Assessment of scatter radiation dose received by comforters and carers during digital breast tomosynthesis mammography

Abstract

Introduction

Mammographic imaging can cause considerable stress and anxiety for some patients and may require someone to remain in the room during the procedure to provide both physical and emotional support. As such, these comforters and carers (C&C) are exposed to ionising radiation. Limited evidence is available stating the radiation dose received during a digital breast tomosynthesis (DBT) examination. This research aims to determine the optimal standing position for a C&C in the mammography room during a DBT mammogram that results in the lowest radiation dose, whilst providing high‐quality imaging, care and comfort to the patient.

Methods

A scatter detector was used to measure the dose at different standing positions of the carer relative to the patient during an examination. A polymethyl methacrylate (PMMA) phantom was also used to model the patient's breast and torso for further scatter dose measurements.

Results

The median air kerma for craniocaudal views posterior to the patient is 0.75 μGy compared with 10.1 μGy to either side. The median air kerma for mediolateral oblique views for posterolateral position is 0.41 μGy compared with 2.6 μGy anterolateral. No significant effect from breast density is noted from the dataset.

Conclusion

The optimal position for the C&C to stand is directly behind the patient in the craniocaudal position, and as far as possible posterolateral to the breast being imaged in the mediolateral oblique position. These two positions will result in the least radiation dose to the C&C.

To receive the least ionising radiation during digital breast tomosynthesis, a comforter or carer should stand posterolateral to the patient on the side of the breast being imaged.

Introduction

Digital breast tomosynthesis (DBT) is routinely used in diagnostic mammography and has shown promising results in breast screening. ¹ , ² Positioning for a mammogram can be very challenging, regardless of the patient's ability level. For this reason, sometimes, it is necessary for a comforter and carer (C&C) to remain in the room and assist the patient during the procedure. Comforter and carers support patients undergoing medical examinations and can be key to the successful performance of a mammographic procedure. Consequently, they may be exposed to ionising radiation whilst assisting the patient. The radiation dose to C&Cs must abide to the as low as reasonably achievable (ALARA) principal, meaning the dose must be minimised as much as reasonably possible, considering societal and economic factors. ³ Therefore, C&Cs should stand in a position that receives the lowest radiation dose without compromising their role assisting the patient.

The International Commission on Radiological Protection (ICRP) recommends that C&Cs limit exposure to 5 mSv radiation dose per episode of care. ³ Variations of this recommendation exist around the world. For example, in Australia, the Australian Radiation Protection and Nuclear Safety Agency (ARPANSA) Code C‐5 recommend 1 mSv per diagnostic radiological examination. ⁴ These dose constraints assist in the goal of reducing radiation doses and any potential associated radiation risk. For a C&C, the ideal position would be to stand behind the operator's protective screen during the mammogram, thereby not being subject to a measurable radiation dose. However, there are many reasons why this may not be possible. Importantly, the Privacy Act of 1988 mandates protection of patient information ⁵ and a C&C standing behind the console may be privileged to personal and sensitive patient information breaching confidentiality agreements. Secondly, to the untrained eye, a normal‐looking mammogram could appear to have a pathology, which could result in undue stress to the C&C until results are received. Thirdly, space is limited inhibiting social distancing in the current COVID‐19 climate. These reasons were the impetus for assessing the scatter radiation dose to C&Cs during DBT in this study, which can be used as a reference to determine the optimal standing position. Additionally, it informs mammography staff with evidenced‐based information about the C&C radiation dose received. This is preferable to the current arbitrary definition of a C&C receiving a ‘very low dose’.

Methods

This study was performed at Austin Health, Repatriation Campus, Melbourne, Australia and granted approval by the Austin Health Human Research and Ethics Committee with the patients' informed consent.

A Hologic Selenia Dimensions (Danbury, USA) mammography system was used for imaging. It is regularly serviced by the manufacturers and meets all Australasian College of Physical Scientists and Engineers in Medicine (ACPSEM)/Royal Australian and New Zealand College of Radiologists (RANZCR) quality assurance standards. Quality Control tests are regularly performed by both mammographers and medical physicists at the recommended ACPSEM and RANZCR frequency. ⁶ , ⁷

The scatter detector used was a Step OD‐02 Survey Meter (STEP Sensortechnik und Elektronik Pockau GmbH, Pockau, Germany ⁸ ), calibrated by The Australian Nuclear Science and Technology Organisation. The scatter detector was clamped onto an intravenous drip pole at the same marked height for each examination, preventing any height discrepancy and allowing for reproducibility of results.

In addition to patient scatter radiation measurements, phantom measurements were also taken. The patient's breast was modelled by rectangular blocks of polymethyl methacrylate (PMMA), which were 24 × 30 cm² in area. Polymethyl methacrylate thicknesses of 20, 40, 60 and 70 mm were used to model a range of equivalent breast thicknesses (21, 45, 75 and 90 mm respectively; ⁹ , ¹⁰ ). These PMMA thicknesses were matched to exposures for breasts of the same equivalent breast thickness. For each PMMA block, the machine breast thickness was set to the equivalent breast thickness and exposed with an air gap. The patient's torso was also partially modelled. A 30 × 30 × 15 cm³ block of PMMA was placed against the breast support platform with the block's short axis in the patient's anteroposterior (AP) direction. This was used to model the attenuation of the scatter radiation provided by the patient's body. To compare these phantom measurements to the patient data, the median of all the patient scatter data with a compressed breast thickness within 10 mm of the patient phantom thickness was found, that is to compare to the 75 mm phantom results, and all patient data from 65 to 85 mm (inclusive) were selected and the median found.

Four DBT images were acquired with C‐view (a synthesised 2D image) as per protocol. This included two craniocaudal images (CC) and two mediolateral oblique (MLO) images. The scatter detector was placed at four different marked floor positions during imaging, with a maximum of four measurements per patient. The detector was placed so the centre of the pole was at the centre of the cross marks on the ground with the scatter detector receiver pointing directly towards the patient's body and the scatter detector cap removed. Figure 1 shows the set up of the room during dose recording and the scatter detector position relative to the patient in each imaging position. Position one collected data during the right CC; the detector was placed next to the patient to represent a C&C standing next to the patient enabling eye contact. Position two was for the left CC; the detector was placed behind the patient, to represent the C&C standing behind the patient enabling physical contact. Position three was for the right MLO; the detector was placed to the left side and slightly in front of the patient, enabling eye contact only. Position four was for the left MLO; the detector was placed behind the patient towards the corner of the room, allowing eye contact only and represented the furthest position a C&C can stand from the patient. These four positions were selected as they represent the most commonly used positions by C&Cs. It is important to note that the imaging protocol was not altered for this project and no patient received extra, unnecessary radiation. Consequently, for each projection, the scattered radiation was measured at one position only.

Figure 1.

Schematic diagram of the room layout in relation to the scatter detector position and patient (A) The scatter detector was placed in one of the four positions for CC views and MLO views, as marked by the crosses. (B) The positioning of the patient and detector during image acquisition. The arrows represent the direction the patient's head is facing. The X‐ray tube head is directly above the detector and breast for all four positions. The distance from the patient for positions 1, 2, 3 and 4 was 77, 75, 86 and 130 cm, respectively.

The scatter detector cap was removed to enhance the uniformity of the detector energy response in the mammography energy range and the detector was zeroed before the first image acquisition of each patient. Mammography breast tissue metrics (density and compressed breast thickness) were also captured to correlate radiation scatter measurements with breast tissue type. Patient images were examined by a Consultant Radiologist and assigned a density score as per the Breast Imaging Reporting and Data System (BI‐RADS). ¹¹ The following factors were also recorded: target and filter combination, kVp, mAs, EI value, paddle size, compressed breast thickness, compression force, mean glandular dose (MGD) and scattered radiation dose.

The temperature for each scatter detector reading was recorded and used to temperature correct the reading of the open‐air ionisation chamber. The scatter detector readings have been corrected for the energy response of the detector using correction factors and the mean beam energy of the primary beam for each exposure. ⁸ , ¹² These corrected ambient dose equivalent values have been converted to air kerma values using conversion factors found in the literature. ¹³

The uncertainty of the scatter radiation measurements was calculated using the Root Mean Square Error. The dosimeter reading is predominately altered by the placement of the dosimeter relative to the radiation source, its calibration and the mammography system output uncertainty. It was found to be ±0.07 μGy. This value is applicable to air kerma measurements in Figures 2, 3, 4, 5 and has not been shown graphically.

Figure 2.

Boxplots of the scatter radiation measured at each position.

Figure 3.

Scatter radiation as a function of compressed breast thickness for each position. Breast density, scored using BI‐RADS, is indicated by the colour of the marker. Note that the vertical axis scale differs between graphs.

Figure 4.

Isodose curves of the weighted average scatter radiation air kerma around the patient phantom up to a radius of 3 m.

Figure 5.

Comparison of scatter radiation measurements in the phantom and the patient.

Results

A total of 80 patients and 297 imaging views were included in this study. For each view, the variation in median compressed breast thickness, mean glandular dose, compression force, and breast density is summarised in Table 1.

Table 1.

Summary of the breast statistics recorded in this study.

	R‐CC (Position 1)	L‐CC (Position 2)	R‐MLO (Position 3)	L‐MLO (Position 4)
Total number of views	70	79	71	77
Density A	9 (13%)	10 (13%)	9 (13%)	9 (12%)
Density B	26 (37%)	29 (37%)	26 (37%)	29 (38%)
Density C	25 (36%)	30 (38%)	26 (37%)	29 (38%)
Density D	10 (14%)	10 (13%)	10 (14%)	10 (13%)
Median (IQR) compressed breast thickness (mm)	64 (20)	63 (20)	62 (20)	63 (24)
Median (IQR) compression force (N)	60 (19)	56 (20)	60 (24)	56 (23)
Median (IQR) MGD (mGy)	2.42 (1.20)	2.33 (1.10)	2.26 (1.29)	2.37 (1.49)
Median (IQR) scattered air kerma to C&C (μGy)	10.1 (6.4)	0.75 (0.57)	2.6 (1.9)	0.41 (0.35)

Abbreviations: IQR = interquartile range. mm = millimetres. N = newtons. MGD = mean glandular dose. mGy = milligray. C&C = comforter and carer. μGy = microgray.

The breasts were assigned a density score and expressed as a percentage of the total views performed for each position. The median value and interquartile range (IQR) is shown here.

A comparison of the scatter radiation measured at each position for the patient cohort is given in Figure 2. In the posterior position (L‐CC), the patient attenuates the scatter radiation by a factor of 13 compared with the patient's side (R‐CC). This demonstrates that unless it is essential for the C&C to be in the patient's eyesight, the C&C should stand behind the patient. The median scatter from the patient's posterolateral for the L‐MLO view (position 4) is over six times lower than the scatter radiation partially in front of the patient (position 3). The scatter radiation as a function of the compressed breast thickness is given in Figure 3 for each of the four positions. In the R‐MLO position, the patient provides some attenuation, but this is inconsistent. This variability in coverage is demonstrated in the large variation in the results, likely due to patient body habitus. Due to the sample size for each compressed breast thickness and density categories (being no more than 30 at maximum) it is difficult to conclude a meaningful relationship between the scatter radiation and the breast tissue density at all positions.

The isodose curves in Figure 4 represent the scatter map around the room for both CC and L‐MLO views. These scatter maps have been prepared using scatter radiation measured with the phantom. The distribution of patient compressed breast thicknesses collected for 1 year (1727 patients) has been used to weight each of the scatter measurements for each modelled breast thickness to create a weighted average scatter measurement; the proportion of patients in each compressed thickness group is multiplied by the scatter radiation measured for that phantom thickness yielding the weighted average scatter measurement. This gives a value for the average phantom scatter for the patient breast thickness population. This distribution is given in Table 2.

Table 2.

Distribution of average compressed breast thicknesses recorded from 12 months of patient imaging.

Equivalent breast compressed thickness group	21 mm	45 mm	75 mm	90 mm
Range of average compressed breast thicknesses included	Less than 33 mm	Between 33 mm and 59 mm	Between 60 mm and 82.5 mm	Greater than 82.5 mm
Number of patients	104	762	786	75
Percentage	6.0%	44.1%	45.5%	4.3%

Figure 5 demonstrates the patient scatter measurements compared with measurements taken using a phantom set up for each position. Phantom measurements correlate well with the patient measurements. However, the magnitudes of the patient and phantom scatter values differ with the phantom measurements almost always significantly larger than the patient values.

Discussion

The results indicate that for both CC views, the C&C should stand posterior to the patient rather than to their sides (position 2 rather than position 1) in order to receive the lowest scatter possible. As for both MLO views, the C&C should stand to the posterolateral of the patient on the side being imaged rather than partially in front of the patient (Figure 2). Note that for the R‐MLO view, the C&C should stand to the patient's posterior right side (mirror position 4 across the gantry). In these MLO positions, although the patient does not provide significant attenuation of the scatter radiation, the larger distance results in a lower scattered radiation dose at the C&C.

The amount of scatter radiation is positively correlated to the patient compressed breast thickness, field size, photon fluence and energy range, and can vary by an order of magnitude depending on the value of the compressed breast thickness (Figure 3). The target filter combination (W/Al) is the same for all compressed breast thicknesses.

The patient doses measured in this study compare favourably to commonly used scatter radiation values found in the literature. ¹⁴ , ¹⁵ The highest scatter at 1 m is found in the R‐CC direction and gives an average value of 7.0 μGy per R‐CC image for the patients imaged in this study compared with 7.6 μGy per image stated in the British Institute of Radiology shielding report for conventional imaging. ¹⁵

The distribution of patient compressed breast thicknesses for each CC and MLO view is similar and allows for fair comparison between the dose a C&C may receive standing at each position. The median compressed breast thickness for each view and median compression force varies over a narrow range from 62 to 64 mm and 56 to 60 N, respectively. These quantities indicate a similar compression for each view and provide confidence in the comparison of the measured scatter radiation. Breast compositions can vary in different populations ¹⁶ and although the overall scatter profile may increase or decrease depending on the populus, the correlation is expected to be akin with peak scatter values at similar positions.

The ICRP provides conversion coefficients for air kerma to effective dose. ¹³ At the mammography energies used in this study (26–43 kVp), this is typically <1. For simplicity, we used a coefficient of 1 for all energies. Consequently, the values presented may be considered an overestimate when compared to the dose constraints and natural background radiation.

In the interest of ALARA, the authors recommend positions 2 and 4 identified in this study to be optimal if a C&C is needed during DBT imaging. A C&C standing in the optimal position will receive a median dose of 0.75 μSv for each CC view and 0.41 μSv for each MLO view, giving a total of 2.3 μSv for a median compressed breast thickness. To put this into perspective, the average daily exposure to natural background radiation in Australia is equivalent to about 5 μSv. ¹⁷ By comparison, a C&C standing in suboptimal positions 1 and 3 will receive a total of 25 μSv from a median compressed breast thickness, about five times the Australian natural daily background level. The dose measured in the optimal and suboptimal positions range from 0.14 to 4.3 μSv and 0.4 to 27 μSv, respectively. Notably, the dose will vary significantly with the compressed breast thickness and breast composition. Of importance, all radiation values measured are below the ARPANSA 1 mSv C&C dose constraint. ⁴

There is a positive correlation between the patient and phantom scatter radiation measurements at equivalent breast thicknesses, although the scatter radiation dose is almost always larger for the phantom. This is attributed to the larger volume of the phantom compared with the actual breast volume resulting in greater amounts of scatter radiation. The phantom measurements were performed with a standard large‐area PMMA phantom (24 × 30 cm²) covering the whole detector. Indeed, previous studies have shown greater scatter radiation from a larger breast phantom compared with a smaller phantom. ¹⁸ , ¹⁹ Simulated breast thicknesses greater than 21 mm resulted in a significant increase in the measured scatter radiation (Figure 5), consistent with other studies. ¹⁶ The discrepancy between patient and phantom scatter radiation measurements is largest for the R‐MLO views, likely being due to the positioning of the patient's body that affords more shielding from the scatter radiation. The body phantom was a block of 30 × 30 cm² PMMA with no arms or shoulders and only partially attenuated the scatter radiation compared with a patient. Although not the aim of this research, it is hypothesised that in addition to patient position, clinical AEC selection in combination with varying breast composition and thickness will contribute towards the difference in the measured patient scatter values versus PMMA.

Although not measured during the patient scatter measurements, it is important to note that it is not preferable to stand on the mirror of position 4 for L‐MLO views, and at position 4 for R‐MLO views, that is anterolateral to the breast being imaged. This is demonstrated by the phantom isodose curves in Figure 4, where the scatter measured in the previously mentioned positions is significantly larger at a given distance than the scatter on the posterolateral on the same side as the imaged breast. It can also be seen in Figure 4 that if the C&C is required to be near the patient, then they should be directed to stand more directly behind the patient to ensure better radiation shielding coverage by the patient.

Limitations

This study addressed scatter radiation at only one height of 130 cm. Commonly, the direction of the highest scatter is found in either the backscatter or forward scatter direction ¹⁸ and the dose distribution in the vertical plane is anisotropic. ¹⁸ , ²⁰

The isodose curves were prepared using the phantom scatter measurements rather than patient data, and they are expected to be fairly conservative and overestimate the scatter radiation for most patient compressed breast thicknesses. Nonetheless, since they use a weighted average of the patient compressed breast thickness, for patients with larger breasts, the actual scatter can be larger than that shown.

Measurements were obtained from one DBT system, and although the scatter profile from other DBT systems may vary slightly, it is expected that the optimal C&C position determined in this study would be an applicable reference to most mammography systems.

Conclusion

This study concludes that C&Cs should stand posterior to the patient for either CC view and to the posterolateral of the breast being imaged for MLO views to receive the lowest scatter radiation dose during DBT, that is behind the patient and to the right for the right MLO and behind the patient and to the left for the left MLO. This can be used as a valuable reference for radiographers.