Abstract
Background: Ultrasound (US) measurement of cross-sectional area (CSA) of the median nerve has emerged as a viable alternative to electromyography/nerve conduction studies (EMG/NCS) for diagnosis of carpal tunnel syndrome (CTS). The purpose of this study is to compare CSA of the median nerve between US and MRI using current MRI and US technology. The null hypothesis is there is no difference between US and MRI CSA measurements. Methods: The study design was an observational cohort, enrolling patients presenting to clinic with MRI of the wrist. Participants with clinical signs and symptoms of CTS were excluded. The CSA measurements of the median nerve on MRI T1-weighted axial images were performed by a hand fellow blinded to results of US measurements, and US measurement of median nerve CSA was performed by a hand fellowship trained surgeon blinded to results of the MRI measurements. Results were analyzed via percent error, Pearson correlation, and t tests. Results: Twenty participants were enrolled with mean age of 29.4 years. Four left wrists and 16 right wrists were measured. The US mean CSA of the median nerve was 6.8 mm2 (±2.330 mm2). The MRI mean CSA of the median nerve was 6.8 mm2 (±2.153 mm2), P = .442. Pearson correlation between modalities was 0.93, suggesting near-perfect correlation. Mean percent error was 8.8%. Conclusions: Results of this study suggest that US is an accurate method to measure CSA of the median nerve at the carpal tunnel inlet. The mean difference between US and MRI was unlikely to be clinically significant.
Keywords: MRI, carpal tunnel, wrist, median nerve, ultrasound
Introduction
Carpal tunnel syndrome (CTS) is a common upper extremity compression neuropathy among adults, present in approximately 1.5% of the population.1,5 After clinical diagnosis, electrodiagnostic testing (a combination of nerve conduction studies [NCS] and electromyography) is commonly used to confirm the diagnosis.17 However, NCS can be invasive, expensive, and time-consuming. Previous studies have shown that other modalities, in particular musculoskeletal ultrasound (US), may offer similar diagnostic accuracy.6 Compression of the median nerve within the carpal tunnel results in nerve swelling proximal and distal to the site of compression. Measurement of the cross-sectional area (CSA) of the median nerve at the level of the pisiform has been an accurate and reproducible test for diagnosis of CTS.8,13 In a prospective cohort, US was found to have a higher specificity (94%) for the diagnosis of CTS when compared with NCS (83%).6 At a prevalence of 6%, US has been shown to be a better confirmatory test for CTS with a very high negative predictive value (99%).6
Buchberger et al used US and MRI to measure the CSA of the median nerve in 28 normal wrists and in 20 wrists with CTS confirmed by NCS.2 The authors found that the mean CSA of the median nerve at the pisiform was significantly greater in wrists with CTS compared with normal wrists.2 The US measurements of the CSA of the median nerve in the 20 wrists with CTS were compared with MRI measurements of the same wrists and the authors found no significant difference between measurements. However, Buchberger et al did not specifically address the mean differences between US and MRI measurements and the strength of correlation was not reported. In addition, although Buchberger et al suggested that MRI has superior diagnostic capability for CTS (including the ability to visualize milder conditions and identify etiologic findings), other studies have found that US is more sensitive than MRI (93% vs 67%) and has equivalent specificity (86%) in imaging nerve pathology when compared with diagnoses made by clinical evaluation.2,18 Other research suggests that depending on the clinical question, MRI or US may be the modality of choice when imaging peripheral nerves.9,15,18 Ultrasound is readily available, quick, noninvasive, and inexpensive. There are fewer contraindications for its use and it is already a complement for neurophysiological assessment in routine practice.16
Although US measurements of peripheral nerves have been validated in other areas of the body, it is important to validate measurement of the CSA of the median nerve at the carpal tunnel inlet if these measurements are to be used to confirm a diagnosis of CTS. The null hypothesis is that there is no significant difference between US and MRI measurements of the median nerve CSA at the level of the pisiform.
Materials and Methods
The study was designed as an observational cohort. Institutional Review Board approval was obtained prior to initiation of the study. We included patients older than the age of 12 who presented to the office with an MRI of the wrist, which included a T1-weighted image of the median nerve at the level of the pisiform. The following exclusion criteria were utilized: (1) MRI was greater than 2 weeks old (to ensure minimal if any changes in nerve CSA between measurements); (2) MRI did not allow accurate measure of CSA due to positioning or artifact; (3) participant showed clinical signs and symptoms of CTS; (4) MRI was ordered for evaluation of the median nerve and/or flexor tendon injury; (5) pregnancy. No patients refused the study.
Measurements of the CSA of median nerve on MRI T1-weighted axial images of the wrist were performed by a hand surgery fellow blinded to the results of the US measurements (Figure 1). The short axis and long axis of the median nerve were measured and CSA was calculated from these measurements using the formula for area of an ellipse (a × b × π). Participants underwent US measurements of the median nerve of the same wrist at the level of the pisiform by a different fellowship trained hand surgeon who was blinded to the results of the MRI measurements. The US measurements were performed within 2 weeks of the MRI to ensure no interval change in nerve CSA. The US measurements (Figure 2) were performed using a 15-6 MHz linear array transducer (SonoSite M Turbo; SonoSite, Bothell, Washington). Participants were seated with their forearm resting on the table, and wrist and fingers in neutral position with mild flexion at the metacarpophalangeal and proximal interphalangeal joints. The US probe was positioned perpendicular to the long axis of the forearm and the CSA of the median nerve was measured just proximal to the level of the pisiform. The electronic ellipse function was used.7,8 The borders of the median nerve were defined as the area within the hyperechoic epineurium.14
Figure 1.
An example of the measurement of the cross-sectional area of median nerve on magnetic resonance imaging T1-weighted axial images of the wrist using long axis B and short axis A.
Figure 2.
An example of ultrasound measurement of the median nerve using a 15-6 MHz linear array transducer (SonoSite M Turbo; SonoSite, Bothell, Washington). The electronic ellipse function was used and borders of the median nerve were defined as the area within the hyperechoic epineurium denoted by letter A.
Figure 3.
US versus MRI CSA. Twenty measurements of CSA of the median nerve at the level of the pisiform. Trend-line demonstrates correlation of 0.93.
Note. US = ultrasound; MRI = magnetic resonance imaging; CSA = cross-sectional area.
Table 1.
Descriptive Statistics.
| Magnetic resonance imaging | Ultrasound | |
|---|---|---|
| Mean | 6.828047283 | 6.8 |
| Standard error | 0.48144143 | 0.52113237 |
| Median | 5.963827186 | 6 |
| Mode | NA | 5 |
| Standard deviation | 2.153071531 | 2.330574811 |
| Sample variance | 4.635717016 | 5.431578947 |
| Kurtosis | 0.734946217 | 1.305298395 |
| Skewness | 1.006919149 | 1.266170208 |
| Range | 8.252211349 | 9 |
| Minimum | 4 | 4 |
| Maximum | 12.25221135 | 13 |
| Sum | 136.5609457 | 136 |
| Count | 20 | 20 |
Table 2.
Paired Two-Sample t test.
| Magnetic resonance imaging | Ultrasound | |
|---|---|---|
| Mean | 6.828047283 | 6.8 |
| Variance | 4.635717016 | 5.431578947 |
| Observations | 20 | 20 |
| Pearson correlation | 0.930982491 | |
| Hypothesized mean difference | 0 | |
| df | 19 | |
| t statistics | 0.147397717 | |
| P(T ≤ t) one-tail | 0.442185499 | |
| t critical one-tail | 1.729132812 | |
| P(T ≤ t) two-tail | 0.884370998 | |
| t critical two-tail | 2.093024054 |
Descriptive statistics were computed. Difference and percent error between US and MRI measurements of CSA of the median nerve were calculated for each participant and mean percent error was calculated. A Pearson correlation coefficient was calculated based on the results of comparing CSA of median nerve on MRI versus US, and a paired two-sample t test was completed.
Results
Twenty participants with a mean age of 29 years (range: 14-57 years) were enrolled. Four left wrists and 16 right wrists were measured. Participants carried the following diagnoses: seven with triangular fibrocartilage complex (TFCC) tear (with 1 of those 7 additionally having ulnocarpal impaction syndrome), 3 with extensor carpi ulnaris (ECU) tendonitis, 3 with ganglion cysts, 3 had synovitis/arthritis, and 1 each of the following: scaphoid fracture, triquetral contusion, scapholunate ligament tear, De Quervain tenosynovitis, and loose bodies.
The mean CSA of the median nerve using US was 6.8 mm2 (range: 4.0-13.0 mm2, SD: 2.330 mm2) with standard error of 0.521 mm2. The mean CSA of the median nerve using MRI was 6.8 mm2 (range: 4.4-12.3 mm2, SD: 2.153 mm2) with standard error of 0.481 mm.2 The Pearson correlation coefficient between MRI and US measurements of the CSA of the median nerve was 0.93, demonstrating near-perfect correlation.12
Discussion
This study confirms and expands the work of Buchberger and colleagues.2 There is a near-perfect correlation between US and MRI measurements of the median nerve at the inlet of the carpal tunnel. The mean percent error is less than 10%, meaning it is unlikely that measurement error would affect treatment decisions. Although the results of this study are not necessarily novel given the results of Buchberger, this study does confirm those results using contemporary US and MRI technology and for the specific location and specific nerve measured for CTS.
When considering the use of US to measure the CSA of the median nerve and therefore diagnose CTS, it is important to understand whether the measurement itself is accurate. The MRI is widely considered the reference standard for measurement of peripheral nerve swelling and therefore was used as the reference in this study.11 Of particular note, the US examination in the current study was performed in the office by a practicing hand surgeon. This study, therefore, documents that a high level of accuracy can be achieved by a nonradiologist using a commonly available US machine. Other authors have reported both US and MRI measurements of the median nerve at the carpal tunnel. Notably, Horng et al compared the diagnostic accuracy of US and MRI at both the carpal tunnel inlet and outlet using both CSA and also bowing of the flexor retinaculum.10 However, the authors did not discuss the agreement between the US and MRI measurements as was performed in the current study.
It is also important to confirm that US is accurate for the specific location and indication for which it is being used. Cartwright et al found a wide variation in accuracy and reliability of US measurements depending on the structure and site that was being measured.3 The authors found that US measurements of the median nerve at the wrist only had a 0.4 correlation coefficient when compared with direct measurement in a cadaver model. The results of the current study are in stark contrast to the results of Cartwright and colleagues. The reason for the difference is unclear, but it is likely that a cadaver model is suboptimal when comparing a direct measurement technique with US as the tissues themselves may have changes in water content and tissue characteristics that do not mimic the in vivo situation.
This study does have several limitations. First, none of the participants in this study had clinical signs and symptoms of CTS. It is possible, although unlikely, that US measurements of the median nerve in patients with CTS would not correlate with MRI measurements. Future studies could attempt to enroll participants with CTS. Second, it is possible that MRI measurements of peripheral nerves are not accurate. In this scenario, US would be compared with an imperfect reference standard. Another limitation is that the US measurements were performed by a hand surgeon with extensive US experience. It is possible that measurements obtained by less experienced examiners would be less accurate. However, Crasto et al found that even inexperienced examiners were able to obtain measurements with a high level of agreement to an experienced examiner with minimal training.4 Finally, this study only examined 1 specific location for US measurement (the inlet of the carpal tunnel) as this site was previously found to be the most sensitive and specific location for diagnosis of CTS.13 It is possible that correlation of US and MRI would not be similar at other locations around the carpal tunnel (hook of hamate, midpoint carpal tunnel, proximal to carpal tunnel).
In summary, we found that US measurement of CSA of the median nerve at level of the pisiform has near-perfect correlation with the reference standard MRI. This confirms work of prior authors using modern technology, and indicates US is a valuable tool that can be used for accurate measurement of the median nerve.
Footnotes
Ethical Approval: This study was approved by our institutional review board.
Statement of Human and Animal Rights: All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1975, as revised in 2008.
Statement of Informed Consent: Informed consent was obtained from all patients enrolled in this study.
Declaration of Conflicting Interests: The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding: The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Funding for this study was via internal funding and a National Institutes of Health grant (UL1-TR-000005). This study was also made possible by a grant from the American Foundation for Surgery of the Hand (AFSH).
ORCID iD: Beverly Hersh 
https://orcid.org/0000-0001-5709-4237
References
- 1. Atroshi I, Gummesson C, Johnsson R, et al. Prevalence of carpal tunnel syndrome in a general population. JAMA. 1999;282(2):153-158. [DOI] [PubMed] [Google Scholar]
- 2. Buchberger W, Judmaier W, Birbamer G, et al. Carpal tunnel syndrome: diagnosis with high-resolution sonography. AJR Am J Roentgenol. 1992;159(4):793-798. [DOI] [PubMed] [Google Scholar]
- 3. Cartwright MS, Demar S, Griffin LP, et al. Validity and reliability of nerve and muscle ultrasound. Muscle Nerve. 2013;47(4):515-521. [DOI] [PubMed] [Google Scholar]
- 4. Crasto JA, Scott ME, Fowler JR. Ultrasound measurement of the cross-sectional area of the median nerve: the effect of teaching on measurement accuracy. HAND. 2017:1558944717731857. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. De Krom M, Kester A, Knipschild P, et al. Risk factors for carpal tunnel syndrome. Am J Epidemiol. 1990;132(6):1102-1110. [DOI] [PubMed] [Google Scholar]
- 6. Fowler JR, Cipolli W, Hanson T. A comparison of three diagnostic tests for carpal tunnel syndrome using latent class analysis. J Bone Joint Surg Am. 2015;97(23):1958-1961. [DOI] [PubMed] [Google Scholar]
- 7. Fowler JR, Gaughan JP, Ilyas AM. The sensitivity and speci-ficity of ultrasound for the diagnosis of carpal tunnel syndrome: a meta-analysis. Clin Orthop Relat Res. 2011;469(4):1089-1094. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Fowler JR, Munsch M, Tosti R, et al. Comparison of ultrasound and electrodiagnostic testing for diagnosis of carpal tunnel syndrome. J Bone Joint Surg Am. 2014;96(17):e148. [DOI] [PubMed] [Google Scholar]
- 9. Grant GA, Britz GW, Goodkin R, et al. The utility of magnetic resonance imaging in evaluating peripheral nerve disorders. Muscle Nerve. 2002;25(3):314-331. [DOI] [PubMed] [Google Scholar]
- 10. Horng Y-S, Chang H-C, Lin K-E, et al. Accuracy of ultrasonography and magnetic resonance imaging in diagnosing carpal tunnel syndrome using rest and grasp positions of the hands. J Hand Surg Am. 2012;37(8):1591-1598. [DOI] [PubMed] [Google Scholar]
- 11. Jeon T, Fung MM, Koch KM, et al. Peripheral nerve diffusion tensor imaging: overview, pitfalls, and future directions. J Magn Reson Imaging. 2018;47(5): 1171-1189. [DOI] [PubMed] [Google Scholar]
- 12. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics. 1977;33:159-174. [PubMed] [Google Scholar]
- 13. Nakamichi K, Tachibana S. Ultrasonographic measurement of median nerve cross-sectional area in idiopathic carpal tunnel syndrome: diagnostic accuracy. Muscle Nerve. 2002;26(6):798-803. [DOI] [PubMed] [Google Scholar]
- 14. Nakamichi KI, Tachibana S. Enlarged median nerve in idiopathic carpal tunnel syndrome. Muscle Nerve. 2000;23(11):1713-1718. [DOI] [PubMed] [Google Scholar]
- 15. Ochi M, Ikuta Y, Watanabe M, et al. The diagnostic value of MRI in traumatic brachial plexus injury. J Hand Surg Br. 1994;19(1):55-59. [DOI] [PubMed] [Google Scholar]
- 16. Padua L, Liotta G, Di Pasquale A, et al. Contribution of ultrasound in the assessment of nerve diseases. Eur J Neurol. 2012;19(1):47-54. [DOI] [PubMed] [Google Scholar]
- 17. Preston D, Shapiro B. Median neuropathy at the wrist. In Electromyography and Neuromuscular Disorders. 2nd ed. Philadelphia, PA: Elsevier Butterworth-Heinemann; 2005:255-279. [Google Scholar]
- 18. Zaidman CM, Seelig MJ, Baker JC, et al. Detection of peripheral nerve pathology Comparison of ultrasound and MRI. Neurology. 2013;80(18):1634-1640. [DOI] [PMC free article] [PubMed] [Google Scholar]