The x-ray image intensifier (II) was the key component of a fluoroscopy system from approximately 1960 to 2010. As some II based systems remain in clinical use, their characteristics will be discussed below. A diagram of an II is shown in Figure 9-3.

There are four principal components in an image intensifier: (1) an input screen that absorbs incident x-rays and converts the image to an electron pattern, (2) the electron optics, which accelerates the electrons and minifies the electron image,
(3) an output screen, which converts the accelerated electrons into a visible light image, and (4) a housing that maintains a vacuum, contains the other components, and provides an unimpeded path for the electrons that carry the image.

The flat panel image receptor replaces the image intensifier tube, optical system, and cameras by directly converting the x-ray images into a digital form. Because FPDs do not operate in a vacuum, they are mechanically protected by carbon fiber covers instead of the stronger metallic input surface of an image intensifier. This reduces x-ray attenuation and improves the quantum detection efficiency compared to the II. FPDs are comprised of thin film transistor (TFT) arrays of individual detector elements (dexels) that are packaged in a square or rectangular area (Fig. 9-5A). Both indirect and direct x-ray conversion modes are used with TFT panels for fluoroscopy (Fig. 9-5B). In both types of systems, each dexel has a capacitor, which accumulates and stores the signal as an electrical charge, and a transistor that serves as a switch. Indirect detection TFT systems have a phosphor layer that absorbs the x-rays and converts a fraction of their energy into light; each dexel has a transistor and a capacitor, in addition to a photodiode that converts the x-ray induced light from the phosphor into a corresponding charge (Fig. 9-5C). For direct detection fluoroscopy detectors, a semiconductor (selenium) produces x-ray induced charge directly, which is collected under an applied voltage to ensure that the signal is captured within the same dexel as the x-ray absorption event. In either detector type, while a frame (image) is being acquired, the electrical charge proportional to the x-ray flux incident on that dexel is accumulated and stored by the capacitor. Readout of the FPD is performed by electronically switching a gate line “on,” which opens all transistor gates for a given row. The charge stored in each dexel flows through its transistor’s drain line to a charge amplifier, and then the amplified signal is digitized. The gate line is then switched “off,” and readout progresses to the next gate line on the panel. Reading the array discharges the capacitors and readies them for acquiring the next frame. This is performed row by row until the entire frame is read. Even in continuous fluoroscopic acquisitions, the charge continues to be stored in each capacitor in a steady-state situation so that x-ray information is always collected as the flat-panel image receptor provides real-time fluoroscopic presentation of image data.

The general construction of all x-ray tubes is similar. However, due to the nature of the procedures, fluoroscopic x-ray tubes are commonly used in a more continuous manner than radiographic tubes. These tubes also operate at radiographic power levels during fluorography. Fluoroscopic tubes have higher heat-storage capacity and faster cooling rates than radiographic tubes. Traditionally, the small focal spot is used for fluoroscopy and the large focal spot for fluorography. Thus, spatial resolution is likely to decrease when going from fluoroscopy to fluorography. Some tubes also have an additional micro-focus focal spot intended for geometrically magnified procedures. Newer systems select the smallest available focal spot that can accommodate the immediate x-ray power level for all modes of operation. This will result in similar spatial resolution for thin body parts and lower resolution for fluorography of thick parts.

Fluoroscopic collimators are designed to dynamically adjust the overall x-ray field size in response to a combination of inputs from both the operator and the system. The intention is to confine the x-ray beam to the smallest area within the active area of the image receptor that is consistent with immediate clinical requirements. Many systems also have additional, operator controlled, x-ray translucent elements (sometimes called wedge filters) that are used to equalize image receptor inputs when anatomical structures providing very different attenuation(e.g., lung and mediastinum) are simultaneously in the beam.

Portions of the x-ray beam that are not seen by the imaging chain do not supply any clinical information. Such irradiation unnecessarily increases both patient and staff radiation risk. Scatter from these areas also degrades image quality in the useful image.

Confinement of the beam to the active FOV is a system function. For both the image intensifier and the flat panel, the x-ray collimator automatically adjusts to limit the x-ray beam to the active FOV whenever the FOV is changed and to accommodate changes in the source-to-image-receptor-distance (SID). Ideally, the system should be set so that an unirradiated area is always seen around the edges of the image.

Collimation of the beam within the active FOV is the operator’s responsibility. Smaller beams produced by collimator adjustment by the operator reduce total
irradiation, thus improving both safety and image quality. The goal is to only irradiate the region of immediate clinical interest. Obtaining a smaller FOV by selection of a magnification mode does reduce the beam size and magnifies the displayed image; however, most fluoroscopes increase the dose-rate when the FOV is reduced by this method. Depending on programming, this increase may diminish or even negate the radiation saving effects of collimation within the FOV. Image magnification of a collimated area within a larger FOV can also be offered by the image processor through pixel replication and interpolation, or by simply viewing the image on a larger monitor. If the resulting image is clinically acceptable, this can reduce radiation dose relative to the use of magnification mode.

By 1990, the DQE of the image intensifier was near 50% and x-ray quantum noise was the major noise source in fluoroscopy. There was only a factor of two available for reducing patient irradiation by improving the efficiency of the image receptor. Radiation levels can also be reduced when the radiological conspicuity of clinically important items, such as contrast media and guidewires, is increased.

One way of improving the conspicuity of higher atomic number (Z) elements is to increase the fraction of x-ray photons in the beam with energies slightly above the K absorption edge of the material of interest (e.g., iodine at 33.2 keV). Spectral shaping can be accomplished by adding a copper filter (0.1-1.0 mm) and simultaneously restricting the x-ray tube’s operating kV (Fig. 9-7). Adding copper reduces the number of photons in the beam below the iodine edge. Reducing the kV limits the number of higher energy photons. For a given amount of electrical power consumed by the tube, such kV-filter combinations decrease the photon fluence available to form the image. Adequate fluence using spectrally shaped beams became feasible when x-ray tubes were developed with high power ratings.

Most modern fluoroscopic systems include variable thickness copper filters in the collimator assembly. The thickness of copper used at any given moment is automatically controlled by the AERC and operator selection of settings. In some systems, the acquisition selection (e.g., low-dose-rate fluoro, standard-dose-rate fluoro) determines the filter thickness. In other systems, the AERC further controls filter thickness
along with the usual x-ray factors. In some cases, for increasing patient thickness or acquisition mode changes, the limits of the x-ray tube power output require the filter thickness to decrease and the kV to increase. In such settings, at the same kV, the half value layer (HVL) will differ between fluoroscopy and fluorography; and the HVL may decrease with an increase in kV if there is a simultaneous decrease in copper thickness. Figure 9-8 illustrates this behavior for a typical interventional fluoroscope.

Fluoroscopes offer at least two fluoroscopic and one fluorographic sub-modes. Often, systems offer literally dozens to hundreds of sub-modes to accommodate a wide range of operator preferences and clinical requirements. Most of these use an AERC system. The purpose of AERC is to deliver a constant x-ray intensity to the image receptor irrespective of tissue thickness in the beam path. This results in a reasonably constant detector SNR over its working range. In a FPD, the AERC collects detected x-ray intensity information observed by a predetermined subset of dexels. Depending on clinical configuration, the subset ranges from dexels in a relatively small central area to all dexels within the collimated beam. If the average signal is too low, the AERC commands the generator to increase x-ray output. Similarly, too high a signal initiates a command to reduce radiation output. AERCs commonly further adjust outputs when the FOV or frame rate is changed. These adjustments stabilize the image noise perceived by the fluoroscopist when these factors vary.

For large patients or steep projection angles, the x-ray air kerma rate may reach the normal fluoroscopy regulatory limit (88 mGy/min). The resulting x-ray flux
reaching the image receptor (II or FP) is lower than desired, and x-ray quantum noise is higher than desired. Necessary additional electronic amplification may add additional noise. The noise in the resulting image may be too high for some clinical purposes. Some fluoroscopes have a specially activated high-level control (HLC), which enables a higher limit (176 mGy/min) to reduce noise. The high-level mode should be only used when clinically necessary.

Under AERC, the system changes the x-ray factors (kV, mA, pulse width), focal spot, and beam filtration in a predetermined manner that varies between systems. As noted above, the same system usually offers user-selectable options with markedly different behaviors. These selections are part of the toolkit used to select the balance between patient dose and image quality appropriate to different types of examinations, and in some cases different operator imaging requirements. The change in output, and the maximum output dose rate when changing from “low” to “normal” to “high-level” at any patient thickness is also included in the system’s configuration. Image processing parameters may also be controlled by the AERC. These changes further influence image presentation to the operator.

Figure 9-9 illustrates some possibilities with a hypothetical system equipped with three fluoroscopic modes, each with a different regulatory exposure-rate limit. In Figure 9-9A, the output is the same until the mode’s limit is exceeded. All three modes produce the same radiation output for thin to medium size patients. The “low” mode is limited to an air kerma rate (for C-arm gantries, at 30 cm from the image receptor assembly) of 44 mGy/min; in this mode all thicker patients receive the limiting dose rate while image quality decreases as patient thickness increases. Similar behavior is seen for the “normal” and “high-level” modes when their limits are reached. Figure 9-9B illustrates a system with three fluoroscopic modes. The lower two fluoroscopic modes converge near the regulatory limit. The high-dose-rate mode is higher for all patient thicknesses and has double the usual regulatory limit.

Figure 9-9C illustrates that the relationship between patient thickness and patient dose for any mode and dose rate limit is programmable. The upper curve is intended to maximize the visibility of iodinated contrast media as much as possible. The generator is programmed to keep the kV low and adjust mA and/or pulse width to provide an appropriate detector signal. The lowest curve is intended to minimize patient irradiation at the possible expense of iodine visualization. It does so by rapidly increasing kV with increasing patient thickness. One or more additional intermediate curves may be provided on a given system.