Carpal Tunnel Syndrome Diagnosis: A Narrative Review of Complementary Roles of Neuromuscular Ultrasound and Electrodiagnostic Studies

Introduction

Carpal tunnel syndrome (CTS) is the most common entrapment neuropathy of the upper extremities, involving compression of the median nerve within the carpal tunnel, which manifests as hand pain, numbness, paresthesia, and weakness [1,2]. The prevalence of CTS in the general population ranges from 3.8% to 7.8% [2,3], with some studies reporting rates as high as 16% [4]. It is more prevalent among females, particularly those over 50 years of age [2,5]. This condition significantly impairs patients’ quality of life and work performance while imposing a considerable burden on healthcare systems. Consequently, early and accurate diagnosis is crucial for timely intervention and enhanced patient outcomes.

Diagnosis of CTS typically begins with a comprehensive medical history and physical examination, including assessment of clinical signs such as Phalen’s and Tinel’s signs [1,6]. To improve diagnostic standardization, validated clinical prediction rules such as the CTS-6 score – which evaluates numbness, paresthesia, thenar atrophy, and two-point discrimination – are also frequently utilized. However, relying exclusively on clinical evaluation may impede the differentiation of CTS from other causes of hand symptoms, including cervical spondylosis, thoracic outlet syndrome, or other wrist nerve entrapments [1]. Furthermore, the nonspecific nature of clinical symptoms and variability in patients’ subjective experiences complicate the diagnostic process [7,8]. For example, a study documented a 22% discrepancy between diagnoses of mild-to-moderate idiopathic median neuropathy at the carpal tunnel (IMNCT) based on signs and symptoms and those obtained from electrodiagnostic studies (EDS) and ultrasound (US) [7]. This discrepancy underscores diagnostic uncertainty, particularly in surgical candidates, and underscores the necessity for additional testing [7]. Another retrospective study, involving 479 hands, found that 13% of cases clinically diagnosed as CTS were reclassified as non-CTS after electrodiagnostic testing (EDX), further illustrating the importance of supplementary diagnostic tools [8].

EDX, particularly nerve conduction studies (NCS), has traditionally been considered the standard method for diagnosing CTS [1,6]. EDX assesses the extent of median nerve damage by measuring functional impairments. However, it is constrained by its invasiveness, time requirements, operator dependency, and potential for false-negative results in early or mild cases [6,9]. In recent years, neuromuscular ultrasound (NMUS), a non-invasive, real-time, portable, and cost-effective imaging technique, has gained prominence in the diagnosis of CTS [10–14]. NMUS enables direct visualization of morphological changes in the median nerve, such as increased cross-sectional area (CSA), alterations in echogenicity, and abnormal dynamic movement, thereby providing essential supplementary diagnostic information [11,12].

This narrative review synthesizes the evidence regarding the diagnostic utility of NMUS and EDX studies in CTS. Although EDX remains the reference standard for functional assessment and severity grading (eg, distal motor latency), its invasiveness is a notable limitation. In contrast, NMUS provides superior detection of anatomical abnormalities (eg, median nerve cross-sectional area [CSA] ≥10 mm2), but with operator dependency. Rather than serving as competing modalities, these 2 techniques exhibit complementary strengths: integrating EDX’s electrophysiological precision with NMUS’s anatomical resolution yields the highest diagnostic accuracy. Emerging trends include the incorporation of artificial intelligence (AI) and the establishment of standardized multimodal protocols to address existing implementation barriers.

This review delineates the synergistic roles of NMUS and EDX in the diagnosis of CTS, with the overarching goal of optimizing patient management. While EDX delivers essential functional information for severity stratification, NMUS offers critical anatomical visualization for identifying structural causes (eg, space-occupying lesions such as cysts or tumors) and facilitating image-guided interventions. Tables 1–3 collectively substantiate the core argument of this review: Table 1 compiles representative median nerve cross-sectional area (MN-CSA) diagnostic thresholds across heterogeneous populations, thereby emphasizing the variability of sonographic cutoff values and the need for population- and context-specific interpretation; Table 2 evaluates the diagnostic performance of principal EDX and NMUS parameters, highlighting their complementary diagnostic value; and Table 3 synthesizes these findings into a practical, stepwise diagnostic algorithm to support integrated clinical decision-making. Furthermore, the present review examines emerging developments, notably the application of AI to ultrasound image analysis and the formulation of standardized multimodal diagnostic protocols, with particular attention to their capacity to establish consistent decision thresholds.

Literature Search Strategy

A targeted narrative literature review was performed to synthesize evidence regarding the diagnostic roles of NMUS and EDX in CTS. The electronic databases searched included PubMed/MEDLINE, Scopus, Web of Science, and Google Scholar, spanning from database inception to October 15, 2025. No initial date restrictions were imposed, thereby permitting inclusion of seminal studies; nevertheless, contemporary evidence published between 2020 and 2025 was prioritized when summarizing technology-dependent parameters (eg, diagnostic thresholds, shear-wave elastography, and advanced Doppler techniques) to reflect current imaging capabilities and clinical practice [15,16]. Seminal studies published before 2020 were selectively included if they provided direct foundational support for contemporary practice, such as early validations of cross-sectional area measurements [17–19].

The search strategy integrated Medical Subject Headings (MeSH) terms with free-text keywords. A representative Boolean search string employed across all databases was: (“carpal tunnel syndrome” OR “median neuropathy”) AND (“neuromuscular ultrasound” OR “musculoskeletal ultrasound” OR “sonography” OR “NMUS”) AND (“electrodiagnosis” OR “nerve conduction studies” OR “electromyography” OR “EMG” OR “EDX”). Additional terms such as “cross-sectional area”, “wrist-to-forearm ratio”, “Doppler vascularity”, and “elastography” were incorporated as required to enhance search precision.

Database searches were complemented by hand-searching the reference lists of key articles, systematic reviews, and meta-analyses to identify additional pertinent publications. Inclusion was limited to full-text articles published in English that demonstrated clinical relevance to CTS diagnosis, methodological rigor, explicit description of ultrasound acquisition protocols, and employment of validated reference standards. This language restriction was implemented to facilitate accurate extraction of technical parameters, ensure consistency with international consensus guidelines, and to maintain practicality for a narrative review; nonetheless, it may have introduced language bias by potentially excluding relevant data from non-English-speaking regions.

Studies were selected according to their contributions to elucidating the complementary roles of NMUS and EDX, with particular emphasis on diagnostic accuracy, anatomical visualization, and emerging applications (eg, artificial intelligence-assisted analysis).

Application of EDX in CTS Diagnosis

PRINCIPLES AND METHODOLOGY OF EDX:

In clinical practice, NCS facilitate the diagnosis of CTS by revealing focal slowing of median sensory conduction across the carpal tunnel, typically demonstrated by comparison with ulnar or radial sensory responses. Three groups of parameters are particularly valuable for clinical decision-making: (1) prolongation of sensory latency or reduction in sensory conduction velocity, which supports diagnosis in mild cases; (2) distal motor latency together with sensory nerve action potential (SNAP) and compound muscle action potential (CMAP) amplitudes, which allow grading of physiologic severity and provide prognostic information; and (3) comparative studies with adjacent nerves, which help localize focal entrapment and exclude generalized polyneuropathy [20–23].

EMG is not required for all patients. It is most informative when severe CTS is suspected (eg, weakness/thenar atrophy), when prognosis is being considered before surgery, or when alternative diagnoses (eg, radiculopathy or plexopathy) need to be excluded. In these settings, EMG evidence of denervation/axonal loss supports more advanced disease and can guide the urgency and expectations of intervention [1,6]. Figure 1 summarizes typical EMG acquisition and representative waveform changes used for electrophysiologic severity grading in CTS.

DIAGNOSTIC CRITERIA AND SEVERITY GRADING WITH EDX:

By quantifying nerve functional deficits, EDX enable objective severity grading of CTS, which is crucial for informing treatment decisions [20,21,24,25]. CTS severity is typically classified as follows: (1) Mild: Primarily characterized by abnormalities in sensory nerve conduction, such as prolonged sensory latency or reduced SCV, with motor conduction remaining normal or minimally affected; (2) Moderate: Involving clear abnormalities in both sensory and motor conduction, including prolonged DML, reduced SCV, and possible reductions in SNAP and CMAP amplitudes; and (3) Severe: Featuring profound impairments in both sensory and motor conduction, which can include the absence of sensory nerve action potentials, significantly prolonged distal motor latency, markedly reduced CMAP amplitude, and evident denervation potentials, along with chronic neurogenic changes on EMG.

Recent studies have demonstrated that comparative indices (such as median–ulnar or median–radial latency differences and their derived parameters) enhance diagnostic sensitivity in cases where routine NCS yield borderline results, whereas motor latency and amplitude-based measures more reliably reflect advanced compression and predict prognosis [20,21]. Given that many reported indices are mathematically interdependent and provide overlapping information, we present representative performance estimates in Table 2 and underscore the importance of selecting the smallest set of parameters capable of altering clinical management (ie, confirming the diagnosis, grading severity, or excluding mimickers).

ADVANTAGES AND LIMITATIONS OF EDX:

As the established standard for diagnosing CTS, EDX offers 3 key advantages. (1) Objective quantification of functional deficits: EDX provides objective metrics, such as nerve conduction velocity, latency, and amplitude, that directly reflect functional damage to the median nerve and aid in grading severity and predicting prognosis [1,20,21]. (2) Differential diagnostic capability: In addition to diagnosing CTS, EDX can distinguish it from other neuropathies causing hand symptoms, including cervical radiculopathy, brachial plexopathy, and ulnar nerve entrapment, thereby preventing misdiagnosis and inappropriate treatment [1,6,26]. For instance, Heard et al (2024) [26] highlighted the role of serial EDX in clarifying diagnosis and prognosis, particularly for reevaluating lesions, confirming progressive disease, and establishing post-injury prognosis. (3) Exclusion of systemic diseases: In certain cases, EDX can identify or rule out systemic neuropathies, such as diabetic peripheral neuropathy, which is essential for developing comprehensive treatment strategies [23,27].

However, EDX also has several notable limitations. (1) Invasiveness and discomfort: NCS involves electrical stimulation, and EMG requires needle insertion, which can cause pain and discomfort to patients, potentially reducing compliance in some cases [6]. (2) Operator dependency: The quality of EDX results depends heavily on the operator’s expertise and skill, which can lead to inter-operator variability[28]. (3) Limited sensitivity in early or mild cases: In early or mild CTS, functional impairments may not exceed electrophysiological detection thresholds, resulting in false-negative findings [6,7]. One study found that 28.8% of patients with clinically suspected CTS had normal EDX results despite presenting with sensory symptoms, indicating that EDX can fail to detect all neuropathies in certain scenarios [28]. (4) Lack of morphological information: EDX evaluates only nerve function and does not allow direct visualization of morphological changes in the median nerve, surrounding tissues, or space-occupying lesions, thus limiting its utility in differential diagnosis [6]. (5) Cost and accessibility: EDX equipment is expensive and requires specialized electrophysiologists and neurologists for operation and interpretation, which may restrict its use in resource-limited settings [6]. (6) Discordance with clinical symptoms: As previously noted, the severity determined by EDX may not align with patient-reported symptoms and functional scores, suggesting that relying solely on EDX may not fully capture the extent of pain and functional limitations [24,25]. (7) Controversy in surgical decision-making: Some studies question the necessity of EDX for surgical decisions in typical CTS cases. A randomized controlled trial by Sucu et al (2023) [9] showed no significant differences in postoperative outcomes for patients with typical CTS, regardless of whether EDX was performed. This suggests that EDX may not be essential for surgical decisions in cases with highly typical symptoms, although it is still valuable for atypical presentations. Furthermore, Bembenek et al (2023) [29] found that, in 454 EDX evaluations, 32.4% revealed new diagnoses or additional injuries, while 25.1% were normal, underscoring discrepancies between initial clinical diagnoses and EDX findings and stressing the importance of thorough interviews and physical examinations prior to EDX.

Although EDX remains essential for diagnosing CTS, its inherent limitations prompt clinicians and researchers to seek more comprehensive and user-friendly adjunctive tools. This shift has promoted the development and adoption of NMUS as a key complement to EDX.

Application of NMUS in CTS Diagnosis

L2PRINCIPLES AND METHODOLOGY OF NMUS:

NMUS is performed with a high-frequency linear transducer to image the median nerve at standardized landmarks, most commonly at the carpal tunnel inlet (pisiform level). Clinically useful outputs include MN-CSA, qualitative echotexture, and identification of local structural contributors (eg, tenosynovitis, cysts, anomalous muscles, bifid nerve). Dynamic assessment can demonstrate restricted nerve gliding, and Doppler or elastography can provide adjunct information in selected settings [11–14].

Ultrasound has limited capability for evaluating intra-osseous pathology. Accordingly, radiography/CT/MRI may be considered in selected atypical scenarios (eg, trauma, suspected inflammatory arthropathy, or other non–soft-tissue mechanisms of tunnel narrowing), rather than as part of routine CTS evaluation [16,30].

KEY PARAMETERS OF NMUS:

NMUS evaluates several parameters in the diagnosis of CTS, the most prominent of which is the cross-sectional area of the median nerve (MN-CSA).

MN-CSA:

The cross-sectional area of MN-CSA at the carpal tunnel inlet (pisiform level) remains the most widely accepted and commonly utilized sonographic parameter for diagnosing CTS. In routine clinical practice, an inlet MN-CSA of approximately 10 mm2 (or a locally validated threshold) is generally considered supportive of CTS in patients with compatible clinical symptoms. Although MN-CSA tends to increase with greater disease severity, substantial overlap among severity categories limits its utility as a standalone indicator of severity. Consequently, when treatment decisions depend on objective physiological severity assessment, EDX studies remain indispensable. Given that diagnostic thresholds vary considerably according to population characteristics and measurement site, representative cutoff values and diagnostic performance metrics from key studies are summarized in Table 1. This section of our report focuses on the role of MN-CSA in clinical screening, diagnostic confirmation, and triage to electrodiagnostic testing for functional severity grading [16,31–36].

WRIST-TO-FOREARM RATIO (WFR):

When absolute CSA values fall within a borderline range or when patient-specific factors (eg, body size, sex, or age) render fixed cutoffs less reliable, relative indices such as WFR or delta CSA (wrist minus forearm) are employed to account for anthropometric variation. Commonly accepted supportive thresholds include a WFR exceeding approximately 1.4 to 1.7 and/or a delta CSA ≥2 mm2, although these values can vary depending on the specific protocol and study population. In clinical practice, these relative measures are particularly valuable for screening purposes and for resolving discrepancies between symptoms and absolute CSA measurements, thereby guiding the decision to proceed to EDX when functional confirmation or severity grading is required [18,31,36,37].

OTHER MORPHOLOGICAL CHANGES:

Several additional morphological features, including reduced echogenicity, nerve flattening, and bulging of the flexor retinaculum, can provide supportive evidence for the diagnosis; however, these signs are less standardized and less reproducible than CSA measurements. Their primary clinical utility lies in the identification of secondary causes or potential mimics of CTS (eg, tenosynovitis, bifid median nerve, persistent median artery, or space-occupying lesions), which can modify the differential diagnosis, facilitate precise targeting of therapeutic injections, or inform surgical strategy [11–14,38,39].

DYNAMIC ULTRASOUND ASSESSMENT:

Dynamic ultrasound visualizes median nerve and tendon motion during wrist and finger movement. In CTS, restricted gliding or abnormal displacement patterns may suggest mechanical tethering, tenosynovitis, or secondary causes that are not fully captured by static imaging. Because quantitative dynamic metrics remain heterogeneous and are not yet standardized, we present dynamic ultrasound as an adjunct for selected clinical questions (eg, suspected mechanical entrapment or space-occupying lesions) rather than a routine requirement [5,11,39–42].

ADVANCED ULTRASOUND TECHNIQUES:

Although B-mode ultrasonography remains the primary screening modality, adjunct techniques such as Doppler ultrasonography and shear-wave or strain elastography can provide additional value in selected equivocal or complex cases by demonstrating intraneural hypervascularity or tissue stiffness. These adjunct findings can assist in refining severity stratification when clinical symptoms and conventional parameters are discordant. However, substantial inter-protocol variability currently restricts their routine application; therefore, we recommend reserving these advanced modalities for addressing specific, targeted questions rather than employing them as first-line screening tools [43–47].

ADVANTAGES AND LIMITATIONS OF NMUS:

NMUS offers 6 advantages in the diagnosis of CTS. (1) Noninvasiveness and Safety: Ultrasound is free of radiation, non-invasive, painless, and highly repeatable, making it suitable for pregnant women (who have a higher prevalence of gestational CTS) [3,12,14,37,48,49]. (2) Real-Time Dynamic Assessment: Ultrasound allows real-time visualization of the morphology and gliding of the median nerve in various positions and during movements, which is crucial for identifying dynamic entrapment and evaluating neural mechanics [11,13,39–42]. (3) Cost-Effectiveness and Accessibility: Compared to MRI and CT, ultrasound equipment is more affordable and portable, facilitating bedside or outpatient examinations and thereby enhancing accessibility [12–14,30,50]. (4) Direct Anatomical Visualization: Ultrasound reveals swelling, flattening, and changes in echogenicity of the median nerve, as illustrated in Figure 2, as well as abnormalities in adjacent tissues (eg, tenosynovitis or space-occupying lesions such as cysts or gouty tophi) [11,12,39,51]. This aids in differential diagnosis. For example, Wang et al (2025) [35] demonstrated that ultrasound is superior to MRI in visualizing variations in the recurrent motor branch (RMB) of the median nerve, with RMB diameter serving as a reliable diagnostic marker for CTS. (5) Monitoring and Follow-Up: Ultrasound can monitor morphological changes in the median nerve after treatment for CTS, such as reductions in CSA following carpal tunnel release, which correlate with symptomatic improvement and EMG recovery [16]. (6) Adjunctive Diagnostic Tool: In cases with ambiguous clinical symptoms and electrophysiological findings, B-mode ultrasound serves as a valuable supplement [12].

However, NMUS also has 5 limitations: (1) Operator Dependency: The acquisition and interpretation of ultrasound images depend heavily on the operator’s expertise, which can lead to inter-operator variability in measurements and compromise standardization and reproducibility [11,16,52–55]. (2) Lack of Standardized Protocols: Although CSA has an established role in diagnosis, protocols for measurement sites, cutoff values, and other parameters need further refinement, limiting the comparability of results across studies [11,16,34]. (3) Limited Sensitivity in Early or Mild CTS: While ultrasound can detect morphological changes, early or mild CTS may not exhibit obvious nerve swelling, reducing its sensitivity [56]. For example, Fryźlewicz et al (2025) [56] found no significant differences in nerve diameter at the site of maximum stenosis (DMS) between patients with mild CTS and controls, suggesting that it is more applicable to advanced CTS. (4) Inability to Assess Nerve Function Directly: Unlike EDX, ultrasound primarily provides morphological data and cannot quantify impairments in nerve conduction, thereby limiting a comprehensive assessment of neuropathy. (5) Image Quality Limitations: Factors such as patient body habitus, probe type, and technique can impair ultrasound image quality, sometimes obscuring deep or complex anatomical structures.

Despite these limitations, the unique advantages of NMUS make it essential in the diagnosis of CTS. The integration of NMUS with EDX has the potential to address their respective limitations, providing more comprehensive and accurate diagnostic information.

Comparative Utility and Integrated Clinical Decision-Making

Building on the modality principles outlined in the Introduction, this section focuses on clinical sequencing and complementarity. Table 2 summarizes comparative diagnostic performance, and Table 3 operationalizes these data into a stepwise diagnostic workflow; accordingly, neither modality is universally superior, and integrated use optimizes diagnostic yield.

To emphasize real-world decision-making (rather than reiterating test mechanics), we outline 4 common clinical decision nodes that determine when to prioritize NMUS, EDX, or both (see Table 3), while referencing the performance trade-offs summarized in Table 2.

Taken together, these decision nodes are summarized in the stepwise diagnostic workflow in Table 3, which translates complementary NMUS and EDX findings into a pragmatic strategy for atypical, early, or complex CTS presentations.

Integrated Diagnostic Strategies and Future Trends

EVIDENCE AND GUIDELINE RECOMMENDATIONS FOR INTEGRATED DIAGNOSIS:

A growing body of evidence supports the combined use of EDX and NMUS, particularly in complex or atypical cases or those with inconclusive results from a single modality [1,6,7,58,59].

INTEGRATION OF EMERGING TECHNOLOGIES AND AI:

Technological advancements are refining CTS diagnosis through emerging technologies, particularly the application of AI) in ultrasound image analysis, which foreshadows significant changes in future diagnostic paradigms.

ELASTOGRAPHY:

SWE, a technique for assessing nerve stiffness, shows considerable potential in the diagnosis and severity grading of CTS [45–47,57,61]. In a prospective study, Kim et al (2025) [46] reported that combining MN-CSA with elasticity parameters (CSA×elasticity) achieves superior diagnostic performance (AUC 0.91) compared to individual parameters and allows for differentiation of CTS severity levels. El-Maghraby et al (2024) [47] confirmed that metrics integrating B-mode ultrasound and SWE outperform EDX in assessing CTS severity. These studies demonstrate that SWE provides biomechanical insights into the nerve, adding new dimensions to CTS diagnosis and severity assessment.

AI AND DEEP LEARNING (DL):

AI and DL technologies have made significant advances in medical image analysis, providing solutions to reduce operator dependency in ultrasound examinations and enhance diagnostic objectivity [55,62–65].

FURTHER DEVELOPMENT OF DYNAMIC ULTRASOUND:

Dynamic ultrasound can assess median nerve motion during wrist and finger movement and can reveal restricted gliding or tendon-related mechanisms not captured by static imaging. Although promising, dynamic metrics are not yet standardized across devices and protocols; we therefore describe dynamic assessment as an adjunct for selected cases (eg, suspected mechanical tethering or secondary CTS) rather than a routine requirement [5,38–42,49].

SUPPLEMENTARY ROLE OF OTHER IMAGING MODALITIES:

Although this review focuses on EDX and NMUS, other modalities such as magnetic resonance neurography (MRN) provide supplementary value in selected complex or atypical CTS cases. MRN offers broader anatomical perspectives and assesses nerve microstructural changes (eg, through diffusion tensor imaging (DTI) and T2 mapping) [35,73–75]. Wang et al (2025) [73] showed that integrating morphological measurements with multiparametric MRN (including DTI and T2 mapping) enhances CTS diagnosis and assessment, improving diagnostic performance (AUC 0.922). However, MRN involves higher costs and limited accessibility, typically serving as a second-line tool following EDX and NMUS.

STANDARDIZATION AND TRAINING:

Diagnostic reliability for both EDX and NMUS is inherently influenced by operator experience and technique. Consequently, the rigorous establishment and dissemination of standardized operational protocols, along with improved clinician training, is essential for maximizing the efficacy of these modalities in clinical practice [10,11,16]. Such protocols should include uniform measurement methods, diagnostic thresholds, and reporting templates to reduce inter-operator variability and enhance the reproducibility and comparability of results. Gasemaltayeb et al (2023) [10] highlighted the efficacy of HRUS in assessing neuromuscular diseases while emphasizing the need for additional data to determine its impact on diagnosis, clinical decision-making, and patient management, as well as its integration with EDX. Battaglia et al (2025) [16] advocated the development of standardized protocols to promote the broader clinical adoption of high-resolution ultrasound in CTS diagnosis and monitoring.

OUTLOOK FOR INTEGRATED DIAGNOSIS:

Future CTS diagnosis will evolve toward multimodal, intelligent, and personalized approaches.

Through these integrated diagnostic strategies and upcoming technological developments, CTS diagnosis will achieve greater precision, objectivity, and efficiency, ultimately leading to improved treatment outcomes and enhanced quality of life for patients.

Conclusions

Reliance on a single modality is frequently insufficient to comprehensively characterize suspected CTS or to determine the optimal treatment pathway. Rather than regarding EDX and NMUS as competing modalities, this narrative review underscores their complementary roles: EDX provides objective functional confirmation and severity grading indispensable for prognosis, whereas NMUS affords anatomical characterization essential for surgical planning, including the identification of structural contributors or anatomical variants. When employed in combination as clinically indicated, these modalities enhance diagnostic confidence, facilitate the interpretation of discordant findings, and help prevent misdiagnosis in atypical or complex presentations. Future research should focus on reducing operator dependency, standardizing protocols, and responsibly integrating artificial intelligence-assisted tools and training. As this is a narrative synthesis without formal evidence grading, the presented statements aim to summarize the existing literature and provide pragmatic considerations rather than serve as formal guidelines.