- Email: [email protected]
Reproducibility of Acoustic Radiation Force Impulse Imaging in Thyroid and Salivary Glands with Experienced and Inexperienced Examiners
Ultrasound in Med. & Biol., Vol. -, No. -, pp. 1–8, 2016 Copyright Ó 2016 World Federation for Ultrasound in Medicine & Biology Printed in the USA. All rights reserved 0301-5629/$ - see front matter
http://dx.doi.org/10.1016/j.ultrasmedbio.2016.06.019
d
Original Contribution REPRODUCIBILITY OF ACOUSTIC RADIATION FORCE IMPULSE IMAGING IN THYROID AND SALIVARY GLANDS WITH EXPERIENCED AND INEXPERIENCED EXAMINERS BENEDIKT HOFAUER,* NAGLAA MANSOUR,* CLEMENS HEISER,* MARKUS WIRTH,* ULRICH STRAßEN,* DENYS LOEFFELBEIN,y MURAT BAS,* and ANDREAS KNOPF* * Otorhinolaryngology/Head and Neck Surgery, Klinikum rechts der Isar, Technical University Munich, Munich, Germany; and y Cranio-Maxillo-Facial Surgery, Klinikum rechts der Isar, Technical University Munich, Munich, Germany (Received 9 January 2016; revised 7 June 2016; in final form 10 June 2016)
Abstract—Acoustic radiation force impulse (ARFI) imaging enables the sonographic measurement of tissue stiffness. The aim of this study was to evaluate if experience in ARFI imaging influences the reproducibility of ARFI imaging of the head and neck. Three experienced sonographers and three inexperienced sonographers performed ARFI imaging of thyroid, submandibular and parotid glands in 10 healthy volunteers. The examination was repeated after 2 wk. Ten single ARFI measurements were done in every gland. Inter-rater and intra-rater reliability was analyzed using the intra-class correlation coefficient (ICC). Moderate agreement was observed between experienced and inexperienced examiners (ICC 5 0.46). In salivary glands, agreement was fair between the groups (ICC 5 0.33), whereas in separate evaluations, inter-rater reliability in the submandibular glands was moderate (ICC 5 0.52), and that in the parotid glands, only poor (ICC 5 0.09). For ARFI imaging of the thyroid gland, there was moderate agreement between the groups (ICC 5 0.50). The intra-rater reliability for the salivary and thyroid glands together and separately was strong in both groups. ARFI imaging of the thyroid and salivary glands did exhibit good reproducibility. ARFI imaging of the thyroid gland reached the highest levels of inter- and intraobserver agreement in both groups. ARFI imaging in salivary glands is only reproducible with experienced examiners. (E-mail: [email protected]) Ó 2016 World Federation for Ultrasound in Medicine & Biology. Key Words: Acoustic radiation force impulse, Thyroid gland, Salivary gland, Reproducibility, Shear wave.
has already been proven to be beneficial in the evaluation of suspicious mammary lesions and in the evaluation of liver fibrosis and cirrhosis (Au et al. 2014; Barr and Zhang 2012; Beland et al. 2014; C ¸ ebi Olgun et al. 2014; Chang et al. 2011; Fleury et al. 2009; Fraquelli et al. 2007; Piccinino et al. 1986; Tozaki et al. 2011). As most organs in the neck have good accessibility for sonographic examination, sonography had already represented the most important imaging method in this area. Various investigations on the diagnostic use of elastographic modalities in the neck have been conducted, predominantly on the evaluation of salivary and thyroid gland pathologies, but also for the assessment of suspicious cervical lymph nodes (Choi et al. 2013; Lyshchik et al. 2007). The application of elastography in the differentiation of thyroid nodules has been investigated by various study groups, and it was reported that elastography not only provides further information on tumor characteristics, but also has good diagnostic performance in discrimination between benign and
INTRODUCTION The evolution of new sonographic modalities, in particular real-time elastography (RTE) and shear wave velocity-based methods, such as Virtual Touch imaging (VTI) and quantification (VTQ, synonymously used in the literature is acoustic radiation force impulse imaging or shear wave elastography) or contrast-enhanced ultrasound (CEUS), facilitated new diagnostic options in the evaluation of pathologies in numerous regions of the human body (Knopf et al. 2012). The ability of shear wave velocity as an approach to the highly localized evaluation of tissue characteristics has been prescribed by Sarvazyan et al. (1998) and further investigated in breast lesions by Nightingale et al. (2000). Sonoelastographic evaluation
Address correspondence to: Benedikt Hofauer, Otorhinolaryngology/Head and Neck Surgery, Klinikum rechts der Isar, Technical University Munich, Ismaningerstrasse 22, 81675 Munich, Germany. E-mail: [email protected] 1
2
Ultrasound in Medicine and Biology
malignant nodules, possibly reducing unnecessary biopsies (Grazhdani et al. 2014; Hou et al. 2013; Lin et al. 2014; Park et al. 2009; Zhang and Han 2013; Zhang et al. 2013, 2014a, 2014b, 2014c, 2015). The evaluation of tissue characteristics with elastography in salivary gland pathologies proved to be beneficial in the assessment of solitary, circumscribed salivary gland lesions and in the evaluation of diffuse salivary gland diseases. Acoustic radiation force impulse (ARFI) imaging reveals alterations in salivary glands after radiation therapy of the head and neck and is a reliable tool for the identification of early disease stages in primary Sj€ ogren’s syndrome (Badea et al. 2013; Knopf et al. 2015; Mansour et al. 2012, 2015). For qualitative elastographic modalities, such as RTE and VTI, interpretation of the generated elastogram by the sonographer is required, whereas ARFI imaging is a quantitative method and is therefore postulated to be less operator dependent (Fukuhara et al. 2014). However, it must be taken into consideration that even though no further interpretation of the elastogram (e.g., with scoring systems) is required, there have been reports on various factors influencing the acquisition of ARFI images. One factor influencing the results of ARFI imaging in salivary glands is the degree of pre-compression (Mantsopoulos et al. 2015). Other limitations of ARFI imaging in the neck are the depth of the selected region of interest (ROI) and tissue characteristics, which can influence shear wave propagation. These findings illustrate that there is in fact operator dependence in ARFI imaging. The aim of this study was to determine if experience in ARFI imaging of thyroid and salivary glands has an influence on the reproducibility of the results generated in a population of healthy volunteers. METHODS Probands Ten healthy volunteers from the Department of Otorhinolaryngology/Head and Neck Surgery were included as probands in this study. There were five female and five male probands with an average age of 28.9 6 3.3 y. The mean body mass index (BMI) was 22.4 6 2.0 kg/m2. The volunteers had no history of thyroid or salivary gland disease, and blood tests revealed no abnormalities in thyroid hormones or antibodies. Ultrasound examination The Acuson S2000 ultrasound system (Siemens Healthcare, Erlangen, Germany) was used for sonographic examinations. ARFI imaging was performed with a B-mode–ARFI combination linear transducer (9 L4, Siemens Medical Solutions). ARFI imaging was performed on the right thyroid lobe, right submandibular
Volume -, Number -, 2016
gland and right parotid gland. The imaging frequency used was 8 MHz. The ARFI push pulse frequency was 4 MHz. With ARFI imaging (synonymous with Virtual Touch quantification), a single focused push pulse is directed to the left side of the region of interest, generating shear waves. The peak of the shear waves is tracked by detection beams across the ROI. The elements of the probe used are dependent on the depth and steered angle of the ROI. Shear wave velocity was obtained. Every examination consisted of 10 single measurements of every gland at a depth of 1.0 cm in the center of the thyroid and submandibular glands and in the center of the caudal pole of the parotid gland, to guarantee a maximum distance to the jawbone or adjacent blood vessels (Fig. 1). The mean value of the 10 single measurements was calculated and used for further analyses. Measurements were obtained with moderate transducer pressure timed to the absence of swallowing maneuvers. This examination protocol was previously prescribed by our study group (Knopf et al. 2015). To avoid circadian variation in ARFI values, the examinations were all performed during the same time period in the morning. The probands were asked not to eat, drink or smoke for 2 h before the examination. Two groups of examiners performed the examination protocol: both groups consisted of sonographers with long experience in head and neck sonography, but one group consisted of three sonographers with at least 4 y of daily practice in ARFI imaging and therefore was the experienced group, whereas the other group consisted of three sonographers without additional experience in the field of ARFI imaging and therefore was the inexperienced group. Before initiation of this study, the inexperienced group received instructions on ARFI imaging (including the correct application, mechanics, physics and potential pitfalls) by a certified medical sonography instructor who was not one of the examiners. The ultrasound examination was repeated by every sonographer after an interval of 2 wk to enable calculation of intrarater reliability, after inter-rater reliability had been calculated. Statistical analysis For statistical analysis, Version 23.0 of the Statistical Package for Social Sciences software (IBM, Armonk, NY USA) was employed. Data are reported as the mean 6 standard deviation if not otherwise stated. Inter-rater and intra-rater reliability was analyzed using the intra-class correlation coefficient (ICC), as ARFI values are continuous variables. Inter-rater agreement was calculated on the ARFI values of the experienced and inexperienced groups. Intra-rater agreement was calculated on the values from the first and the second ARFI evaluations. ICC values can range between 11
Reproducibility of ARFI imaging d B. HOFAUER et al.
3
Fig. 1. Acoustic radiation force impulse images of the center of the right lobe of the thyroid gland (a), of the center of the right submandibular gland (b) and in the center of the caudal pole of the right parotid gland (c).
(100% agreement) and 21 (100% disagreement). Agreement was classified as poor (ICC 5 0.00–0.20), fair (ICC 5 0.21–0.40), moderate (ICC 5 0.41–0.60), strong (ICC 5 0.61–0.80) or excellent (ICC .0.80). ICC values were evaluated for ARFI measurements in all glands or separately for ARFI measurements in salivary glands alone and in thyroid glands alone. Bland–Altman plots of the mean ratings for each patient in the two groups versus the differences in ratings are provided for descriptive purposes (Bland and Altman 1999). For calculation of the Bland–Altman plot, a one-sample t-test was used. Linear regression was conducted to rule out a trend in differences across the mean ratings. p values #0.05 were considered to indicate statistical significance. The local ethics committee (Fakult€at f€ ur Medizin, Ethikkommission, Technische Universit€at M€ unchen) approved this study. Informed consent was obtained from every proband.
RESULTS The six examiners performed 3,600 single ARFI measurements (1,200 measurements of the thyroid gland and 1,200 each of the submandibular and parotid glands), which are the basis for the calculations. The mean values
(6standard deviation) of all measurements are summarized in Table 1. Moderate agreement was observed between experienced and inexperienced examiners for ARFI imaging in thyroid and salivary glands considered together (ICC 5 0.457) (Fig. 2). Within the experienced group, the three examiners reached moderate inter-rater agreement (ICC 5 0.527). Within the inexperienced group, the three examiners achieved moderate inter-rater agreement (ICC 5 0.542). Intra-rater reliability revealed strong agreement in the experienced group (ICC 5 0.764) (Fig. 3) and the inexperienced group (ICC 5 0.680) (Fig. 4). The Bland–Altman plot (Fig. 5) revealed no trend toward systematic over- or underestimation between the two groups. Linear regression analysis did not indicate a trend of difference across the mean ratings. Evaluation of inter-rater agreement only in ARFI imaging of the salivary glands yielded fair agreement (ICC 5 0.327) between the groups. Within the experienced group, the three examiners reached fair interrater agreement (ICC 5 0.301) for ARFI imaging of the salivary glands. Within the inexperienced group, the three examiners achieved fair inter-rater agreement (ICC 5 0.328). The intra-rater reliability again revealed strong agreement in the experienced group
Table 1. Measurements of both experienced and inexperienced examiners together (including runs 1 and 2) and of experienced and inexperienced examiners separately (runs 1 and 2)* Measurement (m/s) Inexperienced examiners
Experienced examiners Glands
All examiners
Run 1
Run 2
Run 1
Run 2
Salivary Parotid Submandibular Thyroid
1.58 6 0.34 1.52 6 0.38 1.65 6 0.27 1.38 6 0.33
1.63 6 0.35 1.57 6 0.39 1.68 6 0.31 1.40 6 0.42
1.55 6 0.30 1.54 6 0.28 1.55 6 0.19 1.35 6 0.32
1.57 6 0.37 1.45 6 0.42 1.68 6 0.29 1.36 6 0.30
1.61 6 0.39 1.52 6 0.45 1.70 6 0.29 1.39 6 0.29
* The initial ultrasound examination (run 1) was repeated by every sonographer after an interval of 2 wk (run 2) to enable calculation of intra-rater reliability. Values are expressed as the mean 6 standard deviation.
4
Ultrasound in Medicine and Biology
Volume -, Number -, 2016
Fig. 2. Inter-observer agreement between a group of experienced examiners and a group of inexperienced examiners.
Fig. 4. Intra-observer agreement within the group of inexperienced examiners.
(ICC 5 0.731) and the inexperienced group (ICC 5 0.645). The Bland–Altman plot (Fig. 6) revealed no trend toward systematic over- or underestimation between the two groups. There was no trend of difference across the mean ratings. Separate evaluation of ARFI imaging of the submandibular and the parotid gland yielded moderate agreement (ICC 5 0.518) between the groups for the submandibular gland. Within the experienced group, the three examiners reached fair inter-rater agreement (ICC 5 0.285) for ARFI imaging of the submandibular gland. Within the inexperienced group, the three examiners achieved no agreement (ICC 5 20.136) for the submandibular gland. For ARFI imaging of the submandibular gland, there was moderate agreement (ICC 5 0.555) in intra-rater reliability within the experi-
enced group and fair agreement (ICC 5 0.380) within the inexperienced group. For the parotid gland, there was poor agreement (ICC 5 0.088) between the groups. Within the experienced group, the three examiners reached moderate agreement (ICC 5 0.436) for ARFI imaging of the parotid gland. Within the inexperienced group, the three examiners achieved no agreement (ICC 5 20.109) for the parotid gland. For ARFI imaging of the parotid gland resulted in strong agreement in intra-rater reliability was strong within the experienced group (ICC 5 0.628) and
Fig. 3. Intra-observer agreement within the group of experienced examiners.
Fig. 5. Bland–Altman plot for acoustic radiation force impulse imaging of the salivary and thyroid glands with a group of experienced examiners and a group of inexperienced examiners. The mean ratings of the two groups are plotted on the x-axis, and the differences in the ratings, on the y-axis. The middle line represents the mean of the difference in ratings (0.021 m/s); the upper and lower lines define the limits of agreement (mean 6 1.96SD, 95% confidence interval).
Reproducibility of ARFI imaging d B. HOFAUER et al.
5
difference across the mean ratings. All reliabilities are summarized in Table 2. The means (6standard deviations) of the 10 single measurements in every gland are summarized in Table 3. DISCUSSION
Fig. 6. Bland–Altman plot for acoustic radiation force impulse imaging of the salivary glands with a group of experienced examiners and a group of inexperienced examiners. The middle line represents the mean of the difference in ratings (0.014 m/s); the upper and lower lines define the limits of agreement.
moderate agreement within the inexperienced group (ICC 5 0.496). Inter-rater agreement on ARFI imaging of the thyroid gland was moderate (ICC 5 0.498). Within the experienced group, the three examiners reached excellent inter-rater agreement (ICC 5 0.811). Within the inexperienced group, the three examiners achieved strong interrater agreement (ICC 5 0.657). Intra-rater reliability was strong t in the experienced group (ICC 5 0.740) and the inexperienced group (ICC 5 0.640). The Bland–Altman plot (Fig. 7) revealed no trend toward systematic over- or underestimation in the two groups. There was no trend of
Fig. 7. Bland–Altman plot for acoustic radiation force impulse imaging of the thyroid gland with a group of experienced examiners and a group of inexperienced examiners. The middle line represents the mean of the difference in ratings (0.027 m/s); the upper and lower lines define the limits of agreement.
Together with contrast-enhanced ultrasound, the development of sonographic methods for the noninvasive evaluation of mechanical tissue properties represents substantial progress in ultrasound in recent years. RTE and VTI are two major sonoelastographic modalities, which differ in the method of image acquisition. In RTE, the tissue displacement required for the elastogram is generated by the pulsation of adjacent blood vessels or the pressure of the sonography probe, whereas in VTI, an acoustic push pulse is used to create tissue displacement. Consequently, these methods can be influenced in two different ways. On the one hand, differences in elastogram acquisition can influence reproducibility (inadequate data acquisition); on the other hand, the elastogram generated has to be interpreted, which leaves room for variation (inaccurate interpretation). The main difference in ARFI imaging (synonymous for VTQ) compared with RTE or VTI is that within a selected region of interest, the shear wave velocity generated by the acoustic push pulse is measured. There is a direct link between shear wave velocity and tissue stiffness (m 5 pc2). Therefore the potentially subjective interpretation of an elastogram is not required, and ARFI imaging is postulated to be observer independent and reproducible (Badea et al. 2013; Fraquelli et al. 2007; Nightingale et al. 2000; Sarvazyan et al. 1998). Examiners must, however, be aware of the fact that there is variability in the generation of ARFI images as there are various influencing factors (5 inadequate data acquisition). A differentiation between organ-specific and general influencing parameters can be made. Values of ARFI imaging in the liver increase with food intake and should therefore be evaluated in the fasting state (Goertz et al. 2012). Reproducibility is significantly reduced in patients with chronic liver disease in case of steatosis, increased BMI and lower degree of hepatic fibrosis (Fraquelli et al. 2007). Jaffer et al. (2012) reported that ARFI values obtained deeper to the liver capsule allow more reliable stiffness quantification. In breast mass evaluation, on the contrary, shallower lesion depth and less breast thickness at the area where the lesion is located are known to influencing elastography (Chang et al. 2011). Pre-compression of the investigated tissue is consistently reported to influence the variation in ARFI imaging in all organs (Barr and Zhang 2012; Mantsopoulos et al. 2015; Wojcinski et al. 2013). Precompression can be created by the sonographer, but
6
Ultrasound in Medicine and Biology
Volume -, Number -, 2016
Table 2. Inter- and intra-rater reliability for all glands and for salivary and thyroid glands separately Reproducibility Inter-rater agreement Between groups Within experienced group Within inexperienced group Intra-rater agreement Experienced group Inexperienced group
All glands
Salivary glands
Submandibular gland
Parotid gland
Thyroid gland
0.457 0.527 0.542
0.327 0.301 0.328
0.518 0.285 20.136
0.088 0.436 20.109
0.498 0.811 0.657
0.764 0.680
0.731 0.645
0.555 0.380
0.628 0.496
0.740 0.640
also by the pulsation of adjacent blood vessels close to the selected region of interest or by patient movement and varying degrees of muscle tension. As there is no standardization of adequate compression, the amount of pressure depends on the performer’s subjective experience and probe operating skills (Yoon et al. 2011). Few studies have evaluated the reproducibility of ARFI imaging in experienced examiners, but no studies have evaluated the possible difference in ARFI imaging of the thyroid and salivary glands between experienced and inexperienced examiners. The most striking discrepancy between the groups became noticeable when the submandibular and parotid glands were examined separately with ARFI. Although agreement was moderate for the submandibular glands, there was only poor agreement on the parotid glands; therefore, the parotid examination was responsible for the impaired inter-rater reliability for the salivary glands compared with the collective results. The reproducibility within the two groups was better. The discrepancy between experienced and inexperienced examiners indicates that ARFI results for parotid glands obtained by examiners with different levels of experience have to be interpreted carefully. When inter-rater agreement within the two groups was taken into consideration, there was moderate agreement within experienced examiners for the parotid glands. As inter-rater agreement within inexperienced examiners for the parotid glands and within both groups for submandibular glands was fair or not detectable, it can be recommended that ARFI imaging of parotid glands should be performed only by experienced examiners, whereas ARFI imaging of submandibular glands does not seem promising for both experienced Table 3. Ten single measurements obtained by experienced and inexperienced examiners in the submandibular, parotid and thyroid glands Reproducibility
All
Experienced examiner
Inexperienced examiner
Submandibular gland Parotid gland Thyroid gland
0.14 0.10 0.11
0.13 0.08 0.09
0.16 0.12 0.12
and inexperienced sonographers. These results agree with the investigation by Knopf et al. (2015) on ARFI imaging of salivary glands in patients with Sj€ogren’s syndrome, where the inter-observer reliability of the findings on submandibular glands was judged unreliable. An explanation for the poor inter-rater agreement between experienced and inexperienced examiners in the parotid gland is its anatomic position, which might impair application of the sonography probe between the mastoid process and the mandible, an usually uneven area. Therefore different amounts of pressure need to be applied, which also explains the better inter-rater reliability of the experienced examiners compared with the inexperienced examiners. The pulsation of the facial artery within the posterior part of the submandibular gland could explain the comparatively low inter-rater reliability within both groups for the findings on the submandibular gland, for which the 10 single measurements also had greater variation (compare Table 3). As depth of the selected region of interest influences ARFI imaging, we agreed to measure at a depth of 1.0 cm, being aware that the measurements are no more reliable for depths inferior to 1.0 cm. This might be a limitation of the applied methods, but reflects the realistic depth of ARFI measurements in thyroid and salivary glands. The mean values of salivary gland measurements did not vary significantly between the two examinations. Only a small number of studies have focused on the use of ARFI imaging in salivary gland lesions. Although Matsuzuka et al. (2015) focused primarily on salivary gland stiffness measured with ARFI imaging in predominantly healthy salivary glands, Mansour et al. (2012) evaluated parotid gland lesions with ARFI imaging as part of a multimodal concept. The latter obtained values of 1.92 and 1.75 m/s (60.64) for healthy parotid glands, similar to our value of 1.52 m/s (60.38). Slightly more research has been conducted on diffuse salivary gland alterations, for example, Badea et al. (2013) in submandibular glands after radiation therapy, Knopf et al. (2015) in patients with Sj€ogren’s syndrome and Kaluzny et al. (2014) again in irradiated salivary glands. Badea et al. obtained a value of 1.82 m/s (60.41) in healthy submandibular glands, similar to our value of 1.65 m/s (60.27). The
Reproducibility of ARFI imaging d B. HOFAUER et al.
only investigation on the reliability of shear wave elastography in the neck, which included salivary glands, was published by Bhatia et al. (2012), with only parts of the patient cohort being evaluated by two radiologists to enable calculation of inter- and intra-rater reliability. According to our classification, strong to excellent reliability was reported. Sonoelastographic modalities in general and ARFI imaging in particular for the differentiation of thyroid nodules have been under thorough investigation. Various authors concluded that ARFI imaging is a promising tool for the identification of suspicious thyroid nodules (Cantisani et al. 2014; Lin et al. 2014; Zhang and Han 2013). Even though sonographic evaluation of thyroid nodules is in wide use, to date there have been no investigations concerning the reliability of ARFI imaging in the thyroid gland, comparing the results of experienced examiners versus those of inexperienced examiners. Our study could indicate that there is moderate inter-rater agreement between experienced and inexperienced sonographers. Within the groups, the examiners reached excellent and strong agreement, respectively. Intra-rater agreement was strong. Interand intra-rater reliability of experienced examiners in ARFI imaging of the thyroid gland has been under investigation. Concordant with our findings, mainly strong agreement has been observed (Cantisani et al. 2015; Grazhdani et al. 2014; Zhang et al. 2014c). We did not observe any significant variation in ARFI values between the two examinations.
CONCLUSIONS The reproducibility of results obtained by ARFI imaging of the thyroid and salivary glands between experienced and inexperienced examiners was good. In particular, the results for ARFI imaging of the thyroid gland had the highest levels of inter- and intra-observer agreement, and even inexperienced sonographers had strong inter- and intra-observer agreement. For the salivary glands, attention must be paid to ARFI imaging of the parotid gland, as inter-rater agreement between the groups was poor. Therefore, we recommend that only experienced examiners should perform ARFI imaging on parotid glands. ARFI imaging in submandibular glands does not seem to be reproducible for both experienced and inexperienced examiners. Acknowledgments—The authors thank the volunteers who contributed to this study as healthy probands for the sonographic examinations: Yumiko Matsuba, Constanze Gahleitner, Magdalena Lenschow, Suzan Badawood, Carolin Ibald, Florian Durst, Markus Brandstetter, Michael Szyper, Michael Sedlmeyer and Christian Winkler.
7
REFERENCES Au FW, Ghai S, Moshonov H, Kahn H, Brennan C, Dua H, Crystal P. Diagnostic performance of quantitative shear wave elastography in the evaluation of solid breast masses: determination of the most discriminatory parameter. AJR Am J Roentgenol 2014;203: W328–W336. Badea AF, Tamas Szora A, Ciuleanu E, Chioreanu I, Baciut¸ G, Lups¸or Platon M, Badea R. ARFI quantitative elastography of the submandibular glands. Normal measurements and the diagnosis value of the method in radiation submaxillitis. Med Ultrason 2013;15:173–179. Barr RG, Zhang Z. Effects of precompression on elasticity imaging of the breast: Development of a clinically useful semiquantitative method of precompression assessment. J Ultrasound Med 2012; 31:895–902. Beland MD, Brown SF, Machan JT, Taliano RJ, Promrat K, Cronan JJ. A pilot study estimating liver fibrosis with ultrasound shear-wave elastography: Does the cause of liver disease or location of measurement affect performance. AJR Am J Roentgenol 2014;203: W267–W273. Bhatia K, Tong CS, Cho CC, Yuen EH, Lee J, Ahuja AT. Reliability of shear wave ultrasound elastography for neck lesions identified in routine clinical practice. Ultraschall Med 2012;33:463–468. Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res 1999;8:135–160. Cantisani V, Grazhdani H, Ricci P, Mortele K, Di Segni M, D’Andrea V, Redler A, Di Rocco G, Giacomelli L, Maggini E, Chiesa C, Erturk SM, Sorrenti S, Catalano C, D’Ambrosio F. Q-Elastosonography of solid thyroid nodules: Assessment of diagnostic efficacy and interobserver variability in a large patient cohort. Eur Radiol 2014;24:143–150. Cantisani V, Lodise P, Di Rocco G, Grazhdani H, Giannotti D, Patrizi G, Medvedyeva E, Olive M, Fioravanti C, Giacomelli L, Chiesa C, Redler A, Catalano C, D’Ambrosio F, Ricci P. Diagnostic accuracy and interobserver agreement of quasistatic ultrasound elastography in the diagnosis of thyroid nodules. Ultraschall Med 2015;36: 162–167. C glu M, ¸ ebi Olgun D, Korkmazer B, Kılıc¸ F, Dikici AS, Velidedeo Aydogan F, Kantarcı F, Yılmaz MH. Use of shear wave elastography to differentiate benign and malignant breast lesions. Diagn Interv Radiol 2014;20:239–244. Chang JM, Moon WK, Cho N, Kim SJ. Breast mass evaluation: Factors influencing the quality of US elastography. Radiology 2011;259: 59–64. Choi YJ, Lee JH, Lim HK, Kim SY, Han MW, Cho KJ, Baek JH. Quantitative shear wave elastography in the evaluation of metastatic cervical lymph nodes. Ultrasound Med Biol 2013;39:935–940. Fleury EF, Fleury JC, Piato S, Roveda D. New elastographic classification of breast lesions during and after compression. Diagn Interv Radiol 2009;15:96–103. Fraquelli M, Rigamonti C, Casazza G, Conte D, Donato MF, Ronchi G, Colombo M. Reproducibility of transient elastography in the evaluation of liver fibrosis in patients with chronic liver disease. Gut 2007; 56:968–973. Fukuhara T, Matsuda E, Fujiwara K, Tanimura C, Izawa S, Kataoka H, Kitano H. Phantom experiment and clinical utility of quantitative shear wave elastography for differentiating thyroid nodules. Endocr J 2014;61:615–621. Goertz RS, Egger C, Neurath MF, Strobel D. Impact of food intake, ultrasound transducer, breathing maneuvers and body position on acoustic radiation force impulse (ARFI) elastometry of the liver. Ultraschall Med 2012;33:380–385. Grazhdani H, Cantisani V, Lodise P, Di Rocco G, Proietto MC, Fioravanti E, Rubini A, Redler A. Prospective evaluation of acoustic radiation force impulse technology in the differentiation of thyroid nodules: Accuracy and interobserver variability assessment. J Ultrasound 2014;17:13–20. Hou XJ, Sun AX, Zhou XL, Ji Q, Wang HB, Wei H, Sun JW, Liu H. The application of Virtual Touch tissue quantification (VTQ) in diagnosis of thyroid lesions: A preliminary study. Eur J Radiol 2013; 82:797–801.
8
Ultrasound in Medicine and Biology
Jaffer OS, Lung PF, Bosanac D, Patel VM, Ryan SM, Heneghan MA, Quaglia A, Sidhu PS. Acoustic radiation force impulse quantification: repeatability of measurements in selected liver segments and influence of age, body mass index and liver capsule-to-box distance. Br J Radiol 2012;85:e858–e863. Ka1u_zny J, Kopec T, Szczepanek-Parulska E, Stangierski A, Gurgul E, Rucha1a M, Milecki P, Wierzbicka M. Shear wave elastography: a new noninvasive tool to assess the intensity of fibrosis of irradiated salivary glands in head and neck cancer patients. Biomed Res Int 2014;2014:157809. Knopf A, Hofauer B, Th€urmel K, Meier R, Stock K, Bas M, Manour N. Diagnostic utility of acoustic radiation force impulse (ARFI) imaging in primary Sjoegren’s syndrome. Eur Radiol 2015;25: 3027–3034. Knopf A, Mansour N, Chaker A, Bas M, Stock K. Multimodal ultrasonographic characterisation of parotid gland lesions—A pilot study. Eur J Radiol 2012;81:3300–3305. Lin P, Chen M, Liu B, Wang S, Li X. Diagnostic performance of shear wave elastography in the identification of malignant thyroid nodules: A meta-analysis. Eur Radiol 2014;24:2729–2738. Lyshchik A, Higashi T, Asato R, Tanaka S, Ito J, Hiraoka M, Insana MF, Brill AB, Saga T, Togashi K. Cervical lymph node metastases: Diagnosis at sonoelastography—Initial experience. Radiology 2007;243: 258–267. Mansour N, Bas M, Stock KF, Strassen U, Hofauer B, Knopf A. Multimodal ultrasonographic pathway of parotid gland lesions. Ultraschall Med 2015 [Epub ahead of print]. Mansour N, Stock KF, Chaker A, Bas M, Knopf A. Evaluation of parotid gland lesions with standard ultrasound, color duplex sonography, sonoelastography, and acoustic radiation force impulse imaging - a pilot study. Ultraschall Med 2012;33:283–288. Mantsopoulos K, Klintworth N, Iro H, Bozzato A. Applicability of shear wave elastography of the major salivary glands: values in healthy patients and effects of gender, smoking and pre-compression. Ultrasound Med Biol 2015;41:2310–2318. Matsuzuka T, Suzuki M, Saijo S, Ikeda M, Matsui T, Nomoto Y, Nomoto M, Imaizumi M, Tada Y, Omori K. Stiffness of salivary gland and tumor measured by new ultrasonic techniques: Virtual Touch quantification and IQ. Auris Nasus Larynx 2015;42:128–133. Nightingale KR, Nightingale RW, Palmeri ML, Trahey GE. A finite element model of remote palpation of breast lesions using radiation force: Factors affecting tissue displacement. Ultrason Imaging 2000; 22:35–54.
Volume -, Number -, 2016 Park SH, Kim SJ, Kim EK, Kim MJ, Son EJ, Kwak JY. Interobserver agreement in assessing the sonographic and elastographic features of malignant thyroid nodules. AJR Am J Roentgenol 2009;193: W416–W423. Piccinino F, Sagnelli E, Pasquale G, Giusti G. Complications following percutaneous liver biopsy: A multicentre retrospective study on 68,276 biopsies. J Hepatol 1986;2:165–173. Sarvazyan AP, Rudenko OV, Swanson SD, Fowlkes JB, Emelianov SY. Shear wave elasticity imaging: A new ultrasonic technology of medical diagnostics. Ultrasound Med Biol 1998;24:1419–1435. Tozaki M, Isobe S, Fukuma E. Preliminary study of ultrasonographic tissue quantification of the breast using the acoustic radiation force impulse (ARFI) technology. Eur J Radiol 2011;80:e182–e187. Wojcinski S, Brandhorst K, Sadigh G, Hillemanns P, Degenhardt F. Acoustic radiation force impulse imaging with Virtual Touch tissue quantification: measurements of normal breast tissue and dependence on the degree of pre-compression. Ultrasound Med Biol 2013;39:2226–2232. Yoon JH, Kim MH, Kim EK, Moon HJ, Kwak JY, Kim MJ. Interobserver variability of ultrasound elastography: how it affects the diagnosis of breast lesions. AJR Am J Roentgenol 2011;196:730–736. Zhang B, Ma X, Wu N, Liu L, Liu X, Zhang J, Yang J, Niu T. Shear wave elastography for differentiation of benign and malignant thyroid nodules: A meta-analysis. J Ultrasound Med 2013;32:2163–2169. Zhang FJ, Han RL. The value of acoustic radiation force impulse (ARFI) in the differential diagnosis of thyroid nodules. Eur J Radiol 2013; 82:e686–e690. Zhang FJ, Han RL, Zhao XM. The value of virtual touch tissue image (VTI) and virtual touch tissue quantification (VTQ) in the differential diagnosis of thyroid nodules. Eur J Radiol 2014a;83:2033–2040. Zhang H, Shi Q, Gu J, Jiang L, Bai M, Liu L, Wu Y, Du L. Combined value of Virtual Touch tissue quantification and conventional sonographic features for differentiating benign and malignant thyroid nodules smaller than 10 mm. J Ultrasound Med 2014b;33:257–264. Zhang YF, He Y, Xu HX, Xu XH, Liu C, Guo LH, Liu LN, Xu JM. Virtual Touch tissue imaging on acoustic radiation force impulse elastography: A new technique for differential diagnosis between benign and malignant thyroid nodules. J Ultrasound Med 2014c;33: 585–595. Zhang YF, Xu HX, Xu JM, Liu C, Guo LH, Liu LN, Zhang J, Xu XH, Qu S, Xing M. Acoustic radiation force impulse elastography in the diagnosis of thyroid nodules: Useful or not useful. Ultrasound Med Biol 2015;41:2585–2593.
http://dx.doi.org/10.1016/j.ultrasmedbio.2016.06.019
d
Original Contribution REPRODUCIBILITY OF ACOUSTIC RADIATION FORCE IMPULSE IMAGING IN THYROID AND SALIVARY GLANDS WITH EXPERIENCED AND INEXPERIENCED EXAMINERS BENEDIKT HOFAUER,* NAGLAA MANSOUR,* CLEMENS HEISER,* MARKUS WIRTH,* ULRICH STRAßEN,* DENYS LOEFFELBEIN,y MURAT BAS,* and ANDREAS KNOPF* * Otorhinolaryngology/Head and Neck Surgery, Klinikum rechts der Isar, Technical University Munich, Munich, Germany; and y Cranio-Maxillo-Facial Surgery, Klinikum rechts der Isar, Technical University Munich, Munich, Germany (Received 9 January 2016; revised 7 June 2016; in final form 10 June 2016)
Abstract—Acoustic radiation force impulse (ARFI) imaging enables the sonographic measurement of tissue stiffness. The aim of this study was to evaluate if experience in ARFI imaging influences the reproducibility of ARFI imaging of the head and neck. Three experienced sonographers and three inexperienced sonographers performed ARFI imaging of thyroid, submandibular and parotid glands in 10 healthy volunteers. The examination was repeated after 2 wk. Ten single ARFI measurements were done in every gland. Inter-rater and intra-rater reliability was analyzed using the intra-class correlation coefficient (ICC). Moderate agreement was observed between experienced and inexperienced examiners (ICC 5 0.46). In salivary glands, agreement was fair between the groups (ICC 5 0.33), whereas in separate evaluations, inter-rater reliability in the submandibular glands was moderate (ICC 5 0.52), and that in the parotid glands, only poor (ICC 5 0.09). For ARFI imaging of the thyroid gland, there was moderate agreement between the groups (ICC 5 0.50). The intra-rater reliability for the salivary and thyroid glands together and separately was strong in both groups. ARFI imaging of the thyroid and salivary glands did exhibit good reproducibility. ARFI imaging of the thyroid gland reached the highest levels of inter- and intraobserver agreement in both groups. ARFI imaging in salivary glands is only reproducible with experienced examiners. (E-mail: [email protected]) Ó 2016 World Federation for Ultrasound in Medicine & Biology. Key Words: Acoustic radiation force impulse, Thyroid gland, Salivary gland, Reproducibility, Shear wave.
has already been proven to be beneficial in the evaluation of suspicious mammary lesions and in the evaluation of liver fibrosis and cirrhosis (Au et al. 2014; Barr and Zhang 2012; Beland et al. 2014; C ¸ ebi Olgun et al. 2014; Chang et al. 2011; Fleury et al. 2009; Fraquelli et al. 2007; Piccinino et al. 1986; Tozaki et al. 2011). As most organs in the neck have good accessibility for sonographic examination, sonography had already represented the most important imaging method in this area. Various investigations on the diagnostic use of elastographic modalities in the neck have been conducted, predominantly on the evaluation of salivary and thyroid gland pathologies, but also for the assessment of suspicious cervical lymph nodes (Choi et al. 2013; Lyshchik et al. 2007). The application of elastography in the differentiation of thyroid nodules has been investigated by various study groups, and it was reported that elastography not only provides further information on tumor characteristics, but also has good diagnostic performance in discrimination between benign and
INTRODUCTION The evolution of new sonographic modalities, in particular real-time elastography (RTE) and shear wave velocity-based methods, such as Virtual Touch imaging (VTI) and quantification (VTQ, synonymously used in the literature is acoustic radiation force impulse imaging or shear wave elastography) or contrast-enhanced ultrasound (CEUS), facilitated new diagnostic options in the evaluation of pathologies in numerous regions of the human body (Knopf et al. 2012). The ability of shear wave velocity as an approach to the highly localized evaluation of tissue characteristics has been prescribed by Sarvazyan et al. (1998) and further investigated in breast lesions by Nightingale et al. (2000). Sonoelastographic evaluation
Address correspondence to: Benedikt Hofauer, Otorhinolaryngology/Head and Neck Surgery, Klinikum rechts der Isar, Technical University Munich, Ismaningerstrasse 22, 81675 Munich, Germany. E-mail: [email protected] 1
2
Ultrasound in Medicine and Biology
malignant nodules, possibly reducing unnecessary biopsies (Grazhdani et al. 2014; Hou et al. 2013; Lin et al. 2014; Park et al. 2009; Zhang and Han 2013; Zhang et al. 2013, 2014a, 2014b, 2014c, 2015). The evaluation of tissue characteristics with elastography in salivary gland pathologies proved to be beneficial in the assessment of solitary, circumscribed salivary gland lesions and in the evaluation of diffuse salivary gland diseases. Acoustic radiation force impulse (ARFI) imaging reveals alterations in salivary glands after radiation therapy of the head and neck and is a reliable tool for the identification of early disease stages in primary Sj€ ogren’s syndrome (Badea et al. 2013; Knopf et al. 2015; Mansour et al. 2012, 2015). For qualitative elastographic modalities, such as RTE and VTI, interpretation of the generated elastogram by the sonographer is required, whereas ARFI imaging is a quantitative method and is therefore postulated to be less operator dependent (Fukuhara et al. 2014). However, it must be taken into consideration that even though no further interpretation of the elastogram (e.g., with scoring systems) is required, there have been reports on various factors influencing the acquisition of ARFI images. One factor influencing the results of ARFI imaging in salivary glands is the degree of pre-compression (Mantsopoulos et al. 2015). Other limitations of ARFI imaging in the neck are the depth of the selected region of interest (ROI) and tissue characteristics, which can influence shear wave propagation. These findings illustrate that there is in fact operator dependence in ARFI imaging. The aim of this study was to determine if experience in ARFI imaging of thyroid and salivary glands has an influence on the reproducibility of the results generated in a population of healthy volunteers. METHODS Probands Ten healthy volunteers from the Department of Otorhinolaryngology/Head and Neck Surgery were included as probands in this study. There were five female and five male probands with an average age of 28.9 6 3.3 y. The mean body mass index (BMI) was 22.4 6 2.0 kg/m2. The volunteers had no history of thyroid or salivary gland disease, and blood tests revealed no abnormalities in thyroid hormones or antibodies. Ultrasound examination The Acuson S2000 ultrasound system (Siemens Healthcare, Erlangen, Germany) was used for sonographic examinations. ARFI imaging was performed with a B-mode–ARFI combination linear transducer (9 L4, Siemens Medical Solutions). ARFI imaging was performed on the right thyroid lobe, right submandibular
Volume -, Number -, 2016
gland and right parotid gland. The imaging frequency used was 8 MHz. The ARFI push pulse frequency was 4 MHz. With ARFI imaging (synonymous with Virtual Touch quantification), a single focused push pulse is directed to the left side of the region of interest, generating shear waves. The peak of the shear waves is tracked by detection beams across the ROI. The elements of the probe used are dependent on the depth and steered angle of the ROI. Shear wave velocity was obtained. Every examination consisted of 10 single measurements of every gland at a depth of 1.0 cm in the center of the thyroid and submandibular glands and in the center of the caudal pole of the parotid gland, to guarantee a maximum distance to the jawbone or adjacent blood vessels (Fig. 1). The mean value of the 10 single measurements was calculated and used for further analyses. Measurements were obtained with moderate transducer pressure timed to the absence of swallowing maneuvers. This examination protocol was previously prescribed by our study group (Knopf et al. 2015). To avoid circadian variation in ARFI values, the examinations were all performed during the same time period in the morning. The probands were asked not to eat, drink or smoke for 2 h before the examination. Two groups of examiners performed the examination protocol: both groups consisted of sonographers with long experience in head and neck sonography, but one group consisted of three sonographers with at least 4 y of daily practice in ARFI imaging and therefore was the experienced group, whereas the other group consisted of three sonographers without additional experience in the field of ARFI imaging and therefore was the inexperienced group. Before initiation of this study, the inexperienced group received instructions on ARFI imaging (including the correct application, mechanics, physics and potential pitfalls) by a certified medical sonography instructor who was not one of the examiners. The ultrasound examination was repeated by every sonographer after an interval of 2 wk to enable calculation of intrarater reliability, after inter-rater reliability had been calculated. Statistical analysis For statistical analysis, Version 23.0 of the Statistical Package for Social Sciences software (IBM, Armonk, NY USA) was employed. Data are reported as the mean 6 standard deviation if not otherwise stated. Inter-rater and intra-rater reliability was analyzed using the intra-class correlation coefficient (ICC), as ARFI values are continuous variables. Inter-rater agreement was calculated on the ARFI values of the experienced and inexperienced groups. Intra-rater agreement was calculated on the values from the first and the second ARFI evaluations. ICC values can range between 11
Reproducibility of ARFI imaging d B. HOFAUER et al.
3
Fig. 1. Acoustic radiation force impulse images of the center of the right lobe of the thyroid gland (a), of the center of the right submandibular gland (b) and in the center of the caudal pole of the right parotid gland (c).
(100% agreement) and 21 (100% disagreement). Agreement was classified as poor (ICC 5 0.00–0.20), fair (ICC 5 0.21–0.40), moderate (ICC 5 0.41–0.60), strong (ICC 5 0.61–0.80) or excellent (ICC .0.80). ICC values were evaluated for ARFI measurements in all glands or separately for ARFI measurements in salivary glands alone and in thyroid glands alone. Bland–Altman plots of the mean ratings for each patient in the two groups versus the differences in ratings are provided for descriptive purposes (Bland and Altman 1999). For calculation of the Bland–Altman plot, a one-sample t-test was used. Linear regression was conducted to rule out a trend in differences across the mean ratings. p values #0.05 were considered to indicate statistical significance. The local ethics committee (Fakult€at f€ ur Medizin, Ethikkommission, Technische Universit€at M€ unchen) approved this study. Informed consent was obtained from every proband.
RESULTS The six examiners performed 3,600 single ARFI measurements (1,200 measurements of the thyroid gland and 1,200 each of the submandibular and parotid glands), which are the basis for the calculations. The mean values
(6standard deviation) of all measurements are summarized in Table 1. Moderate agreement was observed between experienced and inexperienced examiners for ARFI imaging in thyroid and salivary glands considered together (ICC 5 0.457) (Fig. 2). Within the experienced group, the three examiners reached moderate inter-rater agreement (ICC 5 0.527). Within the inexperienced group, the three examiners achieved moderate inter-rater agreement (ICC 5 0.542). Intra-rater reliability revealed strong agreement in the experienced group (ICC 5 0.764) (Fig. 3) and the inexperienced group (ICC 5 0.680) (Fig. 4). The Bland–Altman plot (Fig. 5) revealed no trend toward systematic over- or underestimation between the two groups. Linear regression analysis did not indicate a trend of difference across the mean ratings. Evaluation of inter-rater agreement only in ARFI imaging of the salivary glands yielded fair agreement (ICC 5 0.327) between the groups. Within the experienced group, the three examiners reached fair interrater agreement (ICC 5 0.301) for ARFI imaging of the salivary glands. Within the inexperienced group, the three examiners achieved fair inter-rater agreement (ICC 5 0.328). The intra-rater reliability again revealed strong agreement in the experienced group
Table 1. Measurements of both experienced and inexperienced examiners together (including runs 1 and 2) and of experienced and inexperienced examiners separately (runs 1 and 2)* Measurement (m/s) Inexperienced examiners
Experienced examiners Glands
All examiners
Run 1
Run 2
Run 1
Run 2
Salivary Parotid Submandibular Thyroid
1.58 6 0.34 1.52 6 0.38 1.65 6 0.27 1.38 6 0.33
1.63 6 0.35 1.57 6 0.39 1.68 6 0.31 1.40 6 0.42
1.55 6 0.30 1.54 6 0.28 1.55 6 0.19 1.35 6 0.32
1.57 6 0.37 1.45 6 0.42 1.68 6 0.29 1.36 6 0.30
1.61 6 0.39 1.52 6 0.45 1.70 6 0.29 1.39 6 0.29
* The initial ultrasound examination (run 1) was repeated by every sonographer after an interval of 2 wk (run 2) to enable calculation of intra-rater reliability. Values are expressed as the mean 6 standard deviation.
4
Ultrasound in Medicine and Biology
Volume -, Number -, 2016
Fig. 2. Inter-observer agreement between a group of experienced examiners and a group of inexperienced examiners.
Fig. 4. Intra-observer agreement within the group of inexperienced examiners.
(ICC 5 0.731) and the inexperienced group (ICC 5 0.645). The Bland–Altman plot (Fig. 6) revealed no trend toward systematic over- or underestimation between the two groups. There was no trend of difference across the mean ratings. Separate evaluation of ARFI imaging of the submandibular and the parotid gland yielded moderate agreement (ICC 5 0.518) between the groups for the submandibular gland. Within the experienced group, the three examiners reached fair inter-rater agreement (ICC 5 0.285) for ARFI imaging of the submandibular gland. Within the inexperienced group, the three examiners achieved no agreement (ICC 5 20.136) for the submandibular gland. For ARFI imaging of the submandibular gland, there was moderate agreement (ICC 5 0.555) in intra-rater reliability within the experi-
enced group and fair agreement (ICC 5 0.380) within the inexperienced group. For the parotid gland, there was poor agreement (ICC 5 0.088) between the groups. Within the experienced group, the three examiners reached moderate agreement (ICC 5 0.436) for ARFI imaging of the parotid gland. Within the inexperienced group, the three examiners achieved no agreement (ICC 5 20.109) for the parotid gland. For ARFI imaging of the parotid gland resulted in strong agreement in intra-rater reliability was strong within the experienced group (ICC 5 0.628) and
Fig. 3. Intra-observer agreement within the group of experienced examiners.
Fig. 5. Bland–Altman plot for acoustic radiation force impulse imaging of the salivary and thyroid glands with a group of experienced examiners and a group of inexperienced examiners. The mean ratings of the two groups are plotted on the x-axis, and the differences in the ratings, on the y-axis. The middle line represents the mean of the difference in ratings (0.021 m/s); the upper and lower lines define the limits of agreement (mean 6 1.96SD, 95% confidence interval).
Reproducibility of ARFI imaging d B. HOFAUER et al.
5
difference across the mean ratings. All reliabilities are summarized in Table 2. The means (6standard deviations) of the 10 single measurements in every gland are summarized in Table 3. DISCUSSION
Fig. 6. Bland–Altman plot for acoustic radiation force impulse imaging of the salivary glands with a group of experienced examiners and a group of inexperienced examiners. The middle line represents the mean of the difference in ratings (0.014 m/s); the upper and lower lines define the limits of agreement.
moderate agreement within the inexperienced group (ICC 5 0.496). Inter-rater agreement on ARFI imaging of the thyroid gland was moderate (ICC 5 0.498). Within the experienced group, the three examiners reached excellent inter-rater agreement (ICC 5 0.811). Within the inexperienced group, the three examiners achieved strong interrater agreement (ICC 5 0.657). Intra-rater reliability was strong t in the experienced group (ICC 5 0.740) and the inexperienced group (ICC 5 0.640). The Bland–Altman plot (Fig. 7) revealed no trend toward systematic over- or underestimation in the two groups. There was no trend of
Fig. 7. Bland–Altman plot for acoustic radiation force impulse imaging of the thyroid gland with a group of experienced examiners and a group of inexperienced examiners. The middle line represents the mean of the difference in ratings (0.027 m/s); the upper and lower lines define the limits of agreement.
Together with contrast-enhanced ultrasound, the development of sonographic methods for the noninvasive evaluation of mechanical tissue properties represents substantial progress in ultrasound in recent years. RTE and VTI are two major sonoelastographic modalities, which differ in the method of image acquisition. In RTE, the tissue displacement required for the elastogram is generated by the pulsation of adjacent blood vessels or the pressure of the sonography probe, whereas in VTI, an acoustic push pulse is used to create tissue displacement. Consequently, these methods can be influenced in two different ways. On the one hand, differences in elastogram acquisition can influence reproducibility (inadequate data acquisition); on the other hand, the elastogram generated has to be interpreted, which leaves room for variation (inaccurate interpretation). The main difference in ARFI imaging (synonymous for VTQ) compared with RTE or VTI is that within a selected region of interest, the shear wave velocity generated by the acoustic push pulse is measured. There is a direct link between shear wave velocity and tissue stiffness (m 5 pc2). Therefore the potentially subjective interpretation of an elastogram is not required, and ARFI imaging is postulated to be observer independent and reproducible (Badea et al. 2013; Fraquelli et al. 2007; Nightingale et al. 2000; Sarvazyan et al. 1998). Examiners must, however, be aware of the fact that there is variability in the generation of ARFI images as there are various influencing factors (5 inadequate data acquisition). A differentiation between organ-specific and general influencing parameters can be made. Values of ARFI imaging in the liver increase with food intake and should therefore be evaluated in the fasting state (Goertz et al. 2012). Reproducibility is significantly reduced in patients with chronic liver disease in case of steatosis, increased BMI and lower degree of hepatic fibrosis (Fraquelli et al. 2007). Jaffer et al. (2012) reported that ARFI values obtained deeper to the liver capsule allow more reliable stiffness quantification. In breast mass evaluation, on the contrary, shallower lesion depth and less breast thickness at the area where the lesion is located are known to influencing elastography (Chang et al. 2011). Pre-compression of the investigated tissue is consistently reported to influence the variation in ARFI imaging in all organs (Barr and Zhang 2012; Mantsopoulos et al. 2015; Wojcinski et al. 2013). Precompression can be created by the sonographer, but
6
Ultrasound in Medicine and Biology
Volume -, Number -, 2016
Table 2. Inter- and intra-rater reliability for all glands and for salivary and thyroid glands separately Reproducibility Inter-rater agreement Between groups Within experienced group Within inexperienced group Intra-rater agreement Experienced group Inexperienced group
All glands
Salivary glands
Submandibular gland
Parotid gland
Thyroid gland
0.457 0.527 0.542
0.327 0.301 0.328
0.518 0.285 20.136
0.088 0.436 20.109
0.498 0.811 0.657
0.764 0.680
0.731 0.645
0.555 0.380
0.628 0.496
0.740 0.640
also by the pulsation of adjacent blood vessels close to the selected region of interest or by patient movement and varying degrees of muscle tension. As there is no standardization of adequate compression, the amount of pressure depends on the performer’s subjective experience and probe operating skills (Yoon et al. 2011). Few studies have evaluated the reproducibility of ARFI imaging in experienced examiners, but no studies have evaluated the possible difference in ARFI imaging of the thyroid and salivary glands between experienced and inexperienced examiners. The most striking discrepancy between the groups became noticeable when the submandibular and parotid glands were examined separately with ARFI. Although agreement was moderate for the submandibular glands, there was only poor agreement on the parotid glands; therefore, the parotid examination was responsible for the impaired inter-rater reliability for the salivary glands compared with the collective results. The reproducibility within the two groups was better. The discrepancy between experienced and inexperienced examiners indicates that ARFI results for parotid glands obtained by examiners with different levels of experience have to be interpreted carefully. When inter-rater agreement within the two groups was taken into consideration, there was moderate agreement within experienced examiners for the parotid glands. As inter-rater agreement within inexperienced examiners for the parotid glands and within both groups for submandibular glands was fair or not detectable, it can be recommended that ARFI imaging of parotid glands should be performed only by experienced examiners, whereas ARFI imaging of submandibular glands does not seem promising for both experienced Table 3. Ten single measurements obtained by experienced and inexperienced examiners in the submandibular, parotid and thyroid glands Reproducibility
All
Experienced examiner
Inexperienced examiner
Submandibular gland Parotid gland Thyroid gland
0.14 0.10 0.11
0.13 0.08 0.09
0.16 0.12 0.12
and inexperienced sonographers. These results agree with the investigation by Knopf et al. (2015) on ARFI imaging of salivary glands in patients with Sj€ogren’s syndrome, where the inter-observer reliability of the findings on submandibular glands was judged unreliable. An explanation for the poor inter-rater agreement between experienced and inexperienced examiners in the parotid gland is its anatomic position, which might impair application of the sonography probe between the mastoid process and the mandible, an usually uneven area. Therefore different amounts of pressure need to be applied, which also explains the better inter-rater reliability of the experienced examiners compared with the inexperienced examiners. The pulsation of the facial artery within the posterior part of the submandibular gland could explain the comparatively low inter-rater reliability within both groups for the findings on the submandibular gland, for which the 10 single measurements also had greater variation (compare Table 3). As depth of the selected region of interest influences ARFI imaging, we agreed to measure at a depth of 1.0 cm, being aware that the measurements are no more reliable for depths inferior to 1.0 cm. This might be a limitation of the applied methods, but reflects the realistic depth of ARFI measurements in thyroid and salivary glands. The mean values of salivary gland measurements did not vary significantly between the two examinations. Only a small number of studies have focused on the use of ARFI imaging in salivary gland lesions. Although Matsuzuka et al. (2015) focused primarily on salivary gland stiffness measured with ARFI imaging in predominantly healthy salivary glands, Mansour et al. (2012) evaluated parotid gland lesions with ARFI imaging as part of a multimodal concept. The latter obtained values of 1.92 and 1.75 m/s (60.64) for healthy parotid glands, similar to our value of 1.52 m/s (60.38). Slightly more research has been conducted on diffuse salivary gland alterations, for example, Badea et al. (2013) in submandibular glands after radiation therapy, Knopf et al. (2015) in patients with Sj€ogren’s syndrome and Kaluzny et al. (2014) again in irradiated salivary glands. Badea et al. obtained a value of 1.82 m/s (60.41) in healthy submandibular glands, similar to our value of 1.65 m/s (60.27). The
Reproducibility of ARFI imaging d B. HOFAUER et al.
only investigation on the reliability of shear wave elastography in the neck, which included salivary glands, was published by Bhatia et al. (2012), with only parts of the patient cohort being evaluated by two radiologists to enable calculation of inter- and intra-rater reliability. According to our classification, strong to excellent reliability was reported. Sonoelastographic modalities in general and ARFI imaging in particular for the differentiation of thyroid nodules have been under thorough investigation. Various authors concluded that ARFI imaging is a promising tool for the identification of suspicious thyroid nodules (Cantisani et al. 2014; Lin et al. 2014; Zhang and Han 2013). Even though sonographic evaluation of thyroid nodules is in wide use, to date there have been no investigations concerning the reliability of ARFI imaging in the thyroid gland, comparing the results of experienced examiners versus those of inexperienced examiners. Our study could indicate that there is moderate inter-rater agreement between experienced and inexperienced sonographers. Within the groups, the examiners reached excellent and strong agreement, respectively. Intra-rater agreement was strong. Interand intra-rater reliability of experienced examiners in ARFI imaging of the thyroid gland has been under investigation. Concordant with our findings, mainly strong agreement has been observed (Cantisani et al. 2015; Grazhdani et al. 2014; Zhang et al. 2014c). We did not observe any significant variation in ARFI values between the two examinations.
CONCLUSIONS The reproducibility of results obtained by ARFI imaging of the thyroid and salivary glands between experienced and inexperienced examiners was good. In particular, the results for ARFI imaging of the thyroid gland had the highest levels of inter- and intra-observer agreement, and even inexperienced sonographers had strong inter- and intra-observer agreement. For the salivary glands, attention must be paid to ARFI imaging of the parotid gland, as inter-rater agreement between the groups was poor. Therefore, we recommend that only experienced examiners should perform ARFI imaging on parotid glands. ARFI imaging in submandibular glands does not seem to be reproducible for both experienced and inexperienced examiners. Acknowledgments—The authors thank the volunteers who contributed to this study as healthy probands for the sonographic examinations: Yumiko Matsuba, Constanze Gahleitner, Magdalena Lenschow, Suzan Badawood, Carolin Ibald, Florian Durst, Markus Brandstetter, Michael Szyper, Michael Sedlmeyer and Christian Winkler.
7
REFERENCES Au FW, Ghai S, Moshonov H, Kahn H, Brennan C, Dua H, Crystal P. Diagnostic performance of quantitative shear wave elastography in the evaluation of solid breast masses: determination of the most discriminatory parameter. AJR Am J Roentgenol 2014;203: W328–W336. Badea AF, Tamas Szora A, Ciuleanu E, Chioreanu I, Baciut¸ G, Lups¸or Platon M, Badea R. ARFI quantitative elastography of the submandibular glands. Normal measurements and the diagnosis value of the method in radiation submaxillitis. Med Ultrason 2013;15:173–179. Barr RG, Zhang Z. Effects of precompression on elasticity imaging of the breast: Development of a clinically useful semiquantitative method of precompression assessment. J Ultrasound Med 2012; 31:895–902. Beland MD, Brown SF, Machan JT, Taliano RJ, Promrat K, Cronan JJ. A pilot study estimating liver fibrosis with ultrasound shear-wave elastography: Does the cause of liver disease or location of measurement affect performance. AJR Am J Roentgenol 2014;203: W267–W273. Bhatia K, Tong CS, Cho CC, Yuen EH, Lee J, Ahuja AT. Reliability of shear wave ultrasound elastography for neck lesions identified in routine clinical practice. Ultraschall Med 2012;33:463–468. Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res 1999;8:135–160. Cantisani V, Grazhdani H, Ricci P, Mortele K, Di Segni M, D’Andrea V, Redler A, Di Rocco G, Giacomelli L, Maggini E, Chiesa C, Erturk SM, Sorrenti S, Catalano C, D’Ambrosio F. Q-Elastosonography of solid thyroid nodules: Assessment of diagnostic efficacy and interobserver variability in a large patient cohort. Eur Radiol 2014;24:143–150. Cantisani V, Lodise P, Di Rocco G, Grazhdani H, Giannotti D, Patrizi G, Medvedyeva E, Olive M, Fioravanti C, Giacomelli L, Chiesa C, Redler A, Catalano C, D’Ambrosio F, Ricci P. Diagnostic accuracy and interobserver agreement of quasistatic ultrasound elastography in the diagnosis of thyroid nodules. Ultraschall Med 2015;36: 162–167. C glu M, ¸ ebi Olgun D, Korkmazer B, Kılıc¸ F, Dikici AS, Velidedeo Aydogan F, Kantarcı F, Yılmaz MH. Use of shear wave elastography to differentiate benign and malignant breast lesions. Diagn Interv Radiol 2014;20:239–244. Chang JM, Moon WK, Cho N, Kim SJ. Breast mass evaluation: Factors influencing the quality of US elastography. Radiology 2011;259: 59–64. Choi YJ, Lee JH, Lim HK, Kim SY, Han MW, Cho KJ, Baek JH. Quantitative shear wave elastography in the evaluation of metastatic cervical lymph nodes. Ultrasound Med Biol 2013;39:935–940. Fleury EF, Fleury JC, Piato S, Roveda D. New elastographic classification of breast lesions during and after compression. Diagn Interv Radiol 2009;15:96–103. Fraquelli M, Rigamonti C, Casazza G, Conte D, Donato MF, Ronchi G, Colombo M. Reproducibility of transient elastography in the evaluation of liver fibrosis in patients with chronic liver disease. Gut 2007; 56:968–973. Fukuhara T, Matsuda E, Fujiwara K, Tanimura C, Izawa S, Kataoka H, Kitano H. Phantom experiment and clinical utility of quantitative shear wave elastography for differentiating thyroid nodules. Endocr J 2014;61:615–621. Goertz RS, Egger C, Neurath MF, Strobel D. Impact of food intake, ultrasound transducer, breathing maneuvers and body position on acoustic radiation force impulse (ARFI) elastometry of the liver. Ultraschall Med 2012;33:380–385. Grazhdani H, Cantisani V, Lodise P, Di Rocco G, Proietto MC, Fioravanti E, Rubini A, Redler A. Prospective evaluation of acoustic radiation force impulse technology in the differentiation of thyroid nodules: Accuracy and interobserver variability assessment. J Ultrasound 2014;17:13–20. Hou XJ, Sun AX, Zhou XL, Ji Q, Wang HB, Wei H, Sun JW, Liu H. The application of Virtual Touch tissue quantification (VTQ) in diagnosis of thyroid lesions: A preliminary study. Eur J Radiol 2013; 82:797–801.
8
Ultrasound in Medicine and Biology
Jaffer OS, Lung PF, Bosanac D, Patel VM, Ryan SM, Heneghan MA, Quaglia A, Sidhu PS. Acoustic radiation force impulse quantification: repeatability of measurements in selected liver segments and influence of age, body mass index and liver capsule-to-box distance. Br J Radiol 2012;85:e858–e863. Ka1u_zny J, Kopec T, Szczepanek-Parulska E, Stangierski A, Gurgul E, Rucha1a M, Milecki P, Wierzbicka M. Shear wave elastography: a new noninvasive tool to assess the intensity of fibrosis of irradiated salivary glands in head and neck cancer patients. Biomed Res Int 2014;2014:157809. Knopf A, Hofauer B, Th€urmel K, Meier R, Stock K, Bas M, Manour N. Diagnostic utility of acoustic radiation force impulse (ARFI) imaging in primary Sjoegren’s syndrome. Eur Radiol 2015;25: 3027–3034. Knopf A, Mansour N, Chaker A, Bas M, Stock K. Multimodal ultrasonographic characterisation of parotid gland lesions—A pilot study. Eur J Radiol 2012;81:3300–3305. Lin P, Chen M, Liu B, Wang S, Li X. Diagnostic performance of shear wave elastography in the identification of malignant thyroid nodules: A meta-analysis. Eur Radiol 2014;24:2729–2738. Lyshchik A, Higashi T, Asato R, Tanaka S, Ito J, Hiraoka M, Insana MF, Brill AB, Saga T, Togashi K. Cervical lymph node metastases: Diagnosis at sonoelastography—Initial experience. Radiology 2007;243: 258–267. Mansour N, Bas M, Stock KF, Strassen U, Hofauer B, Knopf A. Multimodal ultrasonographic pathway of parotid gland lesions. Ultraschall Med 2015 [Epub ahead of print]. Mansour N, Stock KF, Chaker A, Bas M, Knopf A. Evaluation of parotid gland lesions with standard ultrasound, color duplex sonography, sonoelastography, and acoustic radiation force impulse imaging - a pilot study. Ultraschall Med 2012;33:283–288. Mantsopoulos K, Klintworth N, Iro H, Bozzato A. Applicability of shear wave elastography of the major salivary glands: values in healthy patients and effects of gender, smoking and pre-compression. Ultrasound Med Biol 2015;41:2310–2318. Matsuzuka T, Suzuki M, Saijo S, Ikeda M, Matsui T, Nomoto Y, Nomoto M, Imaizumi M, Tada Y, Omori K. Stiffness of salivary gland and tumor measured by new ultrasonic techniques: Virtual Touch quantification and IQ. Auris Nasus Larynx 2015;42:128–133. Nightingale KR, Nightingale RW, Palmeri ML, Trahey GE. A finite element model of remote palpation of breast lesions using radiation force: Factors affecting tissue displacement. Ultrason Imaging 2000; 22:35–54.
Volume -, Number -, 2016 Park SH, Kim SJ, Kim EK, Kim MJ, Son EJ, Kwak JY. Interobserver agreement in assessing the sonographic and elastographic features of malignant thyroid nodules. AJR Am J Roentgenol 2009;193: W416–W423. Piccinino F, Sagnelli E, Pasquale G, Giusti G. Complications following percutaneous liver biopsy: A multicentre retrospective study on 68,276 biopsies. J Hepatol 1986;2:165–173. Sarvazyan AP, Rudenko OV, Swanson SD, Fowlkes JB, Emelianov SY. Shear wave elasticity imaging: A new ultrasonic technology of medical diagnostics. Ultrasound Med Biol 1998;24:1419–1435. Tozaki M, Isobe S, Fukuma E. Preliminary study of ultrasonographic tissue quantification of the breast using the acoustic radiation force impulse (ARFI) technology. Eur J Radiol 2011;80:e182–e187. Wojcinski S, Brandhorst K, Sadigh G, Hillemanns P, Degenhardt F. Acoustic radiation force impulse imaging with Virtual Touch tissue quantification: measurements of normal breast tissue and dependence on the degree of pre-compression. Ultrasound Med Biol 2013;39:2226–2232. Yoon JH, Kim MH, Kim EK, Moon HJ, Kwak JY, Kim MJ. Interobserver variability of ultrasound elastography: how it affects the diagnosis of breast lesions. AJR Am J Roentgenol 2011;196:730–736. Zhang B, Ma X, Wu N, Liu L, Liu X, Zhang J, Yang J, Niu T. Shear wave elastography for differentiation of benign and malignant thyroid nodules: A meta-analysis. J Ultrasound Med 2013;32:2163–2169. Zhang FJ, Han RL. The value of acoustic radiation force impulse (ARFI) in the differential diagnosis of thyroid nodules. Eur J Radiol 2013; 82:e686–e690. Zhang FJ, Han RL, Zhao XM. The value of virtual touch tissue image (VTI) and virtual touch tissue quantification (VTQ) in the differential diagnosis of thyroid nodules. Eur J Radiol 2014a;83:2033–2040. Zhang H, Shi Q, Gu J, Jiang L, Bai M, Liu L, Wu Y, Du L. Combined value of Virtual Touch tissue quantification and conventional sonographic features for differentiating benign and malignant thyroid nodules smaller than 10 mm. J Ultrasound Med 2014b;33:257–264. Zhang YF, He Y, Xu HX, Xu XH, Liu C, Guo LH, Liu LN, Xu JM. Virtual Touch tissue imaging on acoustic radiation force impulse elastography: A new technique for differential diagnosis between benign and malignant thyroid nodules. J Ultrasound Med 2014c;33: 585–595. Zhang YF, Xu HX, Xu JM, Liu C, Guo LH, Liu LN, Zhang J, Xu XH, Qu S, Xing M. Acoustic radiation force impulse elastography in the diagnosis of thyroid nodules: Useful or not useful. Ultrasound Med Biol 2015;41:2585–2593.