Stock M, et al. Critical discussion of evaluation parameters for inter-observer variability in target definition for radiation therapy. Strahlenther Onkol 2012;188:160e167. [24] Wong WKH, Leung LH, Kwong DLW. Evaluation and optimization of the parameters used in multiple-atlas-based segmentation of prostate cancers in radiation therapy. Br J Radiol 2016;89:732e739. [25] Prados F, Ashburner J, Blaiotta C, et al. Spinal cord grey matter segmentation challenge. Neuroimage 2017;152:312e329. [26] Altman DG, Bland JM. Diagnostic tests 1: sensitivity and specificity. BMJ 1994;308:1552. [27] Grubbs F. Sample criteria for testing outlying observations. Ann Math Stat 1950;21:27e58.

682

H. Halperin et al. / Brachytherapy 18 (2019) 675e682

[28] Bucci J, Spadinger I, Hilts M, et al. Urethral and periurethral dosimetry in prostate brachytherapy: is there a convenient surrogate? Int J Radiat Oncol Biol Phys 2002;15:1235e1242. [29] Stankovic V, Dzamic Z, Pekmezovic T, et al. Acute and late genitourinary toxicity after 72 Gy of conventionally fractionated conformal radiotherapy for localized prostate cancer: impact of individual and clinical parameters. Clin Oncol 2016;28:577e586.

[30] Nicolae A, Murgic J, Kruljac I, et al. Dosimetric impact of interobserver catheter reconstruction variability in ultrasound-based high-dose-rate prostate brachytherapy. Brachytherapy 2017;17: 306e312. [Epub]. [31] Anderson C, Lowe G, Ostler P, et al. I-125 seed planning: an alternative method of urethra definition. Radiother Oncol 2010;94:24e29.