Comparing the efficacy of multiple quantitative and qualitative ultrasound parameters for the diagnosis of carpal tunnel syndrome

Abstract

Purpose

Carpal tunnel syndrome (CTS) is a compression neuropathy causing significant morbidity. Over the years, ultrasound has been evaluated as an alternative to nerve conduction study (NCS) for diagnosing CTS, however, there is no consensus as to which ultrasound parameter is the best. Our study aimed to determine and compare the efficacy of various ultrasound-based variables for diagnosis of CTS.

Methods

80 patients with clinical suspicion of CTS underwent ultrasound examination with calculation of cross-sectional area (CSA), delta CSA, wrist forearm ratio (WFR), palmer bowing (PB), flattening ratio (FR), flexor retinaculum thickness (FT), and evaluation of echogenicity and vascularity of median nerve. NCS was taken as the gold standard and the diagnostic efficacy of all these variables was compared, followed by receiver operator curve (ROC) analysis.

Results

Delta CSA had the highest accuracy (91.25%), followed by CSAc (80%), WFR (78.75%), and PB (73.75%). Youden’s index and sensitivity were highest for delta CSA (0.783 and 96.15% respectively), while specificity was highest for FT (89.29%). The highest area under the curve was noted for delta CSA (97.1%), followed by WFR (AUC = 87.4%) and CSAc (AUC = 86.0%).

Conclusion

Delta CSA was found to be the best ultrasound parameter for diagnosis of CTS, followed by CSAc, WFR, and PB, and can be used as an alternative to NCS. Using ROC analysis this study also predicted the best cut-off values for these parameters which could improve their diagnostic accuracy and further research is needed to confirm these findings.

Graphical abstract

Introduction

Carpal tunnel syndrome (CTS) is the most common compression neuropathy in the upper limb leading to work disability [1], which occurs due to the entrapment of the median nerve at the level of the wrist [2]. It has a prevalence of 3.8 to 7.8% in the population [3], and is more common in females [4], and the 45 to 60 age-group [5]. The cause of increased carpal tunnel pressure can be idiopathic, or secondary to conditions like- flexor tenosynovitis, ganglion cysts, tumors, etc. [6].

Nerve conduction study (NCS) has long been the gold standard diagnostic measure for CTS [7], however, recipients find it uncomfortable, painful, expensive, time-consuming, and not easily accessible [3]. This led to the evaluation of ultrasound as an alternative diagnostic tool for CTS, as it overcomes these limitations and allows for visual inspection of the median nerve and its surrounding structures [8], helps detect congenital variations, space-occupying lesions, etc. [9]. Ultrasound also triumphs over MRI in being both time and cost-efficient and allows real-time and dynamic assessment of median nerve [10].

Over the last 20 years, many studies have proven that ultrasound has a high accuracy for diagnosing CTS [11, 12], however, all these have used different quantitative and qualitative parameters with different cut-offs, with no single consensus available as to which ultrasound-based parameter gives the best results. In this study we compared the efficacy of six quantitative and two qualitative ultrasound variables for diagnosis of CTS using the cut-offs proposed by the latest studies, taking NCS as the gold standard.

Materials and methods

Study design and participants

This was a prospective, cross-sectional study conducted at a tertiary care center from April 2021 to September 2022, including a total of 80 cases (Fig. 1). In cases where both wrists of a patient were affected, the wrist with more severity of symptoms was included. All adult patients (18 years and above) with clinical suspicion of CTS, were included in the study after taking informed consent. Patients with a history of any previous operation for CTS, any other wrist surgery, local steroid injection received for CTS in the last three months, and known cases of gout and rheumatoid arthritis were excluded.

Fig. 1.

Study flow diagram

Demographic and clinical data collection

A detailed history was taken and the data including name, age, sex, occupation, right/left-handedness, more severely affected hand, duration of symptoms, whether the symptoms were more pronounced at night, and presence of any sensory or motor symptoms was recorded. Sensory symptoms usually precede motor deficit and include pain, tingling, and numbness [13] in median nerve distribution- volar aspect of lateral 3 ½ digits and dorsal aspect of lateral 3 ½ nail beds [14]. Patients with long-standing disease may develop motor weakness with wasting of the muscles of the thenar eminence [15].

Ultrasound equipment and protocol

Real-time grey scale and Doppler ultrasound examination were performed on Canon Aplio a ultrasound machine using 11 MHz and 14 MHz linear probes and a 17 MHz hockey stick probe. Patients were comfortably seated facing the examiner, with their elbows flexed, forearms supinated, wrists extended and fingers partially flexed. The probe was maintained at a perpendicular angle to prevent anisotropy [16].

Qualitative ultrasound parameters

1. Echogenicity of median nerve- Proximal to the site of entrapment, the nerve becomes edematous which is seen as a loss of fascicular pattern and hypoechogenicity [17] (Fig. 2a).

2. Vascularity of median nerve- Compression of the median nerve leads to vascular congestion (Fig. 2b). which is assessed by power Doppler (pulse frequency 6.0 MHz) and is defined as the presence of at least a single neo-vessel of greater than 1 mm in length in the sagittal plane [18, 19].

Fig. 2.

Longitudinal ultrasound image at the level of the wrist in a 32-year-old female with CTS showing median nerve (white arrow) in its sagittal section. Normal fascicular pattern (star) seen proximally in image (a), is lost as the nerve enters the carpal tunnel with associated decrease in its echogenicity (white arrowheads). Increased intraneural vascularity (black arrow) is seen in image (b). C carpals, F flexor tendons, R radius

Quantitative ultrasound parameters

1. Median nerve cross-sectional area (CSA)- In cases of compression neuropathy, there is an enlargement of the nerve proximal to the site of entrapment [20]. Transverse images of the median nerve with CSA measurements were obtained at three levels (Fig. 3)

   • Carpal tunnel inlet (CSAc)- With the external landmark being the distal wrist crease.
   • Pronator quadratus muscle level (CSAp)- The probe is moved 3–4 cm proximal to the distal wrist crease to identify the pronator quadratus muscle with its fibers coursing transversely from distal ulna to radius.
   • Forearm level (CSAf)- 12 cm proximal to the distal wrist crease.
     Measurements were made using a continuous trace method inside the epineurium (in centimeter square). CSAc greater than or equal to 0.10 cm² was considered positive for CTS [21].

2. Delta CSA- Is the difference between CSAc-CSAp. The subject was diagnosed to have CTS if delta CSA was greater than or equal to 0.02 cm² [2].

3. Wrist-forearm ratio (WFR)- Is the ratio of CSAc and CSAf (CSAc/CSAf). A ratio greater than or equal to 1.4 was considered positive for CTS [3, 22].

4. Flexor retinaculum thickness (FT)- The thickness of the flexor retinaculum was measured (in millimeters) as its maximum anteroposterior dimension in the transverse section of the carpal tunnel inlet (Fig. 4a). FT greater than 0.9 mm was taken as positive for CTS [3].

5. Palmer bowing of the flexor retinaculum (PB) (Fig. 4b)- Is calculated as the maximum perpendicular distance between the flexor retinaculum and a line connecting the apices of the scaphoid and pisiform bones. Values were positive when the retinaculum was above the tangential line and negative when it was below. PB of greater than + 2 mm was positive for CTS [6].

6. Flattening ratio (FT)- Using the transverse image of the carpal tunnel inlet- the maximum transverse diameter of the nerve was measured and divided by the maximum anteroposterior diameter (Fig. 4c). A ratio greater than 1.4 was considered positive for CTS [6].

Fig. 3.

Transverse ultrasound images in a 38-year-old female with CTS showing median nerve CSA calculation (blue dashed circle) a At the forearm level- CSAf (0.06 cm²); b At the pronator quadratus muscle level- CSAp (0.06 cm2) and c At the distal wrist crease- CSAc (0.12 cm²); with Delta CSA = CSAc-CSAp = 6 mm² and WFR = CSAc/CSAf = 2, both of which were positive for CTS. Arrowheads flexor retinaculum, Black arrow radial artery, FCU flexor carpi ulnaris, FCR Flexor carpi radialis, FDP flexor digitorum profundus, FDS flexor digitorum superficialis, FPL Flexor pollicis longus, Orange dashed circle pronator quadratus muscle, Ps pisiform, R radius, Sc scaphoid, White solid arrow ulnar artery, White dashed arrow ulnar nerve

Fig. 4.

Transverse ultrasound images of a 42-year-old male patient positive for CTS on NCS, at the level of distal wrist crease, using a linear transducer (a, b) and hockey stick probe (c), showing the evaluation of the hypoechoic flexor retinaculum (white arrow) with the calculation of (a) FT = 0.4 mm (negative for CTS) and (b) PB = 4.4 mm (positive for CTS) and (c) flattening ratio of median nerve (black arrow), FT = major axis (measurement A)/minor axis (measurement B) = 7.8 mm/2.2 mm = 2.9 (positive for CTS). Flex T flexor tendons, Sc scaphoid bone, Ps pisiform bone

Outcome and statistical analysis

NCS is an electrodiagnostic study that evaluates the sensory and motor velocity, latency, and amplitude to help identify sites of abnormality in nerve conduction [23] and was taken as the gold standard. The values of all the ultrasound variables were recorded and labelled as positive or negative for CTS, which was then compared with the NCS results to calculate their sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy. P value < 0.05 was taken as significant. Youden’s index was calculated and receiver operator curve (ROC) analysis was performed.

Results

Demographic characteristics and clinical profile (Table 1)

Table 1.

Frequency distribution of patient characteristics

Variable	Frequency	Percentage (%)
Age (years)
18–35	26	32.5
36–50	27	33.8
> 50	27	33.8
Gender
Male	22	27.5
Female	58	72.5
Occupation
Housewife	43	53.8
Student	9	11.3
Teacher	6	7.5
Others	22	27.5
Hand examined
Right	58	72.5
Left	22	27.5
Diabetes	12	15.0
Hypertension	17	21.3
Hypothyroidism	12	15.0
Clinical profile
Both pain and paraesthesias	35	43.8
Only paraesthesias	34	42.5
Only pain	9	11.2
Motor symptoms	38	47.5
Symptoms more at night	41	51.3
Associated abnormality causing CTS
Detected	25	31.25
Not detected	55	68.75

Among the 80 cases included in our study, 52 cases were positive and 28 cases were negative for CTS (according to NCS results). Almost an equal distribution of patients was noted in all age groups with an average age of 43.1 ± 13.3 years (range- 20–76 years). Most of our subjects were female (72.5%) with a male-to-female ratio measuring 1:2.6.

Mean and standard deviation analysis (Table 2)

Table 2.

Mean and standard deviation analysis

Parameters	NCS result for CTS
	Negative		Positive		Total		T test
	Mean	Standard deviation	Mean	Standard deviation	Mean	Standard deviation	p Value
CSAc (in cm2)	0.08	0.014	0.13	0.084	0.12	0.072	0.004
Delta CSA (in cm2)	0.01	0.006	0.06	0.080	0.04	0.069	0.002
WFR	1.39	0.227	2.35	1.524	2.02	1.315	0.001
PB (in mm)	1.46	0.965	3.10	1.325	2.53	1.439	< 0.001
FR	2.36	0.359	2.74	0.361	2.61	0.402	< 0.001
FT (in mm)	0.40	0.282	0.72	0.354	0.60	0.363	< 0.001

The bold represents the p values less than 0.05 are significant

A statistically significant difference was noted between the mean values of all six quantitative ultrasound variables in CTS positive and negative cases.

Comparison of efficacy of ultrasound parameters for diagnosing CTS (Fig. 5)

Fig. 5.

Graphical representation of sensitivity, specificity, and accuracy of all the ultrasound parameters in comparison to the gold standard (NCS)

• Sensitivity- The highest sensitivity (96.15%) was noted for delta CSA followed by WFR (90.38%) while the least sensitivity was noted for FT- 26.92%.

• Specificity- FT had the highest specificity (89.29%) while FR and hypoechogenicity had the least specificity (53.57%). Delta CSA and CSAc both had second highest specificities (82.14%).

• Accuracy- Delta CSA had the highest accuracy amongst all the variables (91.25%), followed by CSAc (80%) and WFR (78.75%). The lowest accuracy was noted for FT (48.75%). Decreasing order of accuracy- Delta CSA > CSAc > WFR > PB > Hypoechogenicity > FR > Vascularity > FT.

• Youden’s index- It was highest for Delta CSA (0.783) followed by CSAc (0.610), and WFR (0.475) while FT had the least value (0.162).

• Combination of parameters was also assessed as given in Table 3.

Table 3.

Comparison of efficacy of ultrasound-based variables for diagnosing CTS

Test	TP	FP	FN	TN	Sensitivity	Specificity	PPV (%)	NPV (%)	Accuracy (%)	Youden’s index
CSAc	41	5	11	23	78.85%	82.14%	89.13	67.65	80.00	0.610
Delta CSA	50	5	2	23	96.15%	82.14%	90.91	92.00	91.25	0.783
WFR	47	12	5	16	90.38%	57.14%	79.66	76.19	78.75	0.475
PB	40	9	12	19	76.92%	67.86%	81.63	61.29	73.75	0.448
FR	38	13	14	15	73.08%	53.57%	74.51	51.72	66.25	0.266
FT	14	3	38	25	26.92%	89.29%	82.35	39.68	48.75	0.162
Hypoecho-genicity	44	13	8	15	84.62%	53.57%	77.19	65.22	73.75	0.382
Vascularity	27	7	25	21	51.92%	75.00%	79.41	45.65	60.00	0.269
Group 1	49	4	3	24	94.23%	85.71%	92.45	88.89	91.25	0.799
Group 2	45	3	7	25	86.54%	89.29%	93.75	78.12	87.50	0.758
Group 3	48	7	4	21	92.31%	75.00%	87.27	84	86.25	0.673

The bold represents highest values of the efficacy parameters among the ultrasound variables and the groups

Group 1 = Delta CSA, CSAc, WFR (Considered positive when > or = 2 are positive); Group 2 = Delta CSA, CSAc, WFR, PB (Considered positive when > 2 are positive); Group 3 = Delta CSA, CSAc, WFR, PB, Hypoechogenicity (Considered positive when > or = 3 are positive)

FN false negative, FP false positive, TN true negative, TP true positive

ROC analysis

• Area under the curve (AUC) (Fig. 6)- ROC curves for all the quantitative variables were well above the 50% line. AUC was approaching 100% and was found to be the highest for Delta CSA, measuring 97.1% (95% confidence interval: 94.1–100.0, p < 0.001). WFR (AUC = 87.4%) surpassed CSAc (AUC = 86.0%) and had the second-highest AUC value.

• Maximum Youden’s index and new cut-off from ROC curves- ROC analysis was further used to predict the new cut-off values for the quantitative parameters giving the maximum Youden’s index. The predicted new sensitivity and specificity at these cut-offs are also given in Table 4.

Fig. 6.

ROC analysis of quantitative ultrasound-based variables depicting the AUC in decreasing order- a Delta CSA, b WFR, c CSAc, d PB, e FT, and f FR

Table 4.

Predicted new cut-offs for quantitative variables using maximum Youden’s index

	New cut-off	Maximum Youden’s index	Predicted sensitivity	Predicted specificity
Delta CSA	0.025 cm2	0.808	80.77%	100.00%
WFR	1.817	0.615	61.54%	100.00%
CSAc	0.095 cm2	0.610	78.85%	82.14%
PB	2.650 mm	0.599	63.46%	96.43%
FT	0.450 mm	0.538	78.85%	75.00%
FR	2.750	0.503	53.85%	96.43%

Discussion

CTS is the most common compression neuropathy in the upper limb [2]. Even after extensive research, it remains controversial as to which ultrasound parameter and what cut-off should be used for the diagnosis of CTS [24]. This study aimed to fill this lacuna in knowledge by comparing the efficacy of multiple ultrasound parameters for the diagnosis of CTS.

Delta CSA was found to have the highest Youden’s index (0.783), accuracy (91.25%), sensitivity (96.15%), and AUC (97.15%)- making it the best ultrasound parameter for diagnosing CTS. This contrasts with the review conducted by Erickson et al. in 2022 [25], who concluded that the diagnostic accuracy of CSA at the carpal tunnel inlet is the highest. ROC analysis indicated that the accuracy of delta CSA can be further improved by increasing its cut-off to 2.5 mm² (Maximum Youden’s index = 0.808). This is in concordance with the results of the study conducted in 2019 in Saudis [26], where delta CSA (cut-off 2.5 mm²) was found to have the highest sensitivity and specificity for diagnosing CTS.

In our study, CSAc, using a cut-off of 10 mm² as used by Chen et al. in 2022 [27], showed the second-highest Youden’s index (0.610) and accuracy (80.00%) making it the second-best ultrasound parameter for diagnosis of CTS. In 2014, Fowler et al. [21] using a cut-off of  ≥ 10 mm² found CSAc to have a sensitivity of 89% and a specificity of 90%. On using this cut-off for our sample, CSAc showed a relatively lower sensitivity (78.85%) and specificity (82.14%) for diagnosis of CTS. ROC analysis suggested that decreasing its cut-off to 9.50 mm² could further improve its accuracy which agrees with the study by Kang et al. [28].

Using a cut-off of 1.4, Mhoon et al. in 2012 [22] found WFR to have very high sensitivity for diagnosing CTS. In our sample, it had the third-highest Youden’s index (0.475) and accuracy (78.75%), making it the third-best diagnostic variable. But, in ROC analysis, it had the second highest AUC (87.40%), and changing its cut-off to 1.82 improved its Youden’s index to 0.615, surpassing CSAc. Ng AWH et al. [6] demonstrated that using a cut-off for PB of greater than 2 mm at the tunnel inlet resulted in a high diagnostic accuracy of 73.7%. In our study, PB showed a similar accuracy (73.75%) and the fourth-highest Youden’s index (0.448).

FR ranked sixth in accuracy (66.25%) and seventh in Youden’s index (0.266) using a cut-off = 2.5. On ROC analysis FR showed a high AUC = 77.47%, and the best cut-off was 2.75. FT had the lowest accuracy (48.75%) and Youden’s index (0.162), showing no association with NCS results (cut-off > 0.9 mm). However, on ROC analysis, it had an acceptable AUC (78.06%), and it was suggested that decreasing its cut-off to 0.45 mm, could increase its Youden’s index to 0.538, which could improve its diagnostic efficacy.

Hypoechogenicity of the median nerve had good sensitivity (84.62%), but the lowest specificity (53.57%). This may be because it is a subjective assessment, is prone to angle beam artifacts, and cannot be quantified, making interobserver variations high [29, 30]. Similarly, increased vascularity of the median nerve showed a low accuracy (60%) for diagnosing CTS. Our study suggests that these qualitative parameters cannot be used alone for the diagnosis of CTS.

Additionally, we assessed multiple combinations of the parameters and found that combination of the three best ultrasound parameters- delta CSA, CSAc, and WFR, has a higher specificity and Youden’s index compared to delta CSA, however, further research is needed to confirm these findings.

Conclusion

We conclude that Delta CSA has the highest efficacy for the diagnosis of CTS amongst all the eight ultrasound-based variables included in this study. This was followed in decreasing order by CSAc, WFR, and PB, all having very high accuracy for diagnosis of CTS in comparison to NCS substantiating their possible use as an alternative to NCS. FR and FT have low accuracies but FT has an acceptable AUC value and our study predicted that lowering its cut-off could significantly improve its diagnostic capacity. Hypoechogenicity and vascularity of the median nerve had low diagnostic capacity and cannot be used as a sole criterion for diagnosing CTS. Using ROC curve analysis this study predicted the best cut-off values for these ultrasound parameters which could improve their diagnostic accuracy, and further research is needed to analyse these cut-offs in a more representative sample of the population to verify these findings.

Abbreviations

AUC

Area under the curve

CSA

Cross-sectional area of median nerve

CSAc

Cross-sectional area of the median nerve at the distal wrist crease (carpal tunnel inlet)

CSAp

Cross-sectional area of the median nerve at the pronator quadratus muscle level

CSAf

Cross-sectional area of the median nerve in the forearm, 12 cm proximal to the distal wrist crease

CTS

Carpal tunnel syndrome

FR

Flattening ratio

FT

Flexor-retinaculum thickness

NCS

Nerve conduction study

NPV

Negative predictive value

PB

Palmer bowing of flexor-retinaculum

PPV

Positive predictive value

ROC

Receiver operator curve

WFR

Wrist-forearm ratio

Author contributions

“All authors contributed to the study conception and design. Material preparation, data collection and original draft preparation were performed by Dr. Isha Gupta. Writing—review and editing were done by Dr. Kshitij Gupta, Dr. Shashank Sharma, Dr. Vikas Jhanwar and Dr. Parul Gupta, under the supervision of Dr. Naima Mannan and Dr. Meenu Bagarhatta. Statistical analysis was performed by Deepak Gupta and nerve conduction study analysis was done by Jitendra Yadav: All authors read and approved the final manuscript.”

Funding

The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Conflict of interest

The authors have no relevant financial or non-financial interests to disclose.

Ethics approval

Approval was obtained from the ethics committee of SMS Medical college, Jaipur, India. The procedures used in this study adhere to the tenets of the Declaration of Helsinki.

Consent to participate

Informed written consent was obtained from all individual participants included in the study.

Consent to publish

Patients gave informed consent regarding publishing their data and photographs.