Abstract
Objective
To design a device that can support the breast during phase-contrast tomography, and characterise its fit parameterisation and comfort rating.
Methods
27 participants were recruited to trial a system for breast support during simulated phase contrast imaging, including being positioned on a prone imaging table while wearing the device. Participants underwent a photogrammetry analysis to establish the geometric parameterisations. All participants trialled a single-cup design while 14 participants also trialled a double-cup with suction holder and all completed a series of questionnaires to understand subjective comfort.
Results
Photogrammetry revealed significant positive correlations between bra cup volume and measured prone volume (p < 0.001), and between “best fit” single-cup holder volume and measured prone volume (p < 0.005). Both holders were suitable devices in terms of subjective comfort and immobilisation while stationary. However, some re-engineering to allow for quick, easy fitting in future trials where rotation through the radiation beam will occur is necessary. Light suction was well-tolerated when required.
Conclusion
All participants indicated the table and breast support devices were comfortable, and they would continue in the trial.
Advances in knowledge
Phase contrast tomography is an emerging breast imaging modality and clinical trials are commencing internationally. This paper describes the biomedical engineering designs, in parallel with optimal imaging, that are necessary to measure breast volume so that adequate breast support can be achieved. Breast support devices have implications for comfort, motion correction and maximising breast tissue visualisation.
Introduction
What is phase contrast computed tomography for breast imaging?
While conventional X-ray-based imaging technologies rely on differences in X-ray absorption, phase-contrast X-ray imaging techniques have the additional capacity to visualise variations in X-ray refraction in tissue (i.e. phase contrast). 1 For X-ray energies used in biomedical imaging, the refraction effects contain more detailed soft tissue information than the absorption effects, however visualising refraction is less straightforward. Imaging at the Australian Nuclear Science and Technology Organisation (ANSTO) Australian Synchrotron (AS) uses a technique called propagation-based phase-contrast computed tomography (PB-CT), which requires a spatially coherent X-ray beam (also achievable with a commercial microfocus X-ray sources), a relatively long sample-to-detector distance (to improve signal-to-noise ratio), and the use of computational phase-retrieval algorithms as part of the image analysis. 2
Optimisation of the imaging parameters of PB-CT has been an ongoing collaboration between the AS and the Elettra Sincrotrone Trieste (Italy) using surgically excised breast tissue specimens, with other international partners also progressing this technology. 3,4 Our research group has demonstrated that PB-CT technology can produce superior radiological image quality when compared to absorption-based (AB) contrast alone (as acquired in clinical CT), at similar or lower doses than are used in 2D mammography, 3D digital breast tomosynthesis and breast CT. 5 Figure 1 shows the same specimen imaged with PB-CT and AB-CT at the same radiation dose (approximately 5.8 mGy). A further benefit of PB-CT over conventional mammographic imaging techniques is that it does not require compression of the breast.
Figure 1.
Images of DCIS, Grade 2; size: 13 cm x 18 cm at the same dose. (Left) PB-CT image slice: 35 keV and 9 m sample-to-detector distance. (Right) Clinical Koning CBCT image. The images are courtesy of Tavakoli Taba et al (2020). DCIS, ductal carcinoma in situ; PB-CT, propagation-based phase-contrast computed tomography.
Why is support needed?
PB-CT requires a patient support bed to transition through an arc, where the female’s breast (and holder) is rotated through the beam as the target organ in the imaging field of view (FOV); typically, the geometric centre of the breast is maintained in the centre of the FOV. The bed’s movement is controlled by a clinical-grade robot (Figure 2). An aperture is located in the centre of the bed, allowing a single breast to fall freely through when the females is lying prone, similar to prone breast biopsy tables.
Figure 2.
Design schematic of the PB-CT breast table and rotating mechanism set-up at the AS. The green element represents the radiation beam. AS, Australian Synchrotron; PB-CT, propagation-based phase-contrast computed tomography.
The resolution of the PB-CT system is very high (<100 µm) and at least on par or higher than new-to-market AB-CT systems which also use photon-counting technology. 6 The acceleration/deceleration of the PB-CT support table may produce measurable motion artefacts in addition to any involuntary/voluntary motion by the individual, hence motion correction computation may need to be applied. 7 Thus, it is envisaged that light support is required to gently immobilise the breast that will neither compress the breast tissue, nor impede the beam so as to require higher dose. Breast immobilisation devices are classified as Class 1 under the Australian Therapeutic Goods Authority for medical devices, indicating they have the lowest risk of potential harm to human users. 8
Whilst the high spatial resolution of PB-CT increases the likelihood of rotational motion artefact effects, the fast acquisition time for a full breast is less than for MRI protocols. 9 This fast scan time means that acquisition over a single breath-hold is achievable and light immobilisation or physical support of the breast to remain still is likely to be well tolerated. In this paper, we describe the process for designing a breast support device considering the table aperture, comfort, materials and fit.
Imaging and experimentation to date
The breast support device, or breast holder, needs to satisfy four key requirements: it must be compatible with a prone, rotating imaging table; it must immobilise the breast while including as much breast tissue as possible in the imaging FOV; it must have limited attenuation and effect on phase retrieval; and it must be comfortable. Currently, freshly excised wide excision and full mastectomy specimens are imaged at the AS within clinical-grade polyethylene terephthatate glycol (PETG) plastic tubs during scanning. The PETG tubs (µ = 0.085 cm−1) have a negligible impact on either the beam attenuation or phase retrieval, and yielded image quality benchmarks for consideration when developing actual breast support devices. Preliminary work, including from both Australian and Italian synchrotron researchers in designing breast support devices, have been previously published, 10 with results determining that a close fit was of particular importance to prevent air gaps and distortion of the plastic when the breast did not fill the volume and suction was applied.
The research team has explored several options related to breast support, including the manufacturing and materials, including plastics, silicone, three-dimensional (3D) printing, and vacuum forming. The shape of the breast support, from more conical to tear drop-shaped, was considered and the single-cup holder design was based on previously published work, commercially available products in radiation oncology and patents detailing breast volumes determined via breast outlines from CT imaging. 10–12 The work of Hernandez and Boone (2017) 13 discussed the challenges of breast shape and subsequently developed holders which were rotationally symmetrical, while recognising that some breast shapes have larger mediolateral or superior inferior measurements. Similarly, our double-cup holder design used the premise of rotational symmetry, taking into account that the single- and double-cup breast holders would be trialled for fit and comfort in this study, and not likely to be the final design for mass production. Immobilisation including a single-cup breast holder, and a double-cup system that allows for light suction to be applied was also considered. The single- and double-cup systems both had a flange component that allowed smooth insertion into the table aperture, preventing slippage of the holder downwards and this was modelled on a similar concept used in stereotactic radiation therapy. 12 The use of light suction to immoblise and draw breast tissue away from the chest wall was incorporated into the double-cup design, using a similar technique that was previously shown to successfully increase the amount of tissue visualised using MRI. 14 However, in our study the suction function was dispersed across a wide area of skin via small pin holes in the inner plastic ring as opposed to the suction applied directly to the nipple area. 14
Following the information gleaned and non-clinical results from the aforementioned published works, two breast holder designs were fabricated from clear PETG (which is suitable for both printing and thermoplastic 3D forming systems). This paper details findings on how the correct size and fit were determined and reports the results from the survey of participants’ perceptions of comfort when positioned prone on the robotic table.
Methods
27 female participants were recruited for both measurements and experience surveys over 2 days at the Imaging and Medical Beam Line (IMBL) at the AS in May 2021. All participants consented to a trial of the patient bed system, photogrammetry and fitting of breast holders. Ethics was approved by Monash University (Project Number 2021/26399). Participants were greeted at the IMBL clinical rooms where a short demographics survey was taken, including self-reported height, weight and bra size prior to changing into a gown. The clinical researchers (SL, ES, CC, JS) were all registered Diagnostic Radiographers with the Medical Radiation Practice Board of Australia, with experience in breast imaging. The participants were positioned prone on the patient support bed (see Figure 3 (Left)) and the breast was allowed to freely fall through the centre of the aperture where the breast holders (see Figure 3 (Right)) were to be fitted.
Figure 3.
(Left) Patient support bed. (Right) Examples of the single-cup breast holder.
Prior to the fitting, lateral–medial and inferior–superior silhouette photographs were taken (see Figure 4 (Left)) for the purposes of contouring and quantifying the volume of each breast (see Figure 4 (Right)).
Figure 4.
(Left) Breast silhouette photograph. (Right) Contours of the breast silhouettes for all participants. Figure 4 has been de-warped and corrected so the scale in horizontal and vertical is the same, with the markings on the scales on either side of the breast are at 1 cm intervals.
During the fittings, the single-cup holder (see Figure 5 (Left)) was trailed on all 27 participants. The single-cup design was manufactured in 10 different sizes, ranging from 180 to 730 ml, representing an anticipated range of breast cup sizes from A to E. A smaller number of participants (n = 14) also trialled the double-cup holder due to limitations in size availability (see Figure 5 (Right)). The available sizes were: 500 ml, 570 ml, 630 ml and 750 ml. The inner cup of the double-cup system had small holes through the plastic to allow for light suction to be applied through an opening at the apex of the outer cup. Light suction was applied via a portable medical pump (Liberty 7E-A Portable Suction Unit) to draw the breast deeper into the cup. Pressure was capped at 0.08 PSI via a regulator. Figure 5 illustrates the same female fitted with the single-cup and double-cup holders, and the possibility of breast shape changes due to the design of the holders. The design drawings of the two holders are presented in Supplementary Material 1 - Appendix A.
Figure 5.
(Left) A participant testing the single-cup holder design. (Right) The same participant testing the double-cup holder design with under. Both cups shown were fixed to the prone table. The photographs illustrate the challenges of developing a uniform support system and the heterogeneity of breast sizes
The fitting process initially involved a subjective visual appraisal and trialling of the selection of breast holders in the seated position in order to select the optimal size for a given design (single- or double-cup). Each participant was then gently positioned into the optimal holder lying prone and oblique approximately 15o to the side not being imaged; relevant comments relating to the fit were also noted down. Participants were comfortably supported with a series of radiographic sponges to maintain this position.
After the testing, each participant was asked to rate their level of comfort of the procedure, table and holders using a simple Likert scale. Three questions were included about the physical process of having the holders fitted and removed plus the overall experience of being positioned prone with the breast holders in place. Additionally, 14 participants also recorded their experiences of the application of light suction if they were fitted with the double-cup holder. Participants completed the short survey via an iPad and were asked to “choose a ranking “X” using a Likert scale that best fitted with their experience as a volunteer in the project in relation to the breast support (holder, table, positioning)”. The scale included four points, where the responses indicated to the research team the dual and interrelated outcome of comfort and further participation, including one dual negative response (1 = Uncomfortable, it would deter me from further participation,) one neutral-negative/positive response (2 = Mildly uncomfortable, but I would participate in future trials), and two responses that indicated the holder and process was neutral to comfort or comfortable, and the participant would continue as part of the study (3 = Neither uncomfortable or comfortable, I would participate in future trials, 4 = Comfortable, I would participate in future trials). The survey questions can be seen in Supplementary Material 1 - Appendix B.
Results
Participant measurements
The participants were all female with self-reported weights ranging from 50 to 104 kg (mean 69 ± 14 kg) and heights from 164 to 196 cm (mean 167 ± 9 cm). The range of bra sizes nominated by participants spanned 32- to 40-inch chests (UK size) or 10–18 (Australian size), breast cup sizes from “A” to “G” and the range of breast base diameters was between 9.0 and 16.5 cm. The measured prone breast volumes calculated using the photograph-contours ranged from 90 to 1544 cm3. There was a significant correlation between the participants’ bra cup volumes and measured prone volumes: r = 0.99 (p < 0.001); see Figure 6 (Left) where the error bars correspond to one standard deviation and one data point had no error bar as only a single participant had that bra volume. There was also a significant correlation the volumes of the best-fit prone single-cup breast holders and the measured prone volumes: r = 0.84 (p < 0.005); see Figure 6 (Right). There was not a significant correlation for the best-fit prone double-cup breast holders: r = 0.87 (p = 0.12). In Figure 6 (Right), the data points for the single-cup holder are shown by orange squares, whilst those for the double-cup holders are shown by blue circles.
Figure 6.
(Left) Measured prone volume as a function of declared bra cup volume. (Right) Measured prone volume as a function best-fit single-cup breast holder (orange squares) and best-fit double-cup holder (blue circles).
Participant comfort
The results for the patient comfort survey are shown in Figure 7. Mild discomfort was recorded for the double-cup holder for only one participant, with most indicating that they had no distinct feeling as to the comfort or discomfort of this holder type.
Figure 7.
The aggregated Likert scale results from the patient comfort surveys.
Discussion
The challenges of determining prone breast volume were addressed through the use of photogrammetry and correlation to bra size volume and best-fit volume for the available single- and double-cup holders. The limited published data relating to the relationship between bra cup volume and prone breast volume meant that the correlations in Figure 6 were required to be determined in order to facilitate the future implementation of the breast support infrastructure at the IMBL. The range of single-cup holders were shown to be significant in accommodating the participants and account for 71% of the variance in breast size. The range of double-cup holders was not sufficient to accommodate participants with the smallest or largest breasts.
The participants tolerated accessing the elevated table, prone positioning and the application of the breast holders very well with all indicating that they would participate in future. The survey, while concise, provided data that indicated that all participants would continue to participate in future extensions of the trial, with the median responses for Questions 1, 2 and 4 being ‘4’ (Comfortable, I would participate in future trials) and for Question 3 regarding the process of light suction, the median response was ‘3’ (Neither uncomfortable or comfortable, I would participate in future trials). As both the single- and double-cup holders were made of the same plastic and had a similar physical feel when lying prone, the first question “the process of fitting the breast support”, can be interpreted as very positive in terms of physical and emotional comfort. While it was difficult to quantify the additional tissue drawn away from the chest wall, triangulation between photography of a small number of participants, observation between two radiographers and feedback from females participating indicated that the breast was further immobilised and drawn downwards with light suction applied through pinholes around the perimeter of the double-cup holder. The use of distributed holes relative to a single hole at the nipple, as used by Rößler et al. (2014), was observed to draw the breast against the cup walls in an isotropic manner, which would not have been the case for a single hole configuration. This is useful in maintaining a more cylindrical breast volume whilst drawing the breast profile into the FOV, which is more beneficial from a dosimetry perspective due to the shape of the beam which is quite different to that of Rößler et al. (2014).
In terms of feedback regarding the positioning of participants and the application of all holder types, no adverse incidents were recorded. There was a strong correlation between the calculated breast volume and the best-fit of cup size, indicating the radiographers were able to subjectively judge the size of the holder required by the females through visual assessment and drawing upon their mammographic knowledge. The PETG holders were cleaned with a commercially available spray liquid used at BreastScreen Victoria in-between participants. The nature of PETG allows for robust cleaning and reuse between clients/patients and while it is expected that each holder will be used lightly in comparison to other similar clinical plastic surfaces, the holders can be easily reused once cleaned between numerous clients. One of the aims for this study was to help determine the optimal shape design of the holder and the range of sizes that would be necessary to equip a working clinic and part of that consideration was sustainability.
This study was limited by its pilot and progressive nature. It reports on the imaging considerations necessary for PB-CT to be clinically viable. All data were collected over a series of 2 days due to the availability of expert breast radiographers and hence the sizes of the holders was limited to those that had been made for these trial days. Manufacturing of the holders is currently off-site but in the same city as the AS and once the process of optimal design has been finalised, it is expected that the production pipeline will be much faster and linked to bookings of participants.
There was good diversity of body habitus and breast sizes of our participating females; however, it does not represent all females who may need breast imaging; e.g. we did not actively recruit or exclude any females with implants. We anticipate that females with implants would not consider the positioning or comfort of the breast holder uncomfortable to the point of declining to participate in a trial; however, this information is not known. Additionally, while there was good age and physically ability range of females who participated in study, further modifications will need to be considered for those with limited ranges of movement and agility to be positioned on the table.
Conclusion
This current paper describes how a recruited sample of Australian females undertook photogrammetry and trialled two different holders for fit and comfort in order to optimise breast immobilisation for a PB-CT imaging clinical trial. Both single- and double-cup breast holders were able to be fitted to the female participants, with some clear modifications in the availability of sizes in the larger range necessary to allow for imaging when the breast volume is greater than 700 ml. The females tolerated positioning on the table, with a range of holders very well and all participants indicated they would proceed to the next phase of the trial where they would commence table rotation with radiation. Visual determination of the best size of the holders for each female was possible via radiographer decision-making. The plastic material used to manufacture the holders was comfortable and can accommodate new sizing in diameters and depth. As the suction was tolerated well by participants, the double-cup holder will be prioritised for development as maximum breast tissue visualisation is a key consideration of 3D imaging of the breast in the prone position.
Footnotes
Acknowledgment: The authors would like to thank the participants from the ANSTO and Monash University.
Funding: This project was funded by the National Health and Medical Research Council of Australia (NHMRC) Project Grant (2018-2021) and the National Breast Cancer Foundation.
Contributor Information
Matthew Dimmock, Email: matthew.dimmock@monash.edu.au.
Jonathan McKinley, Email: jonathanm@ansto.gov.au.
Adrian Massey, Email: masseya@ansto.gov.au.
Daniel Hausermann, Email: danielh@ansto.gov.au.
Nathan Tam, Email: ntam0057@uni.sydney.edu.au.
Elizabeth Stewart, Email: estewart@breastscreen.org.au.
Cynthia Cowling, Email: cynthia.cowling@monash.edu.au.
Jenny Sim, Email: jenny.sim@monash.edu.au.
Patrick C Brennan, Email: patrick.brennan@sydney.edu.au.
Timur Gureyev, Email: timur.gureyev@unimelb.edu.au.
Seyedamir Tavakoli Taba, Email: amir.tavakoli@sydney.edu.au.
Cindy Schultz-Ferguson, Email: cindyschultzferguson@gmail.com.
Yobelli A Jimenez, Email: yobelli.jimenez@sydney.edu.au.
Sarah Jayne Lewis, Email: sarah.lewis@sydney.edu.au.
REFERENCES
- 1. Wilkins SW, Nesterets Y, Gureyev TE, Mayo S, Pogany A, et al. On the evolution of X‐ray phase‐contrast imaging methods and their relative merits, phil. Trans R Soc A 2014; 372: 20130021. doi: 10.1098/rsta.2013.0021 [DOI] [PubMed] [Google Scholar]
- 2. Nesterets YI, Gureyev TE. Noise propagation in x-ray phase-contrast imaging and computed tomography. J Phys D: Appl Phys 2014; 47: 105402. doi: 10.1088/0022-3727/47/10/105402 [DOI] [Google Scholar]
- 3. Tavakoli Taba S, Baran P, Lewis S, Heard R, Pacile S, et al. Toward improving breast cancer imaging: radiological assessment of propagation-based phase-contrast CT technology. Acad Radiol 2019; 26: e79-89. doi: 10.1016/j.acra.2018.07.008 [DOI] [PubMed] [Google Scholar]
- 4. Longo R, Arfelli F, Bonazza D, Bottigli U, Brombal L, et al. Advancements towards the implementation of clinical phase-contrast breast computed tomography at elettra. J Synchrotron Radiat 2019; 26: 1343–53. doi: 10.1107/S1600577519005502 [DOI] [PubMed] [Google Scholar]
- 5. Tavakoli Taba S, Baran P, Nesterets YI, Pacile S, Wienbeck S, et al. Comparison of propagation-based CT using synchrotron radiation and conventional cone-beam CT for breast imaging. Eur Radiol 2020; 30: 2740–50. doi: 10.1007/s00330-019-06567-0 [DOI] [PubMed] [Google Scholar]
- 6. Berger N, Marcon M, Frauenfelder T, Boss A. Dedicated spiral breast computed tomography with a single photon-counting detector: initial results of the first 300 women. Invest Radiol 2020; 55: 68–72. doi: 10.1097/RLI.0000000000000609 [DOI] [PubMed] [Google Scholar]
- 7. Brombal L, Arana Peña LM, Arfelli F, Longo R, Brun F, et al. Motion artifacts assessment and correction using optical tracking in synchrotron radiation breast CT. Med Phys 2021; 48: 5343–55. doi: 10.1002/mp.15084 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Australian Government Department of Health. Therapeutic Goods Administration. (2019,29 October 11). Australian regulatory guidelines for medical devices (ARGMD). Available from: https://www.tga.gov.au/publication/australian-regulatory-guidelines-medical-devices-argmd
- 9. Mann RM, Cho N, Moy L. Breast MRI: state of the art. Radiology 2019; 292: 520–36. doi: 10.1148/radiol.2019182947 [DOI] [PubMed] [Google Scholar]
- 10. Lewis SJ, Tam N, Arana Pena LM, Juria I, Tavakoli Taba S, et al. Getting a-breast of immobilisation needs for the implementation of phase contrast tomography. Fifteenth International Workshop on Breast Imaging 2020. doi: 10.1117/12.2563572 [DOI] [Google Scholar]
- 11. US7742796B2 - Breast immobilisation device and method of imaging the breast - Google Patents. (2005, October 25). Retrieved from Google.com website: HYPERLINK. Available from: https://patents.google.com/patent/US7742796B2/en
- 12. Yu CX, Shao X, Zhang J, Regine W, Zheng M, et al. GammaPod-a new device dedicated for stereotactic radiotherapy of breast cancer. Med Phys 2013; 40(5): 051703. doi: 10.1118/1.4798961 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Hernandez AM, Boone JM. Average glandular dose coefficients for pendant-geometry breast CT using realistic breast phantoms. Med Phys 2017; 44: 5096–5105. 10.1002/mp.12477 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Rößler AC, Wenkel E, Althoff F, Kalender W. The influence of patient positioning in breast CT on breast tissue coverage and patient comfort. Rofo 2015; 36: 115–22. doi: 10.1055/s-0034-1385208 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.