Search

The biological effectiveness of proton beams varies with depth, spot size and lateral distance from the beam central axis. The aim of this work is to incorporate proton relative biological effectiveness (RBE) and equivalent uniform dose (EUD) considerations into comparisons of broad beam and highly modulated proton minibeams. A Monte Carlo model of a small animal proton beamline is presented. Dose and variable RBE is calculated on a per-voxel basis for a range of energies (30-109 MeV). For an open beam, the RBE values at the beam entrance ranged from 1.02-1.04, at the Bragg peak (BP) from 1.3 to 1.6, and at the distal end of the BP from 1.4 to 2.0. For a 50 MeV proton beam, a minibeam collimator designed to produce uniform dose at the depth of the BP peak, had minimal impact on the open beam RBE values at depth. RBE changes were observed near the surface when the collimator was placed flush with the irradiated object, due to a higher neutron contribution derived from proton interactions with the collimator. For proton minibeams, the relative mean RBE weighted entrance dose (RWD) was ~25% lower than the physical mean dose. A strong dependency of the EUD with fraction size was observed. For 20 Gy fractions, the EUD varied widely depending on the radiosensitivity of the cells. For radiosensitive cells, the difference was up to ~50% in mean dose and ~40% in mean RWD and the EUD trended towards the valley dose rather than the mean dose. For comparative studies of uniform dose with spatially fractionated proton minibeams, EUD derived from a per-voxel RWD distribution is recommended for biological assessments of reproductive cell survival and related endpoints.

The purpose of this study is to design and develop a probabilistic network for detecting errors in radiotherapy plans for use at the time of initial plan verification. Our group has initiated a multi-pronged approach to reduce these errors. We report on our development of Bayesian models of radiotherapy plans. Bayesian networks consist of joint probability distributions that define the probability of one event, given some set of other known information. Using the networks, we find the probability of obtaining certain radiotherapy parameters, given a set of initial clinical information. A low probability in a propagated network then corresponds to potential errors to be flagged for investigation. To build our networks we first interviewed medical physicists and other domain experts to identify the relevant radiotherapy concepts and their associated interdependencies and to construct a network topology. Next, to populate the network's conditional probability tables, we used the Hugin Expert software to learn parameter distributions from a subset of de-identified data derived from a radiation oncology based clinical information database system. These data represent 4990 unique prescription cases over a 5 year period. Under test case scenarios with approximately 1.5% introduced error rates, network performance produced areas under the ROC curve of 0.88, 0.98, and 0.89 for the lung, brain and female breast cancer error detection networks, respectively. Comparison of the brain network to human experts performance (AUC of 0.90 ± 0.01) shows the Bayes network model performs better than domain experts under the same test conditions. Our results demonstrate the feasibility and effectiveness of comprehensive probabilistic models as part of decision support systems for improved detection of errors in initial radiotherapy plan verification procedures.

I have discovered fast X-ray oscillations near 1000 Hz in the flux of the low-mass X-ray binary 4U 06141091 in observations with NASA’s Rossi X-Ray Timing Explorer (RXTE). Such fast oscillations in this and other X-ray binaries (see M. van der Klis et al., Proc. NATO/ASI Ser., in press [1998]) are likely produced very close to the neutron star and provide constraints on the mass and radius of the neutron star and a measurement of its spin period. The quasi-periodic oscillation (QPO) signals are observed as peaks in the power density spectra (see Fig. 1). They are usually present in pairs, the higher frequency QPO appearing between 500 and 1145 Hz with the lower frequency QPO between 480 and 800 Hz. The difference in frequency between these two signals is constant at Hz (Ford, E. C., et 323 5 4 al., 1997, ApJ, 475, L123). The existence of two QPOs with a constant frequency separation is predicted by the “beat-frequency” model. The higher frequency QPO is modulated by the spinning neutron star, producing another QPO simultaneously at lower frequency. The higher frequency QPO signal may be produced by a modulation from gas in Keplerian orbital motion in the inner accretion disk. At the highest frequencies, or smallest inner disk radii, the QPOs are produced closest to the neutron star and serve as probes of this region of strong gravitational fields. Measurements of the fastest QPO frequencies provide upper limits to the mass and radius of the neutron star, since the oscillations must originate outside the radius of the marginally stable orbit predicted by general relativity (Kaaret, P., Ford, E. C., & Chen, K., 1997, ApJ, 482, L167). The upper limits for the neutron star mass and radius in 4U 06141091 are 2.1 M, and 19 km, respectively. I have observed a robust correlation between the frequency of the QPO and the flux of a blackbody component of the Xray spectrum (Ford, E. C., et al., 1997, ApJ, 486, L47). This correlation is explained if the higher frequency QPO is associated with the Keplerian motion at an inner disk edge. An increasing mass accretion rate, reflected in an increasing blackbody flux, causes the inner disk radius to shrink. As the inner disk radius shrinks, the QPO frequency increases. From the QPO frequencies, one can measure the spin period of the neutron star in the low-mass X-ray binary, one of the observational goals of the last two decades. The beat-frequency model predicts that the neutron star spin frequency is equal to the frequency difference between the two QPOs. In 4U 06141091, the frequency separation implies a neutron star spin period of 3.1 ms.