Evaluation of a model-based treatment planning system for dose computations in the kilovoltage energy range.

The ability to determine dose distribution and calculate organ doses from diagnostic x rays has become increasingly important in recent years because of relatively high doses in interventional radiology and cardiology procedures. In an attempt to determine the dose from both diagnostic and orthovoltage x rays, we have used a commercial treatment planning system (Pinnacle, ADAC Laboratories, Milpitas, CA) to calculate the doses in phantoms from kilovoltage x rays. The planning system's capabilities for dose computation have been extended to lower energies by the addition of energy deposition kernels in the 20-110 keV range and modeling of the 60, 80, 100, and 120 kVp beams using the system. We compared the dose calculated by the system with that measured using thermoluminescent dosimeters (TLDs) placed in various positions within several phantoms. The phantoms consisted of a cubical solid water phantom, the solid water phantom with added lung and bone inhomogeneities, and the Rando anthropomorphic phantom. Using Pinnacle, a treatment plan was generated using CT scans of each of these phantoms and point doses at the positions of TLD chips were calculated. Comparisons of measured and computed values show an average difference of less than 2% within materials of atomic number less than and equal to that of water. The algorithm, however, does not produce accurate results in and around bone inhomogeneities and underestimates attenuation of x rays by bone by an average of 145%. A modification to the CT number-to-density conversion table used by the system resulted in significant improvements in the dose calculated to points beyond bone.

Generation and use of photon energy deposition kernels for diagnostic quality x rays.

Accurately determining the dose from low energy x rays is becoming increasingly important. This is especially so because of high doses in interventional radiology procedures and also because of the desire to model accurately the dose around low energy brachytherapy sources. Various methods to estimate the dose from specific procedures are available but they only give a general idea of the true dose to various organs. The use of sophisticated three-dimensional (3D) dose deposition algorithms designed originally for radiation therapy treatment planning can be extended to lower photon energy regions. The majority of modern 3D treatment planning systems use a variation of the convolution algorithm to calculate dose distributions. This could be extended into the diagnostic energy range with the availability of lower energy deposition kernels ( < 100 keV). We have used version four of the Electron Gamma Shower (EGS4) system of Monte Carlo codes to generate photon energy deposition kernels in the energy range of 20-110 keV and have implemented them in a commercial 3D treatment planning system (Pinnacle, ADAC Laboratories, Milpitas, CA). The kernels were generated using the "SCASPH" EGS4 user code by selecting the appropriate transport parameters suitable for the relative low energy of the incident photons. The planning system was subsequently used to model diagnostic quality beams and to calculate depth dose and cross profile curves. Comparisons of the calculated curves have been made with measurements performed in a homogeneous water phantom.

Lung dose calculations at kilovoltage x-ray energies using a model-based treatment planning system.

The determination of the dose to organs from diagnostic x rays has become important because of reports of radiation injury to patients from fluoroscopically guided interventional procedures. We have modified a convolution/superposition-based treatment planning system to compute the dose distribution for kilovoltage beams. We computed lung doses using this system and compared them to those calculated using the CDI3 organ dose calculation program. We also computed average lung doses from a simulated radiofrequency ablation procedure and compared our results to published doses for a similar procedure. Doses calculated using this system were an average of 20% lower for AP beams and 7% higher for PA beams than those obtained using CDI3. The ratio of the average dose to the lungs to the skin dose from the simulated ablation procedure ranged from 25% higher to 15% lower than that determined by other authors. Our results show that a treatment planning system designed for use in the megavoltage energy range can be used for calculating organ doses in the diagnostic energy range. Our doses compare well with those previously reported. Differences are partly due to variations in experimental techniques. Using a three-dimensional (3-D) treatment planning system to calculate dose also allows us to generate dose volume histograms (DVH) and compute normal tissue complication probabilities (NTCP) for diagnostic procedures.

Routine quality control tests for film-screen mammographic systems with automatic exposure control.

Quality control of the contrast and density of mammograms is of extreme importance not only because of patient dose considerations but also because of the need to monitor changes in the breast over extended periods of time. A phantom and test technique has been developed and used at two institutions for monitoring the ability of mammographic generators and phototiming systems to provide consistent contrast and density. The phantom consists of a solid acrylic block and an embedded aluminum step wedge designed specially for low kVp use. Optical densities of various portions of the phantom are used to determine constancy of density and contrast. By minimizing fluctuations due to processing and film handling, normal variations were reduced enough to determine changes in contrast and density due to generator and phototimer changes equivalent to those monitored in processor quality control programs. The data have been correlated with changes in processor function. Changes in density and contrast values have also been related to phototimer malfunction and reduced image quality.

Suitability of laser stimulated TLD arrays as patient dose monitors in high dose x-ray imaging.

Skin entrance doses of patients undergoing interventional x-ray procedures are capable of causing skin damage and should be monitored routinely. Single TLD chips are not suitable because the location of maximum skin exposure cannot be predicted. Most photographic films are too sensitive at diagnostic x-ray energies for dosimetry, exhibit temporal changes in response, and require special packaging by the user. We have investigated the suitability of laser heated MgB4O7 TLDs in a polyimide binder in the range of 0.2-20 Gy. These are available in radioluscent arrays up to 30 x 30 cm for direct measurement of patient skin dose. Dose response was compared with a calibrated ion chamber dosimeter. Exposures were made at 60, 90, and 120 kVp, at low (fluoroscopy) and high (DSA) dose rates, and at different beam incidence angles. Longitudinal reproducibility and response to temperature changes during exposure were also checked. The dose response is linear below approximately 6 Gy where the slope starts to increase 2% per Gy. Errors were less than +/- 2% over a 0-80 degrees range of beam incidence angles; less than +/- 3% for both dose rate variations and kVp differences between 70 and 120 kVp. The response was unaffected by temperature changes between 20 and 37 degrees C. Reproducibility is current +/- 7%. MgB4O7 TLD arrays are suitable for patient dosimetry in high dose fluoroscopy procedures if appropriate calibrations are used. Uncertainty in skin dose measurement is less than 10%, which is substantially better than film dosimetry.

The potential for radiation-induced skin damage in interventional neuroradiological procedures: a review of 522 cases using automated dosimetry.

The Food and Drug Administration (FDA) has recommended the monitoring of radiation skin dose to patients during procedures having the potential for radiation damage. Radiologists need information about typical radiation doses during interventional procedures. The skin doses to patients during 522 interventional neuroradiological procedures have been monitored using an automated dosimetry system. Estimated entrance skin doses (ESD) were binned into 0.5 Gy increments and compared to FDA recommended thresholds for inclusion in the patient record. Percentages of procedures exceeding the above mentioned thresholds are presented. In addition, the percentage of dose in each view and the percentage of dose in fluoroscopic and digital angiographic modes are shown. Six percent of embolization procedures and one percent of cerebral angiograms are estimated to have potential for main erythema (ESD>6 Gy). All types of procedures have potential for temporary erythema and exceed the threshold for inclusion in the patient record (ESD> 1 Gy) at the 95% percentile. The types of procedures with most potential for skin damage also have significant percentages of dose in the digital angiographic mode. Thus, monitoring fluoroscopic time alone underestimates the potential for skin injury. On the other hand, combining the doses in the posterior-anterior and lateral views, tends to overestimate the potential for radiation injury.

Establishing a quality control program for an automated dosimetry system.

Automatic dosimetry systems can provide instantaneous dose and dose-rate information during fluoroscopic procedures as well as long-term records of patient doses. For this information to be useful, it is necessary that the accuracy of such systems be maintained through a rigorous quality control program. Daily and weekly quality control checks were performed on a PEMNET automated dosimetry system to determine its stability and the value of such tests in a quality control program. Weekly tests included monitoring the accuracy of the measured doses under a variety of conditions. The results of the tests indicate possible improvements in test methodology and real and potential sources of system failure and provide a statistical basis for setting quality control limits for future system monitoring.

Measurement of the dose deposition characteristics of x-ray fluoroscopy beams in water.

The purpose of this study was to characterize the x-ray dose distribution of fluoroscopy beams by measuring their percent depth dose curves and lateral dose profiles in a water phantom. Percent depth dose curves were measured near the surface with an Attix parallel plate chamber and deep within the water phantom with a Farmer-type cylindrical chamber. Percent depth dose curves were compared to published data where applicable. Lateral dose profiles were measured at depths of 2, 5, 10, and 15 cm in phantom with a Farmer chamber. Pulsed, 50 mA x-ray beams with peak tube potentials of 60, 80, 100, and 120 kV and half value layers of 1.89, 2.52, 3.20, and 4.09 mm Al, respectively, were investigated.

Radiolucent prosthetic gel.

The successful use of silicone breast implants is complicated by their interference with mammography. We have evaluated clinically available implant filling materials and found that a new Bio-Oncotic gel approximates the radiolucency of normal breast tissue. Silicone implants completely obscure areas of the breast in mammography. Recently proposed as a filler material, peanut oil is significantly more radiolucent than normal tissue. Physiologic saline solution compares favorably as a tissue-density-simulating substance. However, saline's lack of lubricating properties results in leakage, making it less than optimal. Bio-Oncotic gel is biologically compatible. We conclude that this gel is the most appropriate filler for breast prostheses. Clinical studies are indicated.

Temperature response of two photographic films and TLDs suitable for patient dosimetry of high dose fluoroscopic procedures.

When making direct measurements of skin entrance dose for patients undergoing interventional x-ray procedures the dosimeter may be placed near or in direct contact with the skin. The response of such dosimeters to heating during x-ray exposure must be known for accurate dosimetry. Two photographic films and MgB4O7 TLDs used for patient dose monitoring were exposed under different temperature conditions. Film dose response curves were obtained at room temperature and body temperature (22 degrees C and 38 degrees C). The TLD response at room and body temperature was also checked at 14 Gy. Dose response curves of FGP film show at most a 22+/-2% increase in predicted dose at 37 degrees C in the range of 0.3 to 2.8 Gy. LOC4 film showed less than a 2% difference in response between 22 and 37 degrees C over a range of 1 to 11 Gy. The TLDs showed less than a 2% difference at a dose of 14 Gy. Uncertainty in skin dose estimation is less than 2% using LOC4 film and MgB4O7 TLD arrays. Uncertainty in temperature during exposure of Kodak fine grain positive film may lead to errors in dose estimates of +/-11%. Arrays of TLDs and copy film are more suitable for measurements of high doses and are less sensitive to temperature changes during exposure than the fine grain film.

Radiotherapy verification film for estimating cumulative entrance skin exposure for fluoroscopic examinations.

Measurement of skin entrance exposures during fluoroscopic procedures is complicated by the use of automatic exposure control devices and the presence of contrast media. Due to variability in positioning spot films from patient to patient, standard dosimeters, such as thermoluminescent, cannot be properly placed on the skin prior to examination. Prepackaged film of the type used for portal verification in radiation therapy held next to the patient's skin in a specially modified patient examination gown was found to be useful for determining the entrance skin exposure from both fluoroscopy and spot films during air contrast barium enema exams. The usable sensitivity range of this film has been found satisfactory for exposure measurements at exposures and kVps typically used for gastrointestinal fluoroscopic procedures. Errors in exposure estimates due to changes in film speed and contrast with kVp are less than 5% for the range of kVps used. Errors from variations in beam quality due to the adjacency of scattering material are approximately 5%. Entrance exposures determined with film agreed with those determined from TLD measurements to within 21%, with an average difference of 9%.

A survey of films for use as dosimeters in interventional radiology.

Analysis of radiation doses in interventional radiological procedures that can lead to deterministic radiation effects such as erythema and epilation would assist physicians in planning patients care after exposure and in reducing doses. Photographic films used to measure skin exposure in the past are too sensitive for the high doses involved in interventional procedures. Seventeen different types of films, many of which are generally available in hospitals, were surveyed to see if any would meet the demands of interventional radiology. Sensitometric curves obtained demonstrate that most films are inappropriate for high dose procedures. Using Kodak Fine Grain Positive and Dupont duplicating films and automatic processing, doses as high as 2.8 Gy could be measured with reasonable accuracy. Similar results can be obtained by manually processing Kodak XV-2 verification film at room temperature.

Radiation dose in interventional fluoroscopic procedures.

Vascular interventional procedures carried out under fluoroscopic guidance often involve high radiation doses. Above certain thresholds, radiation can cause significant damage to the skin including hair loss and severe necrosis. Such damage has been reported by several investigators. Many attempts have been made to quantitate the radiation doses to the skin involved with these procedures, but dosimetry methods are often flawed. To improve the situation better monitoring of radiation doses, fluoroscopist education, and changes in technology and methods are needed.

Fluoroscopy: recording of fluoroscopic images and automatic exposure control.

Some means of recording images is a necessary part of most fluoroscopic systems. Several methods are available for recording images during fluoroscopy. Screen-film recording methods such as use of spot film devices and automatic film changers provide high-spatial-resolution images. Recording images by using the image intensifier (fluorography) provides film or digital images at relatively lower doses but with poorer spatial resolution. Digitally recorded images have better contrast resolution than analog images but lower spatial resolution and represent a compromise between dose and image quality. Motion picture (cine fluorographic) recording requires extremely high dose rates compared with those of lower-resolution videotape recording of motion. Recording systems in fluoroscopy require automatic exposure control for optimum image quality. The same feedback system used to control fluorographic exposures can be used to control exposure rates during fluoroscopy as well. Automatic brightness control maintains intensifier exposure rates on the basis of subject thickness by adjusting various technique factors. The type of control mechanism depends on the imaging task and the complexity (age and cost) of the equipment. The operator can choose between better image quality (higher contrast) or lower radiation dose.

Composition of mammographic phantom materials.

PURPOSE: To determine the composition of a breast phantom that is most suitable for phototimer testing. MATERIALS AND METHODS: Kilovolt peak and milliampere second values were analyzed for phototimed exposures of phantoms containing 0%, 30%, 50%, 70%, and 100% simulated glandular material at thicknesses of 2, 4, 6, and 8 cm. Data also were analyzed from 1,578 mammograms in 417 patients. The mean effective glandular content of phantoms that simulated patient breasts of varying thicknesses was determined. RESULTS: Examination of mammographic techniques showed that the effective glandular content that simulated the average patient breast was 34%. Effective glandular content was 16% in breasts at least 7 cm thick when compressed; 26%, more than 5 to up to 7 cm thick; 42%, more than 3 to up to 5 cm thick, and 68% in breasts at most 3 cm thick. CONCLUSION: A phantom composed of 30% glandular tissue and 70% adipose tissue allows closer simulation of the phototimer response of the mammographic x-ray unit for the average breast. The phantom currently used contains 16% more glandular tissue than the average breast.

The use of CT scanners in megavoltage photon-beam therapy planning.

Several problems related to the true value of quantitative information from CT scanners in improving estimates of photon-beam isodose distribution remain unanswered. The authors conclude that two widely used correction factors provide satisfactory isodose corrections behind heterogeneities of simple geometry. Lateral perturbations were found to be small for megavoltage photon beams, indicating that complex computational schemes are uncalled for. It appears doubtful that the use of CT information in megavoltage photon-beam isodose curve generation requires (a) larger computers than those currently available either in existing therapy planning systems or on the scanner itself or (b) direct access to the CT numbers.

Potential biological effects following high X-ray dose interventional procedures.

Some interventional procedures can result in very high x-ray doses. Potential biological effects of high x-ray doses are reviewed. Deterministic and stochastic effects in skin, bone, parotid glands, and lung are discussed. Threshold doses for the effects and relevant dosimetric principles are addressed. General principles for minimizing the potential for these effects are presented. Knowledge about these effects and the means to minimize radiation dose can assist the physician in the care of patients undergoing lengthy invasive radiologic procedures.

Improved mammography with a reduced radiation dose.

A 2-cm-high focused air-interspaced grid with a 7:1 ratio and eight lamellae per inch was devised. The grid is arc-shaped and pivots on its focal point, the anode-cathode axis. Milliamperage was 35% less for breast phantom radiographs obtained with the new grid versus those obtained with a conventional 5:1 mammographic grid. In a comparison study of five mammograms, each evaluated for five parameters, the images obtained with the new grid were judged to be of equal or better quality than those obtained with a conventional grid in 17 of 20 comparisons.

Radiation doses during pediatric radiofrequency catheter ablation procedures.

Because RF catheter ablation procedures may be lengthy and are commonly performed in young patients, concern has arisen about radiation dose in this group of patients. This article investigates radiation doses in pediatric patients undergoing RF catheter ablation. Standard fluoroscopic equipment used for diagnostic electrophysiological catheterization studies is technologically capable of dose rates as high as 90 milligray (mGy) per minute to skin and adjacent lung and 260 mGy/min to vertebral bone. Dose rates of this magnitude when used for extended periods of time have been known to cause erythema, pneumonitis, and retardation of bone growth. We measured skin dose rates of nine pediatric patients undergoing RF catheter ablation for tachycardia and calculated doses to the skin using standard dosimetric methods. Fluoroscopic techniques and equipment were studied using a patient simulating phantom. Overlap of fluoroscopic fields was checked using radiotherapy portal verification film, and regions in which doses overlapped from multiple angle exposures were verified using a treatment planning computer. Patient skin dose rates ranged from 6.2-49 mGy/min for patients ranging in age from 2-20 years. Maximum skin doses ranged from 0.09-2.35 Gy. Our data demonstrate the need to directly measure dose rates for individual fluoroscopic equipment and procedural techniques in order to determine whether limitations need to be set for procedural times.

Volumetric high-resolution CT in the diagnosis of interstitial lung disease and bronchiectasis: diagnostic accuracy and radiation dose.

OBJECTIVE: The purpose of this study was to evaluate the diagnostic accuracy and radiation dose of volumetric high-resolution CT in the diagnosis of interstitial lung disease and bronchiectasis when four contiguous sections were acquired at each of three levels. The potential benefits were weighed against the increased radiation dose of multiple scans. SUBJECTS AND METHODS: High-resolution CT scans of four contiguous sections were obtained at each of three locations (the aortic arch, the carina, and 2 cm above the diaphragm) in 50 consecutive patients (mean age, 44 years old) with known or suspected interstitial lung disease or bronchiectasis who were referred for evaluation with high-resolution CT. Each individual scan was analyzed for the presence of motion-induced streaking, blurring, or doubling. The diagnostic information contained in each set of four scans was compared with that contained in the first of the four scans in the set. RESULTS: Motion degraded at least one of the four images in each set in 69 (46%) of 150 volumetric acquisitions. When the full set of four images was considered instead of just the first scan from the set, the number of motion-free studies in patients with suboptimal respiratory suspension was increased by 40% (from 99 to 139). Diagnostic accuracy was improved as more features were identified on contiguous scans: the sensitivity of the first scan compared with that of the complete set of four scans was 84% for the detection of bronchiectasis, 97% for ground-glass opacity, 88% for honeycombing, 88% for septal thickening, and 86% for nodular opacities. Although the integral radiation exposure for a set of four CT scans was 2.8 times that of a single scan obtained with standard technique, peak skin exposure was unchanged. Slightly increased image noise with the reduced technique compromised diagnostic ability in 6% of studies. CONCLUSION: The use of volumetric high-resolution CT increased diagnostic accuracy, particularly for bronchiectasis at the lung bases, without increasing peak skin radiation exposure. With the availability of four contiguous scans per anatomic level, the subjective confidence in interpretation and number of motion-free studies also increased.