Introduction
Hysterectomy is surgery to remove the uterus for gynecological cancer treatment, such as cervical or endometrial cancer [1–4]. However, in some cases, cancer cells may persist or recur despite the removal of the uterus. In those cases, brachytherapy becomes a valuable treatment option. Brachytherapy after hysterectomy delivers focused radiation directly to the specific area where cancer cells may still be present, ultimately eradicating any remaining cancer cells while minimizing radiation exposure to surrounding healthy tissues [5–9]. Brachytherapy can be administered either intracavitary, where the radiation source is placed inside a body cavity like the vagina, or interstitial, involving the insertion of sources directly into the tissues surrounding the affected area. The post-hysterectomy brachytherapy specifically targets the vaginal cuff, which is the upper part of the vagina that remains after a hysterectomy. The necessity for this vaginal cuff brachytherapy is determined by several factors through a comprehensive assessment by the medical team, including the type and stage of cancer, the extent of disease spread, and the patient’s overall health [7, 9–11].
The choice of applicator, a medical device designed to place and control the radiation dose accurately, depends on factors such as the patient’s anatomy, the characteristics of the cancer, and the treatment plan developed by the radiation oncology team [8, 9, 12]. The selection of the appropriate applicator is part of the personalized and targeted approach to cancer treatment. A cylinder applicator is most chosen for vaginal cuff brachytherapy due to its ability to provide a uniform dose distribution, simplicity in use, stability during treatment, and entire dose coverage of the vagina. However, the use of the cylinder applicator always faces the relative underdosing to the ‘dog-ear’ area of the vaginal cuff and overdosing to the bladder and rectum [6, 9, 13, 14]. On the other hand, a two-ovoid applicator treats the upper part of the vaginal cuff by modifying dose laterality, enabling delivery of a relatively higher dose to the ‘dog-ear’ area of the vaginal cuff and a lower dose to the rectum and bladder. The distance between the ovoid applicators depends on the vaginal size. However, the structure inside the vagina may change due to tumor volume decreasing or tissue stretching during treatment [15]. In this case, the dose discrepancy in the tumor and normal tissue may increase compared to the dose expected in the initial treatment plan due to changes in the position of the applicator between treatment sessions [14, 16–19]. For example, if the distance between the ovoid applicators extends during the several-week treatment using the two-ovoid applicator, the dose distribution will be indented at the vaginal apex region. To safely compensate for this unexpected underdose at the vaginal apex, we proposed a three-ovoid applicator, a modified structure from the Fletcher applicator set, to improve dose coverage and the stability of dose delivery, especially to the central region of the target volume, in our previous study [20]. Our previous study showed the possibility of stable and uniform delivery of the prescription dose to the target volume without any excessive dose increase in the surrounding normal tissue using the three-ovoid applicator when an unexpected situation occurs using the two-ovoid applicator. However, the study only compared two-dimensional (2D) image-based brachytherapy plans for different applicators, which highlights the need for quantitative dosimetric evaluation in three-dimensional (3D) volume.
Meanwhile, standardized phantoms utilized for quality assurance in existing brachytherapy practices verify the accuracy of radiation source positional variations for a specific applicator type [21–26]. However, their applicability is limited when evaluating 3D dose distributions due to diverse patients’ internal treatment condition changes or variations in applicator types. Additionally, using water phantoms requires additional fixation devices for individual applicator types, and there are limitations in dose distribution evaluation using tools such as Gafchromic EBT3 film (Ashland Inc.) [27]. In this study, therefore, we fabricated a polymethyl methacrylate (PMMA)-Agar phantom by mixing Agar powder and water with a PMMA phantom. Our objective is to enable 3D dose assessment experiments for brachytherapy cost-effectively, and to utilize the phantom across various treatment conditions. We then verified the applicability of this in-house designed phantom for experimental dosimetric comparison for four different applicators.
Materials and Methods
1. Applicators for Vaginal Cuff Brachytherapy
This study utilized three distinct gynecological applicator types from Eckert & Ziegler BEBIG: the cylinder, the two-ovoid, and the novel three-ovoid, as illustrated in Fig. 1 to visually differentiate their structural designs relevant to dose delivery. The SET0025 cylinder applicator with variable shielding (90° and 180° tungsten shielding segments are available) is designed for easy insertion into the vagina. Four-diameter cylinder caps ensure a reasonable distance between the source and target area to easily adapt to the individual patient’s anatomy. Moreover, the shielding segments can be arranged to protect sensitive areas, such as organs at risk. The two- and three-ovoid applicators were constructed from the SET0109 Fletcher applicator component set (Eckert & Ziegler BEBIG), which is compatible with computed tomography (CT) and conditional for magnetic resonance. The LCT82-02 individual right and left ovoid tubes and the LC330-01 ovoid cap pair, both with four diameters from 15 mm to 30 mm and a length of 30 mm, are specifically designed to be placed on either side of the cervix or vaginal cuff. Once in place, the components interlock and are secured with screws to stabilize them and prevent movement during treatment. The three-ovoid applicator is designed to mimic the presence of three-ovoid tubes by placing the reversed tandem applicator in the middle of the two-ovoid applicator. To mimic the ovoid using the thin rod-shaped tandem, multiple LCZ03-03 holding discs, with a diameter of 15 mm, are attached to the end of the tandem, which prevents perforation of the surgical site of the vaginal cuff.
2. In-house Designed PMMA-Agar Phantom
To facilitate experimental dose verification for various applicators, an in-house PMMA-Agar phantom was designed and manufactured. Fig. 2 provides detailed schematics and photographs illustrating the phantom’s components, assembly, and critical dimensions, which are essential for understanding the experimental setup and ensuring reproducibility. An in-house designed phantom was manufactured to verify the brachytherapy dose for various applicators using PMMA (ρ≈1.18 g/cm3) and Agar materials. A box-shaped PMMA frame was manufactured with dimensions of 30 cm×20 cm×10 cm and a central cylindrical cavity with a diameter of 10 cm. The frame also has multiple holes for GAMMEX 467 tissue-mimicking inserts (Gammex Inc.) and an Exradin A1SL chamber (Standard Imaging Inc., collecting volume 0.053 cm3). Three types of PMMA cylindrical inserts were manufactured with the same dimensions as GAMMEX tissue-mimicking inserts, with 2.8 cm diameter and 7 cm length. These inserts have three types: one without any hole, one with a Farmer-type chamber hole, and one with an A1SL chamber hole. The Agar phantom was designed to be placed in the central hole of the PMMA frame. To achieve a physical density (ρ) of 1 g/cm3, distilled water and Agar powder were mixed in a ratio of 95:5 [28]. The manufacturing process of the Agar phantom involves boiling the Agar solution, which is a mixture of 950 mL of distilled water and 50 g of Agar powder, and then hardening it over 8 hours. The outer dimension of the Agar phantom was shaped using a 10 cm inner diameter pipe mold. To set a consistent virtual target position inside the Agar phantom in various experimental setups with applicators, a half-cylindrical PMMA rod, two cylindrical PMMA rods (one of which is A1SL chamber insertable), and a 5 mm thick small bolus were combined. A square-shaped EBT3 Gafchromic film (Ashland Specialty Ingredients) was placed between the half-cylindrical PMMA rod and the two cylindrical PMMA rods. In this experimental study, a 5 mm thick bolus was used as a virtual target for vaginal cuff brachytherapy, and it was placed over the two cylindrical PMMA rods. This specific target configuration was chosen to create a relatively thin, planar geometry. This design enhances the sensitivity for evaluating and comparing the dose uniformity achievable by the different applicator types across a surface representative of the vaginal cuff target, particularly for assessing dose homogeneity in the central and lateral regions. Each applicator was then placed over the bolus to uniformly deliver the 3 Gy prescription dose, following a treatment plan. Finally, the applicator and virtual target inserted Agar phantom was manufactured by pouring the boiled Agar solution into the pipe mold after placing these PMMA rods, film, bolus, and each applicator in its proper position.
3. Brachytherapy Planning
CT images of the PMMA-Agar phantom were obtained for four different applicators using a Toshiba Aquilion LB CT scanner with 120 kV and 330 mA X-ray settings. The physical quality of the PMMA-Agar phantom was evaluated using the cheese phantom, which has been widely used for dose verification in radiation therapy, and the GAMMEX insert of various materials with known physical properties. We analyzed the correlation between CT values and mass densities of various materials by examining CT values in the volume of interest in GAMMEX insert for each material in heterogeneous cheese and PMMA-Agar phantoms. A 3D vaginal cuff brachytherapy plan was established using HDRPlus software (Eckert & Ziegler) on the CT image of the PMMA-Agar phantom with five water-equivalent GAMMEX inserts (Fig. 3). The treatment plan was optimized for delivering a uniform 3 Gy prescription dose to the 5 mm thick bolus target volume using cylinder, two-ovoid, and three-ovoid applicators. The dose was delivered by a high dose rate (HDR) 62.2 GBq 60Co source, which moved within each applicator at 5 mm intervals. The dwell time and source position are the same in each applicator for both the two-ovoid applicators at wide and normal setups (detailed dwell times for these applicators are provided in Table 1). Quantitative evaluation of the dose delivered to the bolus target and A1SL chamber was performed by contouring the bolus volume and the cavity volume of the chamber on the CT images.
4. Dose Measurement
Fig. 4 illustrates the experimental setup for measuring the dose using EBT3 film and A1SL chamber in the PMMA-Agar phantom during brachytherapy with various types of applicators. The 2D dose distribution was assessed with the EBT3 film using a pre-defined calibration curve within a range of 0–928 cGy (Fig. 5). Dose distributions were analyzed using Vidar Dosimetry Pro Advantage (VIDAR Systems Corporation) red channel film scanner after a 2-hour post-irradiation of EBT3 films. Using the A1SL chamber connected to a PTW UNIDOSE Webline electrometer, the dose was acquired with a calibration coefficient in terms of absorbed dose to water by correcting temperature and air pressure. This calibration coefficient, called ND,w, can be obtained from a nationally accredited institution via 60Co beam irradiation under a source-detector distance of 80 cm and a field size of 10×10 cm2. Principally, ND,w cannot be used for measuring doses in brachytherapy; however, in this study, we utilized it to provide an approximate dose value for reference purposes.
Results
1. Phantom Quality Evaluation
Fourteen different GAMMEX inserts were included in the cheese phantom (Fig. 6A). Among them, lung (0.45 g/cm3), adipose (0.935 g/cm3), breast (0.978 g/cm3), solid water (1.018 g/cm3), and bone (1.150 g/cm3) inserts were inserted into the PMMA-Agar phantom. The CT values in volumes corresponding to each material in the CT images were analyzed, and it was found that identical CT values were observed for the same material in both images. Based on this, the density of Agar and PMMA phantoms could be estimated as 1.01 g/cm3 and 1.18 g/cm3, respectively. Fig. 6B shows a comparison of CT images obtained with cylinder, two-ovoid, and three-ovoid applicators inserted into the Agar phantom. This phantom included four GAMMEX inserts corresponding to solid water materials and one PMMA insert.
2. Plan Quality Evaluation for Various Applicators
Fig. 7 illustrates 3D brachytherapy dose distributions and dose-volume histograms (DVH) on CT images using four different applicators. The cylinder applicator shows difficulty in achieving uniform dose delivery in the peripheral region of the target volume compared to two-ovoid and three-ovoid applicators, as seen in the 3D dose distribution. On the other hand, uniform dose delivery is facilitated with the two-ovoid applicator at the normal setup across the entire bolus target volume. As shown in Fig. 7, a widening gap between the ovoid applicators (i.e., two-ovoid at the wide setup) during treatment can cause a significant dose reduction at the central region of the bolus target, while the three-ovoid applicator allows for stable and uniform delivery of the prescribed dose to the entire target volume. Fig. 7B shows the DVHs, which provide an intuitive comparison between the delivered dose to the bolus target volume and the A1SL chamber cavity volume. Table 1 provides a comparison of specific dosimetric parameters of interest in the DVH. Except for the three-ovoid, the doses delivered by the three applicators had similar values within a maximum of 0.08 Gy at the chamber cavity. However, the doses delivered to the bolus target showed significant variation. Specifically, the doses delivered to 90% of the target volume (D90 Vol%) and a 5 cm3 target volume (D5 cm3), representing the near-minimum dose and the maximum dose, respectively, differed by more than 0.5 Gy. This indicates a potential discrepancy in the dose of over 5 Gy within the target volume in the case of a 10-fraction treatment. When using the three-ovoid applicator, the dose at the chamber cavity was 0.35 Gy higher compared to that using the cylinder applicator. Moreover, the D90 Vol% and D5 cm3, and the mean dose (DMean) at the bolus target were 1.42, 1.23, and 1.97 Gy higher, respectively. In comparison to the two-ovoid applicator at the normal setup, the DMean was 0.78 Gy higher. Further analysis of dose uniformity using the ratio of a high-dose metric (represented by D5 cm3) to a near-minimum dose metric (represented by D99 Vol%) from Table 2 yielded values of 2.08, 1.68, and 1.61 for the cylinder, two-ovoid (normal), and three-ovoid applicators, respectively. This lower ratio for the three-ovoid applicator quantitatively confirms its superior dose homogeneity, indicating a smaller dose range across the target volume compared to the other applicators.
3. Measured Dose Evaluation
Table 2 shows the measured dose for each applicator setup using the A1SL chamber. It was observed that the measured dose generally matched well within the standard deviation, except for the cylinder applicator. Compared to the calculated dose at the chamber cavity, the measured dose for the cylinder applicator showed a difference of 0.26 Gy, which is twice as high as the standard deviation of the calculated dose of 0.13 Gy. This discrepancy can be caused by uncertainty in ND,w, or the easy setup change within the Agar phantom due to the applicator’s structural characteristics compared to other applicators.
Visual comparison between the planned and measured 2D dose distributions on the EBT3 films is presented in Fig. 8 for each applicator. The observed similarity in the shape and extent of the high-dose regions (e.g., ≥3 Gy) between the planned and measured distributions provides qualitative validation of the treatment planning system dose calculation accuracy in predicting the overall spatial dose patterns within the phantom. The dose distribution patterns enable a clear description of each applicator setup. The dose difference distribution illustrated in Fig. 8C indicates that the three-ovoid applicator delivered a higher dose than 1 Gy at the central region of the target volume compared to the two-ovoid at the wide setup. Considering that the DMean delivered to the bolus target volume was evaluated as about 1.54 Gy higher in the dose distribution using the three-ovoid rather than the two-ovoid at the wide setup specified in Table 2, it was a reliable result. This experimental result also indicates that the three-ovoid applicator allows more uniform dose delivery across a broader area of the target volume than conventional applicators.
Discussion
The Agar phantom is versatile owing to tissue-mimicking properties, having a similar density to water, in medical imaging and radiation therapy. It also has the advantages of customizability, stability, imaging compatibility, safety, and cost-effectiveness. These characteristics provide various options for accurate patient dose verification in brachytherapy using various applicator types. The Agar phantom could fix the four different types of applicators stably no matter the complexity of the shape. Also, it had physical properties similar to those of the bolus target, making it difficult to visually distinguish the bolus target volume in the CT image. Due to these characteristics, it could be useful for brachytherapy dose verification than the water phantom.
A key aspect of this study’s design was the use of a simplified, relatively planar target geometry (5 mm thick bolus). This approach was intentionally chosen to provide a sensitive platform for comparing the fundamental capabilities of the cylinder, two-ovoid, and three-ovoid applicators in achieving dose uniformity over a surface-like target, representative of the vaginal cuff mucosa. This geometry effectively highlights the three-ovoid applicator’s ability to maintain central dose coverage while extending laterally, mitigating potential underdosing observed with other configurations like the two-ovoid (wide) setup (Figs. 7–9). However, this simplified approach has limitations. The in-vivo vaginal cuff presents a more complex 3D structure with patient-specific curvature and potential variations in thickness. Consequently, the absolute dose values and DVH parameters obtained in this phantom study may differ from those in a clinical scenario. Adapting treatment plans, including applicator positioning and dwell time optimization, would be essential to achieve conformal dose distributions for more anatomically complex target volumes. Our findings underscore the comparative dosimetric advantages of the three-ovoid applicator regarding uniformity on a defined plane, but further investigation using more anatomically realistic phantoms, as planned in our future work employing 3D printing, is warranted to fully assess its clinical performance across diverse patient geometries.
The cylinder applicator is easy to use and can deliver the dose uniformly to the entire vaginal area, but the shape of the dose distribution is limited. Therefore, it is difficult to focus the prescription dose on the entire vaginal cuff area while minimizing the dose delivered to the bladder and rectum. The two-ovoid applicator is the form that utilizes only ovoid applicators of the Fletcher applicator set, and due to the structural characteristics in which only the end portion of the applicator is bent, its implant position can vary easily in the anterior-posterior direction. On the other hand, the three-ovoid applicator can resolve the 3D structural instability of the two-ovoid applicator owing to an inverted tandem applicator installed in the central below of the two-ovoid. Therefore, increasing the reproducibility of the applicator’s implant position for every treatment session during 3 to 10-fraction treatment is possible. Ultimately, by overcoming the limitations of the cylinder or two-ovoid applicator, the three-ovoid applicator aids in realizing high-quality brachytherapy, uniformly delivering the prescription dose to the vaginal cuff area while minimizing the dose to the normal tissue. Also, these advantages of the three-ovoid applicator could be estimated through experiments measuring dose distribution using the A1SL chamber and EBT3 film in this study.
Conclusion
In this study, we proposed utilizing the three-ovoid applicator for vaginal cuff brachytherapy with highly reproducible and stable applicator implantation, and wide-region dose coverage. We experimentally verified its dosimetric advantage compared to the conventional applicators by using an in-house designed PMMA-Agar phantom. The experimental results measuring the point dose and 2D dose distribution using the A1SL chamber and EBT3 Gafchromic film showed that the HDRPlus treatment planning system can well estimate the patient dose distribution. This work demonstrates the potential of using customized phantoms not only for comparing applicator types but also for patient-specific pre-treatment verification or quality assurance. Furthermore, it highlights the dosimetric benefit of brachytherapy with the three-ovoid applicator, suggesting it could be a valuable option for improving dose homogeneity and reproducibility in selected patient groups, and demonstrates the feasibility of using the custom phantom for verifying inter-fractional dose variations due to applicator repositioning. In the future, we plan to improve the experimental patient dose verification accuracy in 3D, considering the more realistic patient internal organ structure using a 3D printing technique.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation 
Print 
