Imaging in endocrinology is not a simple “support” to biochemistry, but an extension of pathophysiological reasoning: it localizes the site of dysfunction, defines its anatomical relationships, quantifies the mass or volume of functioning tissue, recognizes patterns of invasion or infiltration and guides diagnostic and therapeutic procedures. In many endocrine conditions, clinical decision-making arises from the integration of phenotype, hormonal profile and targeted imaging, because the biochemical signal may indicate an altered axis but does not always identify the structural cause, whereas imaging may reveal common findings and incidentalomas that need to be “contextualized” by physiology.
The value of imaging is highest when it is chosen according to a precise question. Ultrasound can clarify the nature of a superficial lesion and guide fine-needle aspiration, computed tomography (CT) and magnetic resonance imaging (MRI) delineate deep anatomy and relationships with critical structures, and nuclear medicine describes function and receptor targets in addition to morphology. The choice of modality, contrast agent, timing and technical protocol significantly influences sensitivity, specificity and clinical utility. “Non-targeted” imaging instead increases the risk of overdiagnosis and unnecessary diagnostic pathways.
This page provides a systematic overview of the main techniques and their rational use in endocrine diseases, with attention to limitations, artifacts, standardized reporting, management of incidentalomas and integration with laboratory tests and dynamic tests. The objective is to make imaging a tool consistent with physiology, avoiding the most frequent error in modern endocrinology: treating the image as an autonomous truth, disconnected from the biological context.
The first principle of endocrine imaging is to distinguish between morphological imaging and functional imaging. Morphological imaging describes size, density, signal, vascularization and anatomical relationships; it identifies masses, nodules, cysts, calcifications, infiltrations and signs of invasion. Functional imaging, especially in nuclear medicine, highlights tracer uptake reflecting metabolism, receptor expression or specific transporters, allowing characterization of small, multifocal lesions or lesions in difficult sites and defining disease extent in a “biological” as well as anatomical manner.
The second principle is that the clinical utility of imaging depends on the pre-test probability. In endocrinology, the prevalence of incidental findings is high: thyroid nodules, pituitary microadenomas, adrenal nodules, pancreatic cysts and ovarian lesions are common in the general population and increase with age and with the spread of CT and MRI performed for other reasons. If imaging is requested before defining a biochemical and clinical question, the risk of “incidentaloma-driven medicine” increases, in which the next examination is not guided by physiology but by the need to pursue a frequent and often benign finding.
The third principle is that many endocrine diseases are not a single lesion, but a systemic disorder. In primary aldosteronism, adrenal imaging is important, but functional lateralization may not coincide with morphology and often requires integration with dedicated procedures. In mild hypercortisolism, adrenal masses may be present even in the absence of clinically significant autonomous secretion, whereas an axis abnormality may exist with apparently normal glands. In pituitary medicine, microlesions may be invisible on standard MRI or may be multiple; in thyroid medicine, a hypofunctioning nodule on ultrasound may be benign and vice versa. Endocrinology therefore requires an imaging interpretation centered on the mechanism, not on the isolated “shape”.
Finally, the choice of modality must consider safety and repeatability. CT exposes the patient to ionizing radiation and requires caution during pregnancy and in frequent follow-up. MRI avoids radiation and offers excellent soft-tissue contrast, but may have limitations in patients with non-compatible devices, claustrophobia or severe renal insufficiency when gadolinium use is considered. Nuclear medicine adds biological information, but requires appropriateness, standard protocols and specialist interpretation to reduce false positives and avoid non-targeted use.
Ultrasound is the first-line modality for the thyroid because it combines high spatial resolution, absence of radiation, dynamic evaluation and the possibility of guiding procedures. In thyroid medicine, the request for ultrasound must be justified by clinical or laboratory suspicion, not by the mere presence of an incidental finding on CT or MRI, since identification of small nodules is extremely common and may trigger a diagnostic pathway that does not improve clinical outcomes. Ultrasound provides information on parenchymal echotexture, presence of thyroiditis, nodular features, vascularization and lymph nodes of the central and lateral cervical compartments.
The clinical strength of ultrasound derives from its ability to stratify risk and select nodules for fine-needle aspiration or follow-up. Structured reporting systems, such as Thyroid Imaging Reporting and Data System (TI-RADS) systems, formalize a shared language based on composition, echogenicity, margins, shape and calcifications, reducing inter-operator variability and the number of unnecessary biopsies. The modern approach does not aim to “biopsy everything”, but to identify nodules with a combination of risk and size that makes cytological confirmation useful, while maintaining awareness that a significant proportion of thyroid carcinomas is indolent and that diagnostic intensity must be proportional to the real risk and the clinical context.
For the parathyroid glands, ultrasound may identify superficial adenomas, especially when they are inferiorly ectopic or related to the thyroid pole, but sensitivity depends on size, neck anatomy, the presence of concomitant thyroid nodules and operator experience. In primary hyperparathyroidism, ultrasound is often complementary to nuclear medicine and other techniques, because the preoperative goal is to localize the responsible gland and plan a targeted surgical approach when appropriate. Ultrasound also evaluates complications and indirect organ targets, such as superficial vascular calcifications or signs of renal disease if the examination is extended, but its main role remains cervical localization and procedural guidance.
Ultrasound is also useful for salivary and submandibular glands when involved in endocrine and autoimmune diseases, for evaluating neck masses, lymph nodes, subcutaneous adipose tissue in selected metabolic conditions and, in andrology, for testicular assessment when clinical findings suggest secreting tumors or developmental disorders. The main limitation is depth: ultrasound does not replace CT or MRI for mediastinal, retroperitoneal or intracranial structures, and its interpretation requires awareness of operator-dependent variability.
CT and MRI represent the framework of deep morphological imaging. In endocrinology, data quality critically depends on the technical protocol. Small variations in slice thickness, contrast phases, fat suppression or dedicated sequences can transform a “negative” examination into a diagnostic examination. A classic example is the pituitary gland: dedicated pituitary MRI, with thin slices and dynamic acquisitions after contrast, increases the ability to identify microadenomas compared with a generic brain MRI. Similarly, for the adrenal gland, CT with dedicated phases and assessment of washout dynamics can distinguish lipid-rich adenoma from non-adenomatous lesions more robustly than an abdominal CT not protocolled for the adrenal glands.
CT excels in the evaluation of calcifications, bone, thoracoabdominal parenchymal organs and densitometric characterization of some lesions. MRI excels in soft-tissue contrast, assessment of the brain and hypothalamic-pituitary system, evaluation of liver and pancreas in selected contexts and characterization of lesions with hemorrhagic or lipid components. In many endocrine diseases, MRI is preferable when repeated imaging is expected, to reduce cumulative radiation exposure, and when fine definition of relationships with neural or vascular structures is needed.
Contrast agents improve characterization, but in endocrinology they must be used carefully, because appropriateness does not mean “always better with contrast”. In pituitary MRI, contrast is often decisive for microlesions, whereas in some thyroid or neck assessments iodinated contrast may interfere with radioiodine-based therapeutic strategies, requiring temporal planning. The decision must be individualized: the expected diagnostic benefit, the relative risk of contrast and the need to repeat the examination over time are assessed. Best practice is for the clinical request to contain the pathophysiological question, so the radiologist can optimize the protocol.
A cross-cutting element is mechanism-oriented interpretation. An adrenal nodule is not “an incidentaloma” until it is integrated with the biochemistry of autonomous secretion or with the patient’s oncological history. A pituitary lesion must be read in relation to the hormonal pattern and signs of mass effect. A thyroid nodule must be interpreted in light of thyroid-stimulating hormone, history of irradiation, family history and antibody status. This approach reduces false pathways: imaging becomes the tool that confirms, localizes or excludes a mechanism, not an automatic generator of diagnoses.
MRI is the central modality for the hypothalamic-pituitary axis. Its indication arises from compressive symptoms, visual field defects, headache with signs of mass effect, hypopituitarism, documented hormonal hyperproduction or incidental sellar findings. Dedicated pituitary MRI searches for microadenomas, macroadenomas, pituitary stalk abnormalities, cystic lesions such as Rathke cleft cysts, infiltrative and inflammatory processes. The ability to distinguish a microadenoma from physiological variations in enhancement or from artifacts depends on dynamic sequences, contrast quality and interpretative experience.
The management of pituitary incidentalomas requires a balance between surveillance and appropriateness, because many lesions remain stable and clinically silent. Assessment must include size, relationship with the optic chiasm, signs of compression and secretory or deficient profile. Radiological surveillance, when indicated, follows a risk-based logic: macrolesions and lesions close to the chiasm require closer monitoring, whereas stable microlesions may be monitored at longer intervals. The choice of interval must not be automatic, but consistent with expected growth, symptoms, age and therapeutic plan. Repeated imaging without a clinical question risks transforming reasonable surveillance into endless follow-up with unnecessary anxiety and costs.
In some conditions, hypothalamic-pituitary imaging is essential even when the lesion is not an adenoma. In polyuria-polydipsia, assessment of the neurohypophyseal region and stalk may suggest infiltrative or neoplastic diseases. In autoimmune or immunotherapy-related hypophysitis, MRI shows patterns of enhancement and enlargement that, integrated with clinical and laboratory data, guide therapy and monitoring. Here too, the principle is coherence: imaging serves to clarify cause and consequences, not to label every signal variation as an autonomous disease.
The adrenal gland is one of the districts with the highest frequency of incidental findings. CT and MRI allow characterization of lipid-rich adenomas, non-adenomatous lesions, metastases, carcinomas, pheochromocytomas and hemorrhagic lesions. Interpretation cannot be separated from biochemistry: many lesions are non-secreting, whereas some endocrine dysfunctions may occur without evident masses. In particular, mild autonomous hyperfunction may coexist with small lesions and morphological appearance does not always predict hormonal activity.
Morphological characterization uses densitometric and dynamic parameters. On CT, low-attenuation density suggests lipid content typical of adenoma, while washout patterns in delayed phases may strengthen the diagnosis in indeterminate lesions. MRI, with chemical shift techniques, exploits signal loss on out-of-phase imaging to identify intralesional lipid content. These tools reduce the need for biopsies, which in adrenal endocrinology are rarely the first choice and may be contraindicated if pheochromocytoma has not been excluded.
In suspected primary aldosteronism, adrenal imaging has an important but not definitive role in lateralization. Morphology may show unilateral nodules even in the presence of bilateral secretion and vice versa. For this reason, when the objective is to select surgical candidates and distinguish unilateral from bilateral forms, the clinical strategy integrates imaging and functional lateralization procedures in expert centers. In this context, imaging is not the “proof” of origin, but the tool that excludes suspicious masses, defines anatomy and guides the subsequent pathway.
In suspected pheochromocytoma or paraganglioma, CT and MRI localize lesions and describe their vascular relationships, but the extent of multifocal or metastatic disease may require receptor-based or metabolic functional imaging. Here too, modern endocrinology tends to combine morphology and function: the lesion is not defined only by appearance, but by biological behavior and by targets that may influence therapy and follow-up.
In thyroid nodules, ultrasound remains central for risk stratification and lymph node assessment. The presence of suspicious lymph nodes changes the diagnostic and surgical strategy and requires systematic evaluation of the cervical compartments. Ultrasound guides fine-needle aspiration and, when necessary, assessment of thyroglobulin washout or other markers in lymph node aspirate in selected contexts. The logic is to avoid excessive procedures in low-risk nodules and concentrate diagnostic resources on high-risk patterns.
Nuclear medicine adds functional information. Thyroid scintigraphy with radioisotopes is useful in specific contexts, especially when thyroid-stimulating hormone (TSH) is suppressed and the clinical question is to distinguish an autonomously hyperfunctioning nodule from a non-autonomous nodule. In differentiated thyroid carcinomas, radioiodine and related imaging techniques have a role in therapy selection, functional staging and follow-up in appropriate contexts. The choice of timing in relation to iodinated contrast, risk assessment and use of serum biomarkers guide the appropriateness of functional imaging.
In medullary carcinoma and in some neoplasms with different biological behavior, functional imaging may include different tracers and must be chosen according to the disease profile and clinical questions. The general rule is that nuclear medicine does not replace neck ultrasound, but integrates it when evaluation needs to be extended beyond the cervical district or when biochemistry suggests residual disease not localized by conventional methods.
Functional imaging, particularly positron emission tomography/computed tomography (PET/CT), has taken on a growing role in endocrinology, especially for neuroendocrine neoplasms and for some forms of thyroid carcinoma and paragangliomas. PET with 18F-fluorodeoxyglucose (18F-FDG) measures glycolytic metabolism, often higher in aggressive or dedifferentiated lesions. This allows biological stratification: in some neuroendocrine neoplasms, discordance between receptor imaging and FDG may reflect grade heterogeneity and guide therapeutic and prognostic choices. Interpretation requires standardized protocols and awareness that inflammation may produce nonspecific uptake.
For many neuroendocrine neoplasms, somatostatin receptor imaging with PET represents a high-sensitivity tool, useful for localizing disease, assessing extent and selecting receptor-based therapies, including radiometabolic strategies in appropriate contexts. The central point is that receptor PET is not a generalized “screening” examination: it is indicated when pre-test probability is sufficient and when the answer will influence management. In this way, functional imaging becomes an operational biomarker, not a redundant addition.
Standardization of acquisition and reporting is crucial because PET is a quantitative as well as qualitative measure. Data repeatability and comparability over time depend on quality controls, patient preparation, post-injection timing and technical reconstruction parameters. In endocrinology, where response to therapy is often monitored and escalation or de-escalation is decided, the methodological quality of functional imaging is an integral part of clinical decision-making.
Imaging does not only serve to “see”, but also to guide procedures. In thyroid medicine, ultrasound-guided fine-needle aspiration is the most widespread procedure, and its appropriateness depends on ultrasound risk stratification and size. A biopsy performed on low-risk or excessively small nodules increases the probability of indeterminate results and unnecessary surgical pathways. Sample quality depends on technique, operator experience and integration with cytopathology and, when indicated, with molecular testing according to structured pathways.
In the adrenal gland, biopsy is rarely indicated as the first choice and always requires exclusion of pheochromocytoma and confirmation that the procedure will truly change management, typically in selected oncological contexts. The rule is that adrenal endocrinology prioritizes radiological and biochemical characterization; biopsy comes into play only in targeted scenarios, because procedural risk and impact on the diagnostic pathway may exceed benefit if the case is not well selected.
There are highly specialized procedures in which imaging is part of an integrated multidisciplinary pathway. Some functional sampling techniques, when appropriate, serve to define lateralization or source of secretion, integrating interventional radiology and physiology. In these cases, the quality of the result depends on patient preparation, protocol standardization and correct candidate selection, because a technically perfect procedure may be clinically useless if the pre-test probability does not justify the question.
The management of incidentalomas is one of the most complex tasks in modern endocrinology, because the prevalence of findings is high and the probability that they are clinically significant is often low. An incidentaloma must be evaluated across three dimensions: risk of malignancy, risk of autonomous secretion and risk of growth with compressive or functional effects. Imaging contributes especially to the first and third aspects, whereas autonomous secretion always requires biochemical integration. This avoids a common error: classifying a finding as an “endocrine problem” only because it is located in an endocrine gland.
Radiological surveillance must be rational. Excessively frequent follow-up produces anxiety, radiation exposure and secondary incidentalomas; excessively rare follow-up may miss significant growth in selected lesions. The best surveillance is adapted to the expected natural history: initial size, risk features, symptoms, age and comorbidities. In the pituitary gland, for example, proximity to the chiasm and size define the risk of compressive events. In the adrenal gland, stability over time and radiological characterization guide the intensity of monitoring. In the thyroid, low-risk nodules may be followed with ultrasound at appropriate intervals or not followed at all if the risk is truly minimal.
A fundamental concept is temporal coherence. If a patient is followed longitudinally, it is preferable to maintain the same modality and comparable protocols, because a change in platform or technique may simulate growth or regression. Comparability, in endocrinology, is a clinical requirement: imaging becomes a biomarker only if it can be compared reliably over time.
Endocrine imaging often requires repeated examinations. For this reason, clinical sustainability includes minimization of radiation dose when CT and nuclear medicine are used, appropriate selection of MRI when equivalent and prudent use of contrast agents. Risk assessment must not be abstract: it must be related to age, pregnancy, the need for follow-up and the severity of the suspected disease. In a high-risk suspected endocrine malignancy, the priority is accurate staging; in a low-risk incidentaloma, the priority is to avoid unnecessary exposure and investigations.
Iodinated contrast may influence thyroid pathways related to radioiodine and assessment of iodine load, whereas gadolinium requires caution in severe renal insufficiency and in the context of repeated examinations. The best management is to plan the sequence of examinations, avoid duplications and define in advance what is expected to be obtained with contrast. The clinical criterion is simple: if contrast does not change diagnostic probability or management, it should be avoided.
Safety also includes communication of the result. A structured report, with conclusions oriented to the clinical question and follow-up indications consistent with risk and context, reduces overdiagnosis and repetitions. In endocrinology, where many findings are frequent and often benign, report quality is a determinant of appropriateness as much as image quality.