The use of external detectors for in vivo measurement of radioactivity automatically excludes the use of radionuclides which do not emit penetrating radiation such as x- or γ-rays. One exception is positron emitters where annihilation γ-rays can be detected externally. Three types of problems arise in the in vivo detection of such radionuclides: collimation, scattering, and attenuation. All three factors, coupled with the fact that a detector has to be a certain distance away from a radioactive source (the source being in vivo), reduce the geometric efficiency in in vivo detection by two to three orders of magnitude as compared with the in vitro case (see Chapter 9). This is one of the reasons why millicurie (MBq) amounts are used in in vivo studies as compared with only microcurie or less (kBq) in in vitro studies.

Collimation. Because one is interested in the detection of radioactivity present in a small area or volume (e.g., an organ or part of an organ), it is important to exclude all x- or γ-rays originating outside the area or volume of interest from reaching the radiation detector. This is achieved by the use of a collimator, a device that limits the field of view (FOV) of a radiation detector. A variety of collimators are available for different purposes. These are usually made of lead, which is inexpensive and has both, a high-density and high atomic number Z, and therefore, a high-attenuation coefficient for the x- or γ-rays whose energies are of particular interest in nuclear medicine (<500 keV). A diagram of a simple collimator is shown in Figure 10.1. Note that simple collimation does not allow full discrimination against the radioactivity underlying or overlying the volume of interest. Also, the FOV of
such a collimator is determined by two parameters: the length and the radius of the opening (hole) in a collimator. By reducing the radius or increasing the length, one can reduce or narrow the FOV of a collimator to any desired size.

The FOV of a single-hole collimator is related to the spatial resolution and sensitivity of the collimator. Increasing the FOV degrades the spatial resolution and improves the sensitivity, and vice versa (Chapter 12). Also, as can be seen in Figure 10.1, the FOV extent widens with depth; therefore, spatial resolution also gets worse with depth—a troublesome feature to be dealt with when accurate quantification of radioactivity is desired (i.e., single-photon emission computed tomography [SPECT]; Chapter 14).

Scattering. The γ- and x-rays emitted by radionuclidic sources embedded in a mass of matter experience, during interaction with that material, scattering via the Compton process. In Compton scattering (see Chapter 6), an interacting γ- or x-ray loses some of its energy and changes its direction. The change in the direction of many γ-rays, which originate outside the FOV of the collimator, may cause them to be scattered toward the radiation detector, thus defeating the purpose of this device (Fig. 10.2).

Depending on the angle of scattering, an x- or γ-ray loses different amounts of energy during Compton scattering. As a result, x- or γ-rays emerging from a monochromatic source embedded in a tissue volume are no longer monochromatic but are polychromatic, as shown in Figure 10.3A. In the case of a 140-keV γ-ray, the energies of scattered γ-rays range from 90 to 140 keV. This assumes that a γ-ray is scattered only once in the tissue. In the case of double or triple Compton scattering, which are generally negligible but become more probable as the tissue volume embedding the source increases, the lower limit of scattered γ-ray energy (90 keV) becomes even lower. In diagnostic x-ray imaging, scattered x-rays
are reduced by using grids. In nuclear medicine, this can be accomplished by using energy discrimination. Efficiency of scatter rejection with energy discrimination depends on the energy resolution of a detector. When using an NaI(Tl) detector (Fig. 10.3B), pulse-height analyzer (PHA) controls, E and ΔE, are set to accept only those pulses whose height corresponds to the photo peak of the unscattered γ-ray. For an effective job of energy discrimination, the energy resolution of a detector should be very good. For a 140-keV γ-ray, NaI(Tl) scintillation detectors possess moderate energy discrimination capability with full-width at half-maximum (FWHM) of 15 to 20 keV as compared with Ge(Li) semiconductor detectors (Fig. 10.3C) with FWHM of 1 to 2 keV. As a result, a Ge(Li) detector is capable of rejecting almost all scattered radiation, whereas an NaI(Tl) detector does only a partial job. However, despite their better energy discrimination capability, Ge(Li) detectors are seldom used in nuclear medicine, primarily because of their very low sensitivity compared with an NaI(Tl) scintillation detector. The newer detector, cadmium zinc tellurium (CZT), has energy resolution in between the two, NaI(Tl) and Ge(Li), and therefore better scatter rejection ability than NaI. It also has better sensitivity than NaI(Tl) detector (see Chapter 8). That is why it is a prime candidate to replace NaI detector in nuclear medicine, once the technical problems in growing good quality CZT crystals have been overcome.

The amount of scattered radiation rejected by an NaI(Tl) detector depends on the width of the PHA window. The narrower the window, the more scattered radiation is excluded. But narrowing the window to reject scattered radiation exacts a price in terms of reduced sensitivity because more of the primary radiation is also rejected. With NaI(Tl) detectors, there are two other and better ways to reject scattered radiation. One way is the simultaneous use of one more (second) window at a lower energy as shown in Figure 10.3B. The counts or a fraction of counts in the second window are subtracted from the window set at the photo peak. Another way is the simultaneous use of two more windows (second and third) that are only a few keV in width and set just below and above the photo peak as shown in Figure 10.3B. Here, the average of the counts from these two windows is subtracted from the window set at the photo peak. Both methods are better than narrowing the photo peak window because primary radiation is not lost. However, these methods by no means reject scatter completely and accurately or as much as a Ge(Li) detector does as shown in Figure 10.3C.

Attenuation. Because the depth, shape, and size of an organ containing the radioactive substance are unknown beforehand, the attenuation of x- or γ-rays (absorption through the photoelectric effect and the Compton scattering) in the organ and the tissue overlying or underlying the organ is another serious hurdle in the in vivo determination of radioactivity. As can be seen in Figure 10.4, the distribution of radioactivity varies throughout the cross section, and x- or γ-rays emitted by a radioactive source, depending on its location, pass through different thicknesses and often different types of underlying or overlying tissue and therefore are attenuated by different amounts. Hence, a uniform distribution of radioactivity in an organ produces varying number of counts (independent of statistical variations) from different locations of the organ, not a desirable feature in any imaging.

It is possible to minimize attenuation loss by the use of high-energy x- or γ-rays. The attenuation coefficient for x- or γ-rays in tissue drops sharply with an increase in γ-ray energy up to about 100 keV and levels off with the increase in γ-ray energy above 100 keV. Therefore, radionuclides emitting x- or γ-rays with energies above 100 keV are preferred. Because the sensitivity of NaI(Tl) detectors and the collimators used in scanning decreases with the increase in x- or γ-ray energy, the optimum range of energies for in vivo use for single-photon detection (as opposed to coincidence detection, Chapter 15) is between 100 and 300 keV (see Chapter 5, p. 43 and Chapter 6, Figure 6.4).

In organ uptake studies such as thyroid uptake of radioiodine, in addition to the use of high-energy γ-rays, attenuation effects are taken care of by measuring a known amount of radioactivity in a standard phantom that reflects the average size, shape, and depth of an organ in a standard man.

In planar imaging, generally, attenuation is considered a fait accompli and is taken into account only when interpreting a scan. However, digital imaging instruments as commonly used can compensate for some attenuation effects by taking images from two opposite sides and then forming a geometric mean image. Geometric mean G of two images or numbers, N₁ and N₂, is defined as G = . Methods of attenuation correction in SPECT and PET are discussed in Chapters 14 and 15.

An uptake probe consists of two basic parts: an NaI(Tl) detector and a collimator.

NaI(Tl) Detector. In vivo studies, the size of the crystal in an NaI(Tl) detector is an important consideration. Crystal size is determined by the energy of the γ-ray to be detected and the sensitivity requirements of a particular study. For thyroid uptake studies using ¹³¹I, the International Atomic Energy Agency recommends that a crystal less than 1 × 1 inch should not be used. A 1.5 × 1-inch crystal is generally adequate for thyroid uptake measurements of ¹³¹I and also serves as a multipurpose instrument in the nuclear medicine laboratory.

Collimator. The design of a collimator for uptake studies is dictated by its intended application. However, the following general requirements apply in most cases:

Because requirements 1 and 3 oppose each other to some extent, one should make the best of a given situation. A typical collimator used in thyroid uptake measurement is shown in Figure 10.5. The overall sensitivity within the FOV of such a collimator varies inversely as the square of the distance between the source and the detector, but the uniformity across the organ improves as the distance from the detector is increased. For thyroid uptake, a distance of 30 cm is considered optimum.

These probes are used to detect the radioactive uptake in surgical procedures called intraoperative lymphatic mapping or sentinel lymph node biopsy, a minimally invasive technique for evaluating the potential spread of cancer to lymph node tissues and organs. They consist of a small detector (5 to 15 mm diameter and 2 to 5 mm thickness) and a collimator with a narrow FOV or resolution (3 mm to 1.5 cm). Because of surgical requirements, these probes have to be small and sterile. As a result, either semiconductor detector (e.g. CdTe or CZT), or scintillator detectors (e.g. BGO or LSO) coupled with a photodiode instead of a PM tube, are used in these probes. Recently, wireless probes have been introduced to make their use even simpler during surgery (Fig. 10.6).