Integration of kerma-area product and cumulative air kerma determination into a skin dose tracking system for fluoroscopic imaging procedures

Abstract

The skin dose tracking system (DTS) that we developed provides a color-coded mapping of the cumulative skin dose distribution on a 3D graphic of the patient during fluoroscopic procedures in real time. The DTS has now been modified to also calculate the kerma area product (KAP) and cumulative air kerma (CAK) for fluoroscopic interventions using data obtained in real-time from the digital bus on a Toshiba Infinix system. KAP is the integral of air kerma over the beam area and is typically measured with a large-area transmission ionization chamber incorporated into the collimator assembly. In this software, KAP is automatically determined for each x-ray pulse as the product of the air kerma/ mAs from a calibration file for the given kVp and beam filtration times the mAs per pulse times the length and width of the beam times a field nonuniformity correction factor. Field nonuniformity is primarily the result of the heel effect and the correction factor was determined from the beam profile measured using radio-chromic film. Dividing the KAP by the beam area at the interventional reference point provides the area-averaged CAK. The KAP and CAK per x-ray pulse are summed after each pulse to obtain the total procedure values in real-time. The calculated KAP and CAK were compared to the values displayed by the fluoroscopy machine with excellent agreement. The DTS now is able to automatically calculate both KAP and CAK without the need for measurement by an add-on transmission ionization chamber.

1. INTRODUCTION

The skin dose tracking system (DTS) we developed¹ recreates the real-time interventional procedure in a 3D graphical environment with the patient simulated to a closely matching graphical model. The cumulative radiation dose received during the intervention is illustrated by a color-coded representation as shown in Figure 1. OPENGL assisted graphical rendition is performed using the machine technique parameters received through a digital Controller Area Network (CAN) bus.

Figure 1.

Display screen of DTS illustrating the color-coded distribution on an adult patient graphic for a simulated fluoroscopically guided intervention

The DTS system was originally developed to estimate the deterministic risk to the patient’s skin^2-3. The kerma-area product (KAP) closely corresponds to the energy imparted to the patient, and thus is related to the stochastic radiation risk. Most fluoroscopic machines have a large area transmission ionization chamber placed after the collimator assembly to measure the KAP or air kerma integrated over the beam area as indicated in figure 2. The chamber is larger than the collimator opening assembly and intercepts the entire x ray beam for KAP calculation. Typically, on these systems, the cumulative air kerma (CAK) is also determined by dividing the measured KAP by the beam area at the interventional reference point. The FDA requires that all fluoroscopic systems manufactured on or after June 10, 2006 must display the CAK such that it is visible “at the fluoroscopist’s working position”.⁴ This study describes two methods to automatically calculate both KAP and CAK without the need for measurement by a transmission ionization chamber.

Figure 2.

a. Illustration of KAP dosimeter location in the x-ray beam. b Picture of the control room monitor indicating the Cumulative Air Kerma and Kerma Area Product display.

2. MATERIALS AND METHODS

To determine the KAP and CAK by software, a calibration file containing the air kerma per mAs values as a function of kVp and beam filtration was determined. Figure 3 shows the experimental set up for the air kerma measurement with a 6 cc PTW ionization chamber (Type 34069, PTW, Freiburg, Germany) placed at the interventional reference point, 15 cm from the isocenter towards the x-ray tube.

Figure 3.

Experimental set up for the air kerma measurement with the ion chamber placed at reference point with a narrow beam.

The x-ray beam is inhomogeneous primarily due to the heel effect and thus the KAP is not directly proportional to beam area. A correction factor is thus needed for the calculated KAP. Figure 4 shows the experimental arrangement for the determination of the beam’s non-uniformity profile. In order to determine the non-uniformity of the beam, a sheet of Gafchromic® film (XR-RV3, ISP Corp., Wayne, NJ) is used to measure the exposure profile for a wide open beam. The film was placed in air at the reference point with the flat panel detector at its lowest possible SID so the beam could be opened to its largest size. The exposure was delivered at 70 kVp and the correction factor extracted from Gafchromic film dose profile is used for other exposure conditions considering the heel-effect dose profiles are similar for other kVp’s.⁵ After exposure, the film was scanned with an HP 7510 scanner and the red channel reading was converted to exposure using a measured calibration curve. The spatial exposure profile along the anode-cathode direction was obtained to determine the beam non-uniformity.

Figure 4.

Experimental set up for the Gafchromic film determination of field non-uniformity point for a wide beam area.

Two computational methods are employed for the KAP and CAK determination in DTS, one utilizing the CPU and a second method utilizing a Graphics Processing Unit (GPU), which provides accurate results in real time. Method one involves the determination of a correction factor from the radio chromic film beam profile by obtaining an average value for the curve as a function of field dimension. This is then normalized to the central axis value to provide a non-uniformity correction factor. The appropriate air kerma value on the central axis as read from the calibration file times the non-uniformity correction factor times the x-ray field area as determined from the collimator position provides the KAP value. Dividing this value by the x-ray field area provides the area-averaged CAK, which is what would be provided by the transmission chamber measurement method.

In the second method, the KAP and CAK determination is facilitated by the parallel computing power of a GPU in the data processor. This method is effective in providing the outcome when having the beam non uniformities in the x ray filed such as compensation filter and region of interest attenuator. A computer graphic of the KAP transmission ionization chamber is modeled by a collection of vertices in a plane perpendicular to the central axis at a reference distance from the focal spot in the DTS as shown in Figure 5. The dosimeter graphics vertices within the x-ray field are identified by means of a ray-triangle intersection algorithm and summed up to obtain the total vertex count (N_T) lying in the x ray field. The field area (A_beam) at the dosimeter is divided by the vertex count to obtain the “effective area” each vertex occupies (A_n). The air kerma value (K_i) of individual vertices are readout from the system calibration files, corrected by the inverse square of the distance from the focal spot ( K_i (r) ) and multiplied by the heel-effect non-uniformity correction factor (C_i). The non-uniformity correction factor is determined from the radiochromic film beam profile and is applied to each KAP dosimeter vertex at the corresponding angular distance from the central axis. The air kerma of the individual vertex times the area occupied by the vertex generates the vertex-specific KAP (K_i). The summation of K_i over all the vertices (N_{i=1 →} N_T) lying in the x-ray beam provides the total beam KAP (K_net) and, subsequently, the CAK (C_dose)

$$\begin{gathered} A_n=\frac{A_{beam}}{N_T},K_i(r)=\frac{K_i}{r^2} \\ {K}_{net}={\sum }_{i=1}^T{K}_i(r)A_n{C}_i \\ {C}_{dose}=\frac{K_{net}}{A_{beam}} \end{gathered}$$

Figure 5.

Screenshot of the DTS display indicating the KAP dosimeter graphic vertices superimposed on a patient model, shown for reference. Vertices are located at the interventional reference point, in a plane 15 cm below the isocenter. The air kerma is determined for each vertex within the bounds of the collimated field and integrated for KAP determination.

KAP and CAK determination was incorporated into the DTS with the following steps.

System Calibration:

1. Measure air kerma/mAs on central axis at the interventional reference point as a function of kVp and beam filtration. This becomes a calibration file.

2. Measure field non uniformity in the anode-cathode direction using Gafchromic film.

3. Method 1: The air kerma value is averaged across the beam length and is normalized by the air kerma at the central axis. This becomes the correction factor for field non uniformity for that field length for the first method.

4. Method 2: Determine the beam non uniformity correction factor for discrete dosimeter graphic vertices as a function of its angular distance from the beam center along the anode-cathode direction. This becomes the correction factor for the second method.

During a procedure:

1. The kVp, filtration, mA, pulse width, beam collimator dimensions are read from the CAN bus in real-time. The KAP is determined for each pulse as the product of the air kerma/ mAs from the calibration file for the given kVp and filtration times the mAs per pulse times the field area and nonuniformity correction.

2. The area averaged CAK value is determined by dividing the KAP by the field area. (This is what is determined by the transmission ionization chamber method for measuring KAP)

3. KAP and CAK per x-ray pulse are summed after each pulse to obtain the total cumulative procedure values.

To evaluate the software, the calculated values of KAP and CAK were compared to those values measured by the PTW Diamentor system which is indicated on the display for a Toshiba Infinix C-Arm imaging system.

3. RESULTS

Figure 6 illustrates the air kerma per mAs at 1 cm from the focal spot as a function of kVp for all available beam filters on the fluoroscopy machine. The measurements are performed at the interventional reference point using a PTW ionization chamber and the air kerma is extrapolated to the unit distance from focal spot for the calibration file to minimize the computation for subsequent inverse square corrections.

Figure 6.

Air Kerma per mAs calculated at 1 cm from the focal spot for a range of kVp’s for all available beam filters in the fluoroscopy

Figure 7 indicates the dose profile oriented along the anode-cathode direction of the beam calculated using the radio chromic film placed at the interventional reference point measured for 80 kVp; this is the maximum field dimension at that distance. For the second computational method, the correction factor is applied to each vertex and is obtained directly from this beam profile by normalizing the dose value at the point of interest to the dose at the center of beam from figure 7; the dosimeter vertices on the anode side of the field can have a correction factor as low as 0.70 obtained by normalizing the anode side dose (145 mGy) with the central axis dose (210 mGy), while on the cathode side, the correction factor changes from unity to only 0.98 determined by normalizing the cathode side dose (205 mGy) with the central axis dose (210 mGy). The correction factor for the first computational method is determined for various field dimensions in the anode cathode direction and is shown in Figure 8. The total correction due to the heel effect averaged over the entire beam is seen to be at most 5% for the largest field on this imaging system.

Figure 7.

The beam dose profile along the anode cathode direction for a field size of 16 cm at the interventional reference point determined measured using radiochromic film.

Figure 8.

A correction factor derived from the non-uniformity in beam dose profile as a function of field size in the anode cathode direction.

The imaging system displayed KAP value was calibrated using the PTW ionization chamber and a correction factor was determined. KAP and CAK values estimated by the DTS were compared with the corrected values displayed on the x-ray system monitor. Results when using the second computational method over a series of exposures for which the kVp and mAs per pulse as well as the field area was changed are shown in Figure 9 to have close agreement. Validation for the KAP by the DTS is demonstrated in Figure 9a and for the CAK is demonstrated in Figure 9b for a separate run. The two computational methods were used in the DTS for the KAP and CAK calculation and their numerical results were in close agreement. The DTS KAP is seen to agree with the KAP values indicated by the fluoroscopy machine with in less than 1% and the agreement for CAK values were within 3%, indicating the proper functioning of DTS KAP & CAK estimation.

Figure 9.

a) The ratio of DTS estimated KAP to the calibration-corrected fluoroscopy machine indicated KAP. b) the relative DTS estimated CAK to the calibration-corrected value displayed by the fluoroscopy machine for a series of

5. CONCLUSIONS

The DTS has been modified to calculate and display the kerma area product and cumulative air kerma for fluoroscopic procedures in real time with an acceptable accuracy. The ability to provide an accurate determination of KAP and CAK provides a viable alternative to the use of a transmission ionization chamber such as the PTW Diamentor system for fluoroscopy machines enabled with DTS.