Efficacy of Transvaginal Ultrasound vs Pelvic MRI in Preoperative Diagnosis of Pelvic Endometriosis

Introduction

Endometriosis is a gynecological disease that requires diagnosis among reproductive-aged women [1]. It develops when active endometrial tissue proliferates outside the endometrial area [2]. In general, it is reported in the pelvic area of women [3]. Endometriosis generally affects almost 6–10% of women of reproductive age and 25–35% of those who experience infertility [1]. Pathologically, it is detached endometrial tissue that causes a persistent inflammatory response, which is why it is called endometriosis [4]. In recent years, the possibility of performing an early diagnosis of endometriosis has increased thanks to the greater accuracy of diagnostic imaging techniques. However, this accuracy often increases when clinical symptoms, especially painful ones, aid imaging.

There is a wide range of clinical characteristics and diagnostic procedures for endometriosis [1]. Generally, Chinese guidelines recommend diagnostic methods for endometriosis in the presence or absence of clinical characteristics of endometriosis in women to confirm endometriosis [5]. Ultrasound is a cost-effective, easily accessible, and convenient method for the diagnosis of endometriosis [6], but it has limited resolution and specificity for imaging [7]. In addition, transvaginal ultrasound is operator-dependent. Therefore, experts and experience in the medical profession are required for reliable prediction and guidance [8]. However, magnetic resonance imaging (MRI) is a radiation-free and non-invasive method that shows the depth of infiltration and edges of lesions in pelvic images of women [9]. In addition, MRI can show tiny hemorrhagic lesions outside the pelvic cavity if endometriosis of the abdominal wall is present [7]. Cysts and adenomyosis require attention because they are difficult to diagnose with transvaginal ultrasound and MRI [10,11] and the laparoscopic surgical technique, and to visualize the endometrial deposits inside the abdomen is the criterion standard for diagnosis of cysts and adenomyosis, but it is invasive method [4,12].

The objective of this study was to compare the diagnostic parameters of transvaginal ultrasound with pelvic MRI for the diagnosis of endometriosis cyst(s) and adenomyosis (pelvic endometriosis) in women who have clinical manifestations, considering invasive laparoscopy and histopathologic evaluation as reference standards.

Material and Methods

ETHICS APPROVAL AND CONSENT TO PARTICIPATE:

The protocols of the established study were designed by the authors themselves and approved by the Human Ethics Committee of Xidian Group Hospital (approval number 15XGHxch, dated January 15, 2022) and the Chinese Obstetricians and Gynecologists Association. The study followed the 2008 Declaration of Helsinki and the law of China. As this was a retrospective study, the requirements for written informed consent and registration in the Chinese clinical trial registry were waived by the Human Ethics Committee of Xidian Group Hospital.

INCLUSION CRITERIA:

Women with dysmenorrhea, chronic pelvic menstruation, pain or tenderness, dyspareunia, constipation, dysuria, and infertility were included in the study.

EXCLUSION CRITERIA:

Women with missing clinical manifestations, MRI scans, transvaginal ultrasound scans, and/or laparoscopy and histopathologic evaluation data from the hospital records were excluded from the study.

PELVIC EXAMINATION:

This included direct examination of the pelvic region, vagina, and uterus. The matching of the following criteria (Table 1) was considered as endometriosis cyst(s) and adenomyosis (pelvic endometriosis) [2].

TRANSVAGINAL ULTRASOUND EXAMINATIONS:

This was performed using a 7.5 MHz ultrasound transvaginal probe (GE Healthcare, Chicago, IL, USA). On non-menstrual days, in women with partially filled bladders, transvaginal ultrasound (GE Healthcare, Chicago, IL, USA) examinations were performed. Transvaginal ultrasound examination for the detection of pelvic endometriosis followed the recommendations of the International Deep Endometriosis Analysis (IDEA) Group (Figure 1) [13]. Chocolate cysts, myometrial cysts, sub-endometrial microcysts, hypoechoic solid lesions (Figure 2; hyperechoic myometrial spots), nodules, other abnormal tissue growths, uterine enlargement, asymmetry (abnormal junction shape), thickening of the junctional zone, and/or peritoneal thickening was considered as adenomyosis [10]. The authors performed the ultrasound examination.

MAGNETIC RESONANCE IMAGING EXAMINATIONS:

Women underwent a 3.0-Tesla MRI scan (GE Healthcare, Chicago, IL, USA) to examine the pelvic compartments and structures. A lubricant gel was applied to the vaginal and pelvic parts of the body. Sagittal T1- and T2-weighted (Figure 3), axial (Figure 4), and coronal images were examined to evaluate pelvic compartments and structures. In addition, contrast images with and without fat suppression were also evaluated.

Image analyses were performed by radiologists with a minimum of 3 years of experience at institutes in image analyses of endometriosis.

LAPAROSCOPY AND HISTOPATHOLOGIC EVALUATION:

Women underwent laparoscopic interventions by urologists (minimum of 3 years of experience at institutes) to extract endometriotic lesions. The samples were then sent to pathologists for examination. In addition, endometriosis lesions removed by surgeons were sent to pathologists for examination. During surgery, the American Society for Reproductive Medicine #Enzian score and Endometriosis Fertility Index scores were used to classify endometriosis [14].

DIAGNOSTIC PARAMETERS:

Diagnostic parameters are defined as follows [15]:

True-positive was defined as diagnosis by an index test and detection by laparoscopy/surgical procedure and histopathologic evaluation. True-negative was defined as not diagnosed by the index test and not detected by laparoscopy or histopathologic evaluation. False-positive results were defined as those diagnosed by an index test but not detected by laparoscopy and histopathologic evaluation. A false-negative was defined as not being diagnosed by an index test but being detected by laparoscopy and histopathologic evaluation.

Sensitivity was defined as the ratio of true-positive values to the sum of true-positives and false-negatives. Specificity was defined as the ratio of true-negative values to the sum of true-negative and false-positive values. Accuracy was defined as the ratio of true-positive and true-negative values to true-positive, true-negative, false-positive, and false-negative values. Efficacy was defined as sensitivity plus specificity minus 100. Positive clinical utility was defined as sensitivity multiplied by positive predictive value. Negative clinical utility was defined as specificity multiplied by negative predictive value. The positive predictive value was defined as proportions of positive values for that diagnostic test. The negative predictive value was defined as proportions of negative values for that diagnostic test.

CLINICAL THREAT TO MISSED REPORTING OF PELVIC ENDOMETRIOSIS:

Clinical threats to missed reporting of pelvic endometriosis were evaluated as a function of detriment scores. The detrimental scores for each diagnostic test were calculated from the risk of overdiagnosis, as expressed in Eq. 1. The risk of overdiagnosis was defined using a calculation that involved diagnostic confidence below which laparoscopy and histopathologic evaluation was performed (Eq. 2). Diagnostic confidence was defined as a numerical value from 0 to 1. The detriment score of the diagnostic test was the area below the curve of the diagnostic test, and the working area was the area above the curve of the diagnostic test. For both parameters, laparoscopy and histopathological evaluation were used as the reference standards [16].

CLINICAL BENEFITS FOR DETECTION OF BIOMARKER SIGNS OF PELVIC ENDOMETRIOSIS:

The clinical benefits of detecting biomarkers (chocolate cysts, myometrial cysts, sub-endometrial microcysts, hypoechoic solid lesions [hyperechoic myometrial spots], nodules, other abnormal tissue growths, uterine enlargement and asymmetry [abnormal junction shape]), thickening of the junctional zone, and/or peritoneal thickening for the detection of pelvic endometriosis were evaluated as a function of the beneficial scores. The beneficial scores for each diagnostic test were calculated from the risk of underdiagnosis, as expressed in Eq. 3. The risk of underdiagnosis was defined with a calculation that involved diagnostic confidence above which laparoscopy and histopathologic evaluation were performed (Eq. 4). The area above the curve was underdiagnosed to detect biomarker signs for the diagnostic test, and the working area was the area under the curve of the diagnostic test. For both parameters, laparoscopy and histopathological evaluation were used as the reference standards [16].

In both clinical threats to missed reporting of pelvic endometriosis and beneficial scores for detection of biomarker signs of pelvic endometriosis (numerical conversation of image analyses), it was assumed that there were no inconclusive results. It was assumed that the radiologists concluded that either pelvic endometriosis was present or absent.

STATISTICAL ANALYSES:

InStat 3.01 GraphPad Software, San Diego, CA, USA) was used for the statistical analyses. Normal continuous, non-normal continuous, and categorical data and variables are presented as mean±standard deviation (SD), median with Q3-Q1 in parentheses, and frequencies with percentages in parentheses, respectively. A quartile calculator was used to calculate interquartile range values. Fisher’s exact test or chi-square test (when the sample size was more than 5 and the total sample size was more than 40; with Yate’s corrections) was used for categorical data and variables. The results were considered significant if the P value was <0.05.

Results

STUDY POPULATION:

From January 15, 2018, to March 12, 2023, 355 women with dysmenorrhea, chronic pelvic menstruation period pain or tenderness, dyspareunia, constipation, dysuria, and infertility were seen at Xinxiang Central Hospital, Xinxiang, Henan, China and the Xidian Group Hospital, Xi’An, Shaanxi, China. Among them, 11 women had missing clinical manifestations, MRI scans, transvaginal ultrasound scans, and/or laparoscopy and histopathologic evaluation data in the hospital records. Therefore, the data from these women were excluded from the study. Among them, 144 women underwent pelvic MRI, transvaginal ultrasound, laparoscopy, and histopathologic evaluation. The analytic, cohort, retrospective study flow diagram is shown in Figure 5.

DEMOGRAPHICAL AND CLINICAL PARAMETERS:

The demographic and clinical parameters of the 144 enrolled women were collected from hospital records and are reported in Table 2.

DIAGNOSTIC PARAMETERS FOR PELVIC ENDOMETRIOSIS:

In conclusion, for the diagnosis of each index test with demographic and clinical parameters, the index test was considered the main test for the diagnosis of pelvic endometriosis (an endometriosis cyst and adenomyosis), and demographic and clinical parameters were considered to support these results. The sensitivity, specificity, accuracy, efficacy, and positive clinical utility values of demographic and clinical parameters plus pelvic examination plus transvaginal ultrasound examinations were lower than those of demographic and clinical parameters plus pelvic examination plus pelvic MRI examinations in women. Demographic and clinical parameters and pelvic examination of women had no specificity, accuracy, or negative clinical utility value. The negative clinical utility value was higher for demographic and clinical parameters plus pelvic examination plus transvaginal ultrasound examinations than for demographic and clinical parameters plus pelvic examination plus pelvic MRI examinations in women. For all results, laparoscopy and histopathological evaluation values were considered as reference standards. The details of diagnostic parameters of pelvic endometriosis (endometriosis cyst and adenomyosis) per woman at a 95% confidence interval are presented in Table 3. True-negative results were reported in 4 (3%) women on transvaginal ultrasound examinations and 4 (3%) women on pelvic MRI, but demographic and clinical parameters plus pelvic examinations reported these were positive for pelvic endometriosis and subjected to laparoscopy and histopathologic evaluation because of risk of other uterine diseases.

CLINICAL THREAT TO MISSED REPORTING OF PELVIC ENDOMETRIOSIS:

The demographic and clinical parameters plus pelvic examination, the demographic and clinical parameters plus pelvic examination plus transvaginal ultrasound examinations, the demographic and clinical parameters plus pelvic examination plus pelvic MRI examinations, and the laparoscopy and histopathologic evaluation to detect pelvic endometriosis had 0–0.03, 0.03–0.09, 0.03–0.055, and 0 diagnostic confidence, respectively, which had detrimental scores for missing pelvic endometriosis in the working area (Table 4). Details of the working area for various index diagnostic tests for the detection of pelvic endometriosis are shown in Figure 6.

BIOMARKER SIGNS (CHOCOLATE CYSTS, MYOMETRIAL CYSTS, SUB-ENDOMETRIAL MICRO CYSTS, HYPOECHOIC SOLID LESIONS (HYPERECHOIC MYOMETRIAL SPOTS), NODULES, OTHER ABNORMAL TISSUE GROWTHS, THICKENING OF THE JUNCTIONAL ZONE, AND/OR PERITONEAL THICKENING): As the specificity of demographic and clinical parameters plus pelvic examination was zero, diagnostic parameters of different types of biomarker signs were not evaluated separately for demographic and clinical parameters plus pelvic examination. The results of the imaging scans and image analyses of different types of biomarker signs for the analytic, cohort, retrospective study are shown in Figures 7 and 8, respectively.

True-negative values, false-negative values, and sensitivity were identical for the demographic and clinical parameters plus pelvic examination plus transvaginal ultrasound examinations and the demographic and clinical parameters plus pelvic examination plus pelvic MRI examinations for biomarker signs (chocolate cysts, myometrial cysts, sub-endometrial microcysts, hypoechoic solid lesions (hyperechoic myometrial spots), nodules, other abnormal tissue growth, thickening of the junctional zone, and/or peritoneal thickening. There were no statistically significant values for true-positive (45 [83%] vs 47 [86%], p=0.7873, 95% confidence interval (CI): 0.5380 to 1.405, Fisher’s exact test) and false-positive (4 [7%] vs 2 [4%], P=0.6785, CI: 0.7467 to 2.477, Fisher’s exact test) values between the demographic and clinical parameters plus pelvic examination plus transvaginal ultrasound examinations and the demographic and clinical parameters plus pelvic examination plus pelvic MRI examinations for biomarker signs (chocolate cysts, myometrial cysts, sub-endometrial microcysts, hypoechoic solid lesions (hyperechoic myometrial spots), nodules, other abnormal tissue growths, thickening of the junctional zone, and/or peritoneal thickening). Specificity, accuracy, efficacy, and positive clinical utility values were higher for the demographic and clinical parameters plus pelvic examination plus pelvic MRI examinations than for the demographic and clinical parameters plus pelvic examination plus transvaginal ultrasound examinations for biomarker signs (chocolate cysts, myometrial cysts, sub-endometrial microcysts, hypoechoic solid lesions [hyperechoic myometrial spots], nodules, other abnormal tissue growth, thickening of the junctional zone, and/or peritoneal thickening). False-negative results present in demographic and clinical parameters plus pelvic examination and transvaginal ultrasound examinations were reported for uterine fibroids in laparoscopy and histopathologic evaluation. The details of diagnostic parameters of biomarker signs (chocolate cysts, myometrial cysts, sub-endometrial microcysts, hypoechoic solid lesions (hyperechoic myometrial spots), nodules, other abnormal tissue growth, thickening of the junctional zone, and/or peritoneal thickening) per woman at 95% CI are presented in Table 5.

BIOMARKER SIGNS – UTERINE ENLARGEMENT AND ASYMMETRY (ABNORMAL JUNCTION SHAPE): True-positives, sensitivity, specificity, accuracy, efficacy, and positive clinical utility values were lower, and false-positives and false-negatives were higher for the demographic and clinical parameters plus pelvic examination plus transvaginal ultrasound examinations than for the demographic and clinical parameters plus pelvic examination plus pelvic MRI examinations to diagnose biomarker signs (uterine enlargement and asymmetry [abnormal junction shape]). True negatives and negative clinical utility values were the same for the demographic and clinical parameters plus pelvic examination plus transvaginal ultrasound examinations than for the demographic and clinical parameters plus pelvic examination plus pelvic MRI examinations to diagnose biomarker signs (uterine enlargement and asymmetry [abnormal junction shape]). False-negative results present in demographical and clinical parameters, plus pelvic examination and transvaginal ultrasound examinations, were reported as inconclusive at first, and then radiologists declared the absence of pelvic endometriosis. These results were reported as pelvic endometriosis uterine enlargement and asymmetry (abnormal junction shape) on laparoscopy and histopathologic evaluation. The details of diagnostic parameters of biomarker signs (uterine enlargement and asymmetry [abnormal junction shape]) per woman at 95% CI are presented in Table 6.

CLINICAL BENEFITS OF DETECTION OF BIOMARKER SIGNS OF PELVIC ENDOMETRIOSIS: The demographic and clinical parameters plus pelvic examination plus transvaginal ultrasound examinations, the demographic and clinical parameters plus pelvic examination plus pelvic MRI examinations, and the laparoscopy and histopathologic evaluation to detect biomarker signs (chocolate cysts, myometrial cysts, sub-endometrial microcysts, hypoechoic solid lesions (hyperechoic myometrial spots), nodules, other abnormal tissue growths, thickening of the junctional zone, and/or peritoneal thickening) had 0–0.92, 0–0.921, and 0–1 diagnostic confidence, respectively, which had good scores for detecting biomarker signs in the working area (Table 7). The details of the working area for various index diagnostic tests to detect biomarker signs (chocolate cysts, myometrial cysts, sub-endometrial microcysts, hypoechoic solid lesions [hyperechoic myometrial spots], nodules, other abnormal tissue growth, thickening of the junctional zone, and/or peritoneal thickening) are shown in Figure 9.

The demographic and clinical parameters plus pelvic examination plus transvaginal ultrasound examinations, the demographic and clinical parameters plus pelvic examination plus pelvic MRI examinations, and the laparoscopy and histopathologic evaluation to detect biomarker signs (uterine enlargement and asymmetry [abnormal junction shape]) had 0–0.75, 0–0.918, and 0–1 diagnostic confidence, respectively, which had good scores for detecting biomarker signs in the working area (Table 8). The details of the working area for various index diagnostic tests to detect biomarker signs (uterine enlargement and asymmetry (abnormal junction shape)) are shown in Figure 10.

Inter- and intra-observer variabilities were higher for transvaginal ultrasound examinations than for pelvic MRI (data not presented).

Discussion

Diagnostic parameters and clinical threat to missed reporting of pelvic endometriosis values of demographical and clinical parameters plus pelvic examination plus transvaginal ultrasound examinations were inferior to those of demographical and clinical parameters plus pelvic examination plus pelvic MRI examinations in women. The diagnostic parameters and clinical threat to missed reporting of pelvic endometriosis are consistent with those of retrospective longitudinal studies [1], prospective studies [2,6], and reviews and meta-analyses [11]. Transvaginal ultrasound examinations have limited clinical applications compared to pelvic MRI examinations in women for the diagnosis of pelvic endometriosis with clinical manifestations.

Diagnostic parameters and clinical benefits for detection of biomarkers of pelvic endometriosis for chocolate cysts, myometrial cysts, sub-endometrial microcysts, hypoechoic solid lesions (hyperechoic myometrial spots), nodules, other abnormal tissue growth, thickening of the junctional zone, and/or peritoneal thickening biomarker signs in women were identical for transvaginal ultrasound examinations with those of pelvic MRI examinations in women. However, diagnostic parameters and clinical benefits for the detection of biomarkers of pelvic endometriosis for uterine enlargement and asymmetry (abnormal junction shape) were inferior for transvaginal ultrasound examinations compared with those of pelvic MRI examinations in women. The results of the detection of biomarker signs of pelvic endometriosis are consistent with those of retrospective longitudinal studies [1], prospective studies [6], and reviews and meta-analyses [11]. Ultrasound has clinical utility only when there are changes in the chemical composition of an organ. However, physiological changes in the uterus are difficult to detect using ultrasound alone [6,17]. Therefore, it is necessary to diagnose pelvic endometriosis based on biomarker signs. In addition, apart from abnormal junction shape and asymmetry, transvaginal ultrasound examinations have sufficient diagnostic parameters to detect pelvic endometriosis in women with clinical manifestations.

Transvaginal ultrasound examinations reported false-negative values for uterine fibroids and inconclusive results (then declared as negative for pelvic endometriosis). The false-negative results of the transvaginal ultrasound examinations in the current study are consistent with those of a prospective study [6]. Transvaginal ultrasound examinations have poor sensitivity for uterine fibroids and abnormal junction shapes of the uterus [6,11]. Fat, muscle, fascia, and peritoneum are difficult to detect using ultrasound [2]. For negative and inconclusive transvaginal ultrasound results, radiologists and gynecologists should rely on other biomarker signs (indirect criteria, for example, >5 mm heterogeneous myometrial lesion [11]) for the diagnosis of pelvic endometriosis in women who have clinical manifestations similar to direct criteria.

Pelvic MRI examinations have also reported false-negatives for biomarker signs (uterine enlargement and asymmetry [abnormal junction shape]). In these women, the uterine junction zone did not appear properly in the image analysis. In addition, the presence of leiomyomas makes their detection in the uterine junction zone difficult [18]. Similar to transvaginal ultrasound, radiologists and gynecologists should rely on other biomarker signs (indirect criteria, for example, the ratio of the uterine junction zone to the full myometrium thickness, the difference in maximal and minimal uterine thicknesses in both anterior and posterior parts, and the presence or absence of a large smooth regular uterus [11]) for the diagnosis of pelvic endometriosis in women who have clinical manifestations similar to direct criteria.

Both transvaginal ultrasonography and pelvic MRI examinations reported false-positive results. In the current study, abnormal junction shape and thickening of the junctional zone were considered indicative of pelvic endometriosis [10]. However, thickening of the junctional zone and abnormal junction shape are also due to sporadic uterine contractions and/or leiomyomas [11]. Combinations of 1 or 2 biomarker signs are ideal index tests for the diagnosis of pelvic endometriosis.

This study has some limitations, including a small and lack of dynamic study and power calculations for the sample size. In addition, we did not perform a cross-sectional study comparing 2 groups of patients or descriptive studies that measure certain parameters of the individuals of a population at a specific point in time (generally 5 years). During surgery, the American Society for Reproductive Medicine, #Enzian score, and Endometriosis Fertility Index score were used to classify pelvic endometriosis [14] and there is a lack of Chinese surgical and/or pathological guidelines for the diagnosis of pelvic endometriosis.

Conclusions

Transvaginal ultrasound examinations have limited clinical applications compared to pelvic magnetic resonance imaging examinations in women for the diagnosis of pelvic endometriosis with clinical manifestations. Therefore, it is necessary to conclude pelvic endometriosis based on biomarker signs. Except for uterine enlargement and asymmetry, transvaginal ultrasound examinations have sufficient diagnostic parameters to detect pelvic endometriosis in women with clinical manifestations. Combinations of 1 or 2 biomarker signs are ideal for index tests for the diagnosis of pelvic endometriosis in women with positive demographic and clinical parameters and pelvic examination for pelvic endometriosis. Chinese guidelines should consider transvaginal ultrasound for hypoechoic solids and cysts in the diagnosis of pelvic endometriosis.