A multiple regression model for peak skin dose using principal component analysis in interventional radiology

This study addresses the growing concerns of increased radiation doses to patients resulting from the increased complexity of interventional radiology procedures. Despite the importance of dose management, few facilities use dosimetry systems to measure and control patient radiation doses. To aid in patient exposure control, this research aimed to predict the peak skin dose (PSD) using dose parameters from digital imaging and communication in medicine-radiation dose structured reports. The study focused on air kerma (Ka,r) and air kerma area product (KAP) values categorized into fixed dose (radiography and fluoroscopy) and motion dose (rotational digital subtraction angiography) for frontal and lateral biplane devices. Using single and multiple regression analysis, model equations for PSD were developed based on data from a radio-photoluminescence glass dosimeter and five dose parameters. Principal component analysis (PCA) was applied to consolidate the data, and multiple regression models were created using principal component scores. The results showed that rotational digital subtraction angiography had a minimal impact on PSD, whereas the Ka,r value demonstrated higher accuracy in predicting PSD than KAP. The inclusion of PCA in the multiple regression model further improved accuracy, with a root mean squared error of 226, confirming that PCA-enhanced models are more effective in predicting PSD.

Clinical benchmarking of a commercial software for skin dose estimation in cardiac, abdominal, and neurology interventional procedures

BACKGROUND

Radiation exposure from interventional radiology (IR) could lead to potential risk of skin injury in patients. Several dose monitoring software like radiation dose monitor (RDM) were developed to estimate the patient skin dose (PSD) distribution in IR.

PURPOSE

This study benchmarked the accuracy of RDM software in estimating PSD as compared to GafChromic film baseline in-vivo measurements on patients during cardiac, abdominal, and neurology IR procedures.

METHODS

The prospective study conducted in four IR departments included 81 IR procedures (25 cardiac, 31 abdominal, and 25 neurology procedures) on three angiographic systems. PSD and field geometry were measured by placing GafChromic film under the patient's back. Statistical analyses were performed to compare the software estimation and film measurement results in terms of PSD and geometric accuracy.

RESULTS

Median values of measured/calculated PSD were 1140/1005, 591/655.9, and 538/409.7 mGy for neurology, cardiac, and abdominal procedures, respectively. For all angiographic systems, the median (InterQuartile Range, IQR) difference between calculated and measured PSD was -10.2% (-21.8%-5.7%) for neurology, -4.5% (-19.5%-15.5%) for cardiac, and -21.9% (-38.7%--3.6%) for abdominal IR procedures. These differences were not significant for all procedures (p > 0.05). Discrepancies increased up to -82% in lower dose regions where the measurement uncertainties are higher. Regarding the geometric accuracy, RDM correctly reproduced the skin dose map and estimated PSD area dimensions closely matched those registered on films with a median (IQR) difference of 0 cm (-1-0.8 cm).

CONCLUSIONS

RDM is proved to be a useful solution for the estimation of PSD and skin dose distribution during abdominal, cardiac and neurology IR procedures despite a geometry phantom which is not specific to the latter type of IR procedures.

UNCERTAINTY ASSOCIATED WITH THE USE OF SOFTWARE SOLUTIONS UTILIZING DICOM RDSR FOR SKIN DOSE ASSESSMENT IN INTERVENTIONAL RADIOLOGY AND CARDIOLOGY

PURPOSE

The purpose of this work is to provide a comprehensive analysis of uncertainties associated with the use of software solutions utilizing DICOM RDSRs for skin dose assessment in the interventional fluoroscopic environment.

METHODS AND RESULTS

Three different scenarios have been defined for determining the overall uncertainty, each with a specific assumption on the maximum deviations of factors affecting the calculated dose. Relative expanded uncertainty has been calculated using two approaches: the law of propagation of uncertainty and the propagation of distributions based on the Monte Carlo method. According to the propagation of uncertainty, it is estimated that the lowest possible relative expanded uncertainty of ~13% (at the 95% level of confidence, i.e. with the coverage factor of k = 2 assuming normal distribution) could only be achieved if all sources of uncertainties are carefully controlled, whereas maximum relative expanded uncertainty could reach up to 61% if none of the influencing parameters are controlled properly. When the influencing parameters are reasonably well-controlled, realistic relative expanded uncertainty amounts to 28%. Values for the relative expanded uncertainty obtained from the Monte Carlo propagation of distributions concur with the results obtained from the propagation of uncertainty to within 3% in all three considered scenarios, validating the assumption of normality.

CONCLUSIONS

The overall skin dose relative uncertainty has been found to range from 13 to 61%, emphasizing the importance of adequate analysis and control of all relevant uncertainty sources.

Patients' Radiation Shielding in Interventional Radiology Settings: A Systematic Review

As a result of the increasing risk of developing radiation-related complications, many approaches aimed at reducing this risk and enhancing the outcomes of the patient, doctor or device operator have been developed. In this systematic review, we aim to discuss previous investigations that studied patient shielding or protection within the context of selected interventional radiology procedures. We included original studies that used K_a,r, and P_KA for the assessment of the outcomes of two procedures: transjugular intrahepatic portosystemic shunt creation (TIPS) and hepatic arterial chemoembolization (HAE). A thorough search strategy was conducted on relevant databases to identify all relevant studies. We included 13 investigations, including 12 cross-sectional studies and one randomized controlled trial. Significant diversity was found among all these studies in terms of the used modalities, which made them hard to compare. However, almost all studies agreed that using novel imaging and interventional modalities is useful when obtaining better outcomes and reducing patient radiation exposure. The use of ultrasound-guided procedures and providing adequate lead curtains has also been recommended by the identified studies in order to minimize the frequency of radiation exposure. The reported K_a,r, and P_KA were also variable between studies and were discussed within this study. Our findings indicate that unified guidelines for patient radiation shielding should be urgently investigated.

Institutional Diagnostic Reference Levels and Peak Skin Doses in selected diagnostic and therapeutic interventional radiology procedures

PURPOSE

Institutional (local) Diagnostic Reference Levels for Cerebral Angiography (CA), Percutaneous Transhepatic Cholangiography (PTC), Transarterial Chemoembolization (TACE) and Percutaneous Transhepatic Biliary Drainage (PTBD) are reported in this study.

MATERIALS AND METHODS

Data for air kerma-area product (P_KA), air kerma at the patient entrance reference point (K_a,r), fluoroscopy time (FT) and number of images (NI) as well as estimates of Peak Skin Dose (PSD) were collected for 142 patients. Therapeutic procedure complexity was also evaluated, in an attempt to incorporate it into the DRL analysis.

RESULTS

Local P_KA DRL values were 70, 34, 189 and 54 Gy.cm² for CA, PTC, TACE and PTBD respectively. The corresponding DRL values for K_a,r were 494, 194, 1186 and 400 mGy, for FT they were 9.2, 14.2, 27.5 and 22.9 min, for the NI they were 844, 32, 602 and 13 and for PSD they were 254, 256, 1598 and 540 mGy respectively. P_KA for medium complexity PTBD procedures was 2.5 times higher than for simple procedures. For TACE, the corresponding ratio was 1.6. PSD was estimated to be roughly 50% of recorded K_a,r for procedures in the head/neck region and 10% higher than recorded K_a,r for procedures in the body region. In only 5 cases the 2 Gy dose alarm threshold for skin deterministic effects was exceeded.

CONCLUSION

Procedure complexity can differentiate DRLs in Interventional Radiology procedures. PSD could be deduced with reasonable accuracy from values of K_a,r that are reported in every angiography system.

Accuracy of skin dose mapping in interventional cardiology: Comparison of 10 software products following a common protocol

PURPOSE

Online and offline software products can estimate the maximum skin dose (MSD) delivered to the patient during interventional cardiology procedures. The capabilities and accuracy of several skin dose mapping (SDM) software products were assessed on X-ray systems from the main manufacturers following a common protocol.

METHODS

Skin dose was measured on four X-ray systems following a protocol composed of nine fundamental irradiation set-ups and three set-ups simulating short, clinical procedures. Dosimeters/multimeters with semiconductor-based detectors, radiochromic films and thermoluminescent dosimeters were used. Results were compared with up to eight of 10 SDM products, depending on their compatibility.

RESULTS

The MSD estimates generally agreed with the measurements within ± 40% for fundamental irradiation set-ups and simulated procedures. Only three SDM products provided estimates within ± 40% for all tested configurations on at least one compatible X-ray system. No SDM product provided estimates within ± 40% for all combinations of configurations and compatible systems. The accuracy of the MSD estimate for lateral irradiations was variable and could be poor (up to 66% underestimation). Most SDM products produced maps which qualitatively represented the dimensions, the shape and the relative position of the MSD region. Some products, however, missed the MSD region when situated at the intersection of multiple fields, which is of radiation protection concern.

CONCLUSIONS

It is very challenging to establish a common protocol for quality control (QC) and acceptance testing because not all information necessary for accurate MSD calculation is available or standardised in the radiation dose structured reports (RDSRs).

Validation of the MC-GPU Monte Carlo code against the PENELOPE/penEasy code system and benchmarking against experimental conditions for typical radiation qualities and setups in interventional radiology and cardiology

INTRODUCTION

Interventional procedures are associated with potentially high radiation doses to the skin. The 2013/59/EURATOM Directive establishes that the equipment used for interventional radiology must have a device or a feature informing the practitioner of relevant parameters for assessing patient dose at the end of the procedure. Monte Carlo codes of radiation transport are considered to be one of the most reliable tools available to assess doses. However, they are usually too time consuming for use in clinical practice. This work presents the validation of the fast Monte Carlo code MC-GPU for application in interventional radiology.

METHODOLOGIES

MC-GPU calculations were compared against the well-validated Monte Carlo simulation code PENELOPE/penEasy by simulating the organ dose distribution in a voxelized anthropomorphic phantom. In a second phase, the code was compared against thermoluminescent measurements performed on slab phantoms, both in a calibration laboratory and at a hospital.

RESULTS

The results obtained from the two simulation codes show very good agreement, differences in the output were within 1%, whereas the calculation time on the MC-GPU was 2500 times shorter. Comparison with measurements is of the order of 10%, within the associated uncertainty.

CONCLUSIONS

It has been verified that MC-GPU provides good estimates of the dose when compared to PENELOPE program. It is also shown that it presents very good performance when assessing organ doses in very short times, less than one minute, in real clinical set-ups. Future steps would be to simulate complex procedures with several projections.

Vendor-independent skin dose mapping application for interventional radiology and cardiology

PURPOSE

The purpose of this paper is to present and validate an originally developed application SkinCare used for skin dose mapping in interventional procedures, which are associated with relatively high radiation doses to the patient's skin and possible skin reactions.

METHODS

SkinCare is an application tool for generating skin dose maps following interventional radiology and cardiology procedures using the realistic 3D patient models. Skin dose is calculated using data from Digital Imaging and Communications in Medicine (DICOM) Radiation Dose Structured Reports (RDSRs). SkinCare validation was performed by using the data from the Siemens Artis Zee Biplane fluoroscopy system and conducting "Acceptance and quality control protocols for skin dose calculating software solutions in interventional cardiology" developed and tested in the frame of the VERIDIC project. XR-RV3 Gafchromic films were used as dosimeters to compare peak skin doses (PSDs) and dose maps obtained through measurements and calculations. DICOM RDSRs from four fluoroscopy systems of different vendors (Canon, GE, Philips, and Siemens) were used for the development of the SkinCare and for the comparison of skin dose maps generated using SkinCare to skin dose maps generated by different commercial software tools (Dose Tracking System (DTS) from Canon, RadimetricsTM from Bayer and RDM from MEDSQUARE). The same RDSRs generated during a cardiology clinical procedure (percutaneous coronary intervention-PCI) were used for comparison.

RESULTS

Validation performed using VERIDIC's protocols for skin dose calculation software showed that PSD calculated by SkinCare is within 17% and 16% accuracy compared to measurements using XR-RV3 Gafchromic films for fundamental irradiation setups and simplified clinical procedures, respectively. Good visual agreement between dose maps generated by SkinCare and DTS, RadimetricsTM and RDM was obtained.

CONCLUSIONS

SkinCare is proved to be very convenient solution that can be used for monitoring delivered dose following interventional procedures.

Optimization of the Maximum Skin Dose Measurement Technique Using Digital Imaging and Communication in Medicine-Radiation Dose Structured Report Data for Patients Undergoing Cerebral Angiography

Understanding the maximum skin dose is important for avoiding tissue reactions in cerebral angiography. In this study, we devised a method for using digital imaging and communication in medicine-radiation dose structured report (DICOM-RDSR) data to accurately estimate the maximum skin dose from the total air kerma at the patient entrance reference point (Total Ka,r). Using a test data set (n = 50), we defined the mean ratio of the maximum skin dose obtained from measurements with radio-photoluminescence glass dosimeters (RPLGDs) to the Total Ka,r as the conversion factor, CFKa,constant, and compared the accuracy of the estimated maximum skin dose obtained from multiplying Total Ka,r by CFKa,constant (Estimation Model 1) with that of the estimated maximum skin dose obtained from multiplying Total Ka,r by the functional conversion factor CFKa,function (Estimation Model 2). Estimation Model 2, which uses the quadratic function for the ratio of the fluoroscopy Ka,r to the Total Ka,r (Ka,r ratio), provided an estimated maximum skin dose closer to that obtained from direct measurements with RPLGDs than compared with that determined using Estimation Model 1. The same results were obtained for the validation data set (n = 50). It was suggested the quadratic function for the Ka,r ratio provides a more accurate estimate of the maximum skin dose in real time.

Feasibility of dosimetric measurements using Al2O3:C OSL dosimeter during fluoroscopy-guided procedures

This study investigated the feasibility of dosimetric measurements using Al2O3:C optically stimulated luminescence (OSL) dosimeters during fluoroscopy-guided procedures. The linearity and energy dependence of Al2O3:C OSL dosimeters were evaluated, and the air kerma rate at the operator's position was measured. The response of Al2O3:C OSL dosimeters to short, repetitive irradiations was compared to that of long uninterrupted irradiation. The change in response of the Al2O3:C OSL dosimeter under automatic exposure rate control (AERC) was evaluated with the use of various thicknesses of polymethyl-methacrylate (PMMA) plates (15-30 cm). The Al2O3:C OSL dosimeters could detect 5µGy and showed good linearity in doses of ≥10µGy (R²: 0.997-0.999,p< 0.001). The relative response of the Al2O3:C OSL dosimeter normalised to that of 36.8 keV was 0.828-1.101 at the energies investigated (30.6-46.0 keV). The air kerma rate at the operator's position was estimated to be 2.61-7.17µGy min^-1depending on the heights representing different body parts. Repetitive short irradiations had no significant impact on the relative response of the Al2O3:C OSL dosimeters (p> 0.05). Despite a high energy dependence on the low energy beam used in fluoroscopy, the change in relative response of the Al2O3:C OSL dosimeter under AERC was within 5.7% depending on the thickness of the PMMA plates. Dosimetric measurement using Al2O3:C OSL dosimeters for patients and operators is feasible. However, one should be cautious about high standard deviations when measuring small doses of ≤20µGy using Al2O3:C OSL dosimeters. It is essential to perform intensive bleaching before measuring very small doses to minimise pre-irradiation counts.

Optimized radiological alert thresholds based on device dosimetric information and peak skin dose in vascular fluoroscopically guided intervention

OBJECTIVES

The National Council on Radiation Protection (NCRP) report no. 168 recommended that during fluoroscopically guided interventions (FGIs), each patient should be monitored when one of the following thresholds is reached: an air kerma > 5 Gy, a kerma area product (KAP) > 500 Gy.cm², a fluoroscopy time > 60 min, or a peak skin dose (PSD) > 3 Gy. Whereas PSD is the most accurate metric regarding the prevention of radiological risks, it remains the most difficult parameter to assess. We aimed to evaluate the relevance of the other, more accessible metrics and propose new optimized threshold (OT) for improved patient follow-up.

METHODS

Overall, 108 patients who underwent FGI in which at least one NCRP threshold was reached and PSD was measured were considered. The correlation between all metrics was assessed using principal component analysis (PCA). ROC curves and the sensitivity/specificity of both NCRP and OT to predict PSD > 3 Gy were evaluated.

RESULTS

The PCA shows that FGI can be decomposed with two components based on time and dose variables. Only KAP and kerma were correlated with PSD. The overall sensitivity and specificity of the new OT regarding KAP (67.6/93.0), kerma (97.3/81.7), and time (62.2/62.0) were better compared with NCRP thresholds (97.3/16.9, 40.5/95.4, and 21.6/74.7).

CONCLUSIONS

This study shows that fluoroscopy time is not a relevant metric when used to predict PSDs > 3 Gy. By adapting KAP and kerma thresholds to predict PSD over 3 Gy, patient follow-ups following vascular FGI can be improved.

KEY POINTS

• In vascular fluoroscopically guided interventions, principal component analysis demonstrates that between fluoroscopy time, KAP, and kerma, only the two last were correlated to the peak skin dose. • Optimized thresholds replacing NRCP ones obtained with ROC curves analysis were 85,451 μGy.cm², 2938 mGy, and 41 min for KAP, kerma, and fluoroscopy time respectively. • Improvements to trigger patient follow-up after vascular fluoroscopically guided interventions may be obtained by using the optimized thresholds.

BENCHMARKING THE DOSE MAP SOFTWARE FOR CLINICAL IMPLEMENTATION AND ESTABLISHMENT OF A LOCAL FOLLOW-UP PROTOCOL FOR THE MANAGEMENT OF SKIN INJURES FOLLOWING COMPLEX INTERVENTIONAL CARDIOLOGY PROCEDURES

This paper aims to validate the accuracy of the peak skin dose (Dskin,max) computed by the Dose Map software (DMS)-general electric and establish a local follow-up protocol for the management of patient skin injuries following complex interventional cardiology procedures (ICPs). Dskin,max was computed by the DMS and was simultaneously measured by a dense mesh of 72 thermoluminescent dosemeters for 20 ICP. Measured and computed Dskin,max were compared using Lin's concordance coefficient (${\rho}_c$). The implementation of a local follow-up strategy was based on a computed Dskin,max of 2 Gy. After eliminating 2 outliers, the average deviation between the two methods was 6% (range: -36 to +40%). Concordance between the two methods was moderate with ${\rho}_c$ (confidence interval) of 0.9128 (0.8541-0.9486). DMS computes Dskin,max with an acceptable accuracy and can be used to setup an individual follow-up process for patients with high skin exposure and risks.

SKIN DOSE MAPPING IN INTERVENTIONAL CARDIOLOGY: A PRACTICAL SOLUTION

Numerous cases of radiation-induced tissue reactions following interventional cardiology (IC) procedures have been reported, resulting in the need for an optimized and personalized dosimetry. At present, there are many fluoroscopy units without Digital Imaging and Communications in Medicine (DICOM) Radiation Dose Structured Report globally installed. Many of these have not been updated yet, and may never be, therefore, the main objectives of this paper are to develop an offline skin dose mapping application, which uses DICOM headers for the peak skin dose (PSD) assessment and to compare the PSD assessment results to XR-RV3 Gafchromic film for common IC procedures. The mean deviation between the measured and the calculated PSD was 8.7 ± 26.3%. Simulated skin dose map showed good matching with XR-RV3 Gafchromic film. The skin dose mapping application presented in this paper is an elegant solution and a suitable alternative to XR-RV3 Gafchromic film.