- Email: [email protected]
Differentiation between idiopathic granulomatous mastitis and malignant breast lesions using strain ratio on ultrasonic elastography
+Model
ARTICLE IN PRESS
DIII-958; No. of Pages 7
Diagnostic and Interventional Imaging (2017) xxx, xxx—xxx
ORIGINAL ARTICLE /Breast imaging
Differentiation between idiopathic granulomatous mastitis and malignant breast lesions using strain ratio on ultrasonic elastography B. Ya˘ gcı a, I. Erdem Toslak a,b,∗, B. C ¸ekic ¸ a, M. Öz a, ¸ c, M. Akdemir d, S. Yıldız a, D. Süren e, B.R. Karakas D. Bova b a
Department of Radiology, Antalya Training and Research Hospital, Health Sciences University, Antalya, Turkey b Department of Radiology, Loyola University Medical Center, Maywood, Illinois, USA c Department of General Surgery, Antalya Training and Research Hospital, Health Sciences University, Antalya, Turkey d Department of Public Health, Akdeniz University Medical School,Antalya, Turkey e Department of Pathology, Antalya Training and Research Hospital, Health Sciences University, Antalya, Turkey
KEYWORDS Ultrasonography; Elastography; Strain ratio; Granulomatous mastitis; Breast cancer
∗
Abstract Purpose: The goal of this study was to investigate the strain elastography imaging characteristics of idiopathic granulomatous mastitis (IGM) and compare strain ratio values of IGM with those of breast cancer. Material and methods: Twenty-three consecutive women with IGM (mean age, 37.9 ± 6.6 [SD] years; range: 26—52 years) and 45 women with malignant breast tumor (mean age, 52.8 ± 12.0 [SD], range, 32—77 years) who had been scheduled for ultrasound-guided core biopsy were recruited to the study. All had ultrasonography with elastography before biopsy. The strain ratios of lesions were calculated using surrounding normal breast tissue as the reference in both groups and compared between the two groups. Receiver-operating-characteristics (ROC) curves were formed. Sensitivity, specificity, cut-off, and area under curve (AUC) values were calculated.
Abbreviations: UE, ultrasound elastography; IGM, idiopathic granulomatous mastitis; SR, strain ratio. Corresponding author at: Department of Radiology, Antalya Training and Research Hospital, Health Sciences University, Antalya, Turkey. E-mail address: [email protected] (I. Erdem Toslak).
http://dx.doi.org/10.1016/j.diii.2017.06.009 2211-5684/© 2017 Editions franc ¸aises de radiologie. Published by Elsevier Masson SAS. All rights reserved.
Please cite this article in press as: Ya˘ gcı B, et al. Differentiation between idiopathic granulomatous mastitis and malignant breast lesions using strain ratio on ultrasonic elastography. Diagnostic and Interventional Imaging (2017), http://dx.doi.org/10.1016/j.diii.2017.06.009
+Model DIII-958; No. of Pages 7
ARTICLE IN PRESS B. Ya˘ gcı et al.
2
Results: The mean strain ratio on sonoelastography was 1.5 ± 0.8 (SD) (range: 0.2—4.0) for IGM and 5.3 ± 5.2 (SD) (range: 1.4—33) for malignant lesions. Strain ratio values in IGM lesions were significantly lower than in malignant lesions (P < 0.05). ROC test yielded an AUC value of 0.939 (95% confidence interval, 0.882—0.995; P < 0.0001). Optimal cut-off value for strain ratio value was 2.5 yielding 87% sensitivity and 96% specificity for the diagnosis of IGM. Conclusion: Sonoelastographic strain ratio contributes to differentiate IGM from malignant breast lesions, thus has potential to influence clinical decision making for further biopsies. © 2017 Editions franc ¸aises de radiologie. Published by Elsevier Masson SAS. All rights reserved.
Introduction Intrinsic elastic features of the tissues may be altered in certain circumstances; neoplasm is one of the pathophysiologic process influencing the elasticity of the tissues [1]. Generally, stiffer and immobile breast lesions suggest malignancy [1]. On the basis of this principle, ultrasound elastography (UE) has been developed as a more objective method to estimate tissue stiffness. The degree of tissue displacement is greater in harder tissues than in softer tissues; this is represented by an overlapping color-encoded elasticity map of the tissues with a continuous scale, from blue representing hardest to red representing softest tissues. UE can be used in various organs including thyroid, liver, kidney, spleen, muscles, or even can be further used for the detection of endometrium carcinoma and evaluation of fetal lungs [1—4]. UE has been shown beneficial in diagnosis of breast lesions [1]. Breast carcinoma is associated with elastosis, which increases progressively with the severity of the disease [5]. Breast cancer has higher strain ratio values, or elasticity scores, as compared to benign breast lesions [6—9]. However, few studies have compared specific subtypes of benign and malignant breast masses [10,11]. Idiopathic granulomatous mastitis (IGM) is a rare and benign chronic granulomatous inflammatory disease of breast, which is often confused with a malignant process, both clinically and radiographically. Imaging features that suggest the diagnosis of IGM remain non-specific and not always found in all patients. Thus, the final diagnosis usually requires tissue biopsy [12]. Because the current imaging modalities are not sufficient to establish a definitive diagnosis of IGM in most patients, and because prior histochemical and molecular studies of breast cancer reveal increased amounts of elastotic material in cancerous lesions, the question arose if overlap in elasticity values of IGM and breast cancer existed or if these may be helpful in diagnosis and clinical management of this rare entity [13]. Given the increased amount of periductal and stromal elastic fibers in breast carcinoma and lack of this feature in non-neoplastic conditions (except for radial scars), this feature can be used to differentiate IGM lesions from stiffer breast cancer lesions [13,14]. In light of this knowledge, UE may be useful in differentiating IGM from the breast cancer lesions. The goal of this study was to investigate the strain elastography imaging characteristics of IGM and compare strain
ratio (SR) values of IGM with those of malignant breast lesions.
Materials and methods Study design This prospective study was conducted in accordance with the 1964 declaration of Helsinki and was approved by the Institutional Ethics Committee; written informed consent was obtained from all subjects. From May 2014 through December 2015, 103 consecutive women who had been referred to our department for ultrasound-guided core biopsy were examined by sonoelastography. All participants underwent strain elastography imaging following conventional breast ultrasound examinations. Patients less than 18-year-old, with a benign pathological diagnosis other than IGM (fibroadenoma, adenosis, radial sclerosing lesion, fibrocystic disease), with mass lesion at the same location of a previously diagnosed IGM, with prior surgery or biopsy of the mass being examined, or with a non-mass lesion on ultrasound were excluded from analysis. We also excluded precancerous lesions (ductal carcinoma in-situ [DCIS]) and certain histologic types of breast cancer (mucinous and colloid carcinoma) on the basis of previous studies’ results which showed that these lesions had potential to yield false negative imaging characteristics at UE [11,15,16]. Women underwent ultrasound-guided percutaneous core needle biopsy based on the clinical decision of the referring physician. Strain elastography and B-mode gray-scale ultrasound evaluation of the lesions were performed within a week of referral; both were performed within the same session and prior to obtaining tissue diagnosis. The presence of non-caseating granulomas in biopsy samples suggested IGM diagnosis. In patients with multiple lesions, only the largest lesion was considered for the purposes of this study; 5 mm was selected as the lower size limit to be included in the study.
Ultrasonographic imaging All patients underwent a conventional diagnostic breast ultrasound examination, conducted with the same scanner
Please cite this article in press as: Ya˘ gcı B, et al. Differentiation between idiopathic granulomatous mastitis and malignant breast lesions using strain ratio on ultrasonic elastography. Diagnostic and Interventional Imaging (2017), http://dx.doi.org/10.1016/j.diii.2017.06.009
+Model DIII-958; No. of Pages 7
ARTICLE IN PRESS
Sonoleastography of idiopathic granulomatous mastitis (Hi-Vision Prerius; Hitachi medical systems, Japan) equipped with a high-resolution 8—12 MHz linear array transducer and an elastography module. All ultrasound examinations were performed by the same radiologist with 3 years of experience in elastography imaging, blinded to the clinical diagnosis. Both breasts and axilla were scanned in longitudinal and transverse planes. All subjects were examined in supine position. Still and cine images of the palpable mass or area of concern were recorded electronically in the digital imaging and communications in medicine (DICOM) format. Size, shape, margin, echogenicity, and posterior acoustic phenomena of the mass or lesion were assessed while recording images. Free hand elastography and standard sonographic evaluation were performed during the same session. Elastographic images were acquired by using the freehand repeated manual compression technique using the transducer [9]. In this technique, the transducer is gently compressed and released over the lesion periodically with split-screen recording, displaying real-time B-mode image and strain elastography images side-by-side. Care was taken to include subcutaneous fat and pectoralis major muscle within the same frame as the lesion. The transducer was always oriented parallel to the pectoralis anterior margin, with the transducer orthogonal to the chest wall. Colormapping real-time elastography images were obtained on a continuous scale from blue to green to red, corresponding to hard (no strain), intermediate (average strain) to soft (maximum strain) components respectively, depending on the tissue displacement. Pressure and speed of compression were modulated to demonstrate a mixture of red and green color was observed on the subcutaneous fat layer and blue color was observed on the pectoralis muscle layer for standardization. Still elastography images containing the strain graphics were used for quantification. We measured the SR of each lesion as previously described [1,16]. SR was described as the strain value of fat to lesion, with higher lesion stiffness (i.e. higher SRs) increasing the likelihood of malignancy [16]. We drew the first region of interest (ROI) at least 2 mm free margins of the target lesion (A) manually for mass strain. We put the second ROI onto the fat tissue (B) for fat strain, preferably both of similar size, with the largest possible circular ROI obtainable both in the mass and in the fatty tissue, and with a maximum 5 mm depth difference between the measurements in the fat tissue and mass to eliminate possible stress decay caused by increasing depth (Figs. 1 and 2). The SR was calculated automatically by the elastography software available on the sonography unit, as the fraction of the average strain in the reference area divided by the average strain in the lesion (B/A). Screen capture measured images and SR calculations were saved as DICOM image for further analysis (Figs. 1 and 2). The data acquisition time with elastography including SR measurements was approximately 3 to 5 minutes for each lesion.
Statistical analysis Statistical analysis was performed using the Statistical Package for Social Sciences 22.0 for Windows (SPSS Inc., Chicago, IL). P-value < 0.05 was considered statistically significant. Descriptive statistics were used to describe the data and presented as mean, standard deviation (SD), min-
3 imum, maximum, and ratio. Shapiro Wilk test was used for the assessment of normal ranges. Mann-Whitney U test was used to compare differences between two independent groups when the dependent variable is either ordinal or continuous, but not normally distributed. The independentsamples t-test was used to compare the means between two unrelated groups on the same continuous, dependent variable. Chi-square test was used to calculate categorical variables. Fischer exact test was used to assess data with low cell frequencies. Receiver-operating-characteristic (ROC) curve was also analyzed to test UE diagnostic performance in distinguishing IGM from malignancy. For different predictive models, area under the curve (AUC), cut-off points, sensitivity, and specificity values were determined. Data was expressed as mean ± standard deviation [SD]. Best cut-off values of SR were selected to provide optimal sensitivity and specificity. AUC values near to 1.0 represented perfect test results; values less than or equal to 0.5 were equivalent or worse results than expected by random chance. Histology results obtained by percutaneous core biopsy were used as gold standard for our study.
Results From an original pool of 103 women, 35 were excluded from the study on the basis of pathology results (21 with fibroadenoma, 3 with DCIS, 2 with ductal hyperplasia, 3 with adenosis, one with radial sclerosing lesion, 3 with fibrocystic disease, 1 mucinous carcinoma, and 1 colloid carcinoma). Ultimately, 23 women diagnosed with IGM (mean age 37.9 ± 6.6 [SD]; range: 26—52 years) and 45 women diagnosed with malignancy (mean age 52.8 ± 12.0 [SD] years; range: 32—77 years) were included in the analyses. The maximal diameter of lesions determined by ultrasound ranged from 5 to 46 mm (mean, 18.6 ± 12.3 [SD] mm) for IGM, and from 5 to 60 mm (mean, 19.6 ± 10.2 [SD] mm) for breast cancer, the difference between mean diameters was not statistically significant (P = 0.352) (Table 1). At histopathological analysis, 42 women had invasive ductal carcinoma, 2 had invasive lobular carcinoma, and 1 had malignant fibroepithelial tumor. Conventional B-mode ultrasound findings included microlobulated irregular contours in 14 women (61%) with IGM and 17 (38%) with breast cancer, spiculated contours in 2 (9%) with IGM and 28 (62%) with breast cancer, heterogeneous hypoechoic content in 22 (96%) of women with IGM and all women with breast cancer, posterior acoustic shadowing in 2 (9%) women with IGM and 17 (68%) with breast cancer. Nine (39%) of IGM had tubular configuration. Skin thickening was present in 11 (48%) women with IGM; 2 of them had associated draining sinus tracts. The mean SR in women with IGM was 1.5 ± 0.8 (SD) (range: 0.2—4) and 5.3 ± 5.2 (SD) (range: 1.4—33) in women with breast cancer. The mean SR value in women with IGM was significantly lower than that in women with breast cancer (P < 0.001) (Table 1). ROC curve analysis revealed that best cut-off value for discriminating between IGM and malignancy was 2.5, yielding sensitivity and specificity of 87% and 96% respectively for the diagnosis of IGM. ROC test results
Please cite this article in press as: Ya˘ gcı B, et al. Differentiation between idiopathic granulomatous mastitis and malignant breast lesions using strain ratio on ultrasonic elastography. Diagnostic and Interventional Imaging (2017), http://dx.doi.org/10.1016/j.diii.2017.06.009
+Model DIII-958; No. of Pages 7
ARTICLE IN PRESS B. Ya˘ gcı et al.
4
Figure 1. Invasive ductal carcinoma in a 45-year-old woman with a palpable, non-tender mass. Side-by-side strain elastogram (A) and dual B-mode ultrasound (B) images demonstrate a taller than wide solid hypoechoic heterogeneous lesion with irregular margins with green and blue color (relatively hard) compared to surrounding tissue. A relatively higher strain ratio of 5.62 was calculated.
Figure 2. Idiopathic granulomatous mastitis in a 35-year-old woman with pain, erythema, and tenderness in breast. Side-by-side strain elastogram (A) and dual B-mode ultrasound (B) images show an irregular hypoechoic heterogeneous mass showing a predominantly green lesion with some red portions (intermediate hardness). The strain ratio of the lesion was 1.67.
Table 1 cancer.
Comparison of the lesion size and strain ratio values between idiopathic granulomatous mastitis and breast
Variables
Lesion size Lesion SR
Breast cancer
IGM
Mean ± SD (range)
Mean ± SD (range)
19.6 ± 10.2 (5.0—60) 5.3 ± 5.2 (1.4—33)
18.6 ± 12.3 (5.0—46) 1.5 ± 0.8 (0.2—4.0)
P-value
0.352 0.001
IGM: idiopathic granulomatous mastitis; SR: strain ratio.
yielded an area under the curve values of 0.939 (95% CI: 0.882—0.995; P < 0.0001) (Fig. 3).
Discussion In the current study, we evaluated the sonoelastographic features of IGM using strain elastography technique and compared them with those of breast cancer. We used the semi-quantitative SR method for the interpretation of the
UE results, which provide a measurable value about the stiffness of the lesions. Our results showed that SR of IGM was significantly lower than SR of breast cancer and yielded 87% sensitivity and 96% specificity for the best cut-off value of 2.5 for the diagnosis of IGM, similar to prior studies [17,18]. Idiopathic granulomatous mastitis is a rare benign inflammatory breast disease. It usually affects women of child-bearing age and its etiology is still unclear. Since it may mimic breast cancer both clinically and radiologically, the diagnosis of the disease is difficult and many different
Please cite this article in press as: Ya˘ gcı B, et al. Differentiation between idiopathic granulomatous mastitis and malignant breast lesions using strain ratio on ultrasonic elastography. Diagnostic and Interventional Imaging (2017), http://dx.doi.org/10.1016/j.diii.2017.06.009
+Model DIII-958; No. of Pages 7
ARTICLE IN PRESS
Sonoleastography of idiopathic granulomatous mastitis
Figure 3. Diagram shows receiver-operating-characteristic curve for sonoelastographic strain ratio.
diagnostic modalities are frequently applied to confirm the diagnosis [12,19]. Ultrasonography and mammography are frequently used imaging modalities for the initial examination of this rare entity. However, most of the time the imaging findings on ultrasonography and mammography are inconclusive and it is difficult to diagnose this disease from malignancy using these conventional imaging methods [12]. Larsen et al., found an irregular, hypoechoic mass with multiple tubular extensions, a sign considered suggestive of IGM, in 59% of 54 women with IGM [12]. However, this sign was absent in the remaining 41% of women with IGM so that the final diagnosis still relies on tissue sampling and further histopathological analysis [12]. Later studies seeking more reliable diagnostic approaches to IGM with alternative imaging techniques did not yield specific findings or measures to discriminate IGM from other entities [19]. Therefore, there is a need to investigate additional imaging characteristics of IGM which would allow discriminate this benign entity from other less favorable lesions. To date, a few studies have specifically focused on sonoelastographic features of IGM [17,18]. Teke et al. evaluated the sonoelastographic features of IGM and compared them with malignancy using acoustic radiation force impulse imaging [17]. They evaluated the IGM and breast cancer lesions qualitatively using virtual touch tissue imaging (VTI), and quantitatively using virtual touch tissue quantification (VTQ). By using VTI, taking into account the width differences of the lesions between B-mode and VTI, they found a significant difference between the IGM and breast cancer lesions. They also found a significant difference of shear wave velocity measurements between the two groups using VTQ. When combining VTI findings with VTQ, they obtained a 96.7% sensitivity and a 88% specificity [17]. Durur et al.
5 defined UE features of IGM patients using strain imaging [18]. They assessed the elastography images qualitatively using elasticity score and elastic diameter method as well as semi-quantitatively by measuring SR value of the lesions. They reported that all UE findings of IGM were suggestive of a benign etiology. They found SR value of 1.1 ± 0.79 (SD) (range: 0.29—4.00) for IGM. However, they researchers did not have a control group and did not report any threshold SR value or diagnostic performance [18]. Our results were consistent with the results of two past studies in many ways [17,18]. We found that mean SR of IGM was 1.5, significantly lower than the SR of malignant lesions estimated at 5.3. This result was similar to those by Teke et al. who found that UE values of breast cancer were significantly higher than those of IGM with a different elastography technique and measurement metrics [17]. The mean SR value in our study was close to that of Durur et al. (1.5 vs. 1.1, respectively) [18]. Durur et al. used the same elastography technique but did not assess the diagnostic performances of SR in differentiating IGM from breast cancer [18]. We found that using a cut-off value of 2.5, the sensitivity and specificity were 87% and 96% respectively, whereas Teke et al. found higher sensitivity (96.7%) and lower specificity (88%) [17]. This difference might be explained in part by the fact that their study included on average larger masses because of the fixed sized ROI they used and they offered this as one of their study’s limitations. On the other hand, we used adjustable ROI depending on the size of the lesion, allowing us to include smaller lesions. Thus, we had lower sensitivity and higher specificity than Teke et al. had [17]. In the present study, best cut-off SR value to discriminate between IGM and malignancy was 2.5. Previous studies suggested that the best cut-off SR value for discriminating benign and malignant breast lesions ranged from 1.4 to 4.3. This high variability can probably be attributed to the size and types of the lesions, size of the populations and types of the ultrasound scanners [6,20—23]. Our results were similar to those of Cho et al. and Thomas et al., who used cut-off values of 2.24, and 2.45 respectively [6,16]. In our series, the cut-off SR value of 2.5 yields 96% specificity, which is similar to 96% specificity of core biopsy [12]. In the present study we found 87% sensitivity and 96% specificity in agreement with those of other studies utilizing UE to differentiate malignant from benign breast masses, with sensitivities of UE ranging from 63% to 95%, and specificities ranging from 61% to 96% [6,20—23]. Histochemical and molecular studies focused on breast cancer stroma surrounding the tumor revealed that large aggregates of elastin fibers (elastosis) were found in the tumoral stroma. Stromal cells including fibroblasts and myofibroblasts as well as the carcinoma cells itself give rise to aggregation of elastotic material around ducts or small blood vessels which results in an increased elastosis of the mass [13]. Besides changes in the expression of stromal cell types for the synthesis of elastic material by the cancer cells in breast carcinoma, there are several factors that have been correlated with this altered tensional homeostasis including increased cellularity, reorganization of extracellular matrix. These are the characteristics that help detect cancer at physical palpation as a rigid mass nesting within a compliant tissue [24]. Tumor screening by the
Please cite this article in press as: Ya˘ gcı B, et al. Differentiation between idiopathic granulomatous mastitis and malignant breast lesions using strain ratio on ultrasonic elastography. Diagnostic and Interventional Imaging (2017), http://dx.doi.org/10.1016/j.diii.2017.06.009
+Model DIII-958; No. of Pages 7
ARTICLE IN PRESS B. Ya˘ gcı et al.
6 help of tissue stiffness on palpation is a widespread clinical application among the physicians, due to this well-known feature of the breast carcinoma. All these ultrastructural changes in and around the tumor mass explain, and are consistent with, our study results, reflecting lesser stiffness of IGM lesions in comparison to breast malignancies. In light of these results, and given the low specificity of gray-scale ultrasonography in differentiating malignancy form both IGM and other benign conditions [25], we suggest the addition of sonoelastography, when available, as an easily applicable adjunct within the same imaging session. Our results have shown a significant improvement in diagnostic performance of breast ultrasound when coupled with elastography in differentiating IGM lesions from malignant lesions. Our study has some limitations. One is the relatively small number of patients but it can be considered high given to the rarity of IGM. Another limitation of our study is inter-observer reproducibility of sonoelastography was not evaluated. Prior studies have shown a statistically significant inter-observer variability which may negatively affect SR measurements [26,27]. Further, the age groups were not matched in our study, due to the different prevalence of the two diseases in different age groups. However, we postulate that age differences would not influence results in our study, based on results of previous studies [11]. Lastly, we cannot establish the diagnostic accuracy of UE and SR values in relation to those malignant breast lesions known to cause potential for false negatives (medullary and mucinous carcinomas) because of their rare incidence in our cohort. Further studies with greater statistical power will be needed to determine UE performance in discriminating IGM from this subset of malignancies. In conclusion, the results of our study suggest the use of UE as an adjunct to conventional B-mode ultrasonography, in conjunction with the calculated SR, for differentiating IGM from malignant breast lesions. UE may influence clinical decision making for at least re-biopsies with a suspicious lesion on imaging so that re-biopsy may be avoided according to the SR values in patients with IGM. However, as the diagnosis of IGM is extremely difficult to confirm, biopsy is still mandatory. Future studies with larger patient populations may however be needed to confirm our results.
Compliance with ethical standards Funding: this research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Funding None.
Disclosure of interest The authors declare that they have no competing interest.
References [1] Zhi H, Xiao XY, Yang HY, Wen YL, Ou B, Luo BM, et al. Semi-quantitating stiffness of breast solid lesions in ultrasonic elastography. Acad Radiol 2008;15:1347—53. [2] Cassinotto C. Elastography: French innovations in the spotlight. Diagn Interv Imaging 2016;97:1—2. [3] Metin MR, Aydın H, Ünal Ö, Akc ¸ay Y, Duymus¸ M, Türkyılmaz E, et al. Differentiation between endometrial carcinoma and atypical endometrial hyperplasia with transvaginal sonographic elastography. Diagn Interv Imaging 2017;97: 425—31. [4] Quarello E, Lacoste R, Mancini J, Melot—Dusseau S, Gorincour G. ShearWave elastography of fetal lungs in pregnant baboons. Diagn Interv Imaging 2016;97:605—10. [5] Parfrey NA, Doyle CT. Elastosis in benign and malignant breast disease. Hum Pathol 1985;16:674—6. [6] Cho N, Moon WK, Kim HY, Chang JM, Park SH, Lyou CY. Sonoelastographic strain index for differentiation of benign and malignant nonpalpable breast masses. J Ultrasound Med 2010;29:1—7. [7] Chang JM, Won J-K, Lee K-B, Park IA, Yi A, Moon WK. Comparison of shear-wave and strain ultrasound elastography in the differentiation of benign and malignant breast lesions. AJR Am J Roentgenol 2013;201:W347—56. [8] Chang JM, Moon WK, Cho N, Kim SJ. Breast mass evaluation: factors influencing the quality of US elastography. Radiology 2011;259:59—64. [9] Itoh A, Ueno E, Tohno E, Kamma H, Takahashi H, Shiina T, et al. Breast disease: clinical application of US elastography for diagnosis. Radiology 2006;239:341—50. [10] Fleury EFC, Rinaldi JF, Piato S, Fleury JCV, Junior DR. Appearence of breast masses on sonoelastography with special focus on the diagnosis of fibroadenomas. Eur Radiol 2009;19:1337—46. [11] Fleury E, de FC, Assunc ¸ão-Queiros M, do CGA, Roveda D. Breast carcinomas: variations in sonoelastographic appearance. Breast cancer 2014;6:135—43. [12] Larsen LJH, Peyvandi B, Klipfel N, Grant E, Iyengar G. Granulomatous lobular mastitis: Imaging, diagnosis, and treatment. AJR Am J Roentgenol 2009;193:574—81. [13] Chen Y, Klingen TA, Wik E, Aas H, Vigeland E, Liestøl K, et al. Breast cancer stromal elastosis is associated with mammography screening detection, low Ki67 expression and favourable prognosis in a population-based study. Diagn Pathol 2014;19:230. [14] Tamura S, Enjoji M. Elastosis in neoplastic and non-neoplastic tissues from patients with mammary carcinoma. Acta Pathol Jpn 1988;38:1537—46. [15] Grajo JR, Barr RG. Strain elastography for prediction of breast cancer tumor grades. J Ultrasound Med 2014;33:129—34. [16] Thomas A, Degenhardt F, Farrokh A, Wojcinski S, Slowinski T, Fischer T. Significant differentiation of focal breast lesions calculation of strain ratio in breast sonoelastography. Acad Radiol 2010;17:558—63. [17] Teke M, Teke F, Alan B, Türko˘ glu A, Hamidi C, Göya C, et al. Differential diagnosis of idiopathic granulomatous mastitis and breast cancer using acoustic radiation force impulse imaging. J Med Ultrason 2017;44:109—15. [18] Durur-Karakaya A, Durur-Subasi I, Akcay MN, Sipal S, Guvendi B. Sonoelastography findings for idiopathic granulomatous mastitis. Jpn J Radiol 2014;33:33—8. [19] Yilmaz R, Demir AA, Kaplan A, Sahin D, Ozkurt E, Dursun M, et al. Magnetic resonance imaging features of idiopathic granulomatous mastitis: is there any contribution of diffusionweighted imaging in the differential diagnosis? Radiol Med 2016;121:857—66.
Please cite this article in press as: Ya˘ gcı B, et al. Differentiation between idiopathic granulomatous mastitis and malignant breast lesions using strain ratio on ultrasonic elastography. Diagnostic and Interventional Imaging (2017), http://dx.doi.org/10.1016/j.diii.2017.06.009
+Model DIII-958; No. of Pages 7
ARTICLE IN PRESS
Sonoleastography of idiopathic granulomatous mastitis [20] Lee JH, Kim SH, Kang BJ, Choi JJ, Jeong SH, Yim HW, et al. Role and clinical usefulness of elastography in small breast masses. Acad Radiol 2011;18:74—80. [21] Mansour SM, Omar OS. Elastography ultrasound and questionable breast lesions: does it count? Eur J Radiol 2012;81:3234—44. [22] Hyun Yoon J, Kyung Song M, Kim E. Semi-quantitative strain ratio in the differential diagnosis of breast masses: measurements using one region-of-interest. Ultrasound Med Biol 2016;42:1800—6. [23] Yerli H, Yilmaz T, Kaskati T, Gulay H. Qualitative and semiquantitative evaluations of solid breast lesions by sonoelastography. J Ultrasound Med 2011;30:179—86. [24] Paszek MJ, Zahir N, Johnson KR, Lakins JN, Rozenberg GI, Gefen A, et al. Tensional homeostasis and the malignant phenotype. Cancer Cell 2005;8:241—54.
7 [25] Stavros AT, Thickman D, Rapp CL, Dennis MA, Parker SH, Sisney GA. Solid breast nodules: use of sonography to distinguish between benign and malignant lesions. Radiology 1995;196:123—34. [26] Burnside ES, Hall TJ, Sommer AM, Hesley GK, Sisney GA, Svensson WE, et al. Differentiating benign from malignant solid breast masses with US strain imaging. Radiology 2007;245:401—10. [27] Regner DM, Hesley GK, Hangiandreou NJ, Morton MJ, Nordland MR, Meixner DD, et al. Breast lesions: evaluation with US strain imaging—clinical experience of multiple observers. Radiology 2006;238:425—37.
Please cite this article in press as: Ya˘ gcı B, et al. Differentiation between idiopathic granulomatous mastitis and malignant breast lesions using strain ratio on ultrasonic elastography. Diagnostic and Interventional Imaging (2017), http://dx.doi.org/10.1016/j.diii.2017.06.009
ARTICLE IN PRESS
DIII-958; No. of Pages 7
Diagnostic and Interventional Imaging (2017) xxx, xxx—xxx
ORIGINAL ARTICLE /Breast imaging
Differentiation between idiopathic granulomatous mastitis and malignant breast lesions using strain ratio on ultrasonic elastography B. Ya˘ gcı a, I. Erdem Toslak a,b,∗, B. C ¸ekic ¸ a, M. Öz a, ¸ c, M. Akdemir d, S. Yıldız a, D. Süren e, B.R. Karakas D. Bova b a
Department of Radiology, Antalya Training and Research Hospital, Health Sciences University, Antalya, Turkey b Department of Radiology, Loyola University Medical Center, Maywood, Illinois, USA c Department of General Surgery, Antalya Training and Research Hospital, Health Sciences University, Antalya, Turkey d Department of Public Health, Akdeniz University Medical School,Antalya, Turkey e Department of Pathology, Antalya Training and Research Hospital, Health Sciences University, Antalya, Turkey
KEYWORDS Ultrasonography; Elastography; Strain ratio; Granulomatous mastitis; Breast cancer
∗
Abstract Purpose: The goal of this study was to investigate the strain elastography imaging characteristics of idiopathic granulomatous mastitis (IGM) and compare strain ratio values of IGM with those of breast cancer. Material and methods: Twenty-three consecutive women with IGM (mean age, 37.9 ± 6.6 [SD] years; range: 26—52 years) and 45 women with malignant breast tumor (mean age, 52.8 ± 12.0 [SD], range, 32—77 years) who had been scheduled for ultrasound-guided core biopsy were recruited to the study. All had ultrasonography with elastography before biopsy. The strain ratios of lesions were calculated using surrounding normal breast tissue as the reference in both groups and compared between the two groups. Receiver-operating-characteristics (ROC) curves were formed. Sensitivity, specificity, cut-off, and area under curve (AUC) values were calculated.
Abbreviations: UE, ultrasound elastography; IGM, idiopathic granulomatous mastitis; SR, strain ratio. Corresponding author at: Department of Radiology, Antalya Training and Research Hospital, Health Sciences University, Antalya, Turkey. E-mail address: [email protected] (I. Erdem Toslak).
http://dx.doi.org/10.1016/j.diii.2017.06.009 2211-5684/© 2017 Editions franc ¸aises de radiologie. Published by Elsevier Masson SAS. All rights reserved.
Please cite this article in press as: Ya˘ gcı B, et al. Differentiation between idiopathic granulomatous mastitis and malignant breast lesions using strain ratio on ultrasonic elastography. Diagnostic and Interventional Imaging (2017), http://dx.doi.org/10.1016/j.diii.2017.06.009
+Model DIII-958; No. of Pages 7
ARTICLE IN PRESS B. Ya˘ gcı et al.
2
Results: The mean strain ratio on sonoelastography was 1.5 ± 0.8 (SD) (range: 0.2—4.0) for IGM and 5.3 ± 5.2 (SD) (range: 1.4—33) for malignant lesions. Strain ratio values in IGM lesions were significantly lower than in malignant lesions (P < 0.05). ROC test yielded an AUC value of 0.939 (95% confidence interval, 0.882—0.995; P < 0.0001). Optimal cut-off value for strain ratio value was 2.5 yielding 87% sensitivity and 96% specificity for the diagnosis of IGM. Conclusion: Sonoelastographic strain ratio contributes to differentiate IGM from malignant breast lesions, thus has potential to influence clinical decision making for further biopsies. © 2017 Editions franc ¸aises de radiologie. Published by Elsevier Masson SAS. All rights reserved.
Introduction Intrinsic elastic features of the tissues may be altered in certain circumstances; neoplasm is one of the pathophysiologic process influencing the elasticity of the tissues [1]. Generally, stiffer and immobile breast lesions suggest malignancy [1]. On the basis of this principle, ultrasound elastography (UE) has been developed as a more objective method to estimate tissue stiffness. The degree of tissue displacement is greater in harder tissues than in softer tissues; this is represented by an overlapping color-encoded elasticity map of the tissues with a continuous scale, from blue representing hardest to red representing softest tissues. UE can be used in various organs including thyroid, liver, kidney, spleen, muscles, or even can be further used for the detection of endometrium carcinoma and evaluation of fetal lungs [1—4]. UE has been shown beneficial in diagnosis of breast lesions [1]. Breast carcinoma is associated with elastosis, which increases progressively with the severity of the disease [5]. Breast cancer has higher strain ratio values, or elasticity scores, as compared to benign breast lesions [6—9]. However, few studies have compared specific subtypes of benign and malignant breast masses [10,11]. Idiopathic granulomatous mastitis (IGM) is a rare and benign chronic granulomatous inflammatory disease of breast, which is often confused with a malignant process, both clinically and radiographically. Imaging features that suggest the diagnosis of IGM remain non-specific and not always found in all patients. Thus, the final diagnosis usually requires tissue biopsy [12]. Because the current imaging modalities are not sufficient to establish a definitive diagnosis of IGM in most patients, and because prior histochemical and molecular studies of breast cancer reveal increased amounts of elastotic material in cancerous lesions, the question arose if overlap in elasticity values of IGM and breast cancer existed or if these may be helpful in diagnosis and clinical management of this rare entity [13]. Given the increased amount of periductal and stromal elastic fibers in breast carcinoma and lack of this feature in non-neoplastic conditions (except for radial scars), this feature can be used to differentiate IGM lesions from stiffer breast cancer lesions [13,14]. In light of this knowledge, UE may be useful in differentiating IGM from the breast cancer lesions. The goal of this study was to investigate the strain elastography imaging characteristics of IGM and compare strain
ratio (SR) values of IGM with those of malignant breast lesions.
Materials and methods Study design This prospective study was conducted in accordance with the 1964 declaration of Helsinki and was approved by the Institutional Ethics Committee; written informed consent was obtained from all subjects. From May 2014 through December 2015, 103 consecutive women who had been referred to our department for ultrasound-guided core biopsy were examined by sonoelastography. All participants underwent strain elastography imaging following conventional breast ultrasound examinations. Patients less than 18-year-old, with a benign pathological diagnosis other than IGM (fibroadenoma, adenosis, radial sclerosing lesion, fibrocystic disease), with mass lesion at the same location of a previously diagnosed IGM, with prior surgery or biopsy of the mass being examined, or with a non-mass lesion on ultrasound were excluded from analysis. We also excluded precancerous lesions (ductal carcinoma in-situ [DCIS]) and certain histologic types of breast cancer (mucinous and colloid carcinoma) on the basis of previous studies’ results which showed that these lesions had potential to yield false negative imaging characteristics at UE [11,15,16]. Women underwent ultrasound-guided percutaneous core needle biopsy based on the clinical decision of the referring physician. Strain elastography and B-mode gray-scale ultrasound evaluation of the lesions were performed within a week of referral; both were performed within the same session and prior to obtaining tissue diagnosis. The presence of non-caseating granulomas in biopsy samples suggested IGM diagnosis. In patients with multiple lesions, only the largest lesion was considered for the purposes of this study; 5 mm was selected as the lower size limit to be included in the study.
Ultrasonographic imaging All patients underwent a conventional diagnostic breast ultrasound examination, conducted with the same scanner
Please cite this article in press as: Ya˘ gcı B, et al. Differentiation between idiopathic granulomatous mastitis and malignant breast lesions using strain ratio on ultrasonic elastography. Diagnostic and Interventional Imaging (2017), http://dx.doi.org/10.1016/j.diii.2017.06.009
+Model DIII-958; No. of Pages 7
ARTICLE IN PRESS
Sonoleastography of idiopathic granulomatous mastitis (Hi-Vision Prerius; Hitachi medical systems, Japan) equipped with a high-resolution 8—12 MHz linear array transducer and an elastography module. All ultrasound examinations were performed by the same radiologist with 3 years of experience in elastography imaging, blinded to the clinical diagnosis. Both breasts and axilla were scanned in longitudinal and transverse planes. All subjects were examined in supine position. Still and cine images of the palpable mass or area of concern were recorded electronically in the digital imaging and communications in medicine (DICOM) format. Size, shape, margin, echogenicity, and posterior acoustic phenomena of the mass or lesion were assessed while recording images. Free hand elastography and standard sonographic evaluation were performed during the same session. Elastographic images were acquired by using the freehand repeated manual compression technique using the transducer [9]. In this technique, the transducer is gently compressed and released over the lesion periodically with split-screen recording, displaying real-time B-mode image and strain elastography images side-by-side. Care was taken to include subcutaneous fat and pectoralis major muscle within the same frame as the lesion. The transducer was always oriented parallel to the pectoralis anterior margin, with the transducer orthogonal to the chest wall. Colormapping real-time elastography images were obtained on a continuous scale from blue to green to red, corresponding to hard (no strain), intermediate (average strain) to soft (maximum strain) components respectively, depending on the tissue displacement. Pressure and speed of compression were modulated to demonstrate a mixture of red and green color was observed on the subcutaneous fat layer and blue color was observed on the pectoralis muscle layer for standardization. Still elastography images containing the strain graphics were used for quantification. We measured the SR of each lesion as previously described [1,16]. SR was described as the strain value of fat to lesion, with higher lesion stiffness (i.e. higher SRs) increasing the likelihood of malignancy [16]. We drew the first region of interest (ROI) at least 2 mm free margins of the target lesion (A) manually for mass strain. We put the second ROI onto the fat tissue (B) for fat strain, preferably both of similar size, with the largest possible circular ROI obtainable both in the mass and in the fatty tissue, and with a maximum 5 mm depth difference between the measurements in the fat tissue and mass to eliminate possible stress decay caused by increasing depth (Figs. 1 and 2). The SR was calculated automatically by the elastography software available on the sonography unit, as the fraction of the average strain in the reference area divided by the average strain in the lesion (B/A). Screen capture measured images and SR calculations were saved as DICOM image for further analysis (Figs. 1 and 2). The data acquisition time with elastography including SR measurements was approximately 3 to 5 minutes for each lesion.
Statistical analysis Statistical analysis was performed using the Statistical Package for Social Sciences 22.0 for Windows (SPSS Inc., Chicago, IL). P-value < 0.05 was considered statistically significant. Descriptive statistics were used to describe the data and presented as mean, standard deviation (SD), min-
3 imum, maximum, and ratio. Shapiro Wilk test was used for the assessment of normal ranges. Mann-Whitney U test was used to compare differences between two independent groups when the dependent variable is either ordinal or continuous, but not normally distributed. The independentsamples t-test was used to compare the means between two unrelated groups on the same continuous, dependent variable. Chi-square test was used to calculate categorical variables. Fischer exact test was used to assess data with low cell frequencies. Receiver-operating-characteristic (ROC) curve was also analyzed to test UE diagnostic performance in distinguishing IGM from malignancy. For different predictive models, area under the curve (AUC), cut-off points, sensitivity, and specificity values were determined. Data was expressed as mean ± standard deviation [SD]. Best cut-off values of SR were selected to provide optimal sensitivity and specificity. AUC values near to 1.0 represented perfect test results; values less than or equal to 0.5 were equivalent or worse results than expected by random chance. Histology results obtained by percutaneous core biopsy were used as gold standard for our study.
Results From an original pool of 103 women, 35 were excluded from the study on the basis of pathology results (21 with fibroadenoma, 3 with DCIS, 2 with ductal hyperplasia, 3 with adenosis, one with radial sclerosing lesion, 3 with fibrocystic disease, 1 mucinous carcinoma, and 1 colloid carcinoma). Ultimately, 23 women diagnosed with IGM (mean age 37.9 ± 6.6 [SD]; range: 26—52 years) and 45 women diagnosed with malignancy (mean age 52.8 ± 12.0 [SD] years; range: 32—77 years) were included in the analyses. The maximal diameter of lesions determined by ultrasound ranged from 5 to 46 mm (mean, 18.6 ± 12.3 [SD] mm) for IGM, and from 5 to 60 mm (mean, 19.6 ± 10.2 [SD] mm) for breast cancer, the difference between mean diameters was not statistically significant (P = 0.352) (Table 1). At histopathological analysis, 42 women had invasive ductal carcinoma, 2 had invasive lobular carcinoma, and 1 had malignant fibroepithelial tumor. Conventional B-mode ultrasound findings included microlobulated irregular contours in 14 women (61%) with IGM and 17 (38%) with breast cancer, spiculated contours in 2 (9%) with IGM and 28 (62%) with breast cancer, heterogeneous hypoechoic content in 22 (96%) of women with IGM and all women with breast cancer, posterior acoustic shadowing in 2 (9%) women with IGM and 17 (68%) with breast cancer. Nine (39%) of IGM had tubular configuration. Skin thickening was present in 11 (48%) women with IGM; 2 of them had associated draining sinus tracts. The mean SR in women with IGM was 1.5 ± 0.8 (SD) (range: 0.2—4) and 5.3 ± 5.2 (SD) (range: 1.4—33) in women with breast cancer. The mean SR value in women with IGM was significantly lower than that in women with breast cancer (P < 0.001) (Table 1). ROC curve analysis revealed that best cut-off value for discriminating between IGM and malignancy was 2.5, yielding sensitivity and specificity of 87% and 96% respectively for the diagnosis of IGM. ROC test results
Please cite this article in press as: Ya˘ gcı B, et al. Differentiation between idiopathic granulomatous mastitis and malignant breast lesions using strain ratio on ultrasonic elastography. Diagnostic and Interventional Imaging (2017), http://dx.doi.org/10.1016/j.diii.2017.06.009
+Model DIII-958; No. of Pages 7
ARTICLE IN PRESS B. Ya˘ gcı et al.
4
Figure 1. Invasive ductal carcinoma in a 45-year-old woman with a palpable, non-tender mass. Side-by-side strain elastogram (A) and dual B-mode ultrasound (B) images demonstrate a taller than wide solid hypoechoic heterogeneous lesion with irregular margins with green and blue color (relatively hard) compared to surrounding tissue. A relatively higher strain ratio of 5.62 was calculated.
Figure 2. Idiopathic granulomatous mastitis in a 35-year-old woman with pain, erythema, and tenderness in breast. Side-by-side strain elastogram (A) and dual B-mode ultrasound (B) images show an irregular hypoechoic heterogeneous mass showing a predominantly green lesion with some red portions (intermediate hardness). The strain ratio of the lesion was 1.67.
Table 1 cancer.
Comparison of the lesion size and strain ratio values between idiopathic granulomatous mastitis and breast
Variables
Lesion size Lesion SR
Breast cancer
IGM
Mean ± SD (range)
Mean ± SD (range)
19.6 ± 10.2 (5.0—60) 5.3 ± 5.2 (1.4—33)
18.6 ± 12.3 (5.0—46) 1.5 ± 0.8 (0.2—4.0)
P-value
0.352 0.001
IGM: idiopathic granulomatous mastitis; SR: strain ratio.
yielded an area under the curve values of 0.939 (95% CI: 0.882—0.995; P < 0.0001) (Fig. 3).
Discussion In the current study, we evaluated the sonoelastographic features of IGM using strain elastography technique and compared them with those of breast cancer. We used the semi-quantitative SR method for the interpretation of the
UE results, which provide a measurable value about the stiffness of the lesions. Our results showed that SR of IGM was significantly lower than SR of breast cancer and yielded 87% sensitivity and 96% specificity for the best cut-off value of 2.5 for the diagnosis of IGM, similar to prior studies [17,18]. Idiopathic granulomatous mastitis is a rare benign inflammatory breast disease. It usually affects women of child-bearing age and its etiology is still unclear. Since it may mimic breast cancer both clinically and radiologically, the diagnosis of the disease is difficult and many different
Please cite this article in press as: Ya˘ gcı B, et al. Differentiation between idiopathic granulomatous mastitis and malignant breast lesions using strain ratio on ultrasonic elastography. Diagnostic and Interventional Imaging (2017), http://dx.doi.org/10.1016/j.diii.2017.06.009
+Model DIII-958; No. of Pages 7
ARTICLE IN PRESS
Sonoleastography of idiopathic granulomatous mastitis
Figure 3. Diagram shows receiver-operating-characteristic curve for sonoelastographic strain ratio.
diagnostic modalities are frequently applied to confirm the diagnosis [12,19]. Ultrasonography and mammography are frequently used imaging modalities for the initial examination of this rare entity. However, most of the time the imaging findings on ultrasonography and mammography are inconclusive and it is difficult to diagnose this disease from malignancy using these conventional imaging methods [12]. Larsen et al., found an irregular, hypoechoic mass with multiple tubular extensions, a sign considered suggestive of IGM, in 59% of 54 women with IGM [12]. However, this sign was absent in the remaining 41% of women with IGM so that the final diagnosis still relies on tissue sampling and further histopathological analysis [12]. Later studies seeking more reliable diagnostic approaches to IGM with alternative imaging techniques did not yield specific findings or measures to discriminate IGM from other entities [19]. Therefore, there is a need to investigate additional imaging characteristics of IGM which would allow discriminate this benign entity from other less favorable lesions. To date, a few studies have specifically focused on sonoelastographic features of IGM [17,18]. Teke et al. evaluated the sonoelastographic features of IGM and compared them with malignancy using acoustic radiation force impulse imaging [17]. They evaluated the IGM and breast cancer lesions qualitatively using virtual touch tissue imaging (VTI), and quantitatively using virtual touch tissue quantification (VTQ). By using VTI, taking into account the width differences of the lesions between B-mode and VTI, they found a significant difference between the IGM and breast cancer lesions. They also found a significant difference of shear wave velocity measurements between the two groups using VTQ. When combining VTI findings with VTQ, they obtained a 96.7% sensitivity and a 88% specificity [17]. Durur et al.
5 defined UE features of IGM patients using strain imaging [18]. They assessed the elastography images qualitatively using elasticity score and elastic diameter method as well as semi-quantitatively by measuring SR value of the lesions. They reported that all UE findings of IGM were suggestive of a benign etiology. They found SR value of 1.1 ± 0.79 (SD) (range: 0.29—4.00) for IGM. However, they researchers did not have a control group and did not report any threshold SR value or diagnostic performance [18]. Our results were consistent with the results of two past studies in many ways [17,18]. We found that mean SR of IGM was 1.5, significantly lower than the SR of malignant lesions estimated at 5.3. This result was similar to those by Teke et al. who found that UE values of breast cancer were significantly higher than those of IGM with a different elastography technique and measurement metrics [17]. The mean SR value in our study was close to that of Durur et al. (1.5 vs. 1.1, respectively) [18]. Durur et al. used the same elastography technique but did not assess the diagnostic performances of SR in differentiating IGM from breast cancer [18]. We found that using a cut-off value of 2.5, the sensitivity and specificity were 87% and 96% respectively, whereas Teke et al. found higher sensitivity (96.7%) and lower specificity (88%) [17]. This difference might be explained in part by the fact that their study included on average larger masses because of the fixed sized ROI they used and they offered this as one of their study’s limitations. On the other hand, we used adjustable ROI depending on the size of the lesion, allowing us to include smaller lesions. Thus, we had lower sensitivity and higher specificity than Teke et al. had [17]. In the present study, best cut-off SR value to discriminate between IGM and malignancy was 2.5. Previous studies suggested that the best cut-off SR value for discriminating benign and malignant breast lesions ranged from 1.4 to 4.3. This high variability can probably be attributed to the size and types of the lesions, size of the populations and types of the ultrasound scanners [6,20—23]. Our results were similar to those of Cho et al. and Thomas et al., who used cut-off values of 2.24, and 2.45 respectively [6,16]. In our series, the cut-off SR value of 2.5 yields 96% specificity, which is similar to 96% specificity of core biopsy [12]. In the present study we found 87% sensitivity and 96% specificity in agreement with those of other studies utilizing UE to differentiate malignant from benign breast masses, with sensitivities of UE ranging from 63% to 95%, and specificities ranging from 61% to 96% [6,20—23]. Histochemical and molecular studies focused on breast cancer stroma surrounding the tumor revealed that large aggregates of elastin fibers (elastosis) were found in the tumoral stroma. Stromal cells including fibroblasts and myofibroblasts as well as the carcinoma cells itself give rise to aggregation of elastotic material around ducts or small blood vessels which results in an increased elastosis of the mass [13]. Besides changes in the expression of stromal cell types for the synthesis of elastic material by the cancer cells in breast carcinoma, there are several factors that have been correlated with this altered tensional homeostasis including increased cellularity, reorganization of extracellular matrix. These are the characteristics that help detect cancer at physical palpation as a rigid mass nesting within a compliant tissue [24]. Tumor screening by the
Please cite this article in press as: Ya˘ gcı B, et al. Differentiation between idiopathic granulomatous mastitis and malignant breast lesions using strain ratio on ultrasonic elastography. Diagnostic and Interventional Imaging (2017), http://dx.doi.org/10.1016/j.diii.2017.06.009
+Model DIII-958; No. of Pages 7
ARTICLE IN PRESS B. Ya˘ gcı et al.
6 help of tissue stiffness on palpation is a widespread clinical application among the physicians, due to this well-known feature of the breast carcinoma. All these ultrastructural changes in and around the tumor mass explain, and are consistent with, our study results, reflecting lesser stiffness of IGM lesions in comparison to breast malignancies. In light of these results, and given the low specificity of gray-scale ultrasonography in differentiating malignancy form both IGM and other benign conditions [25], we suggest the addition of sonoelastography, when available, as an easily applicable adjunct within the same imaging session. Our results have shown a significant improvement in diagnostic performance of breast ultrasound when coupled with elastography in differentiating IGM lesions from malignant lesions. Our study has some limitations. One is the relatively small number of patients but it can be considered high given to the rarity of IGM. Another limitation of our study is inter-observer reproducibility of sonoelastography was not evaluated. Prior studies have shown a statistically significant inter-observer variability which may negatively affect SR measurements [26,27]. Further, the age groups were not matched in our study, due to the different prevalence of the two diseases in different age groups. However, we postulate that age differences would not influence results in our study, based on results of previous studies [11]. Lastly, we cannot establish the diagnostic accuracy of UE and SR values in relation to those malignant breast lesions known to cause potential for false negatives (medullary and mucinous carcinomas) because of their rare incidence in our cohort. Further studies with greater statistical power will be needed to determine UE performance in discriminating IGM from this subset of malignancies. In conclusion, the results of our study suggest the use of UE as an adjunct to conventional B-mode ultrasonography, in conjunction with the calculated SR, for differentiating IGM from malignant breast lesions. UE may influence clinical decision making for at least re-biopsies with a suspicious lesion on imaging so that re-biopsy may be avoided according to the SR values in patients with IGM. However, as the diagnosis of IGM is extremely difficult to confirm, biopsy is still mandatory. Future studies with larger patient populations may however be needed to confirm our results.
Compliance with ethical standards Funding: this research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Funding None.
Disclosure of interest The authors declare that they have no competing interest.
References [1] Zhi H, Xiao XY, Yang HY, Wen YL, Ou B, Luo BM, et al. Semi-quantitating stiffness of breast solid lesions in ultrasonic elastography. Acad Radiol 2008;15:1347—53. [2] Cassinotto C. Elastography: French innovations in the spotlight. Diagn Interv Imaging 2016;97:1—2. [3] Metin MR, Aydın H, Ünal Ö, Akc ¸ay Y, Duymus¸ M, Türkyılmaz E, et al. Differentiation between endometrial carcinoma and atypical endometrial hyperplasia with transvaginal sonographic elastography. Diagn Interv Imaging 2017;97: 425—31. [4] Quarello E, Lacoste R, Mancini J, Melot—Dusseau S, Gorincour G. ShearWave elastography of fetal lungs in pregnant baboons. Diagn Interv Imaging 2016;97:605—10. [5] Parfrey NA, Doyle CT. Elastosis in benign and malignant breast disease. Hum Pathol 1985;16:674—6. [6] Cho N, Moon WK, Kim HY, Chang JM, Park SH, Lyou CY. Sonoelastographic strain index for differentiation of benign and malignant nonpalpable breast masses. J Ultrasound Med 2010;29:1—7. [7] Chang JM, Won J-K, Lee K-B, Park IA, Yi A, Moon WK. Comparison of shear-wave and strain ultrasound elastography in the differentiation of benign and malignant breast lesions. AJR Am J Roentgenol 2013;201:W347—56. [8] Chang JM, Moon WK, Cho N, Kim SJ. Breast mass evaluation: factors influencing the quality of US elastography. Radiology 2011;259:59—64. [9] Itoh A, Ueno E, Tohno E, Kamma H, Takahashi H, Shiina T, et al. Breast disease: clinical application of US elastography for diagnosis. Radiology 2006;239:341—50. [10] Fleury EFC, Rinaldi JF, Piato S, Fleury JCV, Junior DR. Appearence of breast masses on sonoelastography with special focus on the diagnosis of fibroadenomas. Eur Radiol 2009;19:1337—46. [11] Fleury E, de FC, Assunc ¸ão-Queiros M, do CGA, Roveda D. Breast carcinomas: variations in sonoelastographic appearance. Breast cancer 2014;6:135—43. [12] Larsen LJH, Peyvandi B, Klipfel N, Grant E, Iyengar G. Granulomatous lobular mastitis: Imaging, diagnosis, and treatment. AJR Am J Roentgenol 2009;193:574—81. [13] Chen Y, Klingen TA, Wik E, Aas H, Vigeland E, Liestøl K, et al. Breast cancer stromal elastosis is associated with mammography screening detection, low Ki67 expression and favourable prognosis in a population-based study. Diagn Pathol 2014;19:230. [14] Tamura S, Enjoji M. Elastosis in neoplastic and non-neoplastic tissues from patients with mammary carcinoma. Acta Pathol Jpn 1988;38:1537—46. [15] Grajo JR, Barr RG. Strain elastography for prediction of breast cancer tumor grades. J Ultrasound Med 2014;33:129—34. [16] Thomas A, Degenhardt F, Farrokh A, Wojcinski S, Slowinski T, Fischer T. Significant differentiation of focal breast lesions calculation of strain ratio in breast sonoelastography. Acad Radiol 2010;17:558—63. [17] Teke M, Teke F, Alan B, Türko˘ glu A, Hamidi C, Göya C, et al. Differential diagnosis of idiopathic granulomatous mastitis and breast cancer using acoustic radiation force impulse imaging. J Med Ultrason 2017;44:109—15. [18] Durur-Karakaya A, Durur-Subasi I, Akcay MN, Sipal S, Guvendi B. Sonoelastography findings for idiopathic granulomatous mastitis. Jpn J Radiol 2014;33:33—8. [19] Yilmaz R, Demir AA, Kaplan A, Sahin D, Ozkurt E, Dursun M, et al. Magnetic resonance imaging features of idiopathic granulomatous mastitis: is there any contribution of diffusionweighted imaging in the differential diagnosis? Radiol Med 2016;121:857—66.
Please cite this article in press as: Ya˘ gcı B, et al. Differentiation between idiopathic granulomatous mastitis and malignant breast lesions using strain ratio on ultrasonic elastography. Diagnostic and Interventional Imaging (2017), http://dx.doi.org/10.1016/j.diii.2017.06.009
+Model DIII-958; No. of Pages 7
ARTICLE IN PRESS
Sonoleastography of idiopathic granulomatous mastitis [20] Lee JH, Kim SH, Kang BJ, Choi JJ, Jeong SH, Yim HW, et al. Role and clinical usefulness of elastography in small breast masses. Acad Radiol 2011;18:74—80. [21] Mansour SM, Omar OS. Elastography ultrasound and questionable breast lesions: does it count? Eur J Radiol 2012;81:3234—44. [22] Hyun Yoon J, Kyung Song M, Kim E. Semi-quantitative strain ratio in the differential diagnosis of breast masses: measurements using one region-of-interest. Ultrasound Med Biol 2016;42:1800—6. [23] Yerli H, Yilmaz T, Kaskati T, Gulay H. Qualitative and semiquantitative evaluations of solid breast lesions by sonoelastography. J Ultrasound Med 2011;30:179—86. [24] Paszek MJ, Zahir N, Johnson KR, Lakins JN, Rozenberg GI, Gefen A, et al. Tensional homeostasis and the malignant phenotype. Cancer Cell 2005;8:241—54.
7 [25] Stavros AT, Thickman D, Rapp CL, Dennis MA, Parker SH, Sisney GA. Solid breast nodules: use of sonography to distinguish between benign and malignant lesions. Radiology 1995;196:123—34. [26] Burnside ES, Hall TJ, Sommer AM, Hesley GK, Sisney GA, Svensson WE, et al. Differentiating benign from malignant solid breast masses with US strain imaging. Radiology 2007;245:401—10. [27] Regner DM, Hesley GK, Hangiandreou NJ, Morton MJ, Nordland MR, Meixner DD, et al. Breast lesions: evaluation with US strain imaging—clinical experience of multiple observers. Radiology 2006;238:425—37.
Please cite this article in press as: Ya˘ gcı B, et al. Differentiation between idiopathic granulomatous mastitis and malignant breast lesions using strain ratio on ultrasonic elastography. Diagnostic and Interventional Imaging (2017), http://dx.doi.org/10.1016/j.diii.2017.06.009