X-ray spectra produced by mammographic systems are compared to spectra from conventional diagnostic x-ray systems. Some systems use special anode materials and beam filters to produce x-ray spectra more suitable for mammography. The data show that the spectra produced by some systems are unique; in fact, one using molybdenum for both an anode and beam filtering element can produce an x-ray spectrum having more than 80% of the photons below 20 KeV. Using some typical breast phantom materials as attenuators, the primary x-ray spectra incident upon the imaging system were simulated and displayed. Implications of spectral shaping to image quality and patient dose are discussed.

A high-purity germanium spectrometer system was used to determine primary x-ray spectra over the 45--90-kVp region. Methods were devised for producing and examining spectra stimulating diagnostic conditions without operating the x-ray generator at high current levels. The techniques used to correct the experimental data and produce a photon fluence spectrum are discussed. The results, presented graphically and in tables, have been normalized to yield the relative number of photons per 2-keV interval. Methods for converting a normalized spectrum into a photon fluence spectrum that will produce an exposure of 1 R are presented. The analytical model and procedures used to calculate the K-escape fraction are discussed.

No abstract

The energy which must be absorbed in a CaWO4 x-ray intensifying screen to produce unit net opetical density on a film has been evaluated by measurement and calculation for a screen-film system over a range of beam qualities (1.4--7.4 mm A1 HVL) spanning the diagnostic x-ray region. It was found to be constant within experimental error. The absorbed-energy constant for three additional CaWO4 screens is presented for a single beam quality. To correct the estimation of absorbed energy, the fractional escape of tungsten K x rays has been evaluated and the results are presented as a function of phosphor loading. The absorbed-energy constant is useful for predicting optical density for variable beam conditions; a family of characteristic curves based on exposure is reduced to one curve for a particular film-screen system, expressed as optical density as a function of absorbed energy.

Lack of resolution (unsharpness) can reduce contrast in diagnostic radiography if the proper conditions of magnification and unsharpness are met. To describe this phenomenon, a modification of the contrast reduction factor (CRF) was introduced which used the response function of a semi-opaque edge to predict contrast reduction for small bar-shaped objects. To predict CRF, unsharpness is employed as a single-term description of resolution and is obtained experimentally from the edge response function. The unsharpness term is defined as the distance over which the response goes from 16.5% to 83.5% of the maximum. Measured and predicted CRFs were compared and the CRF concept was found to be an excellent predictor of contrast reduction. The individual components of unsharpness were determined experimentally and the sum-of-squares rule predicted adequately their combination. Three methods to measure unsharpness were compared: (a) the ICRU prescription using pinhole radiographs of the focal spot, (b) one-dimensional integration of the focal-spot pinhole radiograph, and (c) the unsharpness term produced by a semi-opaque edge. The latter two were measured using a microdensitometer.