| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Radiology, Duke University Medical Center, Box 3808, Durham, NC 27710.
Received May 7, 1999;
accepted after revision June 23, 1999.
Address correspondence to E. M. Caoili.
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. During an 18-month period, all pelvic masses discovered on sonography were prospectively examined for recurrent refractory shadows. Sonograms of 222 masses with a definitive final diagnosis were reviewed by two radiologists unaware of the follow-up results. Each mass was analyzed for the presence of three or more recurrent refractory shadows, each of which did not arise from an echogenic lead point. After all masses were rated independently, a consensus rating was reached for any discrepancy between observers.
RESULTS. Of the 222 masses, 152 (68%) were leiomyomas. The remainder resulted from a wide variety of other uterine and adnexal disorders. Agreement between the two observers regarding refractory shadows was excellent. Kappa coefficients were 0.86 at transabdominal and 0.96 at transvaginal approaches. Sensitivity of the refractory shadow pattern for leiomyomas was higher for transvaginal (87%) than for transabdominal (52%) sonography. Specificity of the refractory shadow pattern was 89% for transvaginal and 92% for transabdominal sonography. Using logistic regression analysis, we found a statistically significant association between a refractory shadowing pattern and leiomyoma for both transabdominal and transvaginal sonography (p << 0.001).
CONCLUSION. Recurrent refractory shadowing in a pelvic mass on sonography is highly suggestive of leiomyoma and can be a useful diagnostic clue for distinguishing leiomyomas from other pelvic masses.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Recently, a distinctive pattern of shadowing in which multiple discrete shadows originate from within the mass was described with leiomyomas [1]. Histopathologic correlation revealed that the shadows originated from transition points in the mass corresponding to edges between smooth muscle whorls and boundaries between smooth muscle and fibrous connective tissue. However, the sensitivity and specificity of this finding for the diagnosis of leiomyomas is not known. The goal of this study was to determine the diagnostic usefulness of this sonographic pattern in distinguishing leiomyomas from other pelvic masses.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Sonographic evaluation included both transabdominal and transvaginal sonography for 127 masses, only transabdominal imaging for 66 masses, and only transvaginal imaging for 29 masses. Transabdominal sonography was performed with electronically focused sector, vector, or curvilinear transducers ranging in frequency from 2.0 to 5.0 MHz, and transvaginal sonography with 5.0- or 7.0-MHz electronically focused transvaginal transducers (128 XP, Acuson, Mountain View, CA; and ATL 3000 and 5000, Advanced Technology Laboratories, Bothell, WA). Scanning frequencies were selected to optimize imaging of the masses. In each case a concerted effort was made to obtain images showing the maximum number of shadows per scan plane from both the transabdominal and transvaginal perspectives.
Sonograms were reviewed by two radiologists experienced in sonography who were unaware of the final diagnosis. The size of the mass and the transducer frequency were recorded. In cases in which a broadband frequency transducer (ATL 3000 and 5000) was used, the average frequency was recorded. In addition, the location of each mass in relation to the uterus was evaluated. Masses were categorized as intrauterine if the mass was clearly confined to the uterine contour; extrauterine if the mass was separate from the uterus; and indeterminate if its location could not be definitively categorized. The radiologists independently rated each mass for the presence or absence of shadows at each imaging approach used (transabdominal or transvaginal). The presence of three or more well-defined shadows originating from within a mass was defined as the "recurrent shadowing pattern" (Figs. 1 and 2). Shadows originating from echogenic lead points were not considered because they were presumed to arise from calcifications. For patients with multiple masses, the largest three masses seen at each scanning approach were assessed. When the initial ratings of the two reviewers were discrepant, the images were re-reviewed by the two reviewers together, and a consensus decision regarding the presence or absence of the shadowing pattern was reached.
|
|
Data were analyzed separately for masses seen on transabdominal sonography and masses seen on transvaginal sonography. Additional data analysis was performed for the subgroup of masses in which a mass seen on transabdominal sonography could be confidently established as the same mass on transvaginal sonography. Because many of the patients had more than one mass, statistical analysis used generalized estimating equations to account for this clustering of masses. The presence or absence of the shadowing pattern was compared with the final diagnosis. Agreement between reviewers was measured with Cohen's kappa statistic. For cases with discrepant readings between the two reviewers, the consensus ratings were used for the calculations of sensitivity and specificity of the shadowing pattern for leiomyomas. Logistic regression models were used to evaluate the strength of the association between the recurrent shadowing pattern and leiomyoma at each scanning perspective and to assess the effects of mass size and transducer frequency on the production of such shadows. A p value of less than 0.05 was considered statistically significant.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Using logistic regression analysis, a statistically significant association was noted between the refractory shadowing pattern and leiomyoma (p << 0.001) for both transabdominal and transvaginal sonography. Table 1 summarizes the relative occurrence of shadowing in leiomyomas and other masses. Using these data, sensitivity was 52% for transabdominal and 87% for transvaginal sonography. Specificity was 92% for transabdominal sonography and 89% for transvaginal sonography. The final diagnoses for all masses that exhibited the recurrent shadowing pattern are shown in Table 2.
|
|
One hundred twenty-seven masses imaged from both transabdominal and transvaginal perspectives could be unambiguously correlated between the two sets of images. Seventy-three (57%) of these masses were leiomyomas. Forty-one leiomyomas exhibited the shadowing pattern on transvaginal but not on transabdominal sonography. None of the leiomyomas showed the pattern on transabdominal sonography that did not also show the pattern on transvaginal sonography. The sensitivity and specificity for recurrent shadowing seen on transabdominal and transvaginal imaging were 92% and 90%, respectively.
Agreement between the two independent reviewers for interpreting the presence or absence of recurrent shadows was excellent, with kappa coefficients of 0.86 (transabdominal) and 0.96 (transvaginal). Discrepancy in interpreting the presence of shadows between the reviewers occurred in only 17 (8%) of 222 masses. Interpretive discrepancies between the reviewers were more common with transabdominal (n = 14) than transvaginal sonography (n = 3) Discrepant ratings occurred with leiomyomas (n = 13), adenomyosis (n = 2), gastric metastasis to the ovary (n = 1), and a pelvic abscess (n = 1). All leiomyomas that had initial discrepant ratings between reviewers occurred in patients with multiple pelvic masses.
On transabdominal sonography, the mean size of leiomyomas that showed the shadowing pattern was 5.8 cm, whereas the mean size for leiomyomas without the pattern was 3.9 cm. On transvaginal sonography, leiomyomas that showed the shadowing pattern averaged 4.5 cm, and those that did not averaged 1.7 cm. Logistic regression analysis showed a statistically significant association between larger leiomyoma size and the recurrent shadowing pattern on transabdominal (p = 0.0001) and on transvaginal (p = 0.0001) sonography. The relationship is shown in Table 3, which shows an increase in the percentage of leiomyomas exhibiting the recurrent shadowing pattern with increasing leiomyoma size.
|
Of the 130 leiomyomas detected on transabdominal sonography, 94 (72%) were intrauterine, 20 (15%) were indeterminate, and 16 (12%) appeared extrauterine. The sensitivity of intrauterine location for leiomyoma on transabdominal sonography was 72% and the specificity was 89%. Recurrent shadowing was seen in 16 (44%) of 36 leiomyomas that were indeterminate or extrauterine in location on transabdominal sonography. Of the 95 leiomyomas evaluated on transvaginal sonography, 67 (71%) were intrauterine, 16 (17%) were indeterminate, and 12 (13%) appeared extrauterine. Sensitivity of intrauterine location for leiomyoma on transvaginal sonography was 70% and specificity was 90%. Recurrent shadowing was seen in 25 (89%) of 28 leiomyomas that were indeterminate or extrauterine on transvaginal sonography (Fig. 3).
|
On transabdominal sonography, the range of transducer frequencies varied from 2.0 to 5.0 MHz: 2.0 MHz (n =3), 2.5 MHz (n = 15), 3.0 MHz (n =5), 3.5 MHz (n = 12), 4.0 MHz (n = 29), 4.5 MHz (n = 0), and 5.0 MHz (n = 3). Transvaginal transducer frequency ranged from 5 to 7 MHz: 5.0 MHz (n = 80), and 7.0 MHz (n = 2). No significant relationship was noted between higher frequency transducers and the shadowing pattern on transabdominal sonography (p = 0.3859) or on transvaginal sonography (p = 0.1753).
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
A pattern of discrete recurrent sonographic shadowing was recently described with leiomyomas [1]. Pathologic correlation showed these shadows originate from transitional zones between the margins of smooth muscle whorls and the margins of fibrous connective tissue within the leiomyoma. The ultrasound beam is presumably refracted or distorted when it encounters different tissue types or curved surfaces. Although this pattern is frequently exhibited by leiomyomas, its diagnostic usefulness has not been previously studied. Particularly, the sensitivity and specificity of recurrent shadowing, and the ease of applying this finding in clinical use, has not yet been described to our knowledge.
This investigation shows that the recurrent shadowing pattern is highly associated with leiomyomas and that the pattern can be recognized with high interobserver correlation. A pelvic mass that exhibits this pattern has a high likelihood of being a leiomyoma. Recurrent shadowing can be particularly helpful in suggesting the possibility of leiomyoma in the setting of an exophytic or pedunculated lesion that otherwise might be confused with an adnexal mass. However, we found the absence of the shadowing pattern does not exclude the diagnosis of leiomyoma.
Although the shadowing pattern suggests the diagnosis of leiomyoma, it is not unique to leiomyomas. Recurrent shadowing was occasionally observed in a small number of other disorders including leiomyosarcoma, ovarian fibroma, ovarian metastasis, adenomyosis, and pelvic abscess (Fig. 4). In each of these cases, the underlying histopathology of the lesion suggests the cause of the shadowing pattern. For example, a leiomyosarcoma contains multiple swirling masses of smooth muscle cells producing multiple transition points [13]. Similarly, an ovarian fibroma consists of dense collagenous bundles of fibrous tissue [13], and the interfaces between these bundles could create the refractory shadows. Transition zones may also be found in the conglomerations of carcinomatous cells in metastasis to the ovary and in peritoneal deposits of ovarian carcinoma [13]. Interfaces between ectopic glandular tissue and the surrounding myometrium may produce the recurrent shadowing pattern in cases of adenomyosis [13]. Distinguishing between leiomyoma and other disorders that occasionally produce the shadowing pattern will not always be possible. In these cases, ancillary sonographic features could be helpful in making this distinction. For example, adenomyosis should be suspected when it is difficult to define the margins of the mass and if small myometrial cysts are seen [14, 15, 16]. Alternatively, leiomyosarcoma is notoriously difficult to distinguish from leiomyoma unless evidence of local invasion or metastasis is seen [3].
|
Interobserver disagreement regarding the presence of recurrent shadows was infrequent but was more likely to occur with transabdominal imaging and when several leiomyomas were clustered together. The presence of additional masses made it difficult to unambiguously attribute particular shadows to particular masses. Additionally, the boundaries of any one mass were often difficult to define in a heterogeneous echo texture produced by contiguous shadowing masses.
The shadowing pattern is more commonly seen on transvaginal imaging, possibly because the mass is insonated in close proximity and with greater resolution using higher frequency probes than those used for transabdominal sonography. This said, no significant association between the shadowing pattern and transducer frequency was found when the effect of different probe frequencies was evaluated for the transabdominal and transvaginal shadows separately. In part, the lack of association may be a result of the small number of masses imaged using some of the probe frequencies in each group.
The larger the leiomyoma, the more likely the recurrent shadowing pattern will be evident on both transabdominal and transvaginal sonography, probably because of the greater abundance, density, and number of discrete shadows originating from the larger tumors. Smaller leiomyomas may have similar shadows that cannot be seen with current sonographic equipment, and therefore the recurrent shadowing pattern is less likely to be helpful in smaller masses.
The high prevalence of leiomyomas in our study reflects the high prevalence of leiomyomas in clinical practice. Despite this preponderance of leiomyomas in the study group, statistical analysis clearly shows the strong association of refractory shadowing with the diagnosis of leiomyomas (p << 0.001). In clinical practice, the diagnosis of leiomyomas will depend on several sonographic features, not least of which will be the location of the mass. Most leiomyomas are intrauterine, and most intrauterine masses are leiomyomas [17]. Conceding this, refractory shadows evident with transvaginal imaging signaled the diagnosis of 25 (89%) of 28 extrauterine leiomyomas that were otherwise nonspecific in appearance (Fig. 3).
In summary, we have shown that a pattern of sharp, discrete, recurring shadows is typical of leiomyomas and can be recognized with excellent interobserver agreement. Although a similar pattern can occasionally be produced by a small number of other pelvic masses, this pattern can provide useful corroborative evidence for leiomyomas in cases of diagnostic ambiguity.
Acknowledgments
We thank Susan Murray for assistance with manuscript preparation.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  S. Adusumilli, H. K. Hussain, E. M. Caoili, W. J. Weadock, J. P. Murray, T. D. Johnson, Q. Chen, and B. Desjardins MRI of sonographically indeterminate adnexal masses. Am. J. Roentgenol., September 1, 2006; 187(3): 732 - 740. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. L. Berridge and T. C. Winter Saline Infusion Sonohysterography: Technique, Indications, and Imaging Findings J. Ultrasound Med., January 1, 2004; 23(1): 97 - 112. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
