Strain Elastography for Prediction of Malignancy in Soft Tissue Tumours – Preliminary Results

Abstract

Purpose: To evaluate the ability of strain elastography to predict malignancy in patients with soft tissue tumors, and to compare three evaluation methods of strain elastography: strain ratios, strain histograms and visual scoring.

Materials and Methods: 60 patients with 61 tumors were analyzed in the study. All patients were referred due to suspicion of malignant soft tissue tumors after diagnostic imaging (contrast-enhanced MRI, CT or PET-CT). Ultrasound-guided biopsy was preceded by the recording of strain elastography video clips, which were evaluated in consensus between three investigators. Strain ratio, strain histogram analysis and visual scoring using a five-point visual scale were compared with the final pathology from either biopsy or resection of the tumors.

Results: The difference between the mean strain ratio for malignant and benign tumors was significant (p = 0.043). The mean strain ratios for malignant and benign tumors were 1.94 (95 % CI [0.37; 10.21]) and 1.35 (95 % CI [0.32; 5.63]), respectively. There were no significant differences for strain histograms or visual scoring. Liposarcomas had lower mean strain ratio, strain histogram values, and visual scoring than other malignant tumors. When analyzing a subgroup of patients without fat-containing tumors (n = 46), based on appearance on MRI or CT, the difference between the mean strain ratios for malignant and benign tumors increased (p = 0.014).

Conclusion: The mean strain ratios of malignant tumors were significantly higher than the mean strain ratios of benign tumors. There was no significant difference for strain histograms and visual scoring. Strain ratios may be used as an adjunct in soft tissue tumor diagnosis, possibly minimizing the number of biopsies.

Zusammenfassung

Ziel: Bewertung, ob die Strain-Elastografie in der Lage ist, Malignität bei Patienten mit Weichteiltumoren vorherzusagen, sowie Vergleich von drei Auswertungsmethoden der Strain-Elastografie: Strain-Ratios, Strain-Histogramme und die visuelle Bewertung.

Material und Methoden: In der Studie wurden 60 Patienten mit 61 Tumoren analysiert. Alle Patienten wurden wegen Verdachts auf bösartige Weichteiltumore nach bildgebender Diagnostik (kontrastverstärktes MRT, CT oder PET-CT) aufgenommen. Eine ultraschall-gestützte Biopsie ging der Aufnahme der Strain-Elastografie Video-Clips voran, die von drei Untersuchern im Konsensus ausgewertet wurden. Die Strain-Ratio, die Analyse des Strain-Histogramms und die visuelle Bewertung mittels einer 5-Punkte visuellen Skala wurden mit dem pathologischen Endbefund der Biopsie oder des entfernten Tumors verglichen.

Ergebnisse: Der Unterschied zwischen mittlerer Strain-Ratio von gutartigen und malignen Tumoren war statistisch signifikant (p = 0,043). Die mittlere Strain-Ratio für maligne Tumoren betrug 1,94 (95 % CI [0,37; 10,21]) und die entsprechende für gutartige Tumoren war 1,35 (95 % CI [0,32; 5,63]). Es gab keine signifikanten Unterschiede zwischen Strain-Histogrammen und visueller Bewertung. Liposarkome hatten niedrigere mittlere Werte für Strain-Ratios, Strain-Histogramme und visuelle Bewertungen als andere maligne Tumoren. Wenn man jedoch eine Untergruppe an Patienten analysierte, deren Tumoren aufgrund des Befundes im MRT oder CT kein Fett beinhaltete (n = 46), so war die Differenz der mittleren Strain-Ratios zwischen malignen und benignen Tumoren erhöht (p = 0,014).

Schlussfolgerung: Die mittleren Strain-Ratios von malignen Tumoren waren signifikant höher als diejenigen der gutartigen Tumoren. Es gab keinen signifikanten Unterschied zwischen Strain-Histogrammen und visueller Bewertung. Die Strain-Ratios können als zusätzlicher Marker bei der Diagnostik von Weichteiltumoren dienen und möglicherweise die Zahl der Biopsien reduzieren.

Key words

strain elastography - soft tissue tumors - sarcoma - ultrasound - musculoskeletal tissue

Introduction

Strain elastography is a method for visualizing tissue stiffness by means of changes in tissue strain in response to manual compression. Strain measurements are translated to color-coded strain maps, elastograms, by the ultrasound (US) system[1] [2]. Elastograms are depictions of relative tissue strain, as the manual deformation cannot be quantified. The elastogram illustrates differences in tissue stiffness by colors from red (soft) over yellow and green to blue (stiff/hard). Elastography has been evaluated in a range of anatomical areas, for example the liver[3] [4], the thyroid gland[5] [6] and breast tissue[7] [8], but few studies have evaluated elastography in musculoskeletal tissue[9].

Strain elastography can be assessed by calculating strain ratios, by strain histograms and by visual scoring, the first two being semi-quantitative methods and the last being qualitative. The strain ratio is based on two regions of interest (ROIs), one placed within the focal lesion and one placed in reference tissue, preferably at the same depth as the focal lesion[10]. Assuming that equal stress is applied to both ROIs, the relative difference in strain between the two areas can be calculated. The strain histogram is a bar diagram displaying the distribution of colors in an ROI. Each pixel in the ROI has a color, which corresponds to a value on a scale from 0 to 255. A mean pixel value can be calculated from the strain histogram. A third method is visual scoring of elastograms, which involves the rating of elastograms depending on colors and patterns. Visual scoring is well validated in breast tissue using the Tsukuba Elastography Score proposed by Itoh et al. [7], but there is no validated scoring system for focal lesions in soft tissue, muscles or bone.

Among malignant soft tissue tumors, sarcomas constitute a diagnostic and therapeutic entity. Sarcomas originate from mesenchymal cells and can occur in many anatomical areas[11]. Soft tissue sarcomas develop in muscle, fibrous tissue, fat, blood vessels and other supporting tissues of the body. Osteosarcomas develop in bone. In 2013, approximately 11 400 Americans were diagnosed with soft tissue sarcoma[12]. In Europe, the estimated incidence of sarcoma is 4/100 000/year[13]. Sarcomas may be aggressive cancers and multidisciplinary treatment planning is mandatory in all cases. According to the clinical recommendations of the European Society for Medical Oncology (ESMO), diagnosis and treatment should be carried out in reference centers for sarcomas[13].

The aim of this study was to evaluate the ability of strain elastography to predict malignancy in soft tissue tumors in patients referred for biopsy of a soft tissue tumor with a clinical suspicion of sarcoma or metastasis. Furthermore, the aim was to compare three evaluation methods: strain ratios, strain histograms and visual scoring.

Materials and methods

Patients

The study group consisted of patients referred to the sarcoma center at our hospital due to a clinical suspicion of malignant soft tissue tumor. This center covers a population of 2.5 million people and it is specialized in the diagnosis of sarcoma, including biopsy and treatment. In three months 61 patients (27 males and 34 females) with 62 tumors were consecutively and prospectively included regardless of tumor size and anatomical location. 1 of 61 patients had an inconclusive biopsy and was excluded from the study. In total 61 tumors were included. In a post-hoc analysis this tumor number corresponded to a power of 0.6.

All patients had undergone diagnostic imaging (contrast-enhanced MRI, CT or PET-CT)[14] which was evaluated at a multidisciplinary conference, and all patients were scheduled for ultrasound-guided biopsy ([Fig. 1]). Final pathology served as the gold standard for the patients who underwent surgery. For the patients, who did not undergo surgery, histological biopsy served as the gold standard.

Fig. 1 Inclusion criteria.

Abb. 1 Einschlusskriterien.

The Danish National Committee on Biomedical Research Ethics (Journal: H-2 – 2014-FSP1) and the Danish Data Protection Agency (ID: 02 866, Journal: 30 – 1200) approved the study. All patients gave informed consent before participation.

Imaging

Most scans had been performed in other hospitals due to our center being a national sarcoma center, and the protocols varied. In most cases CT and PET/CT images were available in 3 planes with a slice thickness of 3 mm, kV and mAs varied. According to guidelines for soft tissue sarcoma, a minimum of three MRI sequences was available, often supplemented by more sequences including contrast sequences ([Fig. 2]).

Fig. 2 Axial MRI images of tumor on the right upper thigh (same tumor as in Fig. 4a). The tumor is 1.9 cm (largest diameter). a T1 sequence without contrast and b T1 sequence after contrast administration, where contrast enhancement is seen in the tumor (marked with black horizontal arrow). Histology showed myxoma.

Abb. 2 Axiale MRT-Bilder des Tumors am oberen rechten Oberschenkel (derselbe Tumor wie in Abb. 4a). Der Tumor hat 1,9 cm (größter Durchmesser). a T1-Sequenz ohne Kontrast und b T1-Sequenz nach Kontrastmittelgabe mit sichtbarer Kontrastaufnahme im Tumor (markiert mit schwarzem horizontalem Pfeil). Die Histologie zeigte ein Myxom.

US examinations were performed using a Logiq E9 system (GE Healthcare, Chalfont St. Giles, UK) with one of three different probes (9 L, ML6 – 15, C1 – 5) according to anatomical region and tumor depth. US preset also depended on the anatomical region. B-mode ultrasound was performed initially in all patients to locate the tumor and determine the needle entry point and track for insertion. Elastography was then performed in the area where biopsy was planned to ensure comparability with histology.

One of three different physicians experienced in elastography performed all elastography examinations (IRC, JFC, CE). Manual compression of the tissue at a frequency in the range of 80 – 120/min was applied with the ultrasound transducer. To avoid excessive tissue compression, minimal force was exerted during elastography. Elastogram quality was monitored by use of an on-screen color scale, where green indicated the best quality ([Fig. 3]). All physicians aimed at including reference tissue adjacent to the tumor and the elastography box covered the entire image. The B-mode image and the elastography image were shown side by side and the same color scale was used for all patients. Video clips of 10 s were recorded for all patients and stored in the system.

Fig. 3 Elastogram and B-mode image of a benign tumor, a schwannoma. Tumor ROI is blue and reference ROI is yellow. The color bar, which indicates the quality of the elastogram, is placed in the upper left corner of the elastogram.

Abb. 3 Elastogramm und B-Bild eines gutartigen Tumors, einem Schwannom. Die Tumor-ROI ist blau und die Referenz-ROI gelb. Der Farbbalken, der die Qualität des Elastogramms anzeigt, befindet sich an der oberen linken Ecke des Elastogramms.

After the elastography examination, histological biopsy was performed in local anesthesia (lidocaine 10 mg/ml) using an 18-gauge automatic biopsy gun (BARD, Tempe, USA). At least two histological biopsies were performed in the same area in order to obtain a representative sample.

Evaluation of elastography images

Assessment of the elastograms was performed with blinding to the indication for referral and to the final diagnosis. All elastography videos were evaluated at the end of the study period. For each patient the best quality video clip was chosen by consensus between the three physicians. This video clip was used for calculation of strain ratio, strain histogram analysis and visual scoring.

Strain ratios were calculated from two ROIs placed by two of the three physicians. The ROIs were chosen to cover as much of the tumor and the reference tissue as possible, and were preferably placed at the same depth. Due to the compressions, movements within the elastogram were inevitable. Therefore, tumor ROIs were placed with a small distance to the boundaries of the tumor. In this way, the tumor ROI remained inside the tumor and did not include reference tissue in any cine loop frame. Reference ROIs and tumor ROIs were not necessarily of the same size, which depended on tumor size and location. All ROIs were ellipsoid ([Fig. 3]). Data from each recorded frame were saved from the system and a mean strain ratio was calculated for the entire clip, omitting frames without color code.

Strain histograms were constructed from an ROI placed within the tumor similar to the tumor ROI in the elastogram from which the strain ratios were calculated. The original elastography video was decompressed and imported to ImageJ (downloaded at nih.gov), which constructed histograms for each frame ([Fig. 4a, b]). The histogram data were then exported for the calculation of mean pixel values.

Fig. 4 a Elastogram and B-mode image of a benign muscular tumor, a myxoma (same tumor as in [Fig. 2]). The tumor ROI and calculations of the histogram were made in ImageJ after decompression of the original elastography video. b Histogram for a myxoma (same patient as in [Fig. 4a]). The histogram is a bar chart, which illustrates the distribution of colors in the tumor ROI (yellow oval in [Fig. 4a]). Each pixel in the histogram corresponds to a number on the x-axis. This histogram has two spikes; one at 120 and one at 240. The spikes represent the two most common colors within the tumor ROI. Using the horizontal color bar, the numbers on the x-axis translate into the colors green (120) and blue (240).

Abb. 4 a Elastogramm und B-Bild eines gutartigen muskulären Tumors, einem Myxom (derselbe Tumor wie in [Abb. 2]). Die Tumor-ROI und die Berechnungen des Histogramms wurden im ImageJ nach Dekompression des originalen Elastografie-Videos durchgeführt. b Histogramm eines Myxoms (derselbe Patient wie in [Abb. 4a]). Das Histogramm ist ein Balkendiagramm, das die Verteilung der Farben in der Tumor-ROI darstellt (gelbes Oval in [Abb. 4a]). Jeder Pixel des Histogramms entspricht einer Zahl auf der x-Achse. Dieses Histogramm hat zwei Spitzen, die eine bei 120, die andere bei 240. Die Spitzen stellen die beiden häufigsten Farben innerhalb der Tumor-ROI dar. Mithilfe des horizontalen Farbbalkens werden die Zahlen der x-Achse in die Farben Grün (120) und Blau (240) übersetzt.

Visual scoring was performed by consensus between the three readers, using a five class scoring system, the Tsukuba Elastography Score ([Fig. 5]). Inter- and intraobserver variation was not assessed in this study.

Fig. 5 a Elastogram and B-mode image for visual scoring of a grade III sarcoma. b Tsukuba Elastography Score modified according to Itoh et al. (7). In breast tissue, 4 and 5 are considered indicators of malignancy.

Abb. 5 a Elastogramm und B-Bild für die visuelle Beurteilung eines Grad III Sarkoms. b Tsukuba-Elastografie-Score modifiziert nach Itoh et al. (7). Im Brustgewebe werden 4 und 5 als Zeichen für Malignität gewertet.

Statistical analysis

Descriptive statistics including t-tests were calculated using Microsoft Excel (version 2011, Redmond, WA, USA) and SPSS version 20 (IBM, New York, USA).

In the case of skewness in the distribution of data, data were log-transformed.

The level of significance was set to p < 0.05. The possible effect of clustering from one patient having 2 separate tumors was not included in the calculations.

Results

19 tumors were diagnosed as malignant and 42 were diagnosed as benign. 10 tumors were sarcomas. The final pathology is shown in [Table 1]. The average size of benign tumors was 6.49 cm +/– 4.62 and the average size of malignant tumors was 7.22 +/– 3.96 when measuring the largest diameter.

Table 1 Benign and malignant tumors.

benign tumors
benign fat-containing tumor	n = 11
schwannoma	n = 7
benign vascular tumor	n = 6
fibromatosis/fibrosis	n = 6
non-malignant tumors, no further specification	n = 5
myxoma	n = 2
inflammation	n = 2
non-specific reactive alteration	n = 2
cyst	n = 1
total	n = 42
malignant tumors
liposarcoma	n = 4
sarcoma, undifferentiated	n = 2
sarcoma, grade III	n = 1
chondrosarcoma	n = 1
osteosarcoma	n = 1
myxofibrosarcoma	n = 1
metastasis	n = 5
lymphoma	n = 4
total	n = 19

The mean strain ratios for malignant and benign tumors were 1.94 (95 % CI [0.37; 10.21]) and 1.35 (95 % CI [0.32; 5.63]), respectively. There was a significant difference between the mean strain ratios for malignant and benign tumors (p = 0.043).

For strain histograms, the mean pixel values for malignant and benign tumors were 177.09 (95 % CI [99.86; 254.31]) and 171.59 (95 % CI [86.12; 257.07]), respectively. There was no significant difference between the mean pixel values for malignant and benign tumors ([Table 2]).

Table 2 Mean strain ratio, strain histogram and visual scoring values for each group of tumors investigated and p-values from t-tests comparing the groups.

	strain ratio		strain histogram		visual scoring
	mean	p-value	mean	p-value	mean		p-value
benign (N = 42)	1.35	0.043[1]	172	0.3171	2.43	0.4141
malignant (N = 19)	1.94		177		2.37
fat-containing tumors (N = 15)	1.41	0.345[2]	154	0.4832	1.91	0.3652
liposarcomas (N = 4) myxoliposarcoma (N = 3) highly differentiated sarcoma (N = 1)	1.36		143		1.50
benign without lipomas (N = 27)	1.33	0.014[3]	177	0.3593	2.61	0.3523
malignant without liposarcomas (N = 15)	2.13		186		2.60

¹ p-values are calculated from unpaired t-tests between benign and malignant tumors.

² p-values are calculated from unpaired t-tests between lipomas and liposarcomas.

³ p-values are calculated from unpaired t-tests between benign tumors without lipomas and malignant tumors without liposarcomas.

For visual scoring, the mean scores for malignant and benign tumors were 2.43 (95 % CI [0.40; 4.46]) and 2.37 (95 % CI [0.46; 4.28]), respectively. There was no significant difference between the visual scores for malignant and benign tumors ([Table 2]).

The mean values for strain ratio, strain histograms and visual scoring for liposarcomas were lower than for the entire group of malignant tumors ([Table 2]). In a subgroup analysis of 46 patients who, based on the CT or MRI appearance, had tumors that did not contain fat, the significance between the mean strain ratio values for malignant and benign tumors increased (p = 0.014). For strain histograms and visual scoring, the difference between the means for malignant and benign tumors remained insignificant ([Fig. 6]).

Fig. 6 Box plots of a mean strain ratios, unitless (y-axis) and b mean pixel values, unitless (y-axis) of strain histograms for different tumor groups. Lower and upper box levels are lower and upper quartiles respectively. The horizontal line within the box marks the median. Whiskers indicate the 95 % confidence intervals. c Column charts of the visual scoring in benign vs. malignant soft tissue tumors.

Abb. 6 Box-Plots a der mittleren Strain-Ratios, ohne Einheiten (y-Achse) und b der mittleren Pixelwerte, ohne Einheiten (y-Achse) der Strain-Histogramme jeweils für die verschiedenen Tumorarten. Die unteren und oberen Begrenzungen der Kästen entsprechen jeweils den unteren und oberen Quartilen. Die horizontale Linie im Kasten markiert den Median. Die Whisker zeigen die 95 % Konfidenzintervalle an. c Säulendiagramme der visuellen Bewertung in gutartige vs. maligne Weichteiltumore.

Two benign tumors had SRs that were higher than the upper limit value of the confidence interval for benign tumors, i. e. above 5.63. These were hyalinized fibrosis (SR = 5.90) and an intramuscular, subfascial lipoma in the pectoral muscle (SR = 7.27). Two malignant tumors had SRs that were lower or very close to the lower limit value of the confidence interval for malignant tumors, i. e. close to 0.37. These were a myxoliposarcoma (SR = 0.37) and a liposarcoma (SR = 0.38).

Discussion

To our knowledge, this study is the first to evaluate strain ratios in soft tissue tumors. There was a significant difference between the means for benign and malignant tumors, but the confidence intervals overlapped. Better differentiation between malignant and benign tumors by strain elastography can potentially reduce the number of histological biopsies. The difference in strain ratio between malignant and benign tumors was even more evident when tumors that did not contain fat on CT or MRI were studied alone. Especially well-differentiated liposarcomas had a low SR, which made it difficult to distinguish them from benign tumors. There was no significant difference between malignant and benign tumors, when using strain histograms or visual scoring to assess tumor stiffness.

Four tumors had SR values that were not contained in the benign or malignant confidence interval – two benign and two malignant. One of the benign lesions was hyalinized fibrosis in relation to the quadriceps muscle, which had a high SR. The other benign lesion was an intramuscular lipoma in the pectoral muscle. The tumor appeared hard, maybe due to the superficial location of the tumor or the fact that the tumor was located in close proximity to the soft tissue of the patient’s breast. The hard appearance of this lipoma might also be caused by the location inside an inflexible thoracic muscle in a region with rigid structures. The two malignant tumors with low SR values were a liposarcoma and a myxoliposarcoma. Both derive from fat tissue, which is soft, as do lipomas. There were no calcifications in these, which made it difficult to distinguish them from benign tumors by elastography alone. In these cases considering the tumor appearance on B-mode concurrently with the elastography examination might have helped in the differentiation. In breast tumors the combination of B-mode US and elastography has been shown to improve diagnostic accuracy, whereas elastography on its own has inferior diagnostic value compared with B-mode. Still MRI is considered the imaging modality of choice for soft tissue sarcomas[14].

Musculoskeletal elastograms are highly inhomogeneous due to a large variety of stiffness from bone to soft tissue. To minimize the inhomogeneity, it is important to agree on standard settings for the system, for the scanning procedure and for the evaluation of the images. Also, standardization is important to be able to establish cut-offs between benign and malignant tumors, to allow for comparison and reproduction of elastography results, and to minimize observer dependency. In this study only three physicians performed elastography and in all examinations the elastography color overlay covered the entire image. This approach was chosen to obtain the most representative elastogram. The elastograms and videos used for further analysis were selected by consensus[1] [15] [16].

In order to calculate strain ratios accurately, placing and sizing of the ROIs was essential, but no guidelines were available, and previous studies have had different approaches[17] [18] [19]. The tumor ROI and reference ROI were located at the same depth to avoid different dampening of the acoustic impulses. All ROIs were ellipsoid and were chosen to cover as much tissue as possible to ensure that both the tumor and reference ROI were representative, despite the heterogeneity of the tissues.

The visual score did not show a significant difference between malignant and benign tumors. The Tsukuba score was developed for the classification of breast tumors, and is validated in this setting[20] [21] [22]. It is the most frequently used visual scoring system of strain elastography. Unlike breast tissue, musculoskeletal tissue is very heterogeneous. Furthermore, breast tumors represent a more homogeneous group compared to soft tissue tumors, which arise from various tissues. This might be the reason why the results were insignificant. We did not evaluate the intra- and interobserver variation, but all visual scoring was done by consensus between three physicians. In previous studies on the Achilles tendon, two different scoring systems have been used, one having 3 classes and the other having 2 classes. Recently an article on strain elastography in 32 superficial soft tissue tumors using a five-point visual scoring system was published and showed a very high sensitivity (100 %) and specificity (97 %)[9]. The five-point scale was not described in detail, which makes direct comparison with our study difficult. Furthermore, biopsy was only performed in 12 of 32 tumors, while the remaining 24 tumors were evaluated at follow-up only three months after the elastographic assessments.

Strain histograms did not show a significant difference between benign and malignant lesions, although the method has shown significant results in lymph nodes, the pancreas and breast tissue [22] [23] [24]. This may be due to the heterogeneity of soft tissue tumors.

All calculations were performed at the end of the study period, and evaluators were blinded to the final pathology as well as results from previous imaging to decrease the risk of bias. This is unlike the clinical procedure, where the physician knows all previous results. In some cases knowing the suspected diagnosis would probably have improved the evaluation of the elastograms.

The incidence of sarcoma in the present study population is high (18 %) compared to the general population due to the hospital being a national sarcoma center. The high number of malignant lesions in the present study allowed for evaluation of elastography performance, despite sarcomas being a rare finding in general. Applying elastography in primary or secondary centers will probably yield different results.

A limitation of the study is that the quality of the elastograms was not evaluated systematically. As the acoustic impedance increases along with depth, fewer echoes travel back, and fewer high quality frames are available for deep lying lesions. Therefore, the same number of available frames was not available for calculations of all tumors.

In a few elastograms, it was impossible to place the reference ROI at the same depth or superficial to the same structures as the tumor ROI, which may also have influenced the evaluation, i. e. tissue superficial to bone is depicted softer than identical tissue not in the proximity of bone.

Conclusion

The mean strain ratios of malignant tumors were significantly higher than the strain ratios of benign tumors. When analyzing all patients without fat-containing tumors, the difference between the mean strain ratios for malignant and benign tumors increased. This signifies that strain ratios may in the future be used as an adjunct in the diagnosis of soft tissue tumors. Strain histograms and visual scoring did not show any significant differences between benign and malignant tumors.

• References

• 1 Bamber J, Cosgrove D, Dietrich CF et al. EFSUMB guidelines and recommendations on the clinical use of ultrasound elastography. Part 1: Basic principles and technology. Ultraschall in Med 2013; 34: 169-184
• 2 Cosgrove D, Piscaglia F, Bamber J et al. EFSUMB guidelines and recommendations on the clinical use of ultrasound elastography. Part 2: Clinical applications. Ultraschall in Med 2013; 34: 238-253
• 3 Sporea I, Sirli RL. Hepatic elastography for the assessment of liver fibrosis--present and future. Ultraschall in Med 2012; 33: 550-558
• 4 European Association for the Study of the Liver. EASL Clinical Practice Guidelines: management of hepatitis C virus infection. J Hepatol 2011; 55: 245-264
• 5 Vorländer C, Wolff J, Saalabian S et al. Real-time ultrasound elastography--a noninvasive diagnostic procedure for evaluating dominant thyroid nodules. Langenbecks Arch Surg Dtsch Ges Für Chir 2010; 395: 865-871
• 6 Cantisani V, D’Andrea V, Biancari F et al. Prospective evaluation of multiparametric ultrasound and quantitative elastosonography in the differential diagnosis of benign and malignant thyroid nodules: preliminary experience. Eur J Radiol 2012; 81: 2678-2683
• 7 Itoh A, Ueno E, Tohno E et al. Breast disease: clinical application of US elastography for diagnosis. Radiology 2006; 239: 341-350
• 8 Berg WA, Cosgrove DO, Doré CJ et al. Shear-wave elastography improves the specificity of breast US: the BE1 multinational study of 939 masses. Radiology 2012; 262: 435-449
• 9 Magarelli N, Carducci C, Bucalo C et al. Sonoelastography for qualitative and quantitative evaluation of superficial soft tissue lesions: a feasibility study. Eur Radiol 2014; 24: 566-573
• 10 Havre RF, Waage JR, Gilja OH et al. Real-Time Elastography: Strain Ratio Measurements Are Influenced by the Position of the Reference Area. Ultraschall in Med 2011;
• 11 Nystrom LM, Reimer NB, Reith JD et al. Multidisciplinary management of soft tissue sarcoma. ScientificWorldJournal 2013; 852462 2013
• 12 National Cancer Institute. A Snapshot of Sarcoma. Internet. cited 2014 Jul 3. Available from: http://www.cancer.gov/researchandfunding/snapshots/sarcoma
• 13 Casali PG, Blay JY. Soft tissue sarcomas: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol 2010; 21: v198-v203
• 14 ESMO/European Sarcoma Network Working Group. Soft tissue and visceral sarcomas: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol Off J Eur Soc Med Oncol ESMO 2014; 25: iii102-iii112
• 15 Chino K, Akagi R, Dohi M et al. Reliability and validity of quantifying absolute muscle hardness using ultrasound elastography. PloS One 2012; 7: e45764
• 16 Martínez RR, Galán del Río F. Mechanistic basis of manual therapy in myofascial injuries. Sonoelastographic evolution control. J Bodyw Mov Ther 2013; 17: 221-234
• 17 Park GY, Kwon DR. Sonoelastographic evaluation of medial gastrocnemius muscles intrinsic stiffness after rehabilitation therapy with botulinum toxin a injection in spastic cerebral palsy. Arch Phys Med Rehabil 2012; 93: 2085-2089
• 18 Stachs A, Hartmann S, Stubert J et al. Differentiating between malignant and benign breast masses: factors limiting sonoelastographic strain ratio. Ultraschall in Med 2013; 34: 131-136
• 19 Kwon DR, Park GY, Lee SU et al. Spastic cerebral palsy in children: dynamic sonoelastographic findings of medial gastrocnemius. Radiology 2012; 263: 794-801
• 20 Thomas A, Degenhardt F, Farrokh A et al. Significant differentiation of focal breast lesions: calculation of strain ratio in breast sonoelastography. Acad Radiol 2010; 17: 558-563
• 21 Fischer T, Peisker U, Fiedor S et al. Significant differentiation of focal breast lesions: raw data-based calculation of strain ratio. Ultraschall in Med 2012; 33: 372-379
• 22 Hui Zhi XYX. Ultrasonic elastography in breast cancer diagnosis: strain ratio vs 5-point scale. Acad Radiol 2010; 17: 1227-1233
• 23 Săftoiu A, Vilmann P, Gorunescu F et al. Accuracy of endoscopic ultrasound elastography used for differential diagnosis of focal pancreatic masses: a multicenter study. Endoscopy 2011; 43: 596-603
• 24 Săftoiu A, Vilmann P, Hassan H et al. Analysis of endoscopic ultrasound elastography used for characterisation and differentiation of benign and malignant lymph nodes. Ultraschall in Med 2006; 27: 535-542