Work In Progress

WORK IN PROGRESS

function of the first detector is to respond primarily to the lowenergy photons. This requires that the crystal have a strong photoelectric absorption, and also that it be fairly thin. Sodium iodide was not used because its absorption coefficient is so high, and its mechanical properties are so poor, that the crystal would be too thin and brittle. Calcium fluoride, on the other hand, is a sturdier material with a moderately strong photoelectric absorption, and therefore seemed a better choice. CaF 2 crystals of thickness 4,6, and 8 mm (to allow for experimentation) were fabricated by Harshaw Chemical Company to fit directly into an EMI detector housing. The function of the second detector is to capture the remaining photons. In this case thinness is not a requirement, and Nal seemed a good choice. In order to avoid afterglow problems and to maintain compatibility, we used a conventional EMI crystal from the Mark I scanner. The front half of the crystal was removed to make room for the CaF 2 detector, and the remainder was reclad with aluminum. The system was designed to be fully compatible with and easily installable in an EMI Mark I scanner. The Nal crystal is placed in the" A" housing, where it is exposed to the lower slice; the CaF 2 crystal is placed in the "8" housing, but extends over both the upper and lower slices (Fig. 1). A collimator adapter will be used to block off the "8" slice for patient studies; this was not done for the phantom studies described here. In order to have independent control over the gains of each detector during experimentation, the photomultiplier tubes were powered by separate Fluke 4128 high-voltage supplies. These would not be needed once the configuration was standardized. The complete changeover for phantom scans involved replacing the two crystals and changing the high-voltage cables. No alignment adjustments were needed.

Split-Detector Computed Tomography: A Preliminary Report 1 Rodney A. Brooks, Ph.D., and Giovanni Di Chiro, M.D. A split-detector .system, consisting of a thin CaF2 scintillator followed by a thick Nal crystal, was installed in an EMI Mark I scanner. Energy discrimination is possible because the CaF2 crystal responds primarily to low-energy photons; dual-energy images can be obtained from only one scan. Calculations and preliminary experiments show that the split-detector separates energies better than the 100-140 kVp method. It was possible with the split detector to differentiate between two weak solutions of CaCI2 and KI which had the same computed tomography number (44 H), under normal scanning conditions. INDEX TERMS: Computed tomography, apparatus and equipment. Computed tomography, physics

Radiology 126:255-257, January 1978

Several recent investigations of dual-energy computed tomography (1-5) involve changing the x-ray tube voltage, e.g., from 100 kVp to 140 kVp. 8ecause of differences in chemical composition and effective atomic number, the energy dependence of the computed tomography number will be different for various tissues. Thus, high Z elements like calcium and iodine may be identified, and types of soft tissues and tumors may be differentiated. We have independently developed a split-detector system for accomplishing the same purpose. The advantage of this system is that only one scan is required, with consequent elimination of the image registration problem. In addition, preliminary results and calculations indicate that this system offers better energy discrimination than the dual-energy method using 100 kVp and 140 kVp. Other possible applications are the elimination of the need for a "precontrast" scan before injecting iodinated contrast material, and the prediction of attenuation coefficients for use in radiation therapy. Finally, and perhaps most importantly, the system may be used to implement a suggestion of Alvarez and Macovski (6) for correcting the raw data in order to eliminate beam-hardening artifacts. The split-detector technique requires two detectors backto-back in the path of the x-ray beam. To implement this study on an EMI Mark I water-box scanner, the photomultiplier and electronics of the second slice were used to measure the scintillation of the front detector. This approach utilizes existing electronics but produces only one slice per scan. With a single-slice scanner, or to obtain two slices per scan, additional detectors and electronics would be needed. The method may also be implemented with a gas ionization detector, by splitting the collecting electrode. All data in this report were obtained from an EMI Mark I scanner with a "500" scale, but we have multiplied the EMI numbers by two in order to convert them to the Hounsfield scale.

SPECTRAL RESPONSE Calculations of spectral response were performed in order to understand the performance of the detector system. The emergent x-ray spectrum was calculated assuming a continuous 100 kVp bremsstrahlung spectrum at the anode, followed by 4.S-mm aluminum and 27-cm water filtration. These figures are typical for an EMI Mark I scanner (7). Figure 2 shows the relative response of the two detectors, assuming a 6-mm thick CaF 2 crystal. Approximations used in this calculation are: (a) only photoelectric absorption was considered in the CaF2 crystal, and

X-RAYBEAM

DESCRIPTION OF DETECTORS The detector system is shown schematically in Figure 1. The x-ray beam passes through both crystals sequentially. The

Fig. 1.

255

Schematic diagram of split-detector system.

256

RODNEY

A.

BROOKS AND GIOVANNI 01 CHIRO

January 1978

1.5

1.5

~c:

il:

::)

.~

;:,

oS!

Q)

~ 1.0

.~ .!l

...J

-c z

~ 1.0

o

en

..J

-c Z

a::

0

C)

I-

a:

~

en

uw

t

0

0.5

w

tu0 0.5

o o

L..-.._ _.....r.-.

Split-detector computed tomography: a preliminary report.

Work In Progress WORK IN PROGRESS function of the first detector is to respond primarily to the lowenergy photons. This requires that the crystal ha...
252KB Sizes 0 Downloads 0 Views