| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Obstetric-Gynecologic Imaging |
1 Harvard Medical School, 25 Shattuck St., Boston, MA 02115.
2 Department of Radiology, Beth Israel Deaconess Medical Center, 330 Brookline
Ave., Boston, MA 02215.
3 Department of Obstetrics and Gynecology, Beth Israel Deaconess Medical Center,
Boston, MA 02215.
Received October 3, 2003; accepted after revision March 26, 2004.
Address correspondence to D. Levine.
OBJECTIVE. Our aim was to determine the accuracy of the intradecidual sign for the diagnosis of intrauterine pregnancy and the exclusion of ectopic pregnancy.
CONCLUSION. The intradecidual sign reliably excludes the presence of an ectopic pregnancy. The sensitivity for diagnosis of an intrauterine pregnancy increases when human chorionic gonadotropin levels are equal to or greater than 2,000 mIU/mL or the mean sac diameter is equal to or greater than 3 mm. It is of utmost importance to visualize this sign on multiple views with an unchanging appearance.
The incidence of ectopic pregnancy remains high in the United States, accounting for 2% of reported pregnancies and 9% of all pregnancy-related deaths [1]. Because the rate of heterotopic pregnancy is low—approximately one in 7,000 spontaneous pregnancies [2]—visualization of an intrauterine gestational sac on sonography can effectively minimize the likelihood that an ectopic pregnancy is present.
In the first trimester, before a yolk sac or embryo can be visualized, it is important to determine if an intrauterine fluid collection represents either an intrauterine pregnancy (IUP), a pseudogestational sac associated with an ectopic pregnancy, or a decidual cyst (although associated with ectopic pregnancy, it can also be seen in IUP). Several signs have been suggested to distinguish these possibilities. The first, the double decidual sac sign, was described in 1982 by Bradley et al. [3] as two concentric echogenic rings of tissue surrounding the intrauterine sac that protrude into the uterine cavity. This morphologic appearance differs from and excludes the presence of an ectopic pregnancy. However, a yolk sac is typically present by the time the double decidual sac sign can be properly visualized [4]. The time frame of the double decidual sac sign thus places its utility in question.
The intradecidual sign was described by Yeh et al. [5] in 1986 as an echogenic area embedded in the thickened decidua that is eccentrically located on one side of the uterine cavity, which appears as a well-defined endometrial stripe. This area differs from the pseudogestational sac, which appears as fluid surrounded by the echogenic endometrial lining in an ectopic pregnancy [6]. Yeh et al. reported that the intradecidual sign was shown to reveal implantation as early as 25 days' gestational age, earlier than the effective window for the double decidual sac sign, and was more specific and sensitive than the double decidual sac sign in detecting early IUP.
Because Yeh et al. [5] described the intradecidual sign using transabdominal sonography, Laing et al. [6] retested the reliability of the intradecidual sign, using the improved resolution of transvaginal sonography. Surprisingly, they found that among four reviewers of varying experience, the sensitivity for diagnosis of an IUP ranged from 34% to 66% and the specificity ranged from 55% to 73%. Furthermore, even the categorization of ectopic pregnancies based on the intradecidual sign was inaccurate. This study thus concluded that the intradecidual sign is neither specific nor sensitive as a diagnostic tool for normal IUP.
Recently, two other studies have suggested ways to exclude ectopic pregnancies, although the reliability of each has yet to be established. The first reported that the chorionic rim, the echogenic rim surrounding a gestational sac, was 80% sensitive and 97% specific for IUPs. Two pseudogestational sacs with echogenic margins were incorrectly classified as IUPs in this study [7]. The second reported that endometrial arterial blood flow had a predictive value of 97% for detecting IUP with findings in three false-positive ectopic cases that showed flow [8].
The purpose of our study was to revisit the intradecidual sign because to date, no follow-up studies support the results of either the Yeh et al. [5] or Laing et al. [6] studies, and no new signs have been reported or validated to reliably exclude the presence of an ectopic pregnancy. Ours was a retrospective study to determine whether the intradecidual sign was specific for IUP, using 3 years of obstetric images of the first trimester obtained by transvaginal sonography that were independently scored by three reviewers with varying levels of experience with sonography. This study also attempted to assess whether the human chorionic gonadotropin (ß-hCG) value at the time of the sonography or the mean sac diameter of the fluid collection played a part in the accuracy of the intradecidual sign.
Materials and Methods
This study was approved by our institutional internal review board. We
performed a computer search of 3 years (September 1, 1999–August 31,
2002) of sonographic examinations to identify patients who were pregnant and
whose sonographic findings revealed the presence of an intrauterine fluid
collection associated with either an early IUP of less than 5.5-weeks'
gestational age (mean sac diameter, 
8 mm) or an ectopic pregnancy.
Sonograms were obtained by either sonography technologists or radiology
residents using 5-9–MHz transvaginal transducers on 3000 or 5000
machines (ATL). Scans obtained during the day were reviewed at the time of the
study by an attending radiologist. Scans obtained during the evening hours or
on weekends were initially interpreted by residents and were later interpreted
by attending radiologists. Follow-up was coded as a live IUP if subsequent
sonography revealed a fetal heartbeat or a hospital record reported the
delivery of a live infant corresponding to the index pregnancy, an abnormal
IUP if subsequent dilatation and curettage showed chorionic villi, or as
ectopic pregnancy if surgical records showed an ectopic pregnancy or if the
sonographic findings showed classic findings of an ectopic pregnancy with an
extraovarian adnexal mass in patients treated with methotrexate or showed free
fluid with debris with follow-up ß-hCG levels showing inappropriate
increases over more than 7 days. Cases were excluded if there was no
intrauterine fluid; evidence of both intrauterine and extrauterine pregnancies
(n = 2); no transvaginal sonogram (n = 1); or no definitive
follow-up; or if the images revealed the presence of a yolk sac, embryo, or a
double decidual sac.
We compiled a list of 187 cases according to previously stated criteria: 153 IUPs (104 normal, 49 abnormal) and 34 ectopic pregnancies from a total of 16,142 pelvic and obstetric sonography reports. Of the 34 ectopic pregnancies, 23 were confirmed at surgery, eight had a classic appearance with an extraovarian ringlike mass without free fluid with debris, and three women had inappropriately rising ß-hCG values over 8–10 days (two of whom also had free fluid with debris in the pelvis). The ß-hCG value obtained within 24 hr of the sonography was recorded. This value was available in 140 of 187 pregnancies. The mean sac diameter of the intrauterine fluid collection was recorded for all 187 cases. If only endometrial fluid in an oblong collection was seen in the central endometrial cavity, the size of the collection in three dimensions was averaged, although this was not a true mean sac diameter.
For each case, a radiologist with 10 years' obstetric sonography experience, without knowledge of pregnancy outcome, selected four images that best revealed the character and location of the intrauterine fluid collection. The selected images from the cases were then independently reviewed and graded by three individuals: an experienced obstetric imager with 6 years' experience (beyond residency) in obstetric–gynecologic sonography, an experienced abdominal imager with 1 year's (beyond residency) experience in obstetric–gynecologic sonography, and a women's imaging fellow with 1 month's experience in obstetric–gynecologic sonography. Before viewing the images, the reviewers were given the manuscripts of and Yeh et al. [5] and Laing et al. [6] for illustration of the intradecidual sign. On the basis of the morphologic appearance of the fluid collection and the characterization of the sign in the Yeh and Laing articles, each reviewer graded the visualization of the intradecidual sign according to the following 5-point scale: 1, the intradecidual sign is definitely not present (a fluid collection is located centrally or a fluid collection is present but there is no clear gestational sac, or an endometrial stripe was not visualized); 2, the intradecidual sign is probably not present; 3, findings are indeterminate; 4, the intradecidual sign is probably present; and 5, the intradecidual sign is definitely present (a small fluid collection with an echogenic rim is located on one side of a well-visualized endometrial stripe). Reviewers were blinded to adnexal findings and clinical information.
The scores for each case were analyzed separately for each reviewer and as a median of the three reviewers. The number of cases in which the intradecidual sign was considered present, absent, or indeterminate was noted for IUPs and ectopic pregnancies. The sensitivity for an IUP was then calculated, with a score of 4 or 5 being positive and 1, 2, or 3 being negative for the sake of comparison with the findings of Laing et al. [6]. Interobserver correlations were assessed with the kappa statistic using MINITAB (Release 12.23, MINITAB) for Windows (Microsoft) according to the following scale: kappa value equal to or greater than 0.8 suggests perfect agreement; equal to or greater than 0.6 suggests substantial agreement; equal to or greater than 0.4 suggests moderate agreement; and equal to or greater than 0.2 suggests slight agreement [9].
Subgroup analysis was then performed to determine whether the accuracy of the intradecidual sign changed when cases of IUPs were in the discriminatory zone of ß-hCG value equal to or greater than 2,000 mIU/mL and mean sac diameter equal to or greater than 3 mm [10, 11]. The Fisher's exact test (InStat, GraphPad Software) was used for comparing the contingency table of normal IUPs with ß-hCG values greater than and less than 2,000 mIU/mL. The Kruskal-Wallis test was used to compare median scores in groups with gestational sac size of less than 3 mm compared with equal to or greater than 3 mm using MINITAB. For false-positive cases, in which an ectopic pregnancy was scored at 4 or 5, the four selected images were shown again to the group of reviewers to elicit the reason that the case was incorrectly scored.
Results
Images from cases that were graded 1–5 are shown in Figures 1, 2, 3, 4 and 5. The median scores of the three reviewers for the 187 patients are shown in Table 1. Of the 153 patients with an IUP (normal or abnormal), 107 had an average score of 4 or 5, leading to a sensitivity of 70%. All the ectopic pregnancies had a median score of 1 or 2, yielding a specificity of 100% for the intradecidual sign; the accuracy rate was 75%, the positive predictive value was 100%, and the negative predictive value was 43%.
|
|
|
|
|
|
In cases of ectopic pregnancy, the intradecidual sign was considered definitely or probably absent in all 34 cases by the obstetric imager, 33 (97%) of the 34 cases by the abdominal imager, and 28 (82%) of 34 cases by the women's imaging fellow, who classified five of the ectopic pregnancy cases (15%) as indeterminate. Both the abdominal imager and the fellow had one false-positive finding in which an intradecidual sign was considered present (Table 2).
|
Of the 104 cases of normal IUP, an intradecidual sign was definitely or probably seen in 68 cases (65%) by the obstetric imager, 75 cases (72%) by the abdominal imager, and 73 cases (70%) by the women's imaging fellow (Table 3). In these normal IUPs, the intradecidual sign was considered indeterminate in as few as four cases (4%) by the abdominal imager and as many as 17 cases (16%) by the obstretic imagers. Similarly, in the normal IUPs, the intradecidual sign was considered to be definitively or probably absent in as few as 16 cases by the fellow and as many as 25 cases by the abdominal imager. Of the 49 abnormal IUPs, an intradecidual sign was considered definitely or probably present in 24–31 (49–63%), indeterminate in four to 11 (8–22%), and definitely or probably absent in 10–21 (20–41%) of the cases (Table 4). Considering total IUPs (normal and abnormal), the sensitivity for diagnosis of an IUP ranged from 60% to 68%. The comparative statistical measures by reviewer are summarized in Table 5.
|
|
|
Interreviewer-weighted kappa statistics ranged from 0.75 to 0.79, showing substantial agreement. The reviewers agreed exactly (same number score assigned) in 61 (33%) of 187 patients and had scores within one point of each other for an additional 80 (43%) of 187 patients.
Table 6 shows that the average ß-hCG value of the IUP cases increased relative to the assigned score. A similar trend was present when the mean sac diameters were tabulated with respect to the median score. Of the 73 normal IUPs with ß-hCG values available for comparison, 35 (68.6%) of 51 that were scored 4 or 5 had ß-hCG values equal to or greater than 2,000 mIU/mL compared with two (9.1%) of 22 that were scored equal to or less than 3 (p < 0.0001). The 48 of 187 cases with a mean sac diameter less than 3 mm had a median score of 1, whereas the 139 of 187 cases with a mean sac diameter of 3 mm or greater had a median score of 4 (p < 0.0001).
|
Of the IUPs with ß-hCG values greater than or equal to 2,000 mIU/mL, the intradecidual sign was graded as indeterminate in three cases, as probably present in 23 cases, and as definitely present in 23 cases. Three IUPs had a median score of 1 or 2, all abnormal IUPs. In this subgroup, the intradecidual sign had a sensitivity of 88% for diagnosing an IUP, and the accuracy rate was 91% (Table 7). Weighted kappa values showed substantial agreement (0.62–0.70) in this subgroup.
|
Mean sac diameter was greater than or equal to 3 mm for 133 patients and less than 3 mm for 41 patients. A central oblong fluid collection with an average diameter of 3.8 mm was present in 13 patients. In the subgroup of cases in which the mean sac diameter was greater than or equal to 3 mm, the intradecidual sign had 81% sensitivity for diagnosing an IUP, and the accuracy rate was 82% (Table 8). Weighted kappa values showed moderate to substantial agreement (0.56–0.61).
|
There were two false-positive ectopic pregnancies, each classified by one of the three reviewers. On retrospective review of these cases, each of the reviewers agreed that the intradecidual sign should not have been recorded as present. The images for one of the cases are included for educational purposes to show that the differing morphologic appearance of the fluid among the four selected images negates the presence of the intradecidual sign (Figs. 6A, 6B, 6C, and 6D).
|
|
|
|
Discussion
To our knowledge, only two studies have been published to address the reliability of the intradecidual sign, and the results differed drastically. The transabdominal study of Yeh et al. [5] of 41 patients found a sensitivity of 92%, specificity of 100%, and accuracy of 93%. Using transvaginal sonography, Laing et al. [6] found an overall sensitivity of only 48%, specificity of 66%, and accuracy of 45%. In our study, we attempted to expand on the study of Laing et al. by increasing the number of images reviewed by individuals to four in each case, assessing the role of the ß-hCG value and mean sac diameter, and retrospectively reviewing false-positive cases.
Although our study did not confirm the results of the study by Yeh et al. [5], the sensitivity of 60–68%, specificity of 97–100%, and accuracy of 67–73% in our study were much higher than those reported by Laing et al. [6]. Furthermore, when we focused on the sub-group of cases of IUPs in which ß-hCG levels were greater than or equal to 2,000 mIU/mL, the sensitivity increased to 88%. Similarly, the sensitivity was 81% for the subgroup of cases with a mean sac diameter of 3 mm or greater.
Our accuracy rate was higher than that of Laing et al. [6] possibly because we consistently chose four representative images for review, rather than the two to four used in their study. We hypothesize that using four images allowed better visualization of the changing appearance of fluid in the endometrium in cases of ectopic pregnancy, as shown in Figures 6A, 6B, 6C, and 6D. Using these four images highlights the importance of noting the varying shape and location of the fluid collection when assessing the intradecidual sign. Notably, the experienced obstetric imager correctly recorded the absence of the intradecidual sign in all 34 cases of ectopic pregnancy, with no cases even assigned to the indeterminate category. Thus, experience probably played a role in recognizing the difference between a pseudosac and the intradecidual sign.
A second possible reason that the intradecidual sign was more reliable in our study is that our images were obtained from 1999 to 2002, 4 years after those in the study by Laing et al. [6]. The sonographic equipment used in our study may have allowed improved resolution.
The main limitation of our study was the bias associated with the choice of cases and images by the radiologist. More than 16,000 cases were reviewed to find enough cases that fit the criteria for our study. Some inadvertent selection bias could have occurred in this process. Although the radiologist did not have knowledge of pregnancy outcome at the time of image selection, views of the adnexa were included in the study and therefore could potentially have biased image choice.
It can be difficult to visualize the relationship between small intrauterine fluid collections and the endometrial cavity. Any pregnant patient with symptoms of bleeding or pain or both in whom a definitive IUP has not yet been diagnosed should be followed up on sonography to diagnose a developing yolk sac or embryo or both.
In summary, we found the intradecidual sign reliably excludes ectopic pregnancy. Visualizing the unchanging appearance of this sign on multiple views is of the utmost importance.
References
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  D. Levine Ectopic Pregnancy Radiology, November 1, 2007; 245(2): 385 - 397. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
