We have measured the radiation dose in simple heterogeneous phantoms and compared our results with those obtained by various methods of computation. Dose data were obtained both within and distal to simulated regions of lung in order to test the ratio of tissue-air ratios (TAR), Batho, and equivalent TAR methods. These procedures are used routinely in manual and computer-aided planning of radiation therapy, but have been validated primarily for cobalt-60 radiation. Tests performed with 6- and 15-MV x rays reveal that incorrect doses can be computed within or near to a low-density medium, particularly when the field size is small. In these cases, electronic equilibrium is not achieved in the lateral direction, thereby violating an implicit assumption of all the above calculation methods. We quantify the errors in dose calculation for simple slab phantoms, and support our interpretation with a Monte Carlo simulation in which the energy transported by charged particles away from sites of x-ray interactions is considered directly.

We have reexamined the Batho power law for computing the dose within and beyond lung irradiated with small and large fields of cobalt-60 and 6-MV x rays. Using slab phantoms consisting of two materials, agreement between calculated and measured doses was within 2% inside lung for 6-MV x irradiation, but much poorer (9%) for cobalt-60 irradiation. For cobalt-60 irradiation, tissue-air ratios (TARs) were used initially in the Batho equation, while for 6-MV x rays, tissue-maximum ratios (TMRs) were used. When we substituted TMR values instead of TAR values for cobalt-60, we found marked improvement by nearly 5% in the accuracy of dose calculated within lung. This was confirmed by numerical comparison of the Batho expression with an analytic solution of the primary and first-scattered radiation. We therefore encourage the use of TMRs for cobalt-60 radiation, especially for larger radiation fields, and provide measured data tables for field sizes up to 50 X 50 cm2, and depths up to 30 cm. In addition to unifying the dosimetry for all megavoltage irradiation, this approach improves the accuracy of doses calculated within lung.

Specific (A-, B-and D-) granules as well as multivesicular bodies (MVB) with a dense core from the right and left atrium of a variety of mammals (mouse, rat, hamster, guinea pig, rabbit, young and old cat) were compared by quantitative ultrastructural and cytochemical methods as regards number, size, localization and reactivity. The number of specific granules was greater in the right atrium of the rat and lesser in the same atrium of the guinea pig. In the rat this difference was due t o a greater number of all three types of specific granules in the right atrium. In the guinea pig, both A and B granules were more numerous in the left atrium. A greater number of granules was also found in the left atrium of hamster, rabbit and young and old cat but the difference with the right atrium was not significant. All atria contained the same type of specific granules but A-and B-granules were present in the left atrium not only in the paranuclear area, as in the right atrium, but also throughout the sarcoplasm. In both atria, all specific granules were argentaphobic when stained according to the periodic acid-thiocarbohydrazide-silver proteinate technique of Thiery. They reacted positively t o phosphotungstic acid-hydrochloric acid (PTA) (pH 0.3) as did the cell coat, residual bodies (Cgranules), lysosomes, Z-disks and a small portion of the Golgi complex. The PTA and PAS stains were abolished by acetylation, restored by saponification, unchanged by methylation and greatly diminished by sulfation. The MVB with a dense core, identical to those already noted in the cells of various endocrine glands, and thought to be crinophagic, were rare in the right atrium of all species and absent in that of the rat. They were much more numerous in the left atrium, particularly of hamsters, guinea pigs and cats and, to a lesser degree, of mice, rabbits and rats. They were silver negative. Their dense core reacted to PTA but their matrix, contrary to classical MVB without a dense core, did not.

The dosimetric characteristics of a scanning liquid-filled ionization chamber (SLIC) electronic portal imaging device have been investigated. To assess the system's response in relation to incident radiation beam intensity, a series of characteristic curves are obtained for various field sizes and nominal energies of 6 and 10 MV photons. The response of the imaging system is dependent on incident radiation intensity and can be described to within 1% accuracy on central axis using a square root function. Portal dose measurements with the SLIC at the plane of the detector, on central axis of the beam using homogeneous attenuating phantom materials show that the imaging system is capable of measuring the portal (transmission) dose to within 3% of the ionization chamber results for homogeneous material. For two-dimensional dosimetry applications, the system is calibrated with a 10 cm Perspex block used as beam flattening material on the detector cassette to correct for variations in individual ion chamber sensitivity and the effect of nonuniform beam profiles produced by the flattening filter. Open and wedged dose profiles measured with the SLIC agreed with ion chamber measured profiles to within 3.5% accuracy.