{
  "id": "single-photon-emission-computed-tomography",
  "title": "Single-Photon Emission Computed Tomography",
  "lang": "en",
  "source_lang": "en",
  "summary": "Single-photon emission computed tomography maps radiotracer distributions in three dimensions to assess organ function, molecular processes, and disease.",
  "category": "health",
  "subcategory": "medicine",
  "type": "concept",
  "aliases": [
    "SPECT",
    "single-photon emission tomography",
    "SPET"
  ],
  "generated_at": "2026-11-06",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/single-photon-emission-computed-tomography/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/single-photon-emission-computed-tomography/",
      "json": "https://aiwiki.page/zh/single-photon-emission-computed-tomography.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "Single-photon emission computed tomography (SPECT) is a form of nuclear medicine imaging that produces three-dimensional maps of a radioactive tracer within the body. After a radiopharmaceutical is administered, usually by intravenous injection, detectors record gamma rays emitted from the patient at multiple angles. Computer reconstruction converts these measurements into cross-sectional images. Depending on the tracer, the resulting distribution can represent blood supply, bone formation, receptor availability, or other physiological processes rather than anatomical structure alone.\n\nSPECT is widely used to investigate myocardial perfusion, skeletal disease, and selected neurological conditions. It also supports the localization of abnormal endocrine tissue, infection, and some tumors. Its clinical meaning depends on the biological behavior of the tracer, the imaging protocol, and the relationship between tracer uptake and disease. Hybrid SPECT/CT systems combine functional information with X-ray computed tomography, improving anatomical localization and correction for photon attenuation.\n\nAlthough SPECT and positron emission tomography both image internally administered radioactive substances, they detect emissions differently. SPECT usually determines photon direction with a physical collimator, whereas PET detects pairs of annihilation photons in coincidence. This distinction shapes their instrumentation, sensitivity, resolution, radiopharmaceutical supply chains, and clinical applications."
    },
    {
      "heading": "Physical principles and image formation",
      "text": "SPECT detects individual photons emitted during radionuclide decay. The term “single-photon” distinguishes this detection method from coincidence-based PET; it does not mean that a scan uses only one photon, or that every radionuclide emits exactly one photon per decay. A successful examination records many thousands or millions of events, each contributing statistical information about the tracer distribution.\n\nA detector cannot ordinarily determine the incoming direction of a gamma ray from its interaction position alone. A collimator, generally made of lead or tungsten, therefore sits between the patient and detector. Its holes admit photons traveling along selected directions while absorbing most others. With a parallel-hole collimator, a detector image approximates a projection of the activity distribution along parallel lines. Rotating the detector around the patient supplies the angular information needed to reconstruct depth.\n\nCollimation imposes a fundamental trade-off. Narrower, longer holes provide better directional discrimination but admit fewer photons. Wider holes improve sensitivity at the cost of spatial resolution. Resolution also deteriorates as the emitting tissue moves farther from the collimator. Consequently, keeping detectors close to the body can materially improve image quality. Conventional clinical SPECT commonly achieves reconstructed spatial resolution on the order of 8–15 mm, although this varies substantially with equipment, radionuclide, patient size, reconstruction, and the definition of resolution.\n\nPhotons do not travel through tissue without interference. Some are absorbed or scattered out of their original path, producing attenuation. Others undergo Compton scattering and reach the detector after changing direction and losing energy. For a narrow beam of unscattered photons, transmission along a path is described by\n\n\\[I=I_0\\exp\\left(-\\int_{\\text{path}}\\mu(s)\\,ds\\right),\\]\n\nwhere \\(I_0\\) is the incident intensity and \\(\\mu(s)\\) is the linear attenuation coefficient. In SPECT, unlike transmission CT, the emission point can occur anywhere along the path, so reconstruction must account for different attenuation distances.\n\nThe detected count total is statistical rather than exact. Under a simple Poisson model, the standard deviation of a count \\(N\\) is approximately \\(\\sqrt{N}\\), and relative counting uncertainty decreases as \\(1/\\sqrt{N}\\). Increasing acquisition time or detected activity can reduce noise, but competes with motion, patient tolerance, radioactive decay, and radiation exposure. Image quality therefore reflects a balance among resolution, sensitivity, contrast, and statistical uncertainty rather than a single equipment specification."
    },
    {
      "heading": "Radionuclides and radiopharmaceutical behavior",
      "text": "A radiopharmaceutical combines a radionuclide with a chemical or biological form that determines its distribution. The emitted radiation makes the substance detectable, while its chemistry and pharmacokinetics determine what the image represents. The same radionuclide can support very different examinations when incorporated into different compounds. Technetium-99m, for example, is used in agents for myocardial perfusion, skeletal imaging, renal imaging, and cerebral perfusion.\n\nTechnetium-99m is particularly important because it emits a principal gamma ray near 140 keV and has a physical half-life of approximately six hours. This combination is compatible with gamma-camera detection and many preparation and imaging schedules. Much clinical supply comes from molybdenum-99/technetium-99m generators: molybdenum-99 decays to technetium-99m, which can be eluted at the imaging facility. Production and distribution disruptions can therefore affect access to multiple nuclear medicine examinations, not merely one tracer.\n\nOther radionuclides serve different purposes. Iodine-123 has a half-life of approximately 13 hours and a principal gamma emission near 159 keV; it is used in thyroid-related and selected neurological or neuroendocrine applications. Indium-111 has a half-life of about 2.8 days and prominent emissions near 171 and 245 keV, allowing delayed imaging with labeled cells and some targeting agents. Thallium-201, with a half-life of about 73 hours, has historically been important in myocardial perfusion imaging. Photon energy and emission complexity influence collimator selection, scatter correction, and radiation dose.\n\nPhysical decay is only one cause of declining activity in an organ. Biological clearance also matters. When both processes can be approximated by independent exponential removal, the effective half-life satisfies\n\n\\[\\frac{1}{T_{\\mathrm{eff}}} = \\frac{1}{T_{\\mathrm{phys}}} + \\frac{1}{T_{\\mathrm{bio}}}.\\]\n\nReal tracer behavior is often more complex, involving delivery, reversible binding, metabolism, and excretion across several compartments. Consequently, image timing is part of the measurement itself. Early and delayed acquisitions may answer different biological questions.\n\nTracer uptake is not automatically disease-specific. Technetium-labeled diphosphonate accumulation reflects osteoblastic activity and regional delivery, so increased skeletal uptake can occur in metastases, fractures, degenerative change, or infection. Similarly, reduced myocardial tracer accumulation can reflect impaired perfusion, scar, attenuation, or technical error. Radiochemical purity and preparation quality also matter: free or chemically altered radionuclide may produce unexpected organ uptake. Interpretation therefore requires knowledge of both the intended tracer pathway and recognizable patterns of altered biodistribution."
    },
    {
      "heading": "Instrumentation and historical development",
      "text": "Modern SPECT developed from earlier efforts to image radioactive distributions with mechanically scanned detectors and gamma cameras. In 1958, Hal Anger described a scintillation camera capable of recording event positions over a broad field of view without mechanically scanning a single small detector. David Kuhl and Roy Edwards reported emission tomographic methods in 1963, establishing important foundations for reconstructing internal tracer distributions from multiple views. Practical clinical systems expanded as rotating cameras, digital acquisition, and reconstruction computing became more capable during the following decades.\n\nThe conventional gamma camera uses a sodium iodide crystal activated with thallium. A gamma-ray interaction produces a small flash of visible light, which is detected by photomultiplier tubes. Electronics estimate the interaction position and deposited energy from the pattern and magnitude of the light signals. The detector assigns events to an energy window around the radionuclide’s photopeak, rejecting many scattered photons and other unwanted events. Energy discrimination is imperfect, however, because both the detector response and the scattering process broaden the measured spectrum.\n\nMany general-purpose systems have two large detector heads that rotate around the patient. Their relative configuration depends on the examination: opposed heads are common for full angular sampling, while cardiac protocols may use other arrangements. Some systems have three heads or specialized geometries. The detector follows a circular or body-contouring orbit, and positioning affects both patient clearance and resolution. Parallel-hole collimators are standard for many applications; fan-beam collimators can improve sampling efficiency for the head, while pinhole geometries provide magnification for small organs and preclinical imaging.\n\nAn important alternative is direct-conversion detection using cadmium zinc telluride (CZT). Instead of first converting radiation to light, the semiconductor produces electrical charge when a photon deposits energy. Pixelated CZT detectors can provide improved energy resolution and support compact designs. Dedicated cardiac systems combine these detectors with specialized collimation and stationary or limited-motion geometries. Their performance derives from the complete detector–collimator arrangement, not simply from the detector material.\n\nHybrid SPECT/CT systems entered clinical practice around the turn of the twenty-first century. Early systems often paired emission imaging with relatively low-performance CT intended primarily for attenuation correction and localization. Later systems incorporated more capable CT components. The CT examination may still be low-dose and non-diagnostic, or it may be a full diagnostic study, depending on equipment and protocol. These are not interchangeable categories: a localization CT may lack the image quality, contrast administration, or acquisition technique needed to answer a separate anatomical question."
    },
    {
      "heading": "Acquisition, reconstruction, and image correction",
      "text": "A SPECT acquisition samples the body from a sequence of directions. Typical protocols use tens to more than a hundred projection views, acquired over either a complete rotation or a selected angular arc. Some myocardial studies use approximately 180 degrees of sampling, whereas many brain and general-purpose examinations use 360 degrees. The number of views, acquisition duration, image matrix, angular coverage, and detector distance are selected together. Inadequate angular sampling can produce streaks and other reconstruction artifacts even when the total number of detected events appears sufficient.\n\nFiltered back projection was historically central to reconstruction. It combines projection filtering with back projection to compensate for the blurring inherent in simply spreading each measurement through the image volume. Iterative methods are now widely used because they can model the acquisition process more explicitly. Maximum-likelihood expectation maximization and its accelerated variant, ordered-subsets expectation maximization (OSEM), repeatedly compare predicted projections with measured data and update the estimated activity distribution.\n\nA simplified statistical model is\n\n\\[y_i\\sim\\operatorname{Poisson} \\left(\\sum_j a_{ij}x_j+b_i\\right),\\]\n\nwhere \\(y_i\\) is a measured projection count, \\(x_j\\) is activity in image voxel \\(j\\), \\(a_{ij}\\) describes the probability of detecting activity from that voxel in measurement bin \\(i\\), and \\(b_i\\) represents modeled background contributions. The system matrix can incorporate attenuation, detector geometry, and distance-dependent spatial resolution. More iterations can improve recovery of contrast but also increase noise, so stopping rules, filtering, and regularization influence the final appearance.\n\nCT-based attenuation correction converts CT information into estimates appropriate to the gamma-ray energy being imaged. Scatter correction may use neighboring energy windows or more elaborate physical models. Resolution recovery models the detector–collimator response but cannot restore arbitrarily fine detail that was never adequately sampled. It can also introduce edge effects. Corrected images therefore require inspection alongside acquisition data and, when relevant, uncorrected images rather than automatic acceptance as more accurate in every circumstance.\n\nMotion and registration errors are especially consequential. A patient can move during the emission scan, between SPECT and CT, or with respiration. The CT captures anatomy over a much shorter interval than the emission acquisition, so respiratory positions may differ even without gross movement. Misalignment can create false perfusion defects or incorrect localization. Cardiac gating, based on electrocardiographic timing, separates counts into phases of the cardiac cycle and permits measurements of ventricular function. Irregular rhythms and incorrect beat acceptance can degrade those measurements independently of the underlying perfusion images."
    },
    {
      "heading": "Myocardial perfusion and cardiac applications",
      "text": "Myocardial perfusion imaging is a major clinical application of SPECT. It assesses the regional delivery and retention of tracer in heart muscle under resting conditions and during exercise or pharmacological stress. In coronary artery disease, a narrowed artery may permit adequate resting flow but restrict the increase in flow required during stress. A region with relatively reduced stress uptake that improves at rest is therefore consistent with inducible ischemia, while a persistent defect may reflect scar or another cause of reduced apparent activity.\n\nTechnetium-99m sestamibi and tetrofosmin are widely used perfusion tracers. Their myocardial distributions depend on delivery and cellular retention, but they are not perfectly linear measures of flow across all physiological conditions. Stress can be produced by exercise or by agents that alter coronary blood flow; dobutamine protocols increase cardiac workload through a different mechanism. Stress and rest studies may be performed on one day or separate days, with activities and timing adjusted to limit interference from the earlier injection. Selected protocols omit rest imaging when stress images are unequivocally normal and other criteria are satisfied.\n\nGated perfusion SPECT provides functional information in addition to uptake patterns. Software can estimate left ventricular end-diastolic and end-systolic volumes, ejection fraction, regional wall motion, and wall thickening. These measurements assist interpretation of suspected scar and contribute to prognostic assessment. A fixed uptake defect accompanied by preserved thickening, for example, can support an attenuation explanation, although no single finding settles every case. Comparison with clinical information and other cardiac investigations remains necessary.\n\nRelative perfusion imaging has important limitations. Because images are commonly normalized to the region with the highest uptake, globally reduced flow may be underestimated when multiple coronary territories are similarly affected. This is often described as balanced ischemia. Breast and diaphragmatic attenuation, patient motion, subdiaphragmatic tracer activity, and reconstruction errors can also produce misleading defects. CT attenuation correction, positional imaging, gated findings, and inspection of raw data help distinguish these effects, but can themselves introduce errors if applied incorrectly.\n\nSPECT also contributes to selected non-perfusion cardiac evaluations. Bone-avid tracers can demonstrate myocardial uptake associated with transthyretin cardiac amyloidosis. Tomographic imaging is important for distinguishing true myocardial uptake from residual blood-pool activity or adjacent bone uptake. This use is part of a defined diagnostic pathway that also assesses for a monoclonal protein disorder; tracer uptake alone does not establish the amyloid type. The example illustrates how the same physical imaging platform can answer different questions when paired with different tracers and interpretive criteria."
    },
    {
      "heading": "Brain perfusion and neurochemical imaging",
      "text": "Brain SPECT includes both regional perfusion imaging and imaging of selected molecular targets. Technetium-99m hexamethylpropyleneamine oxime, commonly called HMPAO, and technetium-99m ethyl cysteinate dimer, or ECD, cross the blood–brain barrier and are retained through intracellular processes. Their distributions provide a representation of regional cerebral perfusion around the time of injection. Because acquisition can occur later, the physiological state during injection may matter more than the state during the subsequent scan.\n\nThis temporal property is particularly useful in selected evaluations of epilepsy. When tracer is injected during a seizure, ictal SPECT may show increased perfusion associated with seizure activity. Comparison with an interictal examination can help identify candidate seizure-onset regions during presurgical investigation. Subtraction ictal SPECT co-registered to MRI, often abbreviated SISCOM, maps differences onto anatomical images. Injection delay is a major interpretive variable because seizure activity can propagate rapidly, and the perfusion pattern may represent propagation as well as onset.\n\nPerfusion patterns can also provide supporting information in the evaluation of dementia. Some neurodegenerative disorders produce characteristic regional distributions, but these patterns overlap and are affected by vascular disease, atrophy, medication, and technical factors. SPECT does not directly demonstrate the defining molecular pathology of every dementia syndrome. Its role depends on the clinical question, available alternatives, and local practice. Structural magnetic resonance imaging and other functional or molecular examinations supply different, complementary information rather than interchangeable versions of the same test.\n\nIodine-123 ioflupane imaging targets dopamine transporters in the striatum. Reduced binding can support the presence of presynaptic dopaminergic terminal loss when distinguishing degenerative parkinsonian syndromes from conditions without such loss. It does not reliably distinguish Parkinson’s disease from all other degenerative parkinsonian disorders. The normal pattern, age-related effects, medications that interfere with binding, head positioning, and reconstruction all influence interpretation. Visual assessment may be supplemented by semiquantitative comparisons with reference databases.\n\nCerebral perfusion imaging also has an ancillary role in the determination of brain death under applicable clinical and legal standards. Appropriate tracers and acquisition methods evaluate intracranial perfusion, but technical adequacy and tracer delivery must be established before absent uptake is interpreted. More broadly, brain SPECT is not a general-purpose test that independently establishes psychiatric diagnoses or explains nonspecific cognitive symptoms. An observed regional difference is a physiological finding whose significance depends on validated associations, the clinical setting, and the limitations of the measurement."
    },
    {
      "heading": "Skeletal, endocrine, infectious, and oncological applications",
      "text": "Skeletal SPECT commonly uses technetium-99m-labeled diphosphonates, which accumulate at sites of bone mineral turnover. Tomographic imaging separates overlapping structures that are difficult to distinguish on planar scintigraphy. SPECT/CT can then relate uptake to a vertebral facet joint, fracture line, prosthetic interface, or focal bone lesion. This is especially useful in anatomically complex regions such as the spine, pelvis, feet, and ankles, where several structures may contribute to a single planar focus.\n\nIn patients with cancer, increased bone turnover may indicate skeletal metastases, but uptake is not specific to malignancy. Degenerative change, healing injury, and postoperative remodeling can also be conspicuous. The CT component can reveal sclerosis, destruction, or a benign structural explanation for an abnormal focus. Conversely, some lesions dominated by bone destruction or marrow infiltration produce relatively little osteoblastic response and may be less apparent on conventional bone scintigraphy. The biology of the primary tumor therefore affects the strengths and limitations of the examination.\n\nParathyroid localization is another established application. Technetium-99m sestamibi can be retained in hyperfunctioning parathyroid tissue, with SPECT/CT helping identify its relationship to the thyroid, trachea, and mediastinum. Some protocols compare early and delayed imaging; others use additional tracers or subtraction techniques. The purpose is usually localization after a biochemical diagnosis of hyperparathyroidism, not determining whether the biochemical disorder exists. Thyroid nodules, variable tracer washout, small glands, and multigland disease can complicate interpretation.\n\nIn infection imaging, SPECT/CT may be added to examinations using radiolabeled leukocytes or other tracers. Precise localization can help distinguish uptake in bone from uptake in adjacent soft tissue and can clarify findings around orthopedic hardware. Labeled-leukocyte imaging depends on cell handling, migration, imaging timing, and the clinical setting. Uptake is not a universal direct assay for microorganisms, and physiological marrow activity or altered marrow distribution can create interpretive challenges. The choice of tracer and comparison studies varies with the suspected site and type of infection.\n\nAdditional uses include ventilation–perfusion SPECT for suspected pulmonary embolism, iodine imaging in differentiated thyroid cancer, and selected neuroendocrine tumor examinations. Ventilation and perfusion images are compared to identify segmental perfusion defects without corresponding ventilation abnormalities. In thyroid cancer, SPECT/CT can help distinguish physiological radioiodine activity from nodal or distant disease. For some neuroendocrine indications, receptor-targeted PET has displaced older SPECT approaches where available, but SPECT remains clinically relevant according to tracer access, equipment, and the specific diagnostic or therapeutic pathway."
    },
    {
      "heading": "Radiation exposure, examination conduct, and quality assurance",
      "text": "Radiation exposure from SPECT comes primarily from the administered radiopharmaceutical and, in hybrid examinations, from CT. Dose cannot be inferred from scan duration alone: a patient remains a source of radiation after acquisition ends, while the retained activity declines through decay and biological clearance. Administered activity, radionuclide emissions, organ distribution, and residence time all matter. Effective dose, expressed in sieverts and commonly reported in millisieverts for diagnostic imaging, is a protection quantity based on reference models rather than a precise prediction of risk for an individual patient.\n\nMany common adult SPECT examinations produce effective doses of several millisieverts, while some protocols are lower and others exceed 10 mSv. The CT contribution ranges from a relatively small localization exposure to that of a diagnostic CT examination. A single universal dose figure is therefore misleading. Protocol optimization involves selecting an appropriate tracer activity, acquisition duration, detector configuration, and CT technique while retaining information adequate for the intended clinical question. Pediatric protocols require particular attention to body size and diagnostic requirements.\n\nPregnancy and breastfeeding can affect the selection and conduct of nuclear medicine examinations because radionuclides may reach the fetus or be excreted in breast milk. The relevant considerations are radiopharmaceutical-specific; they cannot be reduced to one rule for all SPECT studies. Adverse reactions to diagnostic radiopharmaceuticals are uncommon, but the examination may involve other agents with separate risks, including pharmacological stress drugs or iodinated CT contrast. Radiopharmaceutical tracers are not equivalent to conventional CT contrast agents, despite both being administered to support imaging.\n\nTechnical quality assurance begins before patient imaging. Gamma-camera checks include energy calibration, detector uniformity, and center-of-rotation assessment. Defects in detector response can generate ring or other reconstruction artifacts. Hybrid systems also require checks of SPECT–CT alignment. Radiopharmaceutical preparation involves identity, activity measurement, radiochemical quality, and contamination control. Incorrect tracer, activity, injection route, or patient identification can invalidate an otherwise technically competent scan, making these processes integral to patient safety.\n\nInterpretation requires review of more than reconstructed slices. Projection images can disclose motion, injection-site retention, unexpected biodistribution, and nearby intense activity. Reconstruction settings and attenuation maps need evaluation when a suspicious finding could be technical. A report distinguishes the observed uptake pattern from its diagnostic interpretation and relates both to the examination’s indication. Negative findings are bounded by tracer biology and achievable resolution: absence of visible uptake does not establish the absence of every form of disease in the imaged region."
    },
    {
      "heading": "Quantification, emerging methods, and comparative limitations",
      "text": "Historically, many SPECT studies were interpreted visually or by relative uptake ratios. Quantitative SPECT seeks to express reconstructed activity in calibrated physical units, such as becquerels per milliliter. This requires a chain of corrections for attenuation, scatter, detector sensitivity, decay, and other effects. A camera calibration factor relates reconstructed counts to known activity under defined acquisition and reconstruction conditions. Cross-calibration with the instrument used to measure administered activity is essential when absolute values are intended to be comparable.\n\nSmall structures present a particular difficulty through the partial-volume effect. Because finite spatial resolution spreads signal across neighboring voxels, a small high-activity lesion may appear larger and less concentrated than it is. Adjacent background activity also influences the measured value. Recovery corrections can reduce bias in controlled circumstances, but depend on lesion size, shape, contrast, and reconstruction. Consequently, a displayed numerical uptake value can be highly precise in its formatting without being equally accurate biologically or physically.\n\nQuantitative imaging has an important connection to radionuclide therapy. Some therapeutic radionuclides emit photons that permit SPECT measurements after treatment. Serial imaging can estimate activity retention in organs and tumors, supporting patient-specific absorbed-dose calculations. This differs from diagnostic effective dose: absorbed dose, measured in grays, describes energy deposited per unit mass in the tissue of interest. Dosimetry additionally requires assumptions about time–activity curves, tissue masses, and energy transport, so image quantification is one component of a larger measurement process.\n\nNew detector designs, multi-pinhole geometries, improved reconstruction, and motion correction aim to increase sensitivity or resolution and reduce acquisition time or administered activity. Dynamic acquisitions on suitable systems can support kinetic measurements, including research and selected clinical approaches to myocardial blood flow. Machine-learning methods are being investigated for denoising, reconstruction, segmentation, and correction tasks. Their performance depends on training data and validation: a visually smooth image can still suppress a real lesion or introduce an apparently plausible feature.\n\nSPECT’s comparative strengths include access to a broad range of tracers, generator-based supply for technetium-99m, and established clinical infrastructure. Its physical collimators reject most emitted photons, limiting sensitivity relative to many PET systems, while photon attenuation and scatter remain substantial correction problems. PET, MRI, CT, and ultrasonography answer partly different questions, and superiority cannot be assigned independently of the indication. Continued evaluation of new SPECT methods therefore concerns not only image sharpness or numerical agreement, but also reproducibility, diagnostic accuracy, effects on clinical decisions, and performance across different patients and institutions."
    }
  ],
  "links": [
    "nuclear-medicine",
    "x-ray-computed-tomography",
    "positron-emission-tomography",
    "photon",
    "pharmacokinetics",
    "regularization",
    "coronary-artery-disease",
    "prognosis",
    "blood-brain-barrier",
    "epilepsy",
    "dementia",
    "magnetic-resonance-imaging",
    "parkinsons-disease",
    "brain-death",
    "cancer",
    "metastasis",
    "pregnancy",
    "contrast-agent",
    "patient-safety",
    "radionuclide-therapy",
    "machine-learning",
    "ultrasonography"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}