ANODE HEEL EFFECT ATTENUATION IN LUMBAR SPINE RADIOGRAPHY: CAN THE USE OF ALUMINIUM FILTERS IMPROVE THE CLINICAL PRACTICE OF RADIOGRAPHERS?

Purpose: The aim of this study was to design an aluminium-based fi lter to reduce the anode heel eff ect in lumbar spine radiographs. Methods: Initially, lumbar spine examinations were observed in a public imaging department to determine standard exposure parameters. Then, the characterization of the anode heel eff ect was made using the Unfors Xi R/F detector and, based on the data collected, aluminium fi lters were designed with a wedge shape and thicknesses ranging from 0.1 to 4.0 mm. The assessment of the entrance skin dose (ESD) reduction was performed on the anthropomorphic phantom with and without fi lters, using the universal dosimeter UNIDOS E equipped with an ionization chamber. Finally, the image quality assessment was performed with the Pehamed Phantom Digrad A+K and image quality surveys were applied to radiographers and radiologists. Results and Discussion: Uniformity of the beam was achieved, especially with fi lter number 2, which presents a signifi cant variation of 9% between the cathodic and anodic side. This fi lter contributes to ESD reduction of 35% and 36% for AP and lateral projection, respectively. Also, according to radiographers and radiologists, it improves the image quality of lumbar spine radiography. Conclusion: The use of aluminium fi lters can be advantageous in the clinical practice of radiographers when performing lumbar spine radiographs since it allows the standardization of the anode heel eff ect, reduces the radiation dose to the patient and does not compromise image quality.


INTRODUCTION
Healthcare is a fundamental human right and is considered to be of major importance for global development as it is crucial for enhancing the quality of life and extending life expectancy (1). Rather than being only involved in primary care, healthcare has evolved and follows social trends, thus striving for quality and excellence. That is why healthcare services have become increasingly accessible, with a corresponding increase in the number of imaging examinations. Medical imaging procedures have been playing a key role in current medical care through their use in a variety of modalities. However, those using x-ray have the disadvantage of the radiation dose received by patients and its potential harms. This may be reduced through technological development, scientifi c research, and proper use of equipment (2). Due to the risks of radiation exposure and a gradual increase in the number of examinations, reducing the exposure dose to the patient without decreasing diagnostic image quality should be a primary concern. In the imaging fi eld, the aim is to provide a diagnostic examination with maximum image quality, while keeping the radiation levels as low as possible (ALARA principle) (3). In this sense, there are several parameters that can easily be adjusted, which allows a balance between patient radiation dose and image quality, such as the voltage (kV), the exposure currenttime product (mAs), the source-to-image receptor distance (SID) and the proper use of collimators. However, there are other factors that can also interfere with image quality that are not possible to control, such as the anode heel eff ect (4). This eff ect is defi ned as "the lower fi eld intensity towards the anode in comparison to the cathode due to lower x-ray emissions from the target material at angles perpendicular to the electron beam" (5). Physical aspects of the construction of the x-ray tube combined with the physical properties of the x-ray radiation cause a nonuniformity of the beam intensity along the anode-cathode axis known as the anode heel eff ect. The intensity of the radiation emitted at the cathode side is higher than at the anode side. In most radiographic examinations, the radiographer can naturally compensate this eff ect and take advantage of it by positioning the patient's thicker region at the cathode end. This contributes to obtaining optimal exposures for certain anatomical structures, although it does not completely cancel out the eff ect and changes in the image quality remain (2). In mammography modality, the x-ray machine features specifi c aluminium fi lters to remove non-useful low intensity x-rays and to enhance contrast sensitivity (4). It also presents a considerable anodic heel eff ect due to the anode target angle and short SID (6). Thus, in order to take maximum advantage of this eff ect, the cathode side is positioned at the chest wall. Incorrect use of this eff ect in radiographic examinations may result in cathode side overexposure and underexposure on the anode side, decreasing the image quality (7)(8)(9)(10)(11)(12). For this reason, issues such as x-ray beam intensity assessment, x-ray beam standardization, dose reduction at patient entrance, and image quality have been continuously under study. There are recent studies that investigate the infl uence of the anode heel eff ect in diff erent anatomical regions that present a greater density divergence along the tube axis (cathode to anode (7,8,13). However, since there is little evidence regarding the attenuation of this eff ect in lumbar spine radiography, the aim of this research was to evaluate the attenuation of anode heel eff ect in the lateral lumbar spine examination using a customized aluminium fi lter, as well as to assess the ESD and image quality.

METHODS
In the fi rst step, an assessment of the baseline exposure parameters for the lumbar spine radiography was conducted. Then, anode heel eff ect was evaluated and data for the design and construction of a customized fi lter were gathered. Finally, the viability of the fi lter, ESD and image quality were assessed.
Step 1: Determination of exposure parameters for lumbar spine examination A survey of the exposure parameters used to perform the lateral lumbar spine examination in a digital radiology imaging room (Philips x-ray tube SRO 2550 ROT 350) from a public hospital was applied. Data collected included the following parameters: kV, mAs, the selection of the automatic exposure control (AEC) mode, room confi guration, x-ray fi eld size at image receptor and SID.

Step 2.1: Measurement of the beam intensity along the longitudinal anodecathode axis
Exposure rates were measured along the longitudinal anodecathode axis, using the Unfors Ray Safe Xi detector based on a solid-state sensor, in order to plot the distribution of radiation intensity. The mean values of kV collected in step 1 were used, and the mAs value was decreased to 1 mAs to prevent X-ray tube overheating due to the high number of exposures. Room confi guration and SID were reproduced. The centre of the exposure fi eld was defi ned as the zero position and all measured values in the cathode direction were considered to correspond to the negative axis and in the direction of the anode to the positive axis. Several measurements were made with an increment of 1 cm along the longitudinal fi eld length.

Step 2.2 -Uniformization of the beam intensity distribution along the longitudinal anode-cathode axis
In order to uniformize the distribution of the beam intensity, the minimum value of the exposure rate obtained in the previous step was taken as the reference value. Again, several measurements were repeatedly made with an increment of 1 cm and adding aluminium half-value length (HVL) fi lters (99.5 % of purity) on top of the detector until the exposure rate values reach the reference value, thus counteracting the behaviour of the anode heel eff ect. At the same time, the required aluminium thickness values were obtained at each longitudinal axis position for subsequent fi lter design. The thickness of the aluminium HVL fi lters ranged from 0.1 mm to 4.0 mm. Exposure rate was measured along the longitudinal anode-cathode axis in order to plot the fi lter design.

Step 2.3: Aluminium fi lter construction
In this step, the technical drawing for the fi lter construction was carried out using Autodesk AutoCAD 210. Since the measurements were taken on top of the patient table, it was necessary to scale it to attach the fi lter to the bottom of the collimator assembly. To design the fi lter, a 4.0 mm thick aluminium plate, alloy 2024 (T351) with a purity of 91% to 95%, was used. Due to budgetary constraints, it was not possible to match the aluminium purity of this plate with the HVL fi lters used in the previous step, knowing that this diff erence would have an impact on the results (10). Regarding the cut of the aluminium plate, two samples were manufactured by CNC turning in two diff erent mechanical workshops facilities that specialized in this type of procedure.

Step 3.1: Evaluation of the anode heel eff ect behaviour
Measurements from the fi rst step were repeated without a fi lter and then with each fi lter sample attached to the bottom of the collimator assembly to check the eff ect of the fi lters regarding the beam intensity distribution along the longitudinal anodecathode axis. All previous confi gurations were reproduced.

Step 3.2: Evaluation of the ESD reduction
In order to evaluate the ESD on the phantom with and without fi lters, three exposures for each type of examination were carried out with an ionization chamber placed on top of the phantom in the centre of X-ray fi eld. An important consideration in this work is that ESD is obtained directly from the measurement of the transmitted radiation through the double-faced plane-parallel ionization chamber, which is in contact with the tissue-equivalent phantom, measuring the incident radiation as well as backscatter radiation.

Step 3.3: Image Quality Assessment
Three exposures were performed to assess the image quality using the Pehamed Phantom Digrad A+K. Spatial resolution, grey scale and contrast level were evaluated with this phantom, at SID of 100 cm and the size of the x-ray fi eld was adjusted to the phantom. Then, a survey was distributed to the radiologists and radiographers of the imaging department to assess the subjective image quality, using the European Guidelines on image criteria for lumbar spine examinations and a visual grading analysis (14,15). The survey included images obtained on the anthropomorphic phantom, with a total of twenty questions.

Determination of exposure parameters for lumbar spine examination
The results obtained from the exposure parameters survey of the AP and Lateral Lumbar spine radiographs were as follows: the use of AEC mode, without additional fi ltration, SID of 100 cm, source-table distance of 90.2 cm, the size of the x-ray fi eld on top of the table of 40 cm x 20 cm, large focus, and a voltage of 85 (AP) and 90 kV (Lateral).

Measurement of the beam intensity along the longitudinal anode-cathode axis
It is known that the variation of the anode heel eff ect in terms of relative intensity can vary between 75% to 120% along the longitudinal anode-cathode axis (3). In this study, as expected, the x-ray beam is less intense on the anode side and more intense on the cathode side. A variation of the relative exposure rate up to 58% occurs (Figure 1).

Uniformization of the beam intensity distribution along the longitudinal anode-cathode axis
The thickness of the aluminium required to compensate the beam intensity distribution along the cathode-anode axis was measured and presented in Figure 2. As expected, higher aluminium thickness is needed on the cathode side.
Step lines of equal aluminium thickness on this side are visible because the minimum HVL thickness of aluminium fi lters available for increment was 0.1 mm.

Aluminium fi lter construction
Based on Figure 2, a technical drawing was made scaling the length of the beam at the patient table to the length of the output of the collimator assembly. Two fi lters were constructed, fi lter 1 ( Figure 3) and fi lter 2 ( Figure 4).

Figure 4: Technical drawing of the aluminium fi lter 2. Top image refers to the cross-sectional view of the piece represents three-dimensional in the image below
There is a signifi cant variation of 17% between both cathode and anode sides. With fi lter 2, the same observation can be made, although with a variation of 9% between both cathode and anode sides. Thus, with the use of both fi lters, an almost complete uniformity of the x-ray beam can be observed. Small variations are the result of the diff erence in the aluminium purity.

Evaluation of the ESD reduction
It is possible to observe an ESD reduction in the AP and lateral projection of the lumbar spine, with the use of both fi lters compared to the examination performed without a fi lter. As seen in Figure 6, the ESD values in AP projection were 67.6 μGy; 36.9 μGy and 43.6 μGy for the confi gurations without a fi lter, with fi lter 1 and with fi lter 2, respectively. The lateral projection values were 109.2 μGy; 57.1 μGy and 69.6 μGy,

Evaluation of the anode heel eff ect behaviour
Exposure and dose rates were measured along the longitudinal anode-cathode axis in order to plot the radiation intensity distribution in diff erent confi gurations: with no fi lter, with fi lter 1 and with fi lter 2, as presented in Table 1. Figure 5 shows the radiation exposure variation along the longitudinal anode-cathode axis at the patient table in the mentioned confi gurations. Considering the results obtained with fi lter 1, a uniformization of the x-ray beam intensity on the cathode side can be observed, and from the central position of the exposure fi eld to the edge of the anode side, the values increase gradually.   respectively. Therefore, as displayed in Figure 7, a reduction of 45% was observed with fi lter 1 in AP projection and 48% in lateral projection, and lower rates are illustrated with the use of fi lter 2 (35% and 36%, respectively).

Image Quality Assessment
Dynamic range, spatial resolution and low contrast detectability were tested with the phantom for the same confi gurations. The results are presented in Table 2. No signifi cant diff erences were identifi ed between the images with fi lter 2 and without a fi lter. Filter 1 presented a higher dynamic range and a better contrast of images, but a lower spatial resolution for diagnostic images. To assess the subjective image quality, a total of 30 questionnaires were obtained from radiologists and radiographers of the imaging department. Based on the visual grading analysis, 22% of them considered that the images without a fi lter had better quality and 49% preferred the images obtained with fi lter 1. The remaining 29% identifi ed no diff erences in image quality.

DISCUSSION
In this study, the exposure parameters adopted for the examination of the lumbar spine experiments were reproduced from the actual conditions used at the department where the research was conducted, and it was found that they are in accordance with the European Guidelines on Quality Criteria for Diagnostic Radiographic Images, as well as similar studies (15)(16). The observed anode heel eff ect allowed a verifi cation of a variation of 58% in beam intensity along the longitudinal anode-cathode axis. Similar results were obtained by Gilboy with a variation of 55% (4). Also, Terry et al. have investigated the non-uniformity of the x-ray beam and they found a decrease of 16% to 40% in the radiation intensity along the anode-cathode (18). It is well known that the appropriate use of this eff ect can reduce the eff ective dose to patients in some common radiological examinations. However, due to the anatomy of the lumbar spine in most patients, the use of specifi c fi lters can enable a uniformity of the radiation beam (19)(20)(21). Aluminium alloy 2024 (T351) was used in this study for the design of the fi lters, which is a cheap material easily found on the market, but it has a lower aluminium purity than the HVL aluminium fi lters used to determine the radiation attenuation thickness along the anode-cathode longitudinal axis. This is the main limitation of the study due to fi nancial constraints. The results obtained with fi lter 1 were not satisfactory considering a maximum variation of the x-ray beam intensity of 17%, since a maximum variation of 10% was expected.
These results are mostly related to the fact that the purity of the aluminium used for fi lter manufacturing is low (between 91% and 95%) compared to the purity of HVL aluminium fi lters (99.5%). The aluminium alloy 2024 (T351) includes 3.8% to 4.9% copper as the primary alloying element, and since copper has much higher density than aluminium, this may explain the obtained results. As mentioned above, fi lter 2 was designed and built under diff erent conditions than fi lter 1, due to budget limitations, and a maximum variation of the intensity of the x-rays of 9% was reached, as initially intended. The ESD reduction with fi lter 1 was higher than with fi lter 2 due to the larger thickness of aluminium. A reduction of 45% and 48% was observed in AP and Lateral projection, respectively, and lower rates were obtained using fi lter 2 (35% and 36%). These results are more favorable to those obtained by Fung and Gilboy, where they only assessed the patient's position in relation to the cathode-anode axis, obtaining ESD variations between 12% to 26% (4). In the study by Karami et al, no meaningful diff erence was found for the measured ESD of pelvis radiography between two groups of patients (anode directed toward the feet of patients, compared to the patients in which the anode was directed toward the head) (22). A reduction on the eff ective dose (from 0.022 mSv to 0.002mSv) was achieved by Lai et al. using 0.3mm Cu fi lter in the lateral lumbar spine radiography, maintaining image quality (23). However, since the dose optimization techniques for the routine AP and Lateral lumbar spine projection have not been fully explored in the current literature, it was possible to verify that the use of specifi c fi lters can be eff ective. In addition, other studies revealed ESD values higher than those obtained in the present study, also indicating a good adequacy of the technical exposure parameters (4,17,24). Regarding the evaluation of the image quality of the radiographs, the results obtained were positive and thus support the use of a fi lter when performing lumbar spine radiographs in the clinical practice of radiographers. Since the diagnostic value of the radiographic images is highly dependent on the image quality, it was possible to successfully observe the image quality control tests for the dynamic range, spatial resolution and low contrast detectability, similarly observed in other studies (23,25,26).

CONCLUSION
It has been proven that both aluminium fi lters reduced the anode heel eff ect, achieving better uniformity of the beam with fi lter 2 (9% variation). The use of fi lters is benefi cial for patients in this kind of procedure and is in compliance with the ALARA principle since a signifi cant reduction in ESD was obtained for both fi lters, without compromising the image quality.
Thus, based on this study, it is recommended that radiographers from this imaging department consider using such fi lters when performing lumbar spine radiographs.