Spectral CT and Photon-counting Detector CT: Principles and Applications in Neuroradiology

Abstract

Spectral CT and photon-counting detector (PCD) CT are recent innovations in imaging that have gained increasing use as diagnostic tools. Both modalities leverage advanced detector technologies—PCD CT uses detectors that count individual x-ray photons and measure their energies, while spectral CT separates x-ray energies at dual or multiple levels. These approaches provide higher spatial resolution, improved contrast-to-noise ratio, and reduced noise and enable multienergy imaging from a single acquisition. These techniques enhance visualization of tissue characteristics and reduce artifacts compared with conventional energy-integrating detector CT. This review article introduces the fundamental principles of spectral CT and PCD CT, their advantages in neuro-oncologic applications, and their potential to improve diagnostic imaging through enhanced image quality and quantitative accuracy.

Keywords: Physics, CT–Photon-counting, Pediatrics, CT-Spectral

© RSNA, 2026

Summary

Photon-counting detector (PCD) CT and spectral CT represent the latest advances in CT imaging technology, providing substantial enhancements over conventional systems through energy-resolved imaging and improved image quality.

Essentials

• ■ Spectral CT uses dual- or multienergy acquisition to enable material-specific imaging and improve tissue characterization and diagnostic accuracy in neuroradiology.

• ■ Photon-counting detector (PCD) CT provides higher spatial resolution and improved energy discrimination than conventional CT, supporting the detection of subtle neurologic abnormalities and small lesions.

• ■ Compared with conventional energy-integrating detector CT, spectral CT and PCD CT reduce artifacts and improve tissue differentiation, with PCD CT providing higher spatial resolution and sensitivity.

• ■ In head and neck imaging, PCD CT improves evaluation in complex clinical settings, including cases with metallic dental work, subtle bone invasion, and posttreatment changes.

• ■ Spectral CT and PCD CT enhance neuro-oncologic imaging of primary central nervous system neoplasms and brain metastases by improving lesion detection and characterization while maintaining or potentially reducing radiation dose.

Introduction

Spectral CT is an umbrella term encompassing CT techniques that acquire images at more than one energy level—either two energy levels, as in dual-energy CT (DECT), or multiple energy levels, as in PCD CT. Both DECT and PCD CT fall under the spectral CT category, and these approaches enable advanced tissue characterization and improved artifact reduction (1–5). While most DECT systems use energy-integrating detectors (EIDs), EID CT alone is not spectral unless it operates in dual-energy mode.

PCD CT uses semiconductor detectors that count individual x-ray photons and precisely measure their energies, resulting in higher spatial resolution, improved contrast-to-noise ratio, and reduced electronic noise compared with traditional EID systems (1,2,4,5). This technology enables multienergy imaging in a single scan, facilitating accurate material differentiation and quantitative imaging (1,4).

Both spectral CT and PCD CT have demonstrated value across a range of clinical applications, from improved visualization of vessels and atherosclerotic plaques and perfusion assessment in cardiovascular imaging (2,6–8) to more precise evaluation of bones and joints and improved differentiation of materials such as urate and calcium in musculoskeletal imaging (6–9).

In neuroradiology, spectral CT enables improved contrast between soft tissue structures such as gray and white matter, with increased signal-to-noise and contrast-to-noise ratios, often achieved with reduced radiation exposure (10,11). These advantages enhance the accuracy of neuroimaging for pathologies such as tumors, acute stroke, and vascular diseases. The ability of spectral CT to help distinguish hemorrhage from iodinated contrast media, generate virtual noncontrast (VNC) images, and improve early detection and staging of neoplasms adds substantial clinical value (12–14). Mobile platforms now enable bedside neuroimaging for critically ill patients, expediting diagnosis and management of acute conditions such as stroke or hemorrhage (15).

Spectral CT improves assessment of arterial and venous diseases in the head, neck, and spine. Applications such as CT angiography of the head and neck, spinal CT, and CT myelography, particularly for detecting cerebrospinal fluid–venous fistulas, benefit from improved vascular visualization and reduced artifacts from calcifications and metallic implants (11,16–18). The ability of PCD CT to perform virtual monoenergetic imaging (VMI) and material decomposition enables improved detection of aneurysms, vessel stenosis, and hemorrhage, increasing diagnostic confidence. In addition, improved artifact reduction facilitates more accurate evaluation around metallic prostheses and postoperative changes (11,18,19).

In spinal imaging, spectral CT increases sensitivity for identifying small vascular structures and pathologies, such as the artery of Adamkiewicz and spinal dural arteriovenous fistulas, while reducing artifacts in patients with spinal hardware (11,17,18). The high spatial resolution of PCD CT also enables detailed assessment of intricate structures such as the temporal bone, aiding detection of subtle pathologies such as otosclerosis and optimizing cochlear implant evaluation (11,17,18).

Oncologic head and neck imaging benefits from both PCD CT and spectral CT, which enable better tumor detection and characterization through enhanced tissue contrast and increased spatial resolution. These advantages are particularly valuable in the complex anatomy of the head and neck, where the detection of small lesions and artifact reduction are critical for accurate diagnosis and treatment planning (7,8,18).

Emerging evidence highlights the advantages of PCD CT and spectral CT in reducing radiation exposure and minimizing beam-hardening artifacts while maintaining or improving diagnostic image quality, features that are particularly beneficial for vulnerable oncologic and pediatric populations (1,7,8,19–21).

Overall, the advent of PCD CT and spectral CT is reshaping neuroimaging, improving diagnostic confidence, patient safety, and workflow efficiency across neuroradiologic practice.

With this review article, we aim to elucidate the fundamental physical concepts of spectral CT and PCD CT and to provide an overview of their clinical applications and benefits in neuroradiology, with a particular emphasis on cancer imaging, including head and neck tumors, primary central nervous system neoplasms, and brain metastases.

Fundamental Physics of Spectral and PCD CT

Spectral CT leverages the energy-dependent attenuation of x-rays in different materials by using rapid kilovoltage peak–switching, dual-source, or detector-based imaging strategies to acquire data at multiple (at least two) photon energies. The underlying physics arises from the distinct contributions of the photoelectric effect and Compton scattering: The photoelectric effect predominates at lower photon energies and in elements with high atomic numbers, whereas Compton scattering predominates at higher photon energies and in materials with lower atomic numbers (22). An important application, K-edge imaging, exploits differences in K-edge energies among elements, enabling differentiation and quantification of various tissue types and contrast agents. Enhanced spectral separation allows more accurate material decomposition, construction of energy-dependent attenuation curves, and mapping of effective atomic numbers. This enables both quantitative and functional imaging, increasing the specificity and clinical utility of CT (22).

A key clinical strength of spectral CT is the generation of iodine maps, which enhance lesion contrast by highlighting iodine uptake within tissues. Increased conspicuity of hypervascular tumors enables clearer visualization of lesion margins, facilitating the differentiation of neoplastic tissue from surrounding parenchyma (23). In addition, iodine maps allow quantitative assessment through direct measurement of iodine concentration within lesions. This capability supports more precise tumor characterization, including the differentiation of tissue types based on iodine uptake (24). Iodine mapping also enables the assessment of vascularity and viability, providing insight into tumor perfusion, vascular patterns, and therapeutic response, including evaluation after antiangiogenic therapy, radiofrequency ablation, or chemoembolization. By improving iodine conspicuity, spectral CT may also contribute to dose optimization, as low–kiloelectron volt VM images can increase iodine attenuation and potentially allow protocol tailoring with lower radiation exposure or reduced contrast material dose in selected clinical scenarios (25). Another important application is the generation of VNC images from contrast-enhanced scans. VNC enables creation of synthetic noncontrast images, eliminating the need for a separate unenhanced scan and thereby reducing radiation exposure, which is particularly beneficial for patients with cancer who require repeated imaging. This approach represents one of the most clinically relevant dose-saving mechanisms of spectral CT, particularly in multiphase or longitudinal imaging protocols, although replacement of true noncontrast images should be validated for the specific clinical indication and scanner platform (25). VNC images provide reliable baseline measurements, supporting accurate densitometric evaluation for distinguishing enhancing from nonenhancing lesions and for tissue characterization, which is critical for tumor assessment and therapy monitoring (23). These images also complement iodine maps and standard contrast images, improving the differentiation of tissue types and resolving diagnostic ambiguities when unenhanced images are required. When interpreted alongside iodine maps, VNC images allow simultaneous evaluation of baseline tissue density and perfusion, enhancing lesion characterization and tumor segmentation.

With each DECT scan, datasets are reconstructed at both low and high tube voltages, allowing the separate evaluation of each. The low-voltage dataset (for example, 80 kVp) typically demonstrates increased iodine attenuation and therefore greater contrast enhancement; however, this benefit comes at the cost of higher image noise. Despite this trade-off, the improved lesion conspicuity and contrast-to-noise ratio provided by the low–kilovoltage peak dataset can be valuable when lesion detection is challenging due to small size or complex anatomy. In contrast, the high-voltage dataset (such as 140 kVp combined with a tin filter) is associated with lower image noise but reduced enhancement, which may limit diagnostic utility. Thus, the complementary strengths of low- and high-voltage acquisitions can be leveraged to optimize lesion detection and characterization (26). In addition, VMI permits reconstruction across a range of energy levels: Low–kiloelectron volt images can improve iodine contrast and lesion conspicuity, whereas higher–kiloelectron volt images may reduce beam-hardening artifacts. Together with automated tube current modulation, optimized tube voltage selection, and iterative or model-based reconstruction, these tools may support radiation dose optimization while preserving diagnostic performance (25,27).

PCD CT enhances these capabilities by using semiconductor PCDs that acquire detailed spectral information at multiple energy levels, rather than only two as in conventional DECT. These detectors register and categorize individual x-ray photons into discrete energy bins, providing direct energy-resolved detection (28). This results in improved material discrimination, reduced artifacts, higher spatial resolution, lower noise, improved contrast, and more robust quantitative tissue evaluation through the preservation of photon energy information. In contrast to conventional EID, PCD CT directly converts incident photons into electrical signals and can reduce the influence of electronic noise, thereby improving dose efficiency and contrast-to-noise performance (29). In our institution, the mean protocol-based CT dose index volume across all PCD CT scanners was 11.1 mGy, compared with 16.3 mGy across the EID CT scanners. Because CT dose index volume is determined by scanner protocol, it does not vary among patients scanned using the same protocol. In contrast, dose-length products depend on the individual z-axis scan length in an individual patient and hence does not reflect the differences between the two imaging techniques.

These combined advancements expand the clinical utility of CT by enabling functional, quantitative, and highly specific anatomic imaging (28). However, although these mechanisms suggest meaningful potential for dose reduction, dedicated clinical data demonstrating dose savings specifically in neuro-oncologic cohorts remain limited, and future prospective studies are needed to define optimized acquisition protocols, diagnostic performance, and safe dose thresholds in this patient population.

The Table summarizes the salient features of the principles and applications of spectral CT and PCD CT in neuroradiology.

Summary of Principles and Applications of Spectral CT and Photon-counting Detector CT

Feature/Application	Spectral CT	Photon-counting Detector CT
Principle	Uses dual-energy or multienergy x-rays to differentiate materials based on energy-dependent attenuation	Detects individual photons and their energy levels, offering higher resolution and material specificity
Material decomposition	Differentiates tissues (eg, gray/white matter, blood, bone) and contrast agents (eg, iodine, calcium)	Enhanced material decomposition (eg, iodine, gadolinium, bone) with higher accuracy due to energy binning
Stroke imaging	Improved detection of ischemic stroke via virtual noncontrast and iodine maps to assess perfusion	Superior visualization of hypoattenuating lesions with improved contrast-to-noise ratio
Hemorrhage detection	Accurate differentiation of hemorrhage from iodinated contrast or calcification using material-specific maps	Potential for precise identification of hemorrhage, even at low volumes, due to high sensitivity to blood and tissue contrast
Tumor characterization	Enhanced differentiation of tumor types (eg, gliomas) via iodine uptake and tissue composition analysis in experimental studies	Improved tumor boundary delineation and characterization with high-resolution imaging and spectral data
Vascular imaging	Better visualization of cerebral vasculature and stenosis using iodine-enhanced angiograms and virtual noncontrast studies	High-resolution angiograms with reduced artifacts, improved detection of small vessel pathology
Metal artifact reduction	Reduces artifacts from coils, clips, or shunts, improving visualization of adjacent brain structures	Superior artifact reduction due to energy-resolved imaging, enhancing clarity near metallic implants
Dose efficiency	Lower radiation dose compared with conventional CT by optimizing energy-specific imaging	Potentially lower radiation dose with improved signal-to-noise ratio and photon detection efficiency
Bone and calcium imaging	Clear separation of bone, calcium, and soft tissue for assessing calcified lesions or plaques	Potential for enhanced detection of microcalcifications or vascular pathology
Contrast agent optimization	Quantifies iodine uptake for better contrast agent use and reduced contrast agent dose in sensitive patients	Improved contrast differentiation, enabling lower contrast agent doses and multicontrast agent imaging
Emerging applications	Quantitative imaging for tissue characterization and perfusion studies in research settings	Ultra-high-resolution imaging for microstructural analysis and potential for K-edge imaging in research

Clinical Applications

Head and Neck Tumors

Spectral CT and PCD CT offer substantial advantages over conventional EID CT for evaluating head and neck tumors. While conventional spectral CT improves material decomposition by acquiring attenuation data at two energy spectra, PCD CT advances this capability by directly counting individual photons and assigning them to multiple energy bins, reducing electronic noise, providing smaller detector pixels and multithreshold energy discrimination. Thus, spectral CT primarily provides quantitative iodine- and effective atomic number (Z-effective)–based tissue characterization, whereas PCD CT adds higher spatial resolution, improved spectral separation, and more robust material discrimination and quantification of iodine, calcium, soft tissue, and hemorrhage, particularly in anatomically complex or artifact-prone regions.

In detail, PCD CT provides markedly higher spatial resolution due to its use of smaller detector pixels, enabling the visualization of finer anatomic detail. This capability is particularly valuable in the head and neck, where accurate delineation of tumor boundaries and the involvement of small structures such as nerves and cartilage is essential for precise diagnosis, T-staging, and treatment planning. The improved spatial resolution of PCD CT also enhances the assessment of tumor extension into structures such as laryngeal cartilage or skull base bone (30–32). In addition, PCD CT reduces the electronic noise associated with conventional EID CT, resulting in clearer images and improved detection of subtle differences between tumor tissue and adjacent normal structures (Figs 1, 2). Importantly, these technologies are not limited to malignancies: Spectral CT and PCD CT are also valuable for evaluating benign lesions, such as parathyroid adenomas, where accurate localization and characterization are essential for effective management (Fig 3).

Figure 1:

(A) Coronal contrast-enhanced CT image of the submandibular region acquired with an EID CT scanner in a 62-year-old male patient with p16-positive squamous cell carcinoma of the right tongue base, after definitive chemoradiation therapy, demonstrates asymmetric nodularity in the left submandibular region (arrow), indeterminate for tumor recurrence. (B) Corresponding PCD CT image reconstructed at 50-keV monoenergetic value with a spatial resolution of 0.2 mm, acquired about 2 months later, clearly depicts multiple discrete nodules in the same location (two arrows, right), as well as a rounded right level 1 lymph node (single arrow, left), all suspicious for tumor recurrence, which was confirmed with tissue sampling. EID = energy-integrating detector, PCD = photon-counting detector.

Figure 2:

Noncontrast axial CT images in a patient followed up for medullary carcinoma of the thyroid gland. (A) Conventional EID CT image reveals a hypoattenuating nodule within the thyroid gland, with notably ill-defined nodule margins and relatively hyperattenuating thyroid parenchyma. (B) Follow-up PCD CT image demonstrates sharp margins of the nodule and an enhanced appearance of the surrounding normal thyroid parenchyma. Spatial resolution of PCD CT was 0.2 mm. EID = energy-integrating detector, PCD = photon-counting detector.

Figure 3:

(A) Contrast-enhanced PCD CT image reconstructed at 50-keV monoenergetic value, acquired in a 69-year-old female patient with primary hyperparathyroidism, demonstrates an arterially enhancing nodule posterior to the right thyroid, suspicious for a parathyroid adenoma (arrow). Spatial resolution was 0.2 mm. (B) A prior CT image acquired a few months earlier with an EID scanner at an outside institution also shows the lesion; however, it is much better delineated on the PCD CT image, and the adjacent fat also shows less streak artifact from the shoulders. The earlier CT scan had been performed for neck nodal evaluation and did not report the suspected adenoma. EID = energy-integrating detector, PCD = photon-counting detector.

Another major advantage of PCD CT is the reduction of metal artifacts from dental implants and fillings. Case reports have shown that PCD CT enables clearer visualization of oropharyngeal tumors adjacent to metal crowns, overcoming streak artifacts that often compromise conventional CT and permitting more confident assessment of tumor invasion and margins, particularly within the oral cavity and oropharynx (32,33). Similarly, PCD CT provides improved soft tissue detail in dental imaging, aiding the assessment of tooth-associated abscesses or dental tumors (34).

Beyond spatial resolution and artifact reduction, spectral CT provides additional quantitative imaging features through iodine maps and spectral attenuation measurements. These parameters serve as imaging surrogates for angiogenesis and perfusion within tumor tissue, biologic processes related to glucose metabolism as measured with fluorodeoxyglucose (FDG) PET (Figs 4, 5) (35). It has been reported that normalized iodine-related attenuation changes in hypopharyngeal carcinoma following neoadjuvant chemotherapy not only reflect increased angiogenesis and tumor perfusion but also strongly correlate with FDG PET standardized uptake values, suggesting a direct link between CT-based vascularity metrics and metabolic activity (35).

Figure 4:

(A) Contrast-enhanced PCD CT image reconstructed at 50-keV monoenergetic value demonstrates a heterogeneously enhancing mass in the left tongue base in this 69-year-old male patient with p16-negative squamous cell carcinoma. Associated left level 2 nodal metastases are also seen in the same images, with more necrosis in the anterior node and a more solid appearance of the posterior node. Spatial resolution was 0.2 mm. (B) Corresponding effective Z map illustrates the higher effective atomic numbers in the primary tumor (solid arrow) and nodal metastases (dotted arrows). (C) These areas correlate well with regions of abnormal fluorine 18 (¹⁸F) fluorodeoxyglucose (FDG) avidity on the fused PET/CT image. Note that the necrotic node does not show increased effective-Z (or any abnormal ¹⁸F-FDG avidity), as expected, due to lack of iodine uptake. PCD = photon-counting detector.

Figure 5:

Axial contrast-enhanced PCD CT image reconstructed at 50-keV monoenergetic value in a 55-year-old female patient with treated stage IV melanoma and history of bilateral kidney cancer demonstrates striations in a free flap along the left anterior neck (solid arrow) and an enhancing nodule along its medial aspect (dotted arrow), immediately posterior to the left carotid artery, suspicious for tumor (which was corroborated with a follow-up fluorine 18 fluorodeoxyglucose PET/CT scan). This scan was performed at a dose-length product of 269 mGy · cm compared with a previous scan with an EID CT scanner at 394 mGy · cm. PCD CT enables the acquisition of excellent diagnostic quality images with a reduction in radiation dose compared with conventional CT. Spatial resolution was 0.2 mm. EID = energy-integrating detector, PCD = photon-counting detector.

Dual-layer spectral CT parameters, including iodine density mapping and Z-effective calculations, have improved both quantitative and qualitative differentiation between upper aerodigestive tract head and neck squamous cell carcinomas (HNSCCs) and normal mucosa (36). Iodine density mapping enables sensitive differentiation of head and neck tumors by quantifying tissue iodine content, a marker of vascularity and perfusion (Figs 6, 7). Malignant tumors such as HNSCCs typically demonstrate higher iodine density than normal tissues or benign lesions because of increased blood supply and contrast media uptake (36). For example, primary HNSCCs often show higher iodine density values (eg, 2.01 mg/mL) compared with normal mucosa or benign tissue (typically below 1.58 mg/mL), enabling accurate detection and characterization of malignancy (36). Additionally, iodine maps and low–kiloelectron volt spectral imaging improve the visualization of tumor margins and infiltration by accentuating regions of high iodine uptake, helping distinguish tumors from adjacent tissues with less contrast enhancement (37).

Figure 6:

(A) Contrast-enhanced photon-counting detector CT image reconstructed at a 70-keV monoenergetic value in a 65-year-old male patient demonstrates an enhancing, heterogenous mass (arrow) involving the right aryepiglottic fold and pyriform sinus and extending into the right strap muscles. (B) This is better appreciated on the 50-keV monoenergetic image and (C) the iodine-enhanced map. Spatial resolution was 0.2 mm.

Figure 7:

(A) Trace diffusion-weighted MR image in a 63-year-old male patient with proven nasopharyngeal cancer shows a large area of restricted diffusion (solid arrow) corresponding to the tumor (anterior dotted arrow shows retained secretions in the left maxillary sinus). (B) After radiation treatment, follow-up trace diffusion image shows resolution of restricted diffusion (arrow), suggestive of good response to therapy. (C) Contrast-enhanced PCD CT image reconstructed at a 50-keV monoenergetic value acquired on the same day as the repeat MRI scan for neck evaluation shows no contrast enhancement in this region (50-keV monoenergetic image) and (D) no iodine overlay (iodine-enhanced map). Spatial resolution was 0.2 mm. This example illustrates how PCD CT can be comparable to MRI with diffusion imaging in assessing treatment response. PCD = photon-counting detector.

The Z-effective value provided by spectral CT offers improved tissue discrimination compared with traditional Hounsfield unit attenuation (29,38). This quantitative parameter, which reflects the average atomic number of tissue, changes with malignant transformation. In HNSCC, tumors exhibit higher Z-effective values than normal mucosa (36). PCD CT–generated Z-effective maps have greater precision and can more reliably differentiate materials, increasing contrast between tumor and normal tissues and improving detection, local staging accuracy, and correlation with histopathology (Fig 4) (39). Beyond distinguishing tumor from normal tissue, Z-effective mapping can differentiate tumor subtypes—such as squamous cell carcinoma, adenocarcinoma, and neuroendocrine tumors—and distinguish recurrent tumor from posttreatment changes (40,41).

The spectral and quantitative capabilities of PCD CT also enhance identification and characterization of metastatic versus benign lymph nodes, even when nodes are small or adjacent to vascular structures (Fig 8). Quantitative iodine and water content measurements, together with the analysis of spectral attenuation curves, provide additional diagnostic criteria by exploiting differences in material composition at various energy levels (14,32,36). For example, metastatic nodes may exhibit lower iodine content than benign or inflamed nodes, enabling more accurate differentiation (42).

Figure 8:

Follow-up imaging studies of the neck in a patient with stable metastatic lymphadenopathy. (A) Energy-integrating detector CT image demonstrates a lymph node (arrow) closely approximated to the left mandible, with ill-defined margins of both the node and the surrounding bone and soft tissue. (B) PCD CT image obtained 6 weeks later shows the node more clearly delineated (arrow), with improved definition of adjacent anatomic structures. Spatial resolution of PCD CT was 0.2 mm. PCD CT = photon-counting detector CT.

In patients who have undergone extensive neck surgery or radiation therapy, where fibrosis, scarring, or surgical materials complicate imaging, PCD CT improves the differentiation of recurrent tumor from posttherapeutic changes through features such as VNC imaging, enhanced material decomposition, and improved contrast resolution (32,43). Multienergy imaging enables repeated functional and quantitative assessment through iodine mapping and other spectral metrics, supporting monitoring treatment response and adaptive radiation therapy planning based on changes in tumor vascularity and volume.

Collectively, spectral CT and PCD CT provide meaningful improvements in diagnostic accuracy over conventional EID CT for head and neck cancer imaging, particularly in complex clinical settings involving metallic dental work, subtle bone invasion, or challenging posttreatment anatomy. These modalities also improve laryngeal tumor staging by enabling more accurate assessment of cartilage invasion, a known limitation of conventional CT due to similar attenuation profiles of tumor and cartilage (44,45).

Primary Central Nervous System Neoplasms

Recent early-phase studies suggest a promising role for spectral CT and PCD CT in the evaluation of primary central nervous system tumors and brain metastases, particularly for lesion characterization, tumor grading, assessment of treatment response, and radiation therapy planning (40,46–48).

A major challenge in neuro-oncologic imaging is differentiating tumor recurrence from posttreatment effects such as pseudoprogression or radiation necrosis. Conventional MRI often cannot reliably be used to make this distinction, as both may manifest with similar patterns of new or increasing contrast enhancement and edema. Perfusion MRI addresses this limitation by exploiting differences in vascularity, with tumor recurrence demonstrating hyperperfusion and pseudoprogression typically showing hypoperfusion. Spectral CT, particularly DECT and PCD CT, provides an analogous approach by enabling direct quantification of iodine and spectral attenuation. DECT offers quantitative vascular metrics, including iodine maps, spectral curves, and Z-effective, that improve identification of recurrent glioma by distinguishing hypervascular tumor from treatment-related changes within enhancing regions (40). Venous phase iodine concentration is often substantially higher in recurrent glioma than in treatment-related changes, reflecting increased perfusion, with one study reporting a sensitivity for recurrence as high as 91% (AUC ~ 0.91) (40). In addition, energy-dependent attenuation curves, Z-effective measurements, and normalized Z-effective are higher in recurrent tumors, further supporting differentiation based on increased iodine uptake and altered tissue composition.

Beyond recurrence assessment, spectral CT and PCD CT are advancing tumor grading and characterization through quantification of microvascularity and tissue composition, both key indicators of glioma aggressiveness (47). Combining perfusion CT–derived cerebral blood volume with DECT-derived electron density improves discrimination between high- and low-grade gliomas compared with either metric alone, with high-grade tumors showing both increased blood volume and electron density (47). Iodine maps from DECT or PCD CT correlate with tumor proliferation and grade, with higher iodine uptake typically observed in more aggressive tumor regions (48). Preclinical models support these findings, demonstrating that monoenergetic CT values, spectral Hounsfield unit curve slopes, and iodine concentrations are highest within solid tumor, intermediate in peritumoral regions, and lowest in normal brain tissue. These metrics may serve as surrogates for proliferation and angiogenesis and may help guide biopsy target selection (49).

PCD CT extends the capabilities of spectral CT through direct photon-counting detection, improved spectral separation, higher spatial resolution, and more robust material-specific quantification. PCD CT has been highlighted for its improved contrast-to-noise ratio, superior soft tissue resolution, reduced beam-hardening artifacts, and enhanced delineation of tumor margins, all of which are critical for assessing infiltrative margins and satellite lesions (46). By recording individual photon energies, PCD CT can generate material-specific maps (eg, iodine, calcium, iron, gadolinium) with improved spectral separation compared with conventional DECT, enabling more accurate quantification of tumor, blood, calcification, and hemorrhagic components.

In addition, DECT and PCD CT provide more precise electron density and stopping-power maps for particle therapy planning, aiding target volume and organ-at-risk delineation based on tissue composition (50). These advanced maps capture tumor heterogeneity and may indirectly inform tumor grading by correlating imaging features with histologic aggressiveness, although further study is needed.

Despite these promising advances in tumor grading and treatment-response assessment, there remains a lack of robust evidence demonstrating that quantitative metrics derived from spectral CT and PCD CT correlate with clinically meaningful outcomes, particularly progression-free survival and overall survival, in neuro-oncologic patients. Future prospective studies are needed to validate these imaging biomarkers and define their prognostic value.

Brain Metastases

Although direct data on spectral CT and PCD CT specifically targeting brain metastases remain limited, early evidence—largely extrapolated from studies of gliomas and general neuro-oncologic CT applications—suggests several promising applications for these advanced imaging modalities.

Spectral CT, particularly VMI at low kiloelectron voltage, enhances iodine conspicuity, improving detection of small enhancing metastatic lesions against background parenchyma (51). Material decomposition imaging enables the separation of iodine signal from hemorrhage and calcification, helping determine whether a small cortical or gray-white matter junction focus represents true enhancement rather than chronic microbleed or calcification.

Iodine maps generated by spectral CT and PCD CT provide quantitative measures of lesion vascularity and permeability, which may differ between metastases from various primary tumors, such as hypervascular renal cell carcinoma or melanoma versus hypovascular lung carcinoma. These maps may also help distinguish viable tumor from treatment-related changes after radiosurgery (52). Spectral separation of bone and iodine reduces beam-hardening artifacts at the skull base, improving the assessment of metastases in regions such as the posterior fossa, skull base, or calvarium (51). Furthermore, electron density and stopping-power maps derived from spectral CT or PCD CT are useful for radiation therapy planning, including stereotactic radiosurgery of brain metastases, by improving dose calculations and aiding the segmentation of organs at risk, similar to applications in glioma imaging (53).

Compared with conventional spectral CT, PCD CT may further improve brain metastasis evaluation by providing higher spatial resolution, lower electronic noise, improved contrast-to-noise ratio, and more precise multienergy material decomposition, which may be particularly beneficial for detecting small enhancing metastases and characterizing hemorrhagic, calcified, or skull-base lesions. For patients presenting acutely with neurologic symptoms during systemic therapy, PCD CT provides high-quality noncontrast and contrast-enhanced imaging, with improved visualization of edema, mass effect, and hemorrhage while maintaining a reasonable radiation dose (10).

In preclinical studies, nanoparticle-based K-edge imaging with spectral CT and PCD CT has emerged as a molecular imaging strategy for detecting and quantifying high-Z nanoparticles, including gold-, gadolinium-, bismuth-, ytterbium-, tantalum-, or hafnium-based agents (25,29). By exploiting element-specific K-edge signatures, these approaches can generate material-specific “color” maps and potentially distinguish nanoparticle contrast material from iodine, calcium, and soft tissue. For metastatic brain lesions, this strategy is conceptually attractive because nanoparticle accumulation may reflect blood-tumor barrier disruption, vascular permeability, or targeted uptake, features that vary substantially across brain metastases and may influence drug or radiosensitizer delivery (54). The most advanced clinical activity in this area involves AGuIX, an ultrasmall gadolinium-based nanoparticle being evaluated primarily as an MRI-visible radiosensitizer in radiation therapy trials for brain metastases, with potential relevance for K-edge CT, as gadolinium has favorable K-edge energy (55). Although clinical outcome data for metastatic disease and PCD CT–based nanoparticle mapping remain limited, these technologies may play an increasing role in neuro-oncologic imaging. However, AGuIX has not yet been clinically validated as a PCD CT K-edge contrast agent, and no FDA-approved high-Z nanoparticle contrast agents are currently indicated specifically for PCD CT–based K-edge imaging. Thus, while nanoparticle-based K-edge imaging represents a promising preclinical and translational approach for metastatic lesion characterization and theranostic treatment planning, its role in neuro-oncology remains investigational. In current practice, spectral CT and PCD CT are best viewed as advanced CT techniques for improving the detection, characterization, and treatment planning of brain metastases, particularly when MRI is limited or as a complementary tool alongside MRI.

Other Neuroradiologic Applications

Beyond neuro-oncologic imaging, PCD CT offers substantial advantages in vascular imaging of the head and neck, enabling precise lumen visualization and improved stent assessment with reduced blooming artifacts, particularly in challenging segments adjacent to bone (56–62).

In temporal bone imaging, PCD CT provides ultra-high-resolution visualization of small anatomic structures, such as the ossicles and semicircular canals, with lower radiation dose and reduced noise compared with conventional CT, enabling improved evaluation of conditions such as cholesteatoma (63–65).

Clinical Applications in Pediatric Oncology Imaging

Spectral CT and PCD CT offer advantages over conventional EID CT in pediatric oncologic imaging by reducing noise, improving spatial and contrast resolution, and providing multispectral capabilities, with the potential for lower radiation exposure (66,67).

High spatial resolution is especially critical in pediatric imaging, where accurate delineation of small structures is essential for diagnostic and therapeutic decision-making. These modalities facilitate improved lesion characterization; for example, quantitative iodine mapping helps distinguish enhancing from cystic or calcified masses and helps assess therapeutic response, including identification of malignancy in pulmonary nodules and monitoring treatment effects (68).

PCD CT improves the detection of subtle pulmonary metastases, such as those from osteosarcoma or neuroblastoma, through enhanced edge definition and earlier identification of subcentimeter nodules, with implications for surgical planning (63). In abdominal masses, VNC images help clarify calcification and hemorrhage, aid margin delineation, and distinguish tumors from contrast-opacified bowel (69,70).

Low–kiloelectron volt VMI improves vascular conspicuity for operative planning in nephroblastoma and enhances the detection of calcification in neuroblastoma and hepatoblastoma (69,70).

In summary, spectral CT and PCD CT improve pediatric oncologic imaging by enhancing lesion detection, characterization, and surgical planning while maintaining or reducing radiation dose.

Limitations and Future Perspectives

While the integration of spectral CT and PCD CT into routine neuroimaging holds substantial promise, several technical challenges must be considered.

Detector-related Artifacts and Limitations

At high photon flux encountered in clinical CT, detector systems are susceptible to pulse pileup, in which multiple photons arriving in rapid succession are misregistered as a single event. If not properly corrected, this effect leads to inaccurate energy assignment and reduced image quality (59,71,72). Charge sharing and cross talk are additional concerns, arising when photon interactions near pixel boundaries produce charge clouds that extend into adjacent pixels. These phenomena introduce spectral contamination, degrade spatial and energy resolution, and impair material differentiation (28,71,72). Moreover, scattering events such as Compton scattering and K-fluorescence can exacerbate these effects, sometimes leading to double counting or loss of counts. Although PCD CT markedly improves spectral capability, the energy resolution of current detectors remains insufficient to reliably distinguish photons with closely similar energies, limiting the detection of subtle differences in tissue composition or contrast agents (72).

Image Noise and Reconstruction

Electronic noise, particularly at low x-ray flux, can compromise image quality, especially for low-energy photons that are essential for contrast in neuroimaging. This degradation manifests as increased image noise and reduced quantitative stability (eg, CT number variability) (28). Beam-hardening artifacts, prominent near dense structures such as bone, also pose challenges for spectral CT by introducing image distortion, a common issue in the head and neck region. Robust material decomposition and artifact correction require advanced reconstruction algorithms (73). The development and clinical validation of such techniques, including emerging deep learning approaches, remain active areas of research.

Technical and Physical Constraints

Clinical CT applications require exceptionally high photon count rates (up to 10⁹ photons/sec/mm²), placing substantial demands on detector electronics and increasing the risk of artifacts and performance degradation (59,71). The hardware complexity inherent to PCD CT, involving advanced semiconductor materials (eg, cadmium telluride) and high-speed electronics, also increases cost and maintenance requirements (28). Additionally, early-generation or prototype systems may lack features such as tube-current modulation or advanced spatial-spectral correction, limiting their utility in complex neuroanatomical regions such as the posterior fossa (59).

Standardization, Clinical Validation, and Workflow

Reliable quantitative imaging with spectral CT and PCD CT requires rigorous validation, system calibration, and artifact correction. Clinical adoption also necessitates the adaptation of imaging protocols, radiologist training, and potentially modified contrast agent administration strategies, factors that may affect workflow efficiency, particularly in high-volume neuroimaging settings (59,72). In addition, real-world implementation will require seamless integration of quantitative outputs from spectral CT and PCD CT into picture archiving and communication system/radiology information system workflows, clear reimbursement pathways, and user-friendly postprocessing tools to avoid adding interpretive burden. Standardized reporting templates that include key spectral CT metrics, along with structured training programs and consensus guidelines, may help facilitate broader clinical adoption and reproducibility across institutions.

To address these limitations, spectral CT and PCD CT technologies are rapidly advancing through the use of higher atomic number detector materials, integration of artificial intelligence (AI)–driven corrections, and development of multicontrast K-edge imaging techniques, with important implications for neurovascular and neuro-oncologic imaging. Deep learning is already being applied to CT image reconstruction, denoising, artifact reduction, and low-dose image enhancement, improving image quality and potentially supporting more dose-efficient protocols (74–76). In spectral CT and PCD CT, AI-based approaches are also being investigated to improve material decomposition, stabilize quantitative iodine assessment, enhance VMI and VNC quality, and reduce artifacts related to beam hardening, photon starvation, motion, and metal implants (29,77). In neuro-oncology, these methods may facilitate automated tumor segmentation, the extraction of quantitative radiomic features from iodine maps, VMI, and Z-effective maps, and more reproducible assessment of lesion vascularity, enhancement heterogeneity, and treatment response (78–80). Looking forward, the integration of deep learning with spectral and PCD CT may enable automated multiparametric workflows that combine morphologic, spectral, and quantitative biomarkers for tumor characterization, response prediction, and longitudinal monitoring. However, it should be noted that most current evidence for AI in neuro-oncology is derived from MRI-based segmentation and radiomics (81), making the validation of AI models specifically on spectral CT or PCD CT datasets in neuro-oncology an urgent research priority.

Commercial PCD CT scanners typically use cadmium telluride or cadmium zinc telluride detectors, which provide high stopping power and excellent energy resolution across the diagnostic range. This enables the acquisition of thin sections and robust multienergy datasets for both neurovascular and oncologic imaging. Variations in detector material composition, thickness, and threshold configuration contribute to differences in count-rate performance, spectral separation, and noise correction strategies, which in turn affect the visualization of small vessels, metal artifact reduction, and soft tissue contrast in neuro-oncology. Emerging detector designs focus on optimizing thickness and pixel geometry to minimize charge sharing and enhance energy discrimination, key requirements for reliable multicontrast K-edge imaging and accurate quantitative mapping of tumor composition at clinically relevant radiation doses (28,82).

True K-edge imaging represents another notable innovation, particularly for molecularly targeted neuroimaging and cancer imaging, although clinical implementation remains in the early stages (20). Gold and other metal-based nanoparticles have been developed for PCD CT, enabling multiplexed molecular and vascular imaging of tumors at low concentrations. PCD CT can uniquely identify and quantify these particles, even in the presence of iodinated contrast agents (83).

AI is increasingly used to correct detector nonidealities and to leverage the rich spectral data for neurologic and cancer imaging applications. Deep learning models are being developed to address pulse pileup, charge sharing, and photon energy misclassification, thereby improving spectral discrimination and image stability for brain and tumor assessment. In addition, deep learning and model-based methods for spectral decomposition reduce noise amplification and cross talk among basis images (such as water, iodine, calcium, and high-Z agents), yielding more reliable iodine maps, VNC images, and parametric images for tumor perfusion and blood-brain barrier assessment (84–86).

Conclusion

In summary, while spectral CT and PCD CT offer advantages in tissue characterization and artifact reduction, widespread clinical implementation remains limited by technical challenges related to detector physics, image reconstruction, hardware complexity and cost, and workflow integration. Continued advances in detector technology, reconstruction algorithms, and clinical validation are needed to support broader adoption of these promising techniques in neuroimaging.

Disclosures of conflicts of interest

A.S. is a deputy editor for Radiology: Imaging Cancer. J.R.B. is an associate editor for Radiology. Please see ICMJE form(s) for author conflicts of interest. These have been provided as supplemental materials.