Chapter 7: Radiation Reduction Strategies in Myocardial Perfusion Imaging

Patient radiation exposure during medical procedures is a growing concern among health care providers, professional organizations as well as the general public. Medical radiation (of all subtypes) has increased by over 700% since 1980. Because of its value in diagnosis and prognosis in patients with known or suspected obstructive coronary disease, radionuclide myocardial perfusion imaging use has also increased over the last 25 years. Nuclear imaging accounts for approximately 25% of medical radiation. Cardiac imaging represents ~50% of all nuclear imaging procedures but is responsible for nearly 85% of all nuclear radiation doses.^1–4

Optimizing radiation exposure for patients is of considerable importance for patient safety and should be taken into account when ordering tests. For the nuclear cardiologist, this impacts choices in testing protocols, equipment, and even tracers. Radiation-reduction strategies should also take into account the value of the testing procedure and should not be performed at the expense of image quality, and thus the value of the examination itself. This chapter will present concepts of the consequences of radiation exposure, methods of measuring radiation exposure in medical imaging and in particular, nuclear cardiology, describe current radiation exposure in common cardiovascular single-photon emission computed tomography (SPECT) and positron emission tomography (PET) myocardial perfusion imaging (MPI) procedures, and discuss methods of reducing radiation exposure through instrumentation changes, protocol changes, and tracer choice.

Radiation Exposure: The Data

The exact consequences of radiation exposure are uncertain. The deterministic effects of direct radiation to an organ system, such as epidermal reactions, are well studied. However, understanding the consequences of radiation exposure to an individual is more obscure and difficult to assess. These stochastic effects of radiation exposure acquired during medical imaging are indeed difficult to apply to an individual's lifetime risk of developing cancers. This is in part due to variations in radiation types, exposure rates and quantities, and tissue susceptibilities as well as timing of procedures. In addition, malignancy generated by radiation exposure is often indistinguishable from those occurring from other causes.² Data estimated from the coronary computed tomography (CT) literature suggest that a 10-mSv radiation exposure increases lifetime risk of developing a fatal malignancy by 0.0005%.⁴ This represents a small, but measureable increase in lifetime risk, but is difficult to quantify when considering an individual patient. For perspective, Table 7-1 illustrates comparative risks of death from both radiation sources and other causes.⁴

Evaluating lifetime risk due to radiation exposure, one must consider age and life expectancy at time of exposure. Certainly, as demonstrated in Figure 7-1, augmenting a patient's lifetime malignancy risk is weighed more heavily at younger decades when many tissues are more rapidly replicating than when older and cell lines tend to be more senescent. Gender differences exist as well, and should also be acknowledged when considering testing in nuclear cardiology.^4–6

Measurement of Radiation Exposure in Nuclear Cardiology and Current Status

Measures of radiation and subsequent exposure are complex and more thoroughly discussed elsewhere (see Chapters 3 and 4). Typically, the Sievert unit (or milliSievert) is the most commonly used stochastic variable to compare radiation exposure levels from different sources. A Sievert (Sv), or milliSievert (mSv), is a unit which estimates the biological effect that 1 joule of radiation energy has on 1 kilogram of body tissue. Background radiation (from sources such as radon) produces exposure on the order of 3 mSv/year. The estimated radiation exposure from a typical rest/stress SPECT MPI is approximately 12 to 15 mSv of radiation. This classifies nuclear cardiac imaging as a high-dose procedure by regulatory bodies.^1,2,4

Table 7-2 represents the recommended radiotracer doses and estimated radiation exposure from recently published American Society of Nuclear Cardiology (ASNC) Guidelines.³ The number of radiopharmaceuticals that are available, combined with a variety of the protocols, create a heterogeneous potential of exposure to US patients undergoing cardiac nuclear testing.

Radiation exposure has been a concern for professional societies for several years. As a result, ASNC published an "Information Statement" in 2010 intended to address radiation exposure.¹ The recommended reduction in nuclear cardiac imaging was to utilize less than 9 mSv per patient in at least 50% of nuclear cardiology studies. The writing committee also recommended methods of reducing radiation to an individual patient, illustrated in Figure 7-2, such as the use of technetium-based tracers rather than thallium-201, stress-only imaging and cardiac PET over SPECT, if available.

Despite the published statements, it is unclear whether these recommendations resonated within the community. A recent study using the Intersocietal Accreditation Commission (IAC) database examined compliance with these recommendations 2 years after the information statement publication.⁵ Utilizing accreditation data submitted from 1047 laboratories in the United States, only 1% of laboratories were adhering to the <9 mSv standard (see Fig. 7-3). Furthermore, 10% of laboratories were still using thallium as their primary imaging agent although strongly discouraged by the information statement, and less than 1% of sites had incorporated stress-only imaging into their protocols to reduce radiation exposure. These findings suggest that other means may be necessary to gain accepted reductions of radiation from nuclear cardiology testing procedures.

Methods to Reduce Radiation Exposure: Patient Testing

Cardiac testing, whether it be for diagnostic or prognostic purposes, is based on risk–benefit analysis. The decision to pursue testing is based on pretest probabilities and estimated benefit from attaining the subsequent diagnosis. The practice of contemporary medicine offers a variety of imaging modalities. Not infrequently, patients undergo multiple testing procedures (for multiple indications) which incur some radiation exposure. For this reason there has been considerable attention paid to which patients would be most likely, and by default, least likely, to benefit from nuclear cardiac imaging for further investigation of obstructive coronary disease. The ASNC/ACC guidelines are important clinical recommendations meant to increase diagnostic yield and reduce patient risk.¹ A review of appropriate-use criteria is discussed elsewhere (see Chapter 13). Recent data from Doukky et al. suggest patients who undergo cardiac nuclear studies for "rarely appropriate" indications have significantly fewer abnormal studies, indicating less value.⁶ In conclusion, appropriate utilization of nuclear cardiac imaging is the first step toward reducing excessive radiation exposure to the individual patient and the population as a whole.

Reducing Radiation Exposure: Imaging Protocols

Various radiopharmaceutical choices and protocols exist for SPECT MPI, utilizing blood flow differential (see Table 7-2). The most widely utilized tracers, those that are technetium (Tc-99m) based, typically require rest and poststress images in order to sufficiently diagnose and risk stratify patients. Tc-99m protocols are preferred over Tl-201 (thallium) or dual-isotope (Tl-201 rest/Tc-99m stress) protocols due to the increased radiation of Tl-201 (>25 mSv). Despite this, IAC data demonstrates that over 1.5 million dual-isotope MPI studies are performed each year in the United States.⁵ Thus, an immediate means of reducing radiation exposure would be to eliminate thallium as a primary imaging tracer in all laboratories, as currently recommended.^1,7,8

Stress-Only or Stress-First Imaging to Reduce Patient Radiation

In more than 95% of laboratories in the United States, a patient undergoes the rest study first, followed by stress. Despite this common protocol, it has been recently estimated that perhaps only 8% to 15% of studies are found to be positive for ischemia or infarction.⁹ Thus in the overwhelming majority of studies, rest imaging is of no clinical value. It has been proposed for several years that performing the stress first, and rest only when necessary would be a reasonable means of increasing laboratory efficiency and reducing patient radiation. The "stress-only" or "stress-first" protocols which utilize poststress MPI have been advocated by European centers since 2005^10–12 as well as editorials and recommendations from ASNC.^1,3 In the United States, studies by Chang, Duvall, Heller, and others have demonstrated the utility of stress-only imaging.^3,7,10–12 Stress-only protocols have been shown to confer similar mortality to similar patients who underwent rest/poststress protocols.¹¹

An important consideration of stress-only imaging is the administered dose. In older camera systems, a low dose of 8 to 15 mCi of Tc-99m may yield an image that is difficult to interpret due to poor image quality and attenuation artifact. Thus, this protocol may require a high-dose stress study but still avoids the rest images in most circumstances. Comparison of cumulative radiation exposure of various tissues during stress-only, standard rest–stress, and dual-isotope protocols are demonstrated in Figure 7-4. Even without high-sensitivity equipment, utilizing stress-only protocols and standard radiopharmaceutical dosing provides decreased radiation exposure that can comply with ASNC recommendations. This reduction is even greater when low-dose stress-only protocols are utilizing appropriate camera systems. The susceptibility to artifact is a significant obstacle to providing high-specificity stress-only images. Significant reduction (by 35–85%) in the need for rest images is seen when attenuation correction (AC) is applied to image processing.¹³ In a study by Mathur et al., using stress only with AC in a chest pain center environment, only 8% required a rest study, and at least half of these patients had evidence of scar.¹⁴ In conclusion, modification of stress protocols to perform stress before ordering rest imaging represents an attainable method to reduce radiation exposure to patients unlikely to benefit from a rest study.

Image Processing and Software for Existing Systems

The generation of high-quality images from SPECT MPI data is paramount to ensuring diagnostic accuracy and utility of testing. To successfully reduce imaging doses as recommended by Cerqueira et al.¹ either new cameras with different technologies (see next section) or software solutions are required. Using software to enhance images can be used to either reduce image acquisition times or use lower doses with standard acquisition times, or sometimes both. For example, improvements to standard methods for image reconstruction (filtered back projection) such as iterative reconstruction (ordered subset expectation maximization or OSEM) can suppress background noise and improve count statistics. Compared with filtered back projection, iterative reconstruction algorithms require 50% to 75% less counts to produce similar images.^1,15,16 Proprietary software algorithms from the large medical technology companies (UltraSPECT Inc., GE Healthcare, Siemens, Philips) can be purchased to augment resolution of images without additional radiation. Wide-beam reconstruction (WBR) is one such algorithm that compensates for the complicated beam angles and spread of counts when reconstructing the image, and can be applied to standard camera systems with standard collimators. This technology provides the ability to reduce radiation exposure on the order of 50% to 75% without impacting image quality.¹⁷ Algorithms by Philips (Astonish) combine multiple levels of processing by utilizing OSEM, noise-reduction algorithms, collimator design, scatter modeling, and AC to produce high-quality images during half-time acquisition. Which software system is applicable to a given camera is dependent upon the age of the camera, the type (manufacturer), and available computer system. Generally, the manufacturer solutions (GE, Philips, Siemens) are limited to newer systems and only for the given company producing the software. WBC (UltraSPECT) and ImagenSPECT (CVIT, Kansas City, MO) technologies are available and use existing cameras from multiple vendors.^14,16

Impact of Technology: Hardware and Cameras

In the vast majority of US SPECT laboratories, standard Anger cameras are used. The design for these cameras was developed over 40 years ago and has changed little since conception. These cameras utilize collimators to physically resolve signals prior to processing. Standard collimators for general purpose studies attempt to balance count sensitivity with spatial resolution on the order of 9 to 10 mm. High-sensitivity collimator designs produce images with spatial resolutions of 8 to 8.5 mm; however, this is at the expense of doubling the radiation dose.^1,3,16 Clinically, 1.0-mm resolution differences in image resolution do not warrant doubling the radiation exposure to patients. However, specialized collimator designs can lead to decreased radiation dose. Custom cardiac design collimators have been designed that focus on increasing spatial resolution and sensitivity for counts from the cardiac region. These collimators are designed to be utilized within a certain camera system and require accompanying reconstruction algorithms.¹

In contrast to sodium iodide (NaI) cameras, new solid detectors have been developed which utilize cadmium zinc telluride (CZT) arrays. These new camera arrays demonstrate better sensitivities in the range of 2.2 to 4.7 kcps compared to 0.5 to 0.7 kcps of standard Anger NaI cameras. Detailed descriptions of SPECT camera systems are discussed in Chapter 4. Data suggest that higher-count sensitivity of CZT cameras can produce images of comparable diagnostic quality with 15% to 30% of amount of standard radiation dose at the same scan duration.^1,18 Table 7-2 demonstrates the substantial reduction in radiation exposure that is possible with "new-technology" camera systems. Again, application of new camera systems to additional radiation-reduction techniques such as stress-only imaging (last column) significantly reduces radiation exposure when compared to standard Tc-99m stress/rest protocols utilizing standard anger cameras.

The development of these high-efficiency cameras and their incorporation into imaging systems have provided high-quality images to be constructed at a fraction of the standard isotope dose. Examples of combination software/hardware systems include the Cardius-3 XPO camera system produced by Digirad, Inc. (Poway, California, USA). This system utilizes triple-head geometry of indirect, solid-state (Csl[Tl]) detectors which are allowed to vary source distance to produce images with spatial resolution of 8.95 mm. This is, in part, accomplished through OSEM methods for reconstruction (discussed previously). Recent studies have also demonstrated that total patient radiation exposure can be reduced to 4 to 9 mSv with very high-quality images.¹³

Cardiovascular PET Radiation Exposure

This review, thus far, has been confined to SPECT radiation exposure. Recently, cardiovascular PET has been entering the clinical arena for myocardial perfusion imaging as well as other applications such as myocardial viability and infection/inflammation imaging. The value of cardiovascular PET as well as instrumentation is described in Chapters 3, 4 and 10.

Cardiovascular PET utilizes three primary tracers for imaging, rubidium and ammonia for perfusion imaging as 18-fluorodeoxyglucose (F-18 FDG) for viability, infection, and inflammation imaging. The radiation exposure for each of these is represented in Table 7-3. PET systems typically employ line source (utilizing gadolinium) or CT sources for AC. It is noteworthy that AC requires the use of additional radiation. This radiation exposure is generally quite small, usually less than 1 mSv.^19,20

The two FDA-approved PET tracers for myocardial perfusion imaging are rubidium (Rb-82) and ammonia (NH₃). Due to the short half-life of these tracers (76 seconds and 9.8 minutes, respectively), the radiopharmaceuticals can be administered at higher doses than SPECT but result in overall lower radiation exposure. The most commonly used PET perfusion tracer is Rb-82. Several studies have examined the human radiation exposure with this tracer.^19–21 In normal subjects undergoing pharmacologic stress with Rb-82, Senthamizhchelvan et al. determined that 20 mCi results in 0.9-mSv exposure.²² Given that a rest and stress dose is 30 to 60 mCi, the total radiation exposure is estimated to be 2.7 to 5.4 mSv per patient, well below ASNC recommendations and current SPECT procedures. Hunter et al. carefully calculated radiation exposure to patients undergoing Rb-82 at the University of Ottawa and found the average to be 2.4 mSv.²⁰ Other advances in cardiac PET imaging include 3D imaging which utilizes the high-sensitivity PET cameras to quickly image low-dose activity without loss of diagnostic sensitivity.^19–22 With 3D imaging, the patient exposure would be half that of 2D imaging (1.8–3.5 mSv).²³

Another albeit, less commonly utilized tracer is ¹³N-ammonia. ¹³N-Ammonia demonstrates a more linear response to blood flow compared not only to SPECT tracers but ⁸²Rb as well. Subsequent studies have confirmed the accuracy of ⁸²Rb with ¹³N-ammonia in the detection of coronary artery disease with low radiation exposure levels of approximately 4 mSv per study utilizing a 20-mCi dose.⁵ With 3D imaging, the radiation exposure is lower (2–3 mSv).

For nonperfusion PET imaging, the myocardium can be imaged directly with F-18-FDG. FDG viability protocols are discussed in Chapter 21. Data suggest radiation exposure levels of approximately 7 mSv per study for viability assessment using the standard 15-mCi dose.²⁴ In addition, FDG imaging has emerging clinical utility in identifying inflammation and infection in conditions such as sarcoidosis and endocarditis, respectively discussed in Chapter 24.²⁵

Cardiac nuclear imaging is a popular and valuable tool in the practice of cardiology.

These noninvasive diagnostic procedures have high diagnostic accuracy when applied to the appropriate patient. Since these techniques use radioactive tracers, the risk of developing cancers due to radiation exposure is present but not fully understood. This risk has become a societal concern and has resulted in safe practice guidelines to mitigate risk. There is a wide variety of protocols, isotopes, and systems utilized in cardiac nuclear testing, each with varying degrees of radiation exposure. Clinicians must balance the risk and benefits in order to practice for safe and fiscally responsible medicine and then the utilization of techniques becomes critical to their appropriate application. Current ASNC recommendations for radiation exposure during cardiac nuclear imaging are below 9 mSv per study. Unfortunately, data have demonstrated that few nuclear laboratories are able to achieve these goals. The current ASNC goal is attainable through application of stress-only protocols, CZT camera, application of PET tracers, and other techniques noted in this chapter.