Chapter 36: ACUTE PELVIC PAIN: TRANSVAGINAL AND DOPPLER EVALUATION

Patients with acute pelvic pain often pose a diagnostic dilemma. This stems from the similar signs and symptoms which many of these patients exhibit, the most common being nausea, vomiting, and leukocytosis. The differential considerations encompass multiple gynecologic causes such as hemorrhagic ovarian cysts, pelvic inflammatory disease, ectopic pregnancy, and ovarian torsion, as well as nongynecologic etiologies which include appendicitis, urinary calculi, inflammatory bowel disease, diverticulitis, and mesenteric adenitis. Transvaginal (TVS) and transabdominal (TAS) pelvic sonography have the ability to narrow the differential diagnosis and are the imaging modalities of choice when a gynecological etiology is suspected. Not only is it noninvasive, radiation-free, and cost-effective, sonography accurately delineates the architecture of the uterus and ovaries.¹ Computed tomography (CT) is used more frequently when a gastrointestinal or genitourinary abnormality is more likely.

The role of pelvic sonography in the evaluation of acute pelvic pain has been well described. Due to better anatomic resolution, TVS rather than TAS should be used whenever possible, although TAS is recommended when uterine and adnexal structures are beyond the field of view of the transvaginal probe, such as above the pelvic brim. The introduction of the vaginal probe into the vagina to determine whether pelvic structures move independently of each other (sliding organ sign) or are the source of the patient's pain are also helpful techniques. In addition, duplex and color or power Doppler imaging can be used to characterize vascularity to the ovaries, adnexal structures, and uterus, which may also be beneficial in narrowing the field of differential considerations.

A serum β human chorionic gonadotropin (hCG) level is usually performed when a female in the reproductive years presents with acute pelvic pain. Knowledge of pregnancy is of utmost importance for diagnostic reasons (determine whether the pain could be associated with pregnancy such as ectopic pregnancy) and safety reasons (whether the patient may have CT, which uses ionizing radiation). A negative hCG essentially excludes the diagnosis of a live intrauterine pregnancy and acute ectopic pregnancy because it becomes positive approximately 9 days after conception. Studies published in the late 1980s, which correlated the presence of a gestational sac using TVS with β-hCG levels, have played a crucial role in our ability to make this diagnosis. These studies documented the presence of a gestational sac by the time the hCG level was 1000 to 2000 mIU/mL (IRP [International Reference Preparation]).^{2,3, and 4} When the hCG level is above this discriminatory zone, one should expect to see a gestational sac, and if not an ectopic pregnancy is very likely. The high specificity of adnexal findings, which include the classic "tubal ring," has been widely reported in the literature. Further discussion of ectopic pregnancy may be found in Chapter 4.

The following is a review of sonographic characteristics of frequent etiologies of acute pelvic pain. Because sonography allows simultaneous evaluation of nongynecologic as well as gynecologic etiologies, some discussion of common nongynecologic causes will also be incorporated in this chapter. These disorders include the following:

1. Adnexal torsion

2. Ovarian functional cysts

3. Pelvic inflammatory disease

4. Ureteral calculi

5. Appendicitis

6. Inflammatory bowel disease (ischemic bowel)

Adnexal Torsion

In the evaluation of gynecological causes of pain, TVS can not only help to identify the likely etiology, but also separate surgical conditions from medical ones. For example, functional ovarian cyst or pelvic inflammatory disease is more compatible with medical management, whereas ovarian torsion is a surgical emergency. The initial role of sonography in the evaluation of patients at risk for ovarian torsion is to help exclude other causes of acute abdominal pain such as appendicitis, pelvic inflammatory disease, ruptured ovarian cyst, and ectopic pregnancy.^5,6 Characteristic findings may be routinely seen by transvaginal and transabdominal sonography, although appearances may be nonspecific. For example, the most common sonographic gray-scale finding of ovarian torsion is an enlarged ovary or ovarian mass complex (Figure 36-1A). The presence of an adnexal mass or enlarged ovary has been documented in all subjects in several studies. Ovarian volumes within these studies have ranged from 26 to 4308 cm³.^{7,8,9, and 10} However, if the twisted ovary is located in the right lower quadrant of the abdomen, it may be mistaken for an appendiceal abscess.⁷

Several distinguishing gray-scale appearances of ovarian torsion have been described using transabdominal sonography in children and adolescents. A large cystic mass with internal debris has been shown to be fairly specific for torsion in neonates.^{11,12,13, and 14} In older children and adolescents, a unilaterally enlarged ovary containing peripheral follicular cysts has been described as characteristic of torsion in as many as 74% of cases (Figure 36-2).^7,10,15 Weed and Collins,¹⁶ in an experimental study using an animal model, produced passive congestion and cystic dilatation of follicles by ligating the ovarian vessels. Follicular enlargement was ascribed at least partially to transudation of fluid into the follicles due to vascular impairment. Reports of cases of massive ovarian edema, a condition attributed to partial torsion of the ovary, describe similar follicular structures in the cortex of an enlarged ovary.¹⁷ It is possible to conclude that this combination of findings in the prepubertal age group in which follicular enlargement is not expected may be specific for ovarian torsion.

In the postmenarchal population, ovarian appearances are more difficult to interpret. An irregular internal texture of the ovary seems to correlate with intraovarian hemorrhage (Figure 36-3). Peripherally placed follicles and homogeneous echoes seen centrally consistent with areas of edema within enlarged ovaries have also been reported,^10,15 although these descriptors are nonspecific. Entities such as endometriomas, hemorrhagic cysts, tubo-ovarian complexes, and hyperstimulated ovaries undergoing ovulation induction are often similarly described. Free fluid within the cul-de-sac is another nonspecific finding frequently described associated with cases of ovarian torsion.¹⁵ The fluid may be a transudate from the ovarian capsule secondary to obstruction of veins and lymphatic vessels.¹⁸ On the other hand, a small amount of fluid is often physiologic in females of reproductive age.

With the advent of color and spectral Doppler analysis of ovarian arterial and venous waveforms, it was assumed that this would prove to be an accurate tool for the evaluation of ovarian torsion. In actuality, studies of Doppler flow patterns in torsion are variable with the Doppler findings varying depending on degree of torsion and its chronicity. In some cases of torsion, there can be focal areas of blood flow, whereas in those cases that are more complete, no intraovarian venous or arterial flow is present (see Figure 36-1B). Lack of arterial and venous Doppler flow should enable confident diagnosis, although false-positive diagnoses may be obtained as a result of the depth of penetration being greater than the capabilities of the ultrasound beam, improper Doppler or gray-scale priority settings, and too high filter pulse repetition frequency (PRF) settings.^{19,20, and 21} Conversely, ovarian color Doppler signal has been frequently reported in cases of surgically proven torsion.^7,8

Multiple studies have supported the use of Doppler interrogation of the ovary in the evaluation of torsion. Ben-Ami et al¹⁹ reported a positive predictive value for torsion in the absence of venous flow of 94%, and that torsion is highly unlikely if Doppler interrogation reveals venous flow. Fleischer et al²⁰ and Lee et al²¹ assert that the presence of arterial and venous flow within adnexa that were found to be torsed upon subsequent surgical exploration was predictive of the ultimate viability of the organs. The study by Lee et al found that gray-scale and color Doppler sonography preoperatively identified twisting of the vascular pedicle in 28 of 32 patients with surgically confirmed torsion (Figure 36-4). The same study described 57% of cases as having both arterial and venous flow within the vascular pedicle, but later concluded, using pathologic findings or clinical follow-up after a procedure to uncoil the twisted vascular pedicle, that 15 of these 16 ovaries with Doppler flow were viable organs. Investigations by Fleischer et al showed no venous Doppler flow centrally within all 10 cases of surgically proven nonviable ovaries. In addition, venous Doppler signal was seen centrally within all 3 surgically proven viable ovaries. However, these findings were not confirmed in another series by Tepper et al.²²

Other investigators have found Doppler findings to be less reliable. Several case reports and retrospective studies of adnexal torsion diagnosed at the time of surgical intervention have described an abnormal appearance of the ovary with normal arterial and venous Doppler signal (Figure 36-5).^7,8,23,24 In a retrospective series, Albayram and Hamper⁸ evaluated 15 pathologically proven torsed ovaries, of which 53% demonstrated some degree of arterial signal and 27% showed evidence of venous signal within the ovary. Interestingly, in 14 out of the 15 cases, abnormal Doppler flow was described when compared to Doppler flow within the normal contralateral ovary. Similar results have also been found in a retrospective series of 39 patients with pathologically proven ovarian torsion performed at Vanderbilt University Medical Center.²⁵ Twenty-one (54%) patients had documented arterial Doppler signal within the torsed ovary and 13 (33%) had documented venous flow.

The variety and discrepancy of Doppler findings in the literature associated with ovarian torsion may be linked to the completeness of obstruction of the vascular supply.²⁶ Venous thrombosis may cause symptoms prior to the development of arterial occlusion, explaining the more frequently demonstrated absence of venous Doppler signal. Persistent Doppler arterial and/or venous flow may also be related to the dual blood supply to the ovary. To date, there appear to be no large-scale studies in the English literature evaluating the predictive value of Doppler flow in cases of ovarian torsion.

Although characteristic sonographic findings of torsion have been reported in the prepubertal age group, there is much less specificity of sonographic findings in the postmenarchal age group. The enlarged, abnormal-appearing ovary does appear to be a common denominator in most, if not all, cases of ovarian torsion. Evidence strongly points to the diagnosis in the presence of an enlarged ovary with absence of associated arterial and venous Doppler signal.^{19,20, and 21} However, multiple case reports and series in the literature suggest the presence of both arterial and venous Doppler signal within the ovary or associated vascular pedicle in cases of ovarian torsion. Furthermore, these cases in which vascularity is still present may be the very patients who would benefit from conservative surgery. Surgical exploration then should not be delayed when clinical suspicion remains high in the setting of an enlarged ovary with Doppler flow. Perhaps investigators should now turn their attention to research that would develop methods to determine the viability of the ovary. The demonstration of the twisted ovarian pedicle may be a promising avenue to explore since its appearance has been shown to be characteristic of ovarian torsion in both the sonographic and CT literature (see Figure 36-4), and the presence or absence of vascular flow within the pedicle may correlate with ovarian viability. Comparing the vascularity of the abnormal-appearing ovary with that of the normal contralateral side may prove helpful. Additionally, further investigation of the use of contrast-enhanced sonography may provide more accurate assessment of ovarian vascular flow and viability.

Ovarian Functional Cyst

Functional ovarian cysts are physiologic and not neoplastic lesions of the ovary, whose appearance and disappearance is hormone driven. These cysts may be simple or hemorrhagic and are commonly divided into 3 categories dependent upon the clinical scenario, the follicular cyst, the corpus luteum cyst, and the theca lutein cyst. Acute abdominal pain is most often caused by internal hemorrhage, rupture, or leakage, but may be simply a result of its large size, which is greater than 2.5 to 3.0 cm. The most common form is the follicular cyst, which is the result of abnormal follicular enlargement due to anovulation. It is often associated with hemorrhage and rapid resorption, with most resolving within 2 months. However, it should be emphasized that resolution may be delayed if the patient continues to be anovulatory, without a normal menstrual cycle, unless cyst rupture occurs. The corpus luteum cyst is a result of enlargement of the corpus luteum of pregnancy with resolution usually by 14 weeks' gestational age. Theca lutein cysts are associated with very high levels of b-hCG often seen with trophoblastic disease and multiple gestations. They are usually very large and numerous, and may persist for weeks after the withdrawal of the stimulus.

The sonographic appearance of the functional cyst without hemorrhage is that of a thin-walled simple cyst, although with hemorrhage appearances vary with chronicity. Acute hemorrhage is hyperechoic and may be suggestive of a solid mass. A diffuse pattern of low-level echoes may be seen, although this pattern is more commonly associated with an endometrioma (Figure 36-6A). With hemolysis and retraction of clot, a reticular network of stranding, septations, or fluid–fluid levels between fluid components and congealed red blood cells is demonstrated (Figure 36-7). Internal echoes with septations are often seen. Leakage of hemorrhagic fluid with associated peritoneal irritation is common.

Color and spectral Doppler evaluation is characteristic with typical findings of peripheral color Doppler signal with arterial low-impedance flow (see Figures 36-6B and 36-8.).²⁷ Since solid components are not present, there is no central vascularity. This is extremely helpful in the differentiation of an echogenic hemorrhagic cyst from a solid mass.

Pelvic Inflammatory Disease

Pelvic inflammatory disease (PID) consists of inflammation of the endometrium, fallopian tubes, pelvic peritoneum, and adjacent structures. The infection is usually ascending and caused by multiple organisms including Neisseria gonorrhoeae, Chlamydia, anaerobic bacteria, and superinfecting organisms from the vagina. Typically, the primary infection is a sexually transmitted disease in the normal pelvis, although with previous disruption of endometrial and tubal tissue due to prior infection or postsurgical or postpartum changes, the patient can be infected by her own vaginal flora. The disease is manifested by tubo-ovarian complexes, peritonitis, and abscess formation. It is usually bilateral but may be unilateral in patients with intrauterine devices.

The clinical diagnosis of pelvic inflammatory disease has been difficult. No simple diagnostic test exists. The accuracy with which signs and symptoms predict the presence of this disease has been evaluated using a laparoscopic "gold standard."²⁸ The data set included women who had an initial diagnosis based on clinical presentation and subsequent laparoscopic evaluation. A total of 623 patients were included in the analysis; 494 patients were laparoscopically confirmed as having PID and 129 were not, giving a false-positive rate of 26%. The 3 variables which significantly influenced the prediction were elevated erythrocyte sedimentation rate (ESR), fever, and adnexal tenderness. However, together they correctly classified only 65% of women with laparoscopically diagnosed pelvic inflammatory disease.

The advent of imaging with transvaginal sonography with color or power Doppler has significantly improved our ability to detect abnormalities, but its sensitivity and specificity in the imaging literature has been highly user and equipment dependent and is determined by the findings which have been used to make the diagnosis. TVS can usually only detect complications of the disease since early changes are often subtle. Increased echogenicity of peritoneal fat and indistinctness of the uterus may be seen early in the disease process but are very difficult to appreciate. Sonographic findings of the fallopian tubes are the most specific and conspicuous indicators of PID. The tubal wall is usually not visualized without intraluminal or peritoneal fluid unless a truly targeted study is performed. With inflammation, the tube swells, endosalpingeal folds thicken, and with progressive inflammation and distal occlusion of the lumen, the tube fills with purulent echogenic material becoming a pyosalpinx (Figure 36-9). In the presence of a fluid-distended fallopian tube, common findings include wall thickness greater than 5 mm, incomplete septa seen as the tube folds back upon itself, and thickening of endosalpingeal folds (cogwheel sign), which is best seen when the tube is viewed in cross section (Figure 36-10).²⁹ Color or power Doppler may also be a sensitive method of detecting hyperemia in the walls and incomplete septi associated with fallopian tube inflammation.³⁰ Nonspecific findings of PID include fluid in the endometrial cavity and/or cul-de-sac, and ovarian enlargement often with numerous small cysts ("polycystic ovary" appearance"; Figure 36-11).^31,32

With progression of disease, there is exudation of pus from the distal fallopian tube and the ovary can become involved. An inflammatory mass including the tube and adjacent ovary is formed. If a separate ovary is still visualized, this indicates a tubo-ovarian complex. A tubo-ovarian abscess results in complete breakdown of tubal and ovarian architecture so that separate structures are no longer identified and there is obscuration of the posterior and lateral margins of the uterus.³³

Although CT is generally better than ultrasound in evaluating pelvic pain likely from the gastrointestinal or urinary tract, transvaginal sonography and color Doppler sonography can be used in the diagnosis of ureteral calculi, appendicitis, and other bowel disorders. Sonography should always be considered when gastrointestinal or urinary tract pathology is suspected in pregnant patients. Despite some diagnostic limitations, sonography has the advantage over CT of sparing the fetus ionizing radiation.

Symptomatology of appendicitis may be suggestive of gynecological conditions, so pelvic sonography may be the initial imaging study. Appendicitis can be recognized by either TAS or TVS by the presence of a thick-walled, tubular, noncompressible, aperistaltic structure measuring over 3 mm in single-wall thickness arising from the tip of the cecum (Figure 36-12). Sonographic detection of an appendicolith within the appendix, regardless of wall diameter, should be considered diagnostic.³⁴ Other findings include surrounding echogenic fat and increased color Doppler vascularity within the luminal wall consistent with inflammation. Although usually best depicted by the transabdominal approach using the graded compression technique first described by Julien Puylaert in 1986,³⁵ the proximity of bowel within the cul-de-sac to the TVS probe may allow diagnosis by this method.³⁶ Additionally, the adjuvant use of a posterior manual compression technique more recently described by Lee et al³⁷ has demonstrated increased sensitivity in diagnosing appendicitis over graded compression alone. Unfortunately, the technique of graded compression required for this diagnosis is often not feasible in the presence of an enlarged gravid uterus. More importantly, in the diagnosis of appendicitis, TAS has demonstrated variable sensitivity (67% to 100%) and specificity (83% to 96%)^38,39 and the normal appendix is visualized in only 13% to 50% patients, even in the absence of pregnancy.³⁹ Unless the normal appendix is visualized, the diagnosis of appendicitis cannot be excluded.³⁸

The distal ureter can be evaluated with TAS or TVS for calculi (Figure 36-13).⁴⁰ Ureteral obstruction by calculi seen within a distended ureter can be suggested by the lack of a urine jet usually seen by Doppler arising from the ureteral orifice. For diagnosis of obstructing ureteral calculi, a wide range of sensitivities (34% to 95%) has been reported with sonography.^41,42 TVS has been recommended for detection of distal ureteral calculi⁴⁰ should TAS prove inconclusive. Sonographic specificity is reduced by the physiologic urinary dilatation common in pregnancy. Nevertheless, a normal renal sonogram can always be useful in excluding renal colic in pregnant as well as nonpregnant patients.

Ischemic bowel appears as an abnormally thickened aperistaltic loop. Color Doppler sonography may show absent signal within the bowel wall.⁴³ Inflammatory bowel diseases such as Crohn disease can also be associated with increased vascularity within the bowel wall.