Preliminary treatment planning and dosimetry for a clinical trial of neutron capture therapy using a fission converter epithermal neutron beam

A Phase I/II clinical trial of neutron capture therapy (NCT) was conducted at Harvard-MIT using a fission converter epithermal neutron beam. This epithermal neutron beam has nearly ideal performance characteristics (high intensity and purity) and is well-suited for clinical use. Six glioblastoma multiforme (GBM) patients were treated with NCT by infusion of the tumor-selective amino acid boronophenylalanine-fructose (BPA-F) at a dose of 14.0 g/m(2) body surface area over 90 min followed by irradiation with epithermal neutrons. Treatments were planned using NCTPlan and an accelerated version of the Monte Carlo radiation transport code MCNP 4B. Treatments were delivered in two fractions with two or three fields. Field order was reversed between fractions to equalize the average blood boron concentration between fields. The initial dose in the dose escalation study was 7.0 RBEGy, prescribed as the mean dose to the whole brain volume. This prescription dose was increased by 10% to 7.7 RBEGy in the second cohort of patients. A pharmacokinetic model was used to predict the blood boron concentration for determination of the required beam monitor units with good accuracy; differences between prescribed and delivered doses were 1.5% or less. Estimates of average tumor doses ranged from 33.7 to 83.4 RBEGy (median 57.8 RBEGy), a substantial improvement over our previous trial where the median value of the average tumor dose was 25.8 RBEGy.

Voxel model in BNCT treatment planning: performance analysis and improvements

In recent years, many efforts have been made to study the performance of treatment planning systems in deriving an accurate dosimetry of the complex radiation fields involved in boron neutron capture therapy (BNCT). The computational model of the patient's anatomy is one of the main factors involved in this subject. This work presents a detailed analysis of the performance of the 1 cm based voxel reconstruction approach. First, a new and improved material assignment algorithm implemented in NCTPlan treatment planning system for BNCT is described. Based on previous works, the performances of the 1 cm based voxel methods used in the MacNCTPlan and NCTPlan treatment planning systems are compared by standard simulation tests. In addition, the NCTPlan voxel model is benchmarked against in-phantom physical dosimetry of the RA-6 reactor of Argentina. This investigation shows the 1 cm resolution to be accurate enough for all reported tests, even in the extreme cases such as a parallelepiped phantom irradiated through one of its sharp edges. This accuracy can be degraded at very shallow depths in which, to improve the estimates, the anatomy images need to be positioned in a suitable way. Rules for this positioning are presented. The skin is considered one of the organs at risk in all BNCT treatments and, in the particular case of cutaneous melanoma of extremities, limits the delivered dose to the patient. Therefore, the performance of the voxel technique is deeply analysed in these shallow regions. A theoretical analysis is carried out to assess the distortion caused by homogenization and material percentage rounding processes. Then, a new strategy for the treatment of surface voxels is proposed and tested using two different irradiation problems. For a parallelepiped phantom perpendicularly irradiated with a 5 keV neutron source, the large thermal neutron fluence deviation present at shallow depths (from 54% at 0 mm depth to 5% at 4 mm depth) is reduced to 2% on average. Reassigning fluence values in the case of this phantom in angular position produced the maximum deviation in the thermal fluence to decrease from 140% to 23% at the surface of the phantom. Thus, even for the largest deviations, obtained by intentionally placing the phantom in the most disadvantageous position with respect to the voxel grid, the reassignment shows very good performance. Since these results substantially improve the performance of the 1 cm based voxel model in surface boundary regions, the proposed strategy will be implemented in future versions of the NCTPlan code.

The Microdosimetry of the10B Reaction in Boron Neutron Capture Therapy: A New Generalized Theory

The microdosimetry of (10)B thermal neutron capture reactions should be considered as an essential step to be followed before studying the radiobiological aspects of boron neutron capture therapy. The boron dose itself is insufficient as the only quantity used to describe the biological effectiveness of the (10)B reaction for two important reasons: the specific microdistribution that the (10)B carrier compound exhibits at the cellular level and the primarily stochastic nature of the energy deposition process, which influences the biological response to the particulate radiation. In this work, these two aspects are analyzed in detail and an innovative rigorous analytical framework is developed in the microdosimetry domain. This formalism provides the necessary microdosimetric tools for more precisely describing the (10)B dose distribution deposited in sensitive microscopic structures and offers improved approaches for analyzing the biological dose--effect relationship of (10)B reactions.

Dose-rate scaling factor estimation of THOR BNCT test beam

In 1998, an epithermal neutron test beam was designed and constructed at the Tsing Hua Open-Pool Reactor (THOR) for the purpose of preliminary dosimetric experiments in boron neutron capture therapy (BNCT). A new epithermal neutron beam was designed at this facility, and is currently under construction, with clinical trials targeted in late 2004. Depth dose-rate distributions for the THOR BNCT test beam have been measured by means of activation foil and dual ion chamber techniques. Neutron and structure-induced gamma spectra measured at the test beam exit were configured into a source function for the Monte Carlo-based treatment planning code NCTPlan. Dose-rate scaling factors (DRSFs) were determined to normalize computationally derived dose-rate distributions with experimental measurements in corresponding mathematical and physical phantoms, and to thus enable accurate treatment planning using the NCTPlan code. A similar approach will be implemented in characterizing the new THOR epithermal beam in preparation for clinical studies. This paper reports the in-phantom calculated and experimental dosimetry comparisons and derived DRSFs obtained with the THOR test beam.

Moderated 252Cf neutron energy spectra in brain tissue and calculated boron neutron capture dose

While there is significant clinical experience using both low- and high-dose (252)Cf brachytherapy, combination therapy using (10)B for neutron capture therapy-enhanced (252)Cf brachytherapy has not been performed. Monte Carlo calculations were performed in a brain phantom (ICRU 44 brain tissue) to evaluate the dose enhancement predicted for a range of (10)B concentrations over a range of distances from a clinical (252)Cf source. These results were compared to experimental measurements and calculations published in the literature. For (10)B concentrations </=50 microg/g, the (10)B neutron capture dose enhancement was small in comparison to the (252)Cf fast neutron dose.

Pharamacokinetic modeling for boronophenylalanine-fructose mediated neutron capture therapy: 10B concentration predictions and dosimetric consequences

A two-compartment open model has been developed for predicting 10B concentrations in blood following intravenous infusion of the L-p-boronophenylalanine-fructose complex in humans and derived from pharmacokinetic studies of 24 patients in Phase I clinical trials of boron neutron capture therapy. The 10B concentration profile in blood exhibits a characteristic rise during the infusion to a peak of approximately 32 microg/g (for infusion of 350 mg/kg over 90 min) followed by a biexponential disposition profile with harmonic mean half-lives of 0.32 +/- 0.08 and 8.2 +/- 2.7 h, most likely due to redistribution and primarily renal elimination, respectively. The mean model rate constants k12, k21, and k10 are (mean +/- SD) 0.0227 +/- 0.0064 min(-1), 0.0099 +/- 0.0027 min(-1), 0.0052 +/- 0.0016 min(-1), respectively, and the central compartment volume of distribution V1 is 0.235 +/- 0.042 L/kg. In anticipation of the initiation of clinical trials using an intense neutron beam with concomitantly short irradiations, the ability of this model to predict, in advance, the average blood 10B concentration during brief irradiations was simulated in a retrospective analysis of the pharmacokinetic data from these patients. The prediction error for blood boron concentration and its effect on simulated dose delivered for each irradiation field are reported for three different prediction strategies. In this simulation, error in delivered dose (or, equivalently, neutron fluence) for a given single irradiation field resulting from error in predicted blood 10B concentration was limited to less than 10%. In practice, lower dose errors can be achieved by delivering each field in two fractions (on two separate days) and by adjusting the second fraction's dose to offset error in the first.

A critical examination of the results from the Harvard-MIT NCT program phase I clinical trial of neutron capture therapy for intracranial disease

A phase I trial was designed to evaluate normal tissue tolerance to neutron capture therapy (NCT); tumor response was also followed as a secondary endpoint. Between July 1996 and May 1999, 24 subjects were entered into a phase I trial evaluating cranial NCT in subjects with primary or metastatic brain tumors. Two subjects were excluded due to a decline in their performance status and 22 subjects were irradiated at the MIT Nuclear Reactor Laboratory. The median age was 56 years (range 24-78). All subjects had a pathologically confirmed diagnosis of either glioblastoma (20) or melanoma (2) and a Karnofsky of 70 or higher. Neutron irradiation was delivered with a 15 cm diameter epithermal beam. Treatment plans varied from 1 to 3 fields depending upon the size and location of the tumor. The 10B carrier, L-p-boronophenylalanine-fructose (BPA-f), was infused through a central venous catheter at doses of 250 mg kg(-1) over 1 h (10 subjects), 300 mg kg(-1) over 1.5 h (two subjects), or 350 mg kg(-1) over 1.5-2 h (10 subjects). The pharmacokinetic profile of 10B in blood was very reproducible and permitted a predictive model to be developed. Cranial NCT can be delivered at doses high enough to exhibit a clinical response with an acceptable level of toxicity. Acute toxicity was primarily associated with increased intracranial pressure; late pulmonary effects were seen in two subjects. Factors such as average brain dose, tumor volume, and skin, mucosa, and lung dose may have a greater impact on tolerance than peak dose alone. Two subjects exhibited a complete radiographic response and 13 of 17 evaluable subjects had a measurable reduction in enhanced tumor volume following NCT.

Treatment planning and dosimetry for the Harvard-MIT Phase I clinical trial of cranial neutron capture therapy

PURPOSE

A Phase I trial of cranial neutron capture therapy (NCT) was conducted at Harvard-MIT. The trial was designed to determine maximum tolerated NCT radiation dose to normal brain.

METHODS AND MATERIALS

Twenty-two patients with brain tumors were treated by infusion of boronophenylalanine-fructose (BPA-f) followed by exposure to epithermal neutrons. The study began with a prescribed biologically weighted dose of 8.8 RBE (relative biologic effectiveness) Gy, escalated in compounding 10% increments, and ended at 14.2 RBE Gy. BPA-f was infused at a dose 250-350 mg/kg body weight. Treatments were planned using MacNCTPlan and MCNP 4B. Irradiations were delivered as one, two, or three fields in one or two fractions.

RESULTS

Peak biologically weighted normal tissue dose ranged from 8.7 to 16.4 RBE Gy. The average dose to brain ranged from 2.7 to 7.4 RBE Gy. Average tumor dose was estimated to range from 14.5 to 43.9 RBE Gy, with a mean of 25.7 RBE Gy.

CONCLUSIONS

We have demonstrated that BPA-f-mediated NCT can be precisely planned and delivered in a carefully controlled manner. Subsequent clinical trials of boron neutron capture therapy at Harvard and MIT will be initiated with a new high-intensity, high-quality epithermal neutron beam.

Reference dosimetry calculations for neutron capture therapy with comparison of analytical and voxel models

As clinical trials of Neutron Capture Therapy (NCT) are initiated in the U.S. and other countries, new treatment planning codes are being developed to calculate detailed dose distributions in patient-specific models. The thorough evaluation and comparison of treatment planning codes is a critical step toward the eventual standardization of dosimetry, which, in turn, is an essential element for the rational comparison of clinical results from different institutions. In this paper we report development of a reference suite of computational test problems for NCT dosimetry and discuss common issues encountered in these calculations to facilitate quantitative evaluations and comparisons of NCT treatment planning codes. Specifically, detailed depth-kerma rate curves were calculated using the Monte Carlo radiation transport code MCNP4B for four different representations of the modified Snyder head phantom, an analytic, multishell, ellipsoidal model, and voxel representations of this model with cubic voxel sizes of 16, 8, and 4 mm. Monoenergetic and monodirectional beams of 0.0253 eV, 1, 2, 10, 100, and 1000 keV neutrons, and 0.2, 0.5, 1, 2, 5, and 10 MeV photons were individually simulated to calculate kerma rates to a statistical uncertainty of <1% (1 std. dev.) in the center of the head model. In addition, a "generic" epithermal neutron beam with a broad neutron spectrum, similar to epithermal beams currently used or proposed for NCT clinical trials, was computed for all models. The thermal neutron, fast neutron, and photon kerma rates calculated with the 4 and 8 mm voxel models were within 2% and 4%, respectively, of those calculated for the analytical model. The 16 mm voxel model produced unacceptably large discrepancies for all dose components. The effects from different kerma data sets and tissue compositions were evaluated. Updating the kerma data from ICRU 46 to ICRU 63 data produced less than 2% difference in kerma rate profiles. The depth-dose profile data, Monte Carlo code input, kerma factors, and model construction files are available electronically to aid in verifying new and existing NCT treatment planning codes.

A theoretical model for the microdosimetry of discontinuous distributions of heavy particle tracks

A novel approach to solving microdosimetry problems using conditional probabilities and geometric concepts has been developed. This approach is valid for cases where a convex site is immersed in uniform or discontinuous distributions of heavy charged particle tracks and assumes no restrictions in site geometry or the kind of randomness. These conditions are relevant to the study of microdosimetry in applications such as neutron capture therapy (NCT), irradiation experiments using heavy ion particle beams, environmental radon, or occupational exposure to radioactive materials. Expressions applicable to the case of surface-distributed sources of tracks are presented that may represent situations such as NCT, where boron compounds are bound to the membranes of cellular nuclei. Microdosimetric spectra, specific energy averages, and mean number of 10B capture reactions for cell inactivation are calculated, showing their dependence on 10B localisation.

A theoretical model for event statistics in microdosimetry. I: Uniform distribution of heavy ion tracks

In this work we describe a novel approach to solving microdosimetry problems using conditional probabilities and geometric concepts. The intersection of a convex site with a field of randomly oriented straight track segments is formulated in terms of the relative overlap between the chord associated with the action line of the track and the track itself. This results in a general formulation that predicts the contribution of crossers, stoppers, starters, and insiders in terms of two separate functions: the chord length distribution (characteristic of the site geometry and the type of randomness) and an independent set of conditional probabilities. A Monte Carlo code was written in order to validate the proposed approach. The code can represent the intersection between an isotropic field of charged particle tracks and a general ellipsoid of unrestricted geometry. This code was used to calculate the event distribution for a sphere as well as the expected mean value and variance of the track length distribution and to compare these against the deterministic calculations. The observed agreement was shown to be very good, within the precision of the Monte Carlo approach. The formulation is used to calculate the event frequency, lineal energy, and frequency mean specific energy for several monoenergetic and isotropic proton fields in a spherical site, as a function of the site diameter, proton energy, and the event type.

A theoretical model for event statistics in microdosimetry. II: Nonuniform distribution of heavy ion tracks

A microdosimetry model, described in Part I, applies to the case of a convex site immersed in a uniform distribution of heavy particle tracks, and assumes no restrictions in site geometry or the kind of randomness. In Part II, this model is extended to include nonuniform distributions of particle tracks. This situation is relevant to the study of microdosimetry, for example, in boron neutron capture, in irradiation experiments using heavy ion particle beams, where the sources of particle tracks are external to the cell, or in irradiation from internally incorporated particle-emitting radionuclides, such as environmental radon or occupational exposure to radioactive materials. The formalism developed permits the calculation of statistical properties, track length distributions, and microdosimetric spectra for convex sites where the "inner" and "outer" concentrations of sources may be different, or for tracks originating on the surface of a convex site. Expressions applicable to the case of surface-distributed sources of tracks are presented that may represent situations such as boron compounds bound to the membrane of a cellular nucleus in boron neutron capture. A series of Monte Carlo calculations and analytical solutions, illustrating the case of spherical site geometry, are presented and compared. Finally, microdosimetric spectra and specific energy averages are calculated for alpha and lithium particles originating from thermal neutron capture in 10B, showing their dependence on 10B localization (extra-site, uniform, intra-site, or surface-distributed).

A pharmacokinetic model for the concentration of 10B in blood after boronophenylalanine-fructose administration in humans

An open two-compartment model has been developed for predicting (10)B concentrations in blood after intravenous infusion of the l-p-boronophenylalanine-fructose complex (BPA-F) in humans and derived from studies of pharmacokinetics in 24 patients in the Harvard-MIT Phase I clinical trials of BNCT. The (10)B concentration profile in blood exhibits a characteristic rise during the infusion to a peak of approximately 32 microg/g (for infusion of 350 mg/kg over 90 min) followed by a biphasic exponential clearance profile with half-lives of 0.34 +/- 0.12 and 9.0 +/- 2.7 h, due to redistribution and primarily renal elimination, respectively. The model rate constants k(1), k(2) and k(3) are 0.0227 +/- 0.0064, 0.0099 +/- 0.0027 and 0.0052 +/- 0.0016 min(-1), respectively, and the central compartment volume of distribution, V(1), is 0.235 +/- 0.042 kg/kg. The validity of this model was demonstrated by successfully predicting the average pharmacokinetic response for a cohort of patients who were administered BPA-F using an infusion schedule different from those used to derive the parameters of the model. Furthermore, the mean parameters of the model do not differ for cohorts of patients infused using different schedules.

Computational validation of the stereology principle applied to the microdosimetry of boron neutron capture therapy

High resolution quantitative autoradiography (HRQAR) is a novel technique that has been developed in our laboratory and applied to the microdosimetry of boron neutron capture therapy (BNCT). High resolution quantitative autoradiography is employed to define the microdistribution of boron-10 atoms within a 1-2 microm frozen tissue section. This microdistribution is used as input to a novel two-dimensional Monte Carlo charged particle transport calculation that computes various microdosimetric parameters, such as the number of nuclear "hits," energy absorbed in the nuclei, etc., within the environment of actual tissue morphology (i.e., cell nuclei, cytoplasm, and intracellular space). Stereological transformation is then implemented to transform the two-dimensional calculations into effectively three-dimensional results. In the present study no seek to demonstrate the validity of the surrogate two-dimensional 2-D computation as being quantitatively equivalent to a hypothetical full 3-D calculation. The results show that within the limitations of the test parameters used the surrogate 2-D and 3-D results are completely equivalent within the statistical constraints of the Monte Carlo calculations. Limitations of this approach also are evaluated, including a Monte Carlo calculation of the influence of the thickness of the histological tissue section and the track detector and the influence of 4He and 7Li particle lateral and range straggling.

Proton nuclear magnetic resonance measurement of p-boronophenylalanine (BPA): a therapeutic agent for boron neutron capture therapy

Noninvasive in vivo quantitation of boron is necessary for obtaining pharmacokinetic data on candidate boronated delivery agents developed for boron neutron capture therapy (BNCT). Such data, in turn, would facilitate the optimization of the temporal sequence of boronated drug infusion and neutron irradiation. Current approaches to obtaining such pharmacokinetic data include: positron emission tomography employing F-18 labeled boronated delivery agents (e.g., p-boronophenylalanine), ex vivo neutron activation analysis of blood (and very occasionally tissue) samples, and nuclear magnetic resonance (NMR) techniques. In general, NMR approaches have been hindered by very poor signal to noise achieved due to the large quadrupole moments of B-10 and B-11 and (in the case of B-10) very low gyromagnetic ratio, combined with low physiological concentrations of these isotopes under clinical conditions. This preliminary study examines the feasibility of proton NMR spectroscopy for such applications. We have utilized proton NMR spectroscopy to investigate the detectability of p-boronophenylalanine fructose (BPA-f) at typical physiological concentrations encountered in BNCT. BPA-f is one of the two boron delivery agents currently undergoing clinical phase-I/II trials in the U.S., Japan, and Europe. This study includes high-resolution 1H spectroscopic characterization of BPA-f to identify useful spectral features for purposes of detection and quantification. The study examines potential interferences, demonstrates a linear NMR signal response with concentration, and presents BPA NMR spectra in ex vivo blood samples and in vivo brain tissues.

Microdosimetry for boron neutron capture therapy: a review

A review of the microdosimetry of boron neutron capture therapy is presented focusing on the progression of key scientific ideas and developments in this field rather than on a comprehensive and inclusive review of the literature. The author concludes that from a microdosimetry perspective the field is highly advanced, but what is lacking is the correlation of the proposed models and results with experimental radiobiological data.

A novel approach to the microdosimetry of neutron capture therapy. Part I. High-resolution quantitative autoradiography applied to microdosimetry in neutron capture therapy

A novel approach to the microdosimetry of neutron capture therapy has been developed using high-resolution quantitative autoradiography (HRQAR) and two-dimensional Monte Carlo simulation. This approach has been applied using actual cell morphology (nuclear and cytoplasmic cell structures) and the measured microdistribution of boron-10 in a transplanted murine brain tumor (GL261) containing p-boronophenylalanine (BPA) as the boron compound. The 2D Monte Carlo transport code for the alpha and 7Li charged particles from the 10B(n,alpha)7Li reactions has been developed as a surrogate to a full 3D approach to calculate a variety of different microdosimetric parameters. The HRQAR method and the surrogate 2D Monte Carlo approach are described in detail and examples of their use are presented.

Boron neutron capture therapy for murine malignant gliomas

Boron neutron capture therapy (BNCT) involves administration of a boron compound followed by neutron irradiation of the target organ. The boron atom captures a neutron, which results in the release of densely ionizing helium and lithium ions that are highly damaging and usually lethal to cells within their combined track length of approximately 12 microns. Prior to Phase I clinical trials for patients with malignant gliomas, mice with glioma 261 intracerebral tumors were fed D,L-3-(p-boronophenyl)alanine and irradiated with total tumor doses of 1000-5000 RBE-cGy of single fraction thermal neutrons to determine the maximum tolerated dose and effect on survival. These mice were compared to mice that received D,L-3-(p-boronophenyl)alanine alone, neutron irradiation alone, photon irradiation alone, or no treatment. Additional normal mice received escalating doses of neutron irradiation to determine its toxicity to normal brain. BNCT caused a dose-dependent, statistically significant prolongation in survival at 1000-5000 RBE-cGy. At 3000 RBE-cGy, median survival rates of the BNCT and untreated control groups were 68 and 22 days, respectively, with a long-term survival rate of 33%. At 4000 RBE-cGy, median survival was 72 and 21 days, respectively, with a long-term survival rate of 43%. At lower radiation doses, the extended survival was comparable between the BNCT and photon-irradiated mice; however, at 3000 and 4000 RBE-cGy the median survival of BNCT-treated mice was significantly greater than photon-irradiated mice. The maximum tolerated single fraction dose to normal brain was approximately 2000 RBE-cGy.

Dosimetry of 252Cf sources for neutron radiotherapy with and without augmentation by boron neutron capture therapy

Interstitial and intracavity 252Cf sources have been used to treat a number of tumor types with encouraging results. In particular these tumors include a variety of cervical, head-and-neck, and oral-cavity cancers and possible malignant gliomas. As a neutron source, 252Cf offers certain theoretical advantages over photon therapy (i.e., in treating tumors with significant hypoxic or necrotic components). With the recent availability of 10B-labeled tumor-seeking compounds, the usefulness of 252Cf may be further improved by augmenting the 252Cf dose to the tumor with an additional dose due to the fission (following thermal neutron capture) of 10B located in the tumor itself. While the high mean neutron energy permits 252Cf to deliver a high-LET, low-OER dose to the tumor on a macroscopic scale, thermalization of neutrons followed by 10B capture may augment this dose at the cellular level if adequate loading of tumor cells with 10B is possible. This paper presents results of a Monte Carlo simulation study investigating the dosimetric characteristics of linear 252Cf sources both with and without the quantitative increase in tumor dose possible with the addition of 10B. Results are displayed in the form of "along and away" tables and dose profiles in a water phantom. Comparisons of Monte Carlo results with experimental and analytical dosimetry data available in the literature are also presented.

Clinical considerations for neutron capture therapy of brain tumors

The radiotherapeutic management of primary brain tumors and metastatic melanoma in brain has had disappointing clinical results for many years. Although neutron capture therapy was tried in the United States in the 1950s and 1960s, the results were not as hoped. However, with the newly developed capability to measure boron concentrations in blood and tissue both quickly and accurately, and with the advent of epithermal neutron beams obviating the need for scalp and skull reflection, it should now be possible to mount such a clinical trial of NCT again and avoid serious complications. As a prerequisite, it will be important to demonstrate the differential uptake of boron compound in brain tumor as compared with normal brain and its blood supply. If this can be done, then a trial of boron neutron capture therapy for brain tumors should be feasible. Because boronated phenylalanine has been demonstrated to be preferentially taken up by melanoma cells through the biosynthetic pathway for melanin, there is special interest in a trial of boron neutron capture therapy for metastatic melanoma in brain. Again, the use of an epithermal beam would make this a practical possibility. However, because any epithermal (or thermal) beam must contain a certain contaminating level of gamma rays, and because even a pure neutron beam causes gamma rays to be generated when it interacts with tissue, we think that it is essential to deliver treatments with an epithermal beam for boron neutron capture therapy in fractions in order to minimize the late-effects of low-LET gamma rays in the normal tissue. I look forward to the remainder of this Workshop, which will detail recent progress in the development of epithermal, as well as thermal, beams and new methods for tracking and measuring the uptake of boron in normal and tumor tissues.

Neutron capture therapy beams at the MIT Research Reactor

Several neutron beams that could be used for neutron capture therapy at MITR-II are dosimetrically characterized and their suitability for the treatment of glioblastoma multiforme and other types of tumors are described. The types of neutron beams studied are: 1) those filtered by various thicknesses of cadmium, D2O, 6Li, and bismuth; and 2) epithermal beams achieved by filtration with aluminum, sulfur, cadmium, 6Li, and bismuth. Measured dose vs. depth data are presented in polyethylene phantom with references to what can be expected in brain. The results indicate that both types of neutron beams are useful for neutron capture therapy. The first type of neutron beams have good therapeutic advantage depths (approximately 5 cm) and excellent in-phantom ratios of therapeutic dose to background dose. Such beams would be useful for treating tumors located at relatively shallow depths in the brain. On the other hand, the second type of neutron beams have superior therapeutic advantage depths (greater than 6 cm) and good in-phantom therapeutic advantage ratios. Such beams, when used along with bilateral irradiation schemes, would be able to treat tumors at any depth in the brain. Numerical examples of what could be achieved with these beams, using RBEs, fractionated-dose delivery, unilateral, and bilateral irradiation are presented in the paper. Finally, additional plans for further neutron beam development at MITR-II are discussed.

Monte Carlo based dosimetry and treatment planning for neutron capture therapy of brain tumors

Monte Carlo based dosimetry and computer-aided treatment planning for neutron capture therapy have been developed to provide the necessary link between physical dosimetric measurements performed on the MITR-II epithermal-neutron beams and the need of the radiation oncologist to synthesize large amounts of dosimetric data into a clinically meaningful treatment plan for each individual patient. Monte Carlo simulation has been employed to characterize the spatial dose distributions within a skull/brain model irradiated by an epithermal-neutron beam designed for neutron capture therapy applications. The geometry and elemental composition employed for the mathematical skull/brain model and the neutron and photon fluence-to-dose conversion formalism are presented. A treatment planning program, NCTPLAN, developed specifically for neutron capture therapy, is described. Examples are presented illustrating both one and two-dimensional dose distributions obtainable within the brain with an experimental epithermal-neutron beam, together with beam quality and treatment plan efficacy criteria which have been formulated for neutron capture therapy. The incorporation of three-dimensional computed tomographic image data into the treatment planning procedure is illustrated. The experimental epithermal-neutron beam has a maximum usable circular diameter of 20 cm, and with 30 ppm of B-10 in tumor and 3 ppm of B-10 in blood, it produces (with RBE weighting) a beam-axis advantage depth of 7.4 cm, a beam-axis advantage ratio of 1.83, a global advantage ratio of 1.70, and an advantage depth RBE-dose rate to tumor of 20.6 RBE-cGy/min (cJ/kg-min). These characteristics make this beam well suited for clinical applications, enabling an RBE-dose of 2,000 RBE-cGy/min (cJ/kg-min) to be delivered to tumor at brain midline in six fractions with a treatment time of approximately 16 minutes per fraction. With parallel-opposed lateral irradiation, the planar advantage depth contour for this beam (with the B-10 distribution defined above) encompasses nearly the whole brain. Experimental calibration techniques for the conversion of normalized to absolute treatment plans are described.

Evaluation of a commercial mammography image-enhancement system

A commercial mammography image-enhancement system manufactured by Damon Corporation (Needham, MA) is evaluated. Using a dedicated computer, the system implements a real-time video local adaptive image processing algorithm based on the Wallis equation. Radiographs of a mammographic QA phantom (Nuclear Associates Model 76-001-4) containing five groups of simulated breast microcalcifications ranging in diameter from 0.12 to 0.35 mm were viewed by four investigators under three viewing conditions: on a light box with the unaided eye, on the image enhancer in magnified "bypass" (unenhanced) mode, and on the enhancer using all features for optimum enhancement. A mammogram was then overlaid on the radiographs, and the composite images were viewed under the same three conditions. Using the enhancer, as compared to using a light box alone, average increases of 1.4 and 1.1 microcalcifications per radiograph were observed for the phantom and phantom-with-mammogram radiographs, respectively. High-contrast resolution and spatial distortion were also measured.

An improved method for estimating the entrance exposure in diagnostic radiographic examinations

There is currently a widespread consensus on the importance of monitoring patient radiation exposures during radiographic examinations. Diagnostic facilities under federal jurisdiction already legislate maximum patient exposure limits for various diagnostic radiologic examinations, while an increasing number of state legislatures have instituted such regulations. Compliance requires that institutions be capable of assessing each patient's entrance exposures. A method is proposed that would facilitate the acquisition of such patient exposure information in a relatively straightforward and accurate manner, requiring a minimum number of measurements and access to a suitable programmable calculator. A standardized set of exposure measurements obtained on an accurately calibrated three-phase radiographic unit has been fitted by an analytic function. The average accuracy of the fit between the limits of 40-140 kVp and 2.5- to 6.0-mm aluminum filtration was 0.3%. The concept of linear scaling was employed to allow the analytic function to accurately reproduce the exposure outputs of different radiographic units. Validation experiments on patients indicated that an overall accuracy of 10% can be expected when using well-calibrated radiographic equipment. The method described permits institutions to verify their compliance with federal and/or state regulations and to confirm that their radiation exposures are consistent with national averages.

Photon activation analysis as a new technique for body composition studies

A study was undertaken to demonstrate the usefulness of the recently developed photon activation analysis (PAA) technique for in vivo body composition studies. PAA can be used for direct measurement of total-body oxygen, nitrogen, and carbon. Sequential measurements were made on rats fed diets of 0%, 4.2%, or 20% protein for 6 1/2 wk, and significant changes in body composition were noted. In addition, rats of different ages, strains, nutritional states, and degrees of obesity were included in a comparison of PAA results in vivo with results from chemical analysis after sacrifice of the animals. High positive correlations were found between PAA measurements of carbon and chemical analysis measurements of fat and between PAA measurements of oxygen and chemical analysis measurements of total-body water. A low positive correlation was found between PAA measurements of nitrogen and chemical analysis measurements of protein.

Measurement of total-body oxygen, nitrogen, and carbon in vivo by photon activation analysis

A method has been developed to measure total-body oxygen, nitrogen, and carbon in vivo using the x-ray beam of a 45-MV betatron and a whole-body counter. Following x-ray irradiation of living tissue, the positron emitting activation products 15O, 11C, and 13N are produced. The decay of these radionuclides has been measured in both phantoms and animals, and a computer curve-fitting algorithm used to resolve the decay curve into separate contributions from 15O, 11C, and 13N. The decay curve was corrected for interfering activity from 30P, 38K, and 34mCl, and in the case of live animals, also corrected for a substantial fraction of 11C lost through exhalation. Activation uniformity profiles have been measured for phantoms up to 30 cm in thickness. With a radiation dose of 20 cGy, total-body O, N, and C were measured in dead rats with estimated accuracies of +/- 1.4%, +/- 4.5%, and +/- 1.5% [1 standard deviation (SD)], respectively. With a radiation dose of 40 cGy, total-body O, N, and C were measured in living rats with estimated accuracies of +/- 1.4%, +/- 6.9%, and +/- 1.5% (1 SD), respectively. It is anticipated that total-body O, N, and C similarly could be measured in human subjects with a radiation dose of 1-2 cGy and with accuracies comparable to those obtained in rats. Although most of the measurements were made using a beam energy of 45 MV, we have shown that useful results may be achievable with a beam energy as low as 25 MV. This accurate, convenient, and safe technique for total-body O, C, and N measurement should have applications in the study of nutritional status in health and disease, both in human subjects and in animals.

Mammography. Part 2. Evaluation of equipment and guidelines for quality assurance

Whether existing equipment is modified for use in mammography or a dedicated system is purchased, the services of a qualified medical physicist are necessary. Factors to consider in modifying existing equipment and in purchasing new equipment are detailed, as well as features concerning patient safety. Guidelines for quality assurance tests and their respective time intervals are given.

Mammography. Part 1: Physical principles

In this final part of a two-part article, fundamental imaging principles are related to the design features of mammography equipment, to give readers information helpful in making equipment decisions and selections.

Determination of the chemical composition of urinary calculi by computerized tomography

The clinical management of renal calculi would be aided if a direct in vivo determination of stone chemical composition could be made. We investigated the possibility of obtaining this information by a quantitative analysis of the computerized tomography scan images of 80 urinary calculi. Our results show that by using an appropriately calibrated computerized tomography scanner the differentiation of stone chemical composition can be made on the basis of 3 parameters, namely, absolute computerized tomography value at a single x-ray energy, the difference between computerized tomography values measured at 2 different x-ray energies, and computerized tomography value-frequency histograms (pixel patterns) of the stones. Uric acid stones were differentiated from all other stones at a significance level of p less than 0.001. Cystine was differentiated from calcium oxalate and brushite at the same significance level. Using pixel patterns cystine and struvite were separated from each other correctly with 70 per cent accuracy. Struvite stones of low or moderate calcium phosphate content were identified correctly with 80 per cent accuracy. Struvite stones of high calcium phosphate content could not be differentiated from calcium oxalate or brushite. Calcium oxalate and brushite could not be separated. The minimum stone size that allowed chemical identification was established for each stone type. In addition, we demonstrated that all the urinary calculi examined were visible on computerized tomography scan regardless of chemical composition or size.

The optimization of signal detectability in digital fluoroscopy

An optimization analysis of digital fluoroscopy has been carried out using (signal/noise)2/dose as a figure of merit. Three signal detection tasks have been considered: (i) small iodine-containing blood vessels; (ii) regions of fat; and (iii) microcalcifications. The effects of x-ray beam filtration and kilovoltage on the figure of merit have been parameterized and the associated x-ray tube power loadings have been defined. Results are presented in a fashion that facilitates the choice of parameters (amount of aluminum filtration, kilovoltage, and x-ray tube power loading) such that an acceptable balance can be achieved between signal detection sensitivity, patient dose, and x-ray tube power loading.

Boron neutron capture radiation therapy of cerebral gliomas: an analysis of the possible use of boron-loaded tumor-specific antibodies for the selective concentration of boron in gliomas

The possibility of achieving a therapeutically useful tissue boron distribution for boron neutron capture therapy (BNCT) of cerebral gliomas with boron loaded tumor-specific antibodies is discussed. Using a theoretical tumor-immunological model and RBE dose-depth calculations, the effects of various parameters, e.g. antibody-antigen association constant, antigen site density, number of boron atoms per antibody molecule, etc., on the advantage depth, a relative measure of the resulting radiation dose distributions, are determined. It is shown that with this model a maximum in the advantage depth as a function of the blood boron concentration occurs, the position of which is dependent on the value of the parameters used. Frequently this maximum corresponds to a blood boron-10 concentration range of between 0.1 to 0.5 microgram 10B/g blood. It is concluded that given the pharmacodynamic properties of potentially useful antibody preparations for this type of tumor therapy, advantage depths significantly greater than those obtainable with existing "blood-brain-barrier" compounds are not likely to be easily achieved.

A feasibility study of prompt capture gamma in vivo neutron activation analysis

The feasibility of using the information contained in the radiative capture gamma ray spectrum of the neutron-irradiated human body to measure quantitatively total body elemental composition in vivo has been investigated. Results of time dependent Monte Carlo simulations have shown that spectral interference of nonradiative capture origin can be completely eliminated by pulsing the detector/spectrometer system in anticoincidence with the neutron source. Calculations based on the results of the Monte Carlo simulation and on an experimental measurement of the efficiency versus energy characteristics of a Ge(Li) detector suggest that the primary limitation of the proposed technique would be inter-element spectral interference rather than inadequate detector sensitivity. Experimental measurements using a pulsed 14-MeV neutron generator and Ge(Li) gamma-ray spectrometer have produced results that are consistent with the predictions of the theoretical model. A radiative capture gamma-ray spectrum of a tissue-equivalent phantom was measured in pulsed mode and analyzed offline using a computerized spectral analysis procedure. The results were scaled to a proposed facility consisting of a 2.5-MeV pulsed neutron source and a detection system comprising two 50-cm3 (Ge(Li) detectros past which the subject would be scanned. It has been shown that in principle the elements hydrogen, chlorine, calcium, and nitrogen [the latter using large NaI(T1) detectors] could be measured with such a facility at an average body dose level of 0.1 rad. At this dose level the coefficients of variation based on counting statistics alone would be +/- 2%, or better, for these four elements. With an improvement in the detector/spectrometer energy resolution, the elements sulfur and carbon might also be measurable. It is also shown that by modifying the pulsing sequence appropriately and using 14-MeV neutrons, total body oxygen could also be measured at the 0.1 rad dose level via its inelastic neutron scattering deexcitation gamma activity.

An interesting form of high-energy gamma-ray--detector interaction and its effect on spectral peak centroid determination

The pulse-height continuum produced in a radiation detector following the interaction of photons by pair-production is discussed. The event when one annihilation photon escapes from the detector without interaction while the other escapes after one Compton scattering interaction produces a characteristic component in the pulse-height spectrum. The significance of this component in influencing the accurate determination of the centroid of the double-escape peak is examined and shown to be negligibly small.

Computed tomography scanning for the measurement of bone mineral in the human spine

An investigation into the feasibility of using X-ray computed tomography (CT) to measure disease induced changes in bone mineral content of the human spine is described. A theoretical study of this type of measurement has been made using a mathematical model of osteomalacia. The measured EMI number changes linearly with the mineral content, and the sensitivity is shown to be 1.2 EMI units (EU: 500 scale)/1% change in mineral content in vertebral bone. The physical sensitivity to an equal mineral change in cortical bone is found to be 8.6 times greater. The mineral selectivity of the CT method is such that only about half the change in the EMI number arising from progressive osteomalacia reflects change in actual mineral content, while half is due merely to changes in bone density that accompany demineralization. In addition, the perturbing effects of beam hardening on measurement accuracy are evaluated and shown to be significant. Finally, an experimental measurement indicates that in practice the reproducibility of such measurements would be about 1 EU, and it is shown that the measured parameter correlates well with the calculated total linear attenuation coefficients.

[32P] diphosphonate dose determination in patients with bone metastases from prostatic carcinoma

In an initial safety study, phosphorus-32 (as diphosphonate) was administered intravenously to five patients with painful bone metastases from prostatic carcinoma; two patients received 9 mCi and three were given 3 mCi. Hematological, biochemical, ECG, x-ray, bone-scan data, and clinical observation, were followed for 2 mo. At both dose levels, bone-marrow depression was noted. One of the patients, who received 9 mCi, had only a slight dip in the levels of circulating white blood cells and platelets. The other 9-mCi patient was the only one with discrete metastases by bone scan; he had bone-marrow depression, from which he recovered, and was the only one of the five who had relief of bone pain.

Boron neutron capture therapy for the treatment of cerebral gliomas. I. Theoretical evaluation of the efficacy of various neutron beams

The technique of boron neutron capture therapy in the treatment of cerebral gliomas depends upon the selective loading of the tumor with a 10B-enriched compound and subsequent irradiation of the brain with low-energy neutrons. The charged particles produced in the 10B (n,alpha) 7Li reaction have ranges in tissue of less than 10 mum so that the dose distribution closely follows the 10B distribution even to the cellular level. The effectiveness of this therapy procedure is dependent not only on the 10B compound but on the spectral characteristics of the neutron source as well. Hence, an optimization of these characteristics will increase the chances of therapeutic success. Transport calculations using a neutral particle transport code have been made to determine the dose-depth distributions within a simple head phantom for five different incident neutron beams. Comparison of these beams to determine their relative therapeutic efficacy was made by the use of a maximum useable depth criterion. In particular, with presently available compounds, the MIT reactor (MITR) therapy beam (a) is not inferior to a pure thermal neutron beam, (b) would be marginally improved if its gamma-ray contamination were eliminated, (c) is superior to a partially 10B-filtered MITR beam, and (d) produces a maximum useable depth which is strongly dependent upon the tumor-to-blood ratio of 10B concentrations and weakly dependent upon the absolute 10B concentration in tumor. A pure epithermal neutron beam with a mean energy of 37 eV is shown to have close to the optimal characteristics for boron neutron capture therapy. Futhermore, these optimal characteristics can be approximated by a judiciously D2O moderated and 10B-filtered 252Cf neutron source. This tailored 252Cf source would have at least a 1.5 cm greater maximum useable depth than the MITR therapy beam for realistic 10B concentrations. However, at least one gram of 252Cf would be needed to make this a practical therapy source. If the moderated 252Cf source is not 10B filtered, the resultant neutron beam has characteristics similar to those of the MITR beam with no gamma-ray contamination. For usch a beam, 100 mg of 252Cf would produce a flux of 2.4 X 10(8) neutrons/(cm2 sec), which is an intensity suitable for therapy applications.