Comparison of diagnostic values of transvaginal sonography with laparoscopic and histological results in the evaluation of uterosacral ligaments' involvement in endometriosis patients

Study setting: This is a cross-sectional study performed on patients with suspected clinical symptoms of endometriosis (chronic pelvic pain, dysmenorrhea, or dyspareunia) at Rasoul Akram Hospital in Tehran during 2019–2020. It was approved by the Ethics Committee of Iran University of Medical Sciences (IR.IUMS.FMD.REC1398.383). A total of 80 patients were included in this study. Inclusion criteria included all patients with high clinical suspicion of endometriosis and aged 18–49 years, and exclusion criteria included virginity or any conditions in which the patient cannot undergo transvaginal ultrasound or laparoscopic surgery. Patients were evaluated for dysmenorrhea, pelvic pain, and dyspareunia in terms of intensity (scoring from 0 to 10 based on the visual analog scale) and duration (months or years) of pain. All patients were confirmed by their physicians before participating in the study. Also, all patients' clinical information was collected from their files based on the relevant checklist that had been prepared in advance. Sampling method in this study is random sampling and the sample size was calculated based on the formula below 80 people. n = Z 1 - α 2 + Z 1 - β 2 ( α 1 2 + α 2 2 ) d 2 (In this formula, the probability of the first type error is 0.05 and the probability of the second type error is 0.2(. Transvaginal sonography evaluation: An expert radiologist in endometriosis diagnosis, examined the patients with vaginal ultrasound in OB presentation with a GE VOLOSON sonography device with a transvaginal probe. Ultrasound findings include nodular aggregations or the presence of irregular hypoechoic nodules, thickening of the wall, or retractable masses and hypoechoic points, which are reported mainly in the animals, left–right USLs, rectosigmoid region, bladder, and ureter. Laparoscopic evaluation and histology: The right and left USLs were examined laparoscopically, resection of uterosacral endometriosis was performed, the samples were transferred to the pathology ward, then laparoscopic findings were compared with the results of pathology report. In the absence of visible lesions, a biopsy of the USL was performed. Uterosacral samples were then sent to pathology for histological confirmation. The results of sonography and pathology were compared and analyzed. Statistical analysis: To describe the data, mean, standard deviation, median, amplitude, frequency, and percentage were used. To compare the results of the two methods, chi-square statistical test along with sensitivity, specificity, and positive and negative predictive values were utilized and analyzed using the SPSS 23 statistical software.

Patient data: This study was performed on 80 patients with mean age 34.47±5.94 (mean±SD) years, with the lowest and highest ages being 20 and 50 years, respectively. Mean height, weight, and body mass index (BMI) were 164±5.94 cm, 65.06±1.08 kg, and 24.09±3.91 kg/m 2 (mean±SD), respectively. Of the total number of patients, 54 (67.5%) were married, and 26 (32.5%) were single (none of the patients were virgins). A total of 43 (53.8%) patients were nulliparous, and 37 (46.2%) patients were multiparous. A total of 10 patients from the nulliparous group and 11 patients from the multiparous group had a history of infertility ( table 1 ). A total of 3 (3.8%) patients had no dysmenorrhea, and another 77 (96.2%) subjects complained of menstrual pain. A total of 40 (50%) patients had pelvic pain, whereas an equal number of patients (50%) did not have it. Dyspareunia was found in 45 (56.2% patients), and 35 (43.8%) patients did not have it. The mean severity of menstrual pain, pelvic pain, and dyspareunia was 7.36, 2.67, and 3.76, respectively. Also, the mean duration of dysmenorrhea, pelvic pain, and dyspareunia was evaluated to be 5.03, 1.02, and 1.8 years, respectively. The study found that 37 (46.2%) patients suffered from pain during defecation, 8 (10%) patients complained of dysuria, and 24 (30%) patients complained of pain due to probe pressure during ultrasound. Ultrasound and laparoscopic findings, including increased thickness of right and left USLs, the presence of nodules in these ligaments, the presence of ovarian endometrioma, and involvement of the rectum, bladder, and ureter were examined. Regarding the increase in thickness of USLs, as detected in ultrasound, in 60 (75%) patients, it was found in both right and left, in 1 (1.2%) patient, only in the left, and 19 (23.8%) patients did not have this increase in thickness. In contrast, as detected in laparoscopy, the increase in thickness of USLs was found in both the right and the left ligaments in 46 patients (57.5%), 9 (11.2%) patients had this only in the right side, 2 (2.5%) patients had this only in the left side, and it was not found at all 23 (28.8%) patients. Compared to the standard method (laparoscopy histopathology), ultrasound as the gold standard method was found to have sensitivity, specificity, and positive and negative predictive values of 93%, 65%, and 87% and 78.9%, respectively, in detecting increase in USL thickness ( table 2 ) ( figure1 ). Demographic data (age, height, weight, parity, and fertility) Increased uterosacral ligament thickness on ultrasound and laparoscopy Thickness laparo is a more reliable diagnostic criterion in determining thickness due to p- significance <0.05 compared to ultrasound method and is also a weak biomarker due to the curved surface area of 66% Transvaginal sonography findings: Examination of nodules in the right and left USLs on ultrasound showed that 27 (33.8%) patients had nodules in both right and left, 23 (28.8%) had them only in the right, 15 (18.8%) only in the left, and another 15 (18.8%) did not have them. On laparoscopy, nodules were found in both the right and the left in 30 (37.5%) patients, only 26 (32.5%) patients had them only in the right side, 23 (28.8%) patients only in the left side, and no nodules were found in 1 (1.2%) patient. Compared to the standard method (laparoscopy), ultrasound has sensitivity, specificity, and positive and negative predictive values of 82%, 100%, and 100% and 6.66%, respectively, in the diagnosis of nodules in the USL ( table 3 ) ( figure2 ). Endometrioma in ovaries was reported on ultrasound in 23 (28.7%) patients in both sides, 27 (33.8%) patients had them only in the right, 17 (21.2%) patients only in the left, and 13 (16.2%) did not have them; laparoscopy reported endometrioma in ovaries in 26 (32.5%) patients in both sides, 24 (30%) patients had them only in the right side, 17 (21.2%) patients only in the left side, and 13 cases (16.2%) did not have endometrioma. Compared to laparoscopy, ultrasound as the gold standard method has sensitivity, specificity, and positive and negative predictive values of all 100% in the diagnosis of endometrioma ( table 4 ) ( figure 3 ). With the involvement of the rectum, bladder, and ureter, on ultrasound, rectum involvement was reported in 31 (38.8%) patients, bladder involvement in 3 (3.8%) patients, simultaneous rectum and bladder involvement in 3 (3.8%) patients, simultaneous rectum and ureter involvement in 1 (1.2%) patient, and no involvement was found in 42 (52.5%) patients. Laparoscopic findings: On laparoscopy, rectal involvement was found in 27 (33.8%) patients, bladder involvement in 5 (6.2%) patients, both rectal and bladder involvement was found in 16 (20%) patients, both rectal and ureteral involvement in 3 (3.8%) patients, all rectal, bladder and ureter involvement in 1 (1.2%) patient, and 28 (35%) patients were negative for this variable, that is, no involvement found. Compared to laparoscopy, ultrasound as the gold standard method has sensitivity, specificity, and positive and negative predictive values of 71%, 96.4%, and 97.3%, and 64.2%, respectively, in the diagnosis of visceral involvement (rectum, urinary bladder, and ureter). Nodules in uterosacral ligament on ultrasound and laparoscopy Nodule Lapar and nodule Sono are excellent biomarkers for detecting the number of nodules in patients with 93 and 72% below the rock curve, and laparoscopy is a more accurate diagnostic criterion. Both criteria have a p-value <0.001 and are significant None of them are significant and are not a good diagnostic criterion for endometriosis Endometrioma in ovaries via ultrasound and laparoscopy

In our study, the diagnostic sensitivity, specificity, positive and negative predictive values of TVS were compared with laparoscopy (confirmed with pathology). In the diagnosis of increase in USL thickness, the diagnostic sensitivity, specificity, positive and negative predictive values were 93%, 65%, and 87%, and 78.9%, respectively. Zhou et al. declared that the sensitivity, specificity, positive and negative predictive values for USL sonography diagnosis were 65%, 92%, 7.8% and 0.38%, respectively ( 22 ). They also reported that TVS provides an excellent comprehensive diagnostic performance for DIE which is similar to our result. The TVS and laparoscopic data by Pattanasri et al. in 2020 revealed that the specificity and sensitivity in the diagnosis of deep infiltrative endometriosis (DIE) were 46.1% and 83.3%. Also, the sensitivity of TVS in diagnosis of uterosacral ligament DIE was 70.9% ( 23 ). Similar to our study, Alborzi et al. achieved sensitivity, specificity, and positive and negative predictive values of 83.3%, 46.1%, and 85.7% and 41.6%, respectively, for DIE ( 24 ). The diagnostic sensitivity, specificity, positive and negative predictive values of uterosacral endometriosis in our study were 82%, 100%, 100% and 6.66%, respectively. Hudelist et al. have stated that TVS has a sensitivity of 40% and a specificity of 95.6% in the diagnosis of uterosacral endometriosis ( 25 ). Probable reasons for finding lower sensitivity compared to our results may be the lower skill set of the operator, use of transrectal sonography in some patients, and less severity of lesions. Zhang et al. found sensitivity, specificity, and positive and negative predictive values of 90.9%, 96.4% and 88.2% and 94%, respectively, of TVS in the diagnosis of USL endometriosis ( 19 ). Nisenblat et al. in 2016 concluded that the laparoscopic diagnostic of pelvic endometriosis as a golden standard is more accurate than TVS, clinical data and serum biomarkers alone ( 26 ). However, Ghatresamani et al. evaluated transvaginal–transrectal sonography in DIE, which showed a diagnostic sensitivity of 100% in pelvic masses, and sensitivity and specificity of 50% and 100%, respectively, in the diagnosis of bladder involvement ( 20 ). Also, Holland et al. reported that TVS has a high specificity in the diagnosis of endometriosis and false-positive results are rare ( 21 ). The lower sensitivity may be due to the lower skill set of the operator, smaller size, location and number of lesions, the use of a lower model of ultrasonography or not using combination of diagnostic techniques with each other. Tammaa et al. in 2015 discovered that using a trained specialist dramatically increases the accuracy and reproducibility of TVS for diagnosing DIE ( 27 ). Saba et al. in 2012 indicated that combination of TVS with MRI increase the accuracy of rectosigmoid endometriosis diagnosis up to 95% ( 28 ). All results in these reports in accordance to the findings in our study that the diagnostic accuracy of sonography depends on the location and number of lesions. Our data revealed that the diagnosis value of all four parameters for ovarian endometrioma were 100%. Vesical endometriosis was examined by TVS in a study by Savelli et al., where they reported sensitivity, specificity, and positive and negative predictive values of 44%, 100%, and 100% and 95%, respectively ( 29 ). Again, the lower sensitivity may be a result of the lower skill set of the operator, or smaller and less severe lesions. Yazbeck et al. reported sensitivity and specificity of 44% and 98% of TVS in the evaluation of severe pelvic adhesions ( 30 ). However, Ayachi et al., in 2017 estimate the sensitivity (96.3%) and specificity (92.6%,) of transvaginal sonographic (TVS) in women with previous abdominopelvic surgery ( 31 ). Their results also showed that there is a close relationship between adhesion and TVS sliding sign. The ultrasound method in the diagnosis of visceral involvement has sensitivity, specificity, positive and negative predictive values of 71%, 96.4%, 97.3%, and 64.2%, respectively. In a study by Saccardi et al., sensitivity and specificity of TVS in USL assessment were found to be 88.9% and 95.6%, respectively, which is similar to the findings in our study ( 32 ). Abrao et al. reported the sensitivity and specificity of TVS in the diagnosis of retro-cervical endometriosis 95% and 98%, respectively. They reported the positive predictive value at 98% ( 33 ). They also found the sensitivity, specificity, and positive predictive value of TVS in the evaluation of rectosigmoid involvement at 98%, 100%, and 100%, respectively. These higher values compared to our study may be due to the better skill set of the operator, more severity of involvement, etc. Goncalves et al. reported 81% sensitivity and 99% specificity of TVS in the evaluation of rectosigmoid endometriosis ( 34 ). On the other hand, Ferrero et al. reported that the accuracy of TVS in diagnosing the presence of rectosigmoid endometriosis was 92.3% ( 35 ). In conclusion we conclude that TVS has an acceptable sensitivity in the diagnosis of uterosacral involvement: 93% in evaluating thickening and 82.5% i n the evaluation of hypo-echo nodules; however, specificity in these two cases was 65% and 100%, respectively. Despite acceptable sensitivity and specificity in the diagnosis of adnexal endometriomas, and urinary bladder and rectal involvement, involvement cannot be ruled out in the presence of negative results with a sensitivity of 71%.