Nuclear medicine is a medical specialty that uses small quantities of radioactive substances to investigate biological processes and, in selected diseases, to deliver treatment. Its principal agents, called radiopharmaceuticals, combine a radioactive isotope with a chemical form that determines where the substance travels and accumulates. Radiation emitted from within the body can be detected to produce images, measure organ function, or estimate the distribution of a therapeutic agent.
Unlike imaging methods that primarily depict anatomy, nuclear medicine often measures aspects of physiology and metabolism, including tissue perfusion, glucose utilization, receptor expression, and excretion. It therefore complements computed tomography (CT), magnetic resonance imaging (MRI), and ultrasonography. Its therapeutic applications include radioiodine treatment of thyroid disease and radioligand treatment of selected cancers. The field brings together nuclear physics, radiochemistry, instrumentation, clinical interpretation, radiation protection, and quantitative measurement.
Biological tracers and the scope of the specialty
The central principle of diagnostic nuclear medicine is the tracer method: a detectable substance is introduced into a biological system, and its movement or transformation provides information about that system. A radiotracer may follow a naturally occurring chemical element, resemble a physiological substrate, or bind to a particular molecular target. The administered chemical mass is often very small, allowing measurement without substantially altering the process under investigation. This assumption is not automatic, however; receptor-binding agents can change their own distribution if enough nonradioactive material is administered to occupy a significant fraction of available binding sites.
Different tracers answer different questions. Radioiodide enters thyroid cells through the sodium–iodide symporter and can subsequently be incorporated into thyroid compounds. Technetium-99m–labeled diphosphonates bind to bone mineral, particularly at sites of increased remodeling. Fluorine-18 fluorodeoxyglucose, usually abbreviated FDG, enters cells through glucose transporters and is phosphorylated by hexokinase. Because its further metabolism is limited, intracellular accumulation provides an indirect measure of glucose utilization over the uptake period. None of these mechanisms is inherently specific to malignancy: bone repair, infection, and activated inflammatory cells may also produce increased tracer uptake.
Diagnostic investigations include external imaging, nonimaging measurements, and some laboratory procedures. A thyroid uptake test measures the fraction of administered radioiodine present in the thyroid at a specified time. Renal investigations can estimate differential kidney function or clearance. Historically, radioimmunoassay and other radioisotope-based laboratory tests were also important components of the specialty, although many have been replaced by nonradioactive analytical methods. The boundary between nuclear medicine and radiology varies between health systems, as do training pathways and responsibility for hybrid imaging.
Therapeutic nuclear medicine uses radionuclides that deposit enough energy in target tissues to damage cells. It differs from external-beam radiation therapy because the radiation source is distributed within the patient rather than delivered from an external machine. Distribution depends on pharmacokinetics, target expression, blood supply, tissue permeability, and elimination. Some treatments are systemic, whereas others are delivered regionally, such as radioactive microspheres introduced into the hepatic arterial circulation.
A nuclear medicine image is therefore a map of a particular biological interaction, not a direct photograph of disease. Interpretation depends on the tracer, acquisition timing, physiological background, medication effects, and clinical question. Abnormal uptake may identify a process that requires further characterization, while absent uptake may reflect either absent disease or a biological property that makes the disease invisible to that tracer.
Historical development
The foundations of nuclear medicine emerged from discoveries in radioactivity and the realization that radioactive atoms could serve as chemical labels. Henri Becquerel discovered spontaneous radioactivity in 1896, and Marie and Pierre Curie identified polonium and radium in 1898. These discoveries initially supported both physical research and the therapeutic use of radioactive materials, but the biological tracer method required the additional recognition that radioactive and stable forms of an element could participate in the same chemical processes.
George de Hevesy developed this approach during the early twentieth century. His studies used radioactive substances to follow chemical and biological movements that were otherwise difficult to measure. He received the 1943 Nobel Prize in Chemistry for his work on isotopes as tracers in the study of chemical processes. The development of particle accelerators and the discovery of artificial radioactivity by Irène and Frédéric Joliot-Curie in 1934 subsequently expanded the range of radionuclides available for investigation.
Clinical applications grew during the 1930s and 1940s. John Lawrence used phosphorus-32 therapeutically in hematological disease in 1936. Saul Hertz and colleagues investigated radioiodine uptake by the thyroid, and Hertz and Arthur Roberts administered radioiodine to treat hyperthyroidism in 1941. After World War II, reactor-produced isotopes became more widely available, making radioisotope research and clinical services possible beyond the small number of institutions that possessed specialized accelerators.
Imaging developed from point-by-point counting to spatially resolved detection. Benedict Cassen developed the rectilinear scanner around 1950, allowing a detector to traverse the body and build an image of tracer distribution. Hal Anger’s scintillation camera, introduced in the 1950s, detected radiation across a broad field simultaneously. Technetium-99m became increasingly important during the following decades because its physical properties and generator-based availability suited clinical imaging. Tomographic methods later reconstructed cross-sectional distributions from multiple projections.
Positron imaging developed through several generations of coincidence detectors and reconstruction systems. Clinical PET expanded substantially with FDG and, from the late 1990s onward, integrated PET/CT. SPECT/CT and PET/MRI further strengthened the connection between molecular information and anatomical localization. Therapeutic development followed a parallel trajectory: radioiodine remained a major treatment, while radiolabeled antibodies, peptides, and small molecules created new ways to deliver radiation selectively. The field’s history is consequently not a simple replacement of older techniques by newer ones; long-established thyroid and renal procedures remain useful alongside increasingly specialized molecular imaging and therapy.
Radioactive decay, radiation emissions, and dose
A radionuclide is an unstable form of an atom whose nucleus transforms spontaneously. Different isotopes of an element have the same number of protons but different numbers of neutrons. Radioactive decay is probabilistic for an individual nucleus but predictable for a large population. If is the initial number of radioactive atoms, the number remaining after time is
where is the decay constant. The physical half-life is . Activity, measured in becquerels, is the number of nuclear transformations per second; one megabecquerel represents one million transformations per second.
The radiation emitted determines how a radionuclide can be used. Technetium-99m emits a principal gamma photon of approximately 140 kiloelectronvolts and has a physical half-life of about six hours. Fluorine-18 has a half-life of approximately 110 minutes and decays by positron emission. After traveling a short distance, a positron annihilates with an electron, usually producing two photons of 511 kiloelectronvolts traveling in nearly opposite directions. Detection of these paired photons underlies positron emission tomography (PET).
Therapeutic radionuclides commonly emit beta particles, alpha particles, or electrons produced in atomic rearrangements. Beta particles distribute energy over distances ranging from fractions of a millimeter to several millimeters or more, depending on their energy and the tissue. This can irradiate neighboring tumor cells through a crossfire effect. Alpha particles travel only tens of micrometers but deposit energy densely along their tracks. Their effectiveness depends on microscopic distribution: delivering activity near the relevant cells is crucial, and uptake averaged across an organ may not describe cellular exposure adequately.
Physical decay and biological elimination both determine how long activity remains in a tissue. For a simplified single-compartment model with exponential clearance, the effective half-life satisfies
Actual uptake and clearance often require several compartments or measured time–activity curves rather than one exponential. A tracer may accumulate slowly, redistribute, or produce radioactive metabolites with different destinations.
Activity is not the same as radiation dose. Absorbed dose is energy deposited per unit mass and is measured in grays, where . Radiation dosimetry combines activity over time with emission energies and models of energy deposition. Effective dose, expressed in sieverts, is a radiation-protection quantity based on reference tissue weighting; it is not a direct measurement of an individual patient’s probability of injury.
Radiopharmaceutical production and quality control
A radiopharmaceutical must satisfy two linked requirements: its radioactive component must have appropriate decay properties, and its chemical component must produce a useful biological distribution. Some agents are chemically simple, such as sodium iodide containing iodine-123 or iodine-131. Others combine a radionuclide, a chelator, a linker, and a targeting molecule. A chelator binds a radioactive metal, while the targeting component may be a peptide, small molecule, or monoclonal antibody. Chemical stability matters because radionuclide release can redirect activity to unintended organs.
Radionuclides are produced mainly in nuclear reactors, particle accelerators, or generator systems. Reactors support neutron activation and production from nuclear fission. A cyclotron accelerates charged particles into a target material, producing radionuclides through nuclear reactions. Fluorine-18 is commonly produced by irradiating oxygen-18–enriched water with protons. Carbon-11, nitrogen-13, and oxygen-15 have substantially shorter half-lives, making nearby production especially important for their clinical and research use.
A radionuclide generator separates a useful daughter radionuclide from a longer-lived parent. The molybdenum-99/technetium-99m generator has been central to routine practice: technetium-99m is eluted and used to label a range of compounds. Germanium-68/gallium-68 generators provide another example, although gallium-68 can also be produced in cyclotrons. Generator operation requires control of parent breakthrough, chemical contaminants, and the interval between elutions, which affects the available daughter activity.
Radiochemistry must proceed within a limited time because activity is continuously decaying. Production commonly uses shielded enclosures, automated synthesis modules, and validated procedures that reduce contamination and occupational exposure. Relevant quality attributes include identity, radiochemical purity, radionuclidic purity, chemical purity, pH, bacterial endotoxin content, and sterility. Radiochemical purity describes the fraction of activity in the intended chemical form; radionuclidic purity describes the fraction attributable to the intended radionuclide. These are different properties and require different tests.
For short-lived preparations, some microbiological results become available only after administration, so manufacturing controls, validated aseptic methods, and release procedures are particularly important. The amount of radioactive material also changes continuously, requiring activity measurements to be referenced to a specified calibration time. Molar activity—the activity per amount of substance—can influence receptor occupancy and image quality. Reliable supply therefore depends not just on isotope production but on chemistry, quality assurance, transport scheduling, and coordination with patient appointments. A delay of several hours may be manageable for one radionuclide and make another preparation unusable.
Gamma cameras, SPECT, PET, and hybrid imaging
The conventional gamma camera detects gamma photons with a large scintillation crystal, historically most often sodium iodide activated with thallium. An absorbed photon produces light, and photodetectors estimate its energy and interaction position. A lead collimator in front of the detector restricts the directions from which photons can arrive. This provides spatial information but rejects most emitted photons, creating a fundamental compromise between sensitivity and spatial resolution. Collimator design must also match photon energy: inadequate septal thickness allows high-energy photons to penetrate and degrade the image.
Planar scintigraphy records two-dimensional projections. Single-photon emission computed tomography (SPECT) acquires projections around the patient and reconstructs a three-dimensional activity distribution. It improves localization and separates structures that overlap on planar images. Detector systems using cadmium zinc telluride can convert radiation directly into electrical signals and are used in some dedicated cardiac and general-purpose instruments. Their performance depends on detector configuration, collimation, and the imaging task rather than on semiconductor material alone.
PET uses electronic coincidence detection rather than a mechanical collimator to identify photon pairs consistent with a positron annihilation event. A detected pair defines a line along which the event probably occurred. Time-of-flight PET measures the small difference between the photons’ arrival times, narrowing the probable event location along that line. Modern systems commonly use lutetium-based scintillators and silicon photomultipliers. Timing resolution, detector dimensions, sensitivity, and count-rate performance all affect image quality.
Hybrid systems acquire molecular and anatomical information in a coordinated examination. PET/CT and SPECT/CT use CT for anatomical localization and attenuation correction. The CT component may be a low-dose localization scan or a fully diagnostic examination, depending on its purpose. PET/MRI combines PET with MRI’s soft-tissue contrast and specialized sequences, but requires different attenuation-correction methods because MRI signal does not directly describe photon attenuation. Metal implants, truncated anatomy, and tissue classification errors can affect the resulting PET measurements.
Instrument performance places limits on what can be observed. Clinical spatial resolution is generally several millimeters and depends on the system, tracer, reconstruction, and patient motion. Small lesions may appear less active than they truly are because their signal is spread into surrounding tissue. High count rates can cause event losses, while low counts produce statistical noise. Longer acquisition or more administered activity can improve some aspects of image quality, but neither eliminates every limitation. Equipment calibration, correction methods, and acquisition design are therefore integral to the biological validity of the image.
Quantification, image interpretation, and sources of error
A nuclear medicine image contains measurement data, but quantitative interpretation requires corrections for processes that separate detected counts from the activity actually present in tissue. Photons may be absorbed before reaching the detector, scattered into an incorrect direction, or lost because electronics cannot process closely spaced events. PET also detects random coincidences between unrelated photons. Corrections for these effects, together with detector normalization and accurate activity calibration, support quantitative imaging. Errors in any part of the measurement chain can create apparent biological differences.
A widely used PET measurement is the standardized uptake value, or SUV. In a common body-weight normalization,
with activity quantities corrected to a consistent time reference. SUV is useful because it provides a practical summary of uptake, but it is not an intrinsic property of a tumor. It varies with uptake time, blood glucose, body composition, reconstruction, scanner calibration, lesion size, and how the region of interest is defined. Lean-body-mass normalization reduces some effects of body composition.
Dynamic imaging can characterize tracer delivery and exchange more directly. Serial images describe tissue activity over time, sometimes combined with blood sampling or an image-derived arterial input function. Compartmental models represent processes such as transport from plasma, binding, phosphorylation, and release. For suitable tracers, these models estimate quantities such as metabolic rate or distribution volume. Graphical methods, including Patlak and Logan analysis, simplify particular classes of kinetic behavior but depend on assumptions that must be appropriate to the tracer and acquisition.
Interpretation also depends on pattern recognition and clinical context. Physiological activity in the brain, myocardium, urinary tract, bowel, or salivary glands can resemble or conceal disease. Injection-site leakage may reduce the effective administered activity and produce unusual lymphatic uptake. Respiratory motion can misalign PET and CT, especially near the diaphragm. Metallic implants and dense contrast material may introduce attenuation-correction artifacts. Reviewing non–attenuation-corrected images and the anatomical component can help identify such discrepancies.
Diagnostic performance is described by sensitivity and specificity, but neither is fixed across all patients and settings. Lesion size, disease subtype, treatment, and the reference standard influence apparent accuracy. Positive and negative predictive values also depend on disease frequency in the tested population. Standardized acquisition protocols and response criteria improve comparability, yet a numerical threshold rarely replaces interpretation. Quantitative change is most persuasive when differences in preparation, timing, scanner performance, and analysis have been controlled.
Cancer imaging and treatment-response assessment
Oncology is a major area of nuclear medicine because tumors can differ from surrounding tissues in metabolism, transport, receptor expression, and cellular turnover. FDG PET/CT is widely used in selected settings for staging, identifying suspected recurrence, defining radiation-treatment volumes, and assessing response. It is especially established in many forms of lymphoma and in several indications involving lung cancer, head and neck cancer, and other solid tumors. Its usefulness is disease-specific rather than universal.
FDG accumulation reflects glucose transport and phosphorylation, not the presence of cancer itself. Activated immune cells can be strongly FDG-avid, producing uptake in inflammation, infection, granulomatous disease, and healing tissue. Conversely, some low-grade tumors and certain well-differentiated malignancies have relatively low uptake. Small lesions may be missed because of spatial resolution, and intense normal activity can reduce conspicuity. Interpretation after surgery, radiation treatment, or systemic therapy must account for treatment-related changes and the timing of the examination.
Target-specific imaging extends beyond FDG. Radiolabeled somatostatin analogues identify many well-differentiated neuroendocrine tumors through their somatostatin receptors. Ligands targeting prostate-specific membrane antigen (PSMA) have become important in selected stages of prostate cancer, including investigation of biochemical recurrence and staging of higher-risk disease. PSMA expression is not exclusive to prostate cancer, and benign conditions or other tumors may show uptake. Other specialized tracers investigate amino-acid transport, estrogen receptors, or catecholamine-handling pathways.
Imaging can reveal disease throughout the body and thereby change stage or identify a site for biopsy, but it does not generally replace histopathology for establishing tumor identity. Different lesions within one patient may have different biological characteristics. A target-positive scan can demonstrate receptor expression in some deposits while failing to represent target-negative components. This heterogeneity matters when imaging is used to select patients for targeted radionuclide treatment.
Response assessment uses both anatomical and functional information. In lymphoma, the Deauville five-point scale compares FDG uptake with reference tissues, particularly mediastinal blood pool and liver. Other frameworks evaluate metabolic change quantitatively or combine it with changes in tumor dimensions. A decrease in uptake may precede shrinkage, while residual masses may persist after active tumor has been eliminated. Nevertheless, an imaging response is a surrogate endpoint in some settings rather than direct proof of improved survival or symptoms. Its clinical meaning depends on validation within the particular disease and treatment pathway.
Cardiovascular and pulmonary applications
Myocardial perfusion imaging evaluates how a radiotracer is delivered to heart muscle at rest and during increased coronary flow. SPECT commonly uses technetium-99m sestamibi or tetrofosmin, while PET perfusion tracers include rubidium-82 and nitrogen-13 ammonia. Exercise or pharmacological stress creates conditions in which a flow-limiting coronary lesion may produce a regional perfusion difference. A reversible defect suggests inducible ischemia, whereas a persistent defect may reflect scar, although artifacts and other explanations must be considered.
Electrocardiographically gated acquisition relates images to the cardiac cycle and estimates ventricular volumes, wall motion, and ejection fraction. These measurements complement the perfusion pattern. Attenuation from the diaphragm or breast tissue can produce apparent defects, and patient motion can distort regional uptake. Attenuation correction, alternative imaging positions, and review of gated function help distinguish artifacts from disease. Nuclear perfusion imaging assesses the physiological consequences of coronary artery disease rather than directly measuring the anatomical degree of arterial narrowing.
PET can quantify myocardial blood flow and myocardial flow reserve, the ratio of stress flow to resting flow. These measurements may reveal diffuse vascular disease or coronary microvascular dysfunction even when regional perfusion appears relatively uniform. Conversely, relative perfusion methods can underestimate the extent of disease when blood flow is reduced in all major coronary territories, a situation often described as balanced ischemia. Flow measurements require tracer-specific models, reliable timing, and attention to hemodynamic and technical factors.
Other cardiac applications exploit different biological processes. FDG imaging can identify viable myocardium with preserved metabolism despite reduced perfusion and can contribute to the assessment of selected inflammatory cardiac disorders. For suspected cardiac sarcoidosis, normal myocardial glucose utilization must be suppressed because physiological uptake can obscure inflammation. Bone-seeking tracers can show characteristic cardiac uptake in transthyretin amyloid cardiomyopathy. In the appropriate diagnostic pathway, strong uptake combined with exclusion of a monoclonal protein disorder can establish transthyretin cardiac amyloidosis without cardiac biopsy; tracer uptake alone is insufficient.
Ventilation–perfusion lung imaging investigates the regional relationship between air delivery and pulmonary blood flow. In pulmonary embolism, an obstructed pulmonary artery may produce a perfusion defect in a region with preserved ventilation. Ventilation can be imaged with radioactive gases or aerosols, while perfusion commonly uses technetium-99m macroaggregated albumin. These particles transiently lodge in a small fraction of pulmonary capillaries, mapping blood flow. Planar and SPECT approaches are used, sometimes with CT, and interpretation incorporates defect shape, anatomical findings, and the clinical probability of embolism.
Brain imaging and neurological disease
Nuclear medicine investigates cerebral metabolism, perfusion, neurotransmitter systems, and pathological protein accumulation. Normal brain tissue has high glucose utilization, so FDG PET produces a characteristic regional pattern. Neurological interpretation often focuses on relative reductions in uptake rather than a single focal “hot spot.” Changes in neuronal activity and synaptic function can alter metabolism before advanced structural loss becomes apparent, but metabolic abnormalities are not uniquely diagnostic of one disease.
In cognitive disorders, FDG PET can support the differential diagnosis of selected causes of dementia. Alzheimer-type disease commonly involves posterior temporoparietal and posterior cingulate regions, whereas frontotemporal degeneration more often affects frontal and anterior temporal networks. Patterns overlap and vary with disease stage. Imaging must be interpreted alongside symptoms, neuropsychological findings, structural imaging, and other laboratory or fluid biomarkers. It is not a stand-alone assessment of cognitive capacity.
Amyloid PET detects fibrillar amyloid-beta deposits using specific fluorine-18–labeled tracers. A positive scan establishes substantial amyloid pathology but does not by itself prove that amyloid is the sole cause of a person’s symptoms; amyloid can also occur in cognitively unimpaired older adults. Tau PET investigates a different pathological process and has tracer-specific binding characteristics and limitations. These methods illustrate the distinction between identifying a molecular lesion and diagnosing a complete clinical syndrome. Their use in treatment eligibility and research depends on the relevant clinical and regulatory framework.
Dopamine-transporter SPECT, commonly performed with iodine-123 ioflupane, assesses presynaptic dopaminergic terminals in the striatum. It can help distinguish disorders with nigrostriatal dopaminergic loss from some conditions without that deficit. It does not reliably separate Parkinson’s disease from all atypical degenerative parkinsonian syndromes, because several share reduced transporter binding. Medications that interfere with tracer binding and technical factors can also affect interpretation.
In epilepsy, interictal FDG PET may identify a region of reduced metabolism associated with the epileptogenic network. Ictal perfusion SPECT instead depends on rapid tracer injection during a seizure, capturing the perfusion pattern around that time. These tests can contribute to presurgical evaluation when considered with electroencephalography, MRI, and clinical seizure characteristics. Cerebral perfusion studies also have specialized roles in assessing absent intracranial perfusion within established brain-death protocols. In research, receptor ligands and kinetic PET methods investigate neurotransmission and drug occupancy, but many such measurements require specialized facilities and are not routine clinical examinations.
Bone, renal, hepatobiliary, and infection imaging
Bone scintigraphy uses radiolabeled diphosphonates to map mineral turnover and local blood supply. It can identify multiple sites of increased osteoblastic activity across the skeleton in one examination. Applications include selected evaluations of skeletal metastases, occult fractures, and bone or prosthetic-joint complications. Because uptake reflects the reaction of bone rather than direct visualization of tumor cells, purely lytic lesions or marrow-confined disease may be less conspicuous. Degenerative change, trauma, and healing can also cause increased uptake.
Renal scintigraphy addresses function and drainage. Technetium-99m MAG3 is predominantly handled by tubular secretion and is useful for dynamic assessment, including in patients with reduced renal function. Technetium-99m DTPA is filtered at the glomerulus and can support filtration measurements under appropriate protocols. Technetium-99m DMSA accumulates in renal cortex and is used for cortical imaging. Dynamic studies can estimate the relative contribution of each kidney and assess drainage after a diuretic, but delayed emptying does not automatically establish clinically significant obstruction. Hydration, bladder fullness, renal impairment, and collecting-system size influence the result.
Hepatobiliary scintigraphy commonly uses technetium-99m–labeled iminodiacetic acid derivatives, which are taken up by hepatocytes and excreted into bile. It can assess biliary transit, investigate bile leaks, and support evaluation of acute cholecystitis. Nonvisualization of the gallbladder has to be interpreted in relation to preparation, liver function, and the imaging protocol. Gastric emptying scintigraphy measures the movement of a standardized radiolabeled meal from the stomach; meal composition and acquisition duration are essential to comparability. Technetium-99m pertechnetate imaging can detect functioning ectopic gastric mucosa in selected cases of suspected Meckel diverticulum.
Infection and inflammation imaging use several strategies. FDG accumulates in many activated inflammatory cells and can help map certain deep or multifocal processes. Alternatively, a patient’s leukocytes can be separated, radiolabeled, and reinjected to follow cell migration. This approach requires careful blood handling and identification procedures. Tracer selection and interpretation depend on the suspected process, its location, prior treatment, and the normal distribution of labeled cells.
Lymphatic mapping introduces a small amount of tracer near a lesion or anatomical site and follows drainage to regional lymph nodes. Sentinel-node localization supports selected surgical pathways, notably in breast cancer and melanoma. Lymphoscintigraphy can also characterize impaired lymph drainage in lymphedema. These varied examinations share a functional basis, but their protocols are not interchangeable: each requires a tracer whose biological handling matches the particular question.
Radioiodine and established radionuclide treatments
Radioiodine is one of the longest-established examples of biologically targeted radiation treatment. Thyroid follicular cells concentrate iodide through the sodium–iodide symporter, allowing iodine-131 to deliver radiation preferentially to iodine-avid thyroid tissue. Iodine-131 has a physical half-life of approximately eight days and emits both beta particles, which provide most of the therapeutic effect, and gamma photons, which permit imaging and contribute to external exposure. The same uptake mechanism supports treatment of both benign thyroid hyperfunction and selected differentiated thyroid cancers, although the clinical goals and administered activities differ.
In hyperthyroidism, radioiodine reduces functioning thyroid tissue. It is used in conditions including Graves’ disease and toxic nodular thyroid disease. The eventual effect develops over weeks to months as irradiated cells lose function. Permanent hypothyroidism is common, particularly after treatment of Graves’ disease, and is an expected outcome in many treatment strategies rather than an immediate measurement of treatment failure. Eye disease associated with Graves’ disease requires separate consideration because radioiodine can influence its course in some clinical circumstances.
In differentiated thyroid cancer, radioiodine may be used after surgery to ablate remnant tissue, treat suspected microscopic disease, or treat known iodine-avid lesions. These are distinct purposes, and not every patient with thyroid cancer receives radioiodine. Uptake generally requires sufficient stimulation by thyroid-stimulating hormone, achieved through thyroid-hormone withdrawal or recombinant human TSH in appropriate settings. Iodine exposure from medications or contrast agents can reduce uptake. Some cancers lose iodine-concentrating ability during progression and consequently do not respond adequately despite the presence of tumor.
Other established systemic radionuclide treatments exploit skeletal targeting. Strontium-89 and samarium-153–labeled compounds have been used to palliate pain from selected osteoblastic bone metastases. Radium-223 behaves as a calcium analogue and emits alpha particles near sites of active bone formation. Its clinical role includes selected patients with metastatic castration-resistant prostate cancer involving bone. Different bone-targeted agents have different evidence, eligibility conditions, marrow effects, and effects on clinical outcomes; they should not be treated as interchangeable formulations.
Regional radionuclide treatment includes hepatic radioembolization with yttrium-90 microspheres for selected liver tumors. Microspheres lodge in the tumor-associated arterial microcirculation and deliver beta radiation locally. Treatment planning evaluates arterial anatomy, possible extrahepatic deposition, and lung shunting. Outcomes depend on tumor characteristics, liver reserve, microsphere distribution, and absorbed dose. This procedure combines interventional radiology and nuclear medicine and differs fundamentally from a freely circulating receptor-binding radiopharmaceutical.
Targeted radioligand therapy and theranostics
Theranostics links diagnostic characterization with a related targeted treatment. In nuclear medicine, an imaging agent demonstrates whether a biological target is present and accessible, while a corresponding therapeutic agent carries a radionuclide that deposits cytotoxic energy. The diagnostic and therapeutic agents may use the same targeting molecule with different radionuclides, or closely related molecules with similar biological behavior. Imaging can support patient selection, map disease, identify important target-negative lesions, and sometimes estimate treatment distribution.
Peptide receptor radionuclide therapy illustrates this approach. Many well-differentiated neuroendocrine tumors express somatostatin receptors and can be imaged with radiolabeled somatostatin analogues. Lutetium-177 dotatate binds these receptors and is internalized, delivering beta radiation. Lutetium-177 has a physical half-life of about 6.6 days and emits gamma photons that can support post-treatment imaging. The NETTER-1 randomized trial established a progression-free-survival benefit for lutetium-177 dotatate in a defined population with progressive, somatostatin-receptor-positive midgut neuroendocrine tumors.
PSMA-targeted therapy uses small-molecule ligands to deliver radioactivity to PSMA-expressing disease. Lutetium-177 vipivotide tetraxetan, also known as lutetium-177–PSMA-617, demonstrated improved outcomes in the VISION trial in a defined population with previously treated metastatic castration-resistant prostate cancer. Imaging eligibility is important because expression varies between patients and lesions. Uptake also occurs in normal organs, particularly salivary and lacrimal glands and kidneys, creating potential limits to the therapeutic window.
Treatment toxicity reflects both target biology and clearance. Bone marrow exposure can cause cytopenias, while renal exposure is relevant for several peptide-based agents. Salivary dysfunction can occur with PSMA-directed treatments. Amino-acid infusions used with some peptide therapies reduce renal tubular reabsorption of the radiolabeled peptide and thereby lower kidney exposure. Longer-term risks, including therapy-related myeloid neoplasms in some settings, require follow-up and careful interpretation because many patients have also received other marrow-damaging treatments.
Alpha-emitting therapies, including actinium-225–labeled compounds, are an active area of development. Their high linear energy transfer may overcome some limitations of beta irradiation, but short particle range, daughter-radionuclide redistribution, supply constraints, and organ toxicity complicate translation. A positive target scan alone does not guarantee benefit: insufficient absorbed dose, heterogeneous expression, poor penetration, resistant cell populations, and competing normal-organ exposure can all limit response. Precision medicine in this context depends on biological selection, dosimetry, clinical evidence, and practical treatment delivery rather than target expression alone.
Radiation protection and patient safety
Radiation protection in nuclear medicine addresses the patient, workers, accompanying persons, and the environment. Diagnostic procedures use radiation to obtain information that may improve management, while therapy deliberately delivers a biologically significant dose. Justification and optimization are therefore distinct tasks: an examination must address a meaningful clinical question, and its protocol must provide adequate information without unnecessary exposure. Patient medical exposures are not managed through the same fixed dose limits used for occupational and public exposure, because a strict limit could prevent a clinically justified procedure.
Diagnostic doses vary substantially by radiopharmaceutical, administered activity, patient size, and elimination. Hybrid imaging adds a separate contribution from CT, which may be small for attenuation correction or larger for a diagnostic acquisition. A single generic dose value cannot represent all nuclear medicine examinations. Children require particular attention to body-size-adjusted activity, equipment sensitivity, and acquisition duration. Lower activity is not automatically safer overall if insufficient image quality leads to repeat imaging or a missed clinically important finding.
Pregnancy and breastfeeding create tracer-specific issues. Radionuclides may cross the placenta, and fetal dose depends on maternal distribution, gestational age, and the tracer’s biological behavior. Radioiodine is especially important because the fetal thyroid can concentrate iodine after it becomes functional; therapeutic iodine-131 is contraindicated during pregnancy. Some radiopharmaceuticals enter breast milk, requiring substance-specific interruption or cessation arrangements. These decisions cannot be generalized from one isotope to every chemical preparation containing that isotope.
Occupational protection relies on time, distance, shielding, contamination control, and secure handling. Shielding must match the emission: the penetrating 511-kiloelectronvolt photons associated with PET require different arrangements from lower-energy gamma emitters. Syringe shields, automated dispensing, remote handling, personal dosimeters, and shielded treatment rooms reduce exposure. Monitoring distinguishes external irradiation from contamination on surfaces or skin. For high-energy beta emitters, shielding design also considers bremsstrahlung generated when electrons are slowed in matter.
Patients retain radioactive material after administration, and therapeutic patients may require temporary restrictions on close contact and arrangements for radioactive excreta. Release criteria and waste rules vary by jurisdiction and treatment. Additional safety risks are not radiological: identification errors, incorrect radiopharmaceutical selection, dosing mistakes, intravenous extravasation, falls, and contrast-related complications can also occur. Patient safety therefore includes verification of the agent, activity, route, clinical indication, and patient identity, alongside appropriate emergency procedures and documentation.
Clinical organization, evidence, and emerging methods
A nuclear medicine service depends on coordinated work by physicians, technologists, radiopharmacists or radiochemists, medical physicists, nurses, and administrative staff. Therapeutic programs often involve oncologists, endocrinologists, surgeons, interventional radiologists, and radiation oncologists. Responsibilities differ between jurisdictions, but all services require reliable activity measurement, equipment quality control, validated preparation procedures, and systems for reporting and follow-up. Hybrid imaging also requires competence in the anatomical findings that may fall outside the original indication.
Supply chains impose distinctive constraints. Short physical half-lives make delays clinically and economically consequential, while some radionuclides depend on a limited number of production facilities. Disruptions in molybdenum-99 supply have periodically affected technetium-99m services. Expanding radioligand therapy increases demand not only for isotopes but also for treatment rooms, waste handling, trained staff, and follow-up capacity. Access is uneven between and within countries, reflecting differences in infrastructure, reimbursement, workforce, and transport networks. These factors influence the place of nuclear medicine within a health system.
Evidence for a diagnostic tracer involves more than demonstrating attractive images or high lesion detection rates. Studies must address reproducibility, diagnostic accuracy, changes in management, and, where feasible, patient outcomes. A more sensitive test can alter apparent stage without necessarily improving survival, a phenomenon related to stage migration. Therapeutic agents require evaluation of meaningful endpoints such as survival, symptoms, quality of life, and toxicity. Randomized controlled trials provide strong evidence for specific treatment questions, but their eligibility criteria also define the populations to which results most directly apply.
Technical development includes longer-axial-field-of-view PET systems, which detect activity across a larger portion of the body simultaneously. Their increased sensitivity can support shorter scans, reduced administered activity, imaging at later times, or dynamic measurement of multiple organs. The best use depends on the clinical task and local workflow. More accurate quantitative SPECT and serial post-therapy imaging are also making individualized dosimetry more practicable, although the relationship between measured absorbed dose and treatment outcome remains incompletely established for several therapies.
Artificial intelligence is being investigated for reconstruction, attenuation correction, organ and lesion segmentation, image interpretation, and dose estimation. These methods require external validation because scanner differences, patient populations, and acquisition protocols can change performance. Other research targets include fibroblast activation protein, immune-cell markers, hypoxia, and additional receptors or transporters. For each proposed agent, biological specificity, manufacturing reliability, normal-organ exposure, and demonstrated clinical utility determine whether a promising molecular signal becomes a useful medical procedure.