IRM – avantaje şi limite în diagnosticul şi tratamentul endometriozei

Introduction

Endometriosis is a common multifocal gynecologic disease that manifests during the reproductive years, often causing chronic pelvic pain and infertility⁽¹⁾, de­fined as adenomyosis externa, mostly presents as a single nodule, larger than 1 cm in diameter, in the vesi­co­ute­rine fold or close to the lower 20 cm of the bowel. When diag­nosed, most nodules are no longer progressive. In more than 95% of cases, deep endometriosis is associated with very severe pain. It may occur as invasive peri­to­neal fibrotic nodules and adhesions or as ovarian cysts with hemorrhagic content. Although the findings at physical exa­mi­nation may be suggestive, imaging is necessary for the de­fi­nitive diagnosis, patient counseling and treatment plan­ning. The imaging techniques that are most useful for preoperative disease mapping are transvaginal ultra­so­nography (US) after bowel preparation and magnetic resonance imaging (MRI).

MRI is indicated as a complementary examination in complex cases of endometriosis with extensive adhesions and ureteral involvement⁽²⁾.

Endometriosis may also manifest as multiple, homo­geneously hyperintense cysts on T1-weighted images. The involvement of the alimentary tract or bladder can appear as areas of high signal intensity. Although mag­netic resonance imaging is limited in its ability to de­pict small endometrial implants and adhesions, the advantages of MRI over laparoscopy include the ability to characterize endometriotic lesions and to evaluate extraperitoneal sites of involvement, contents of a pelvic mass, or lesions hidden by dense adhesions. The roles of the two modalities are therefore complementary. The know­ledge of the variety of magnetic resonance imaging ap­pearances of endometriosis and of the organ involvement with­in the pelvis is important for guiding a subsequent la­pa­roscopic examination⁽³⁾.

Deep pelvic endometriosis is defined as subperitoneal infiltration of endometrial implants in the uterosacral ligaments, rectum, rectovaginal septum, vagina or bladder⁽¹⁾.

Materials and method

We conducted a retrospective study between 2017 and 2021 on a group of 99 patients, aged between 25 and 50 years old, operated in the “Prof. Dr. Panait Sîrbu” Clinical Hospital of Obstetrics and Gynecology, Euroclinic Hospital – Private Health Network, and Monza Hospital, Bucharest.

Exclusion criteria:

• patients who did not perform the MRI investigation;

• patients in whom intraoperative data were missing.

Statistical analysis and results

The statistical analysis was performed using IBM SPSS Statistics 25 and Microsoft Office Excel/Word 2013. The quantitative variables were tested for distribution using the Shapiro-Wilk test and were expressed as means with standard deviations or medians with interpercentile intervals. The categorical variables were expressed in absolute or percentage form and were tested using Fisher’s Exact Test.

The existing correlations were made using the Pearson correlation coefficient, respectively Spearman’s rho, depending on the distribution of the quantitative variables.

The data in Table 1 represent the characteristics of the studied group. The following are observed:

• The average age was 32.66 ± 5.52 years old, with a median of 33 years old.

• The most common age category was 30-39 years old (65.3%).

• 2020 (38.4%) and 2019 (37.4%) were the years in which most patients were included.

• Most patients came from Monza (58.6%) or Giuleşti (32.3%).

• Most patients had changes observed at the clinical examination (83.8%).

• The mean value of preoperative AMH was 1.803 ± 2.833 ng/mL, with a median of 0.665 ng/mL.

• The average total AFS-R score was 3.46 ± 0.747 points, with a median of 4 points.

• Most patients had no complications (99%); only one patient had postoperative complications (postoperative fever).

The data in Table 2 represent the description of the pathologies observed on MRI. The following are ob­served:

• 29.6% of the patients had parametric lesions, more frequently bilateral (11.2%).

• 27.1% of the patients had lesions of the rectovaginal septum.

• 44.9% of the patients had uterosacral ligament lesions, more frequently bilateral (27.6%).

• 29.6% of the patients had rectal nodules, more frequently unique (28.6%).

• 13.4% of the patients had sigmoid nodules, more frequently unique (11.3%).

• 2% of the patients had ileal nodules.

• 14.3% of the patients had bladder invasion.

• 67.3% of the patients had other locations of endometriosis.

• The average size of the rectal nodules was 21.86 ± 8.883 mm, with a median of 20 mm.

• The average size of the sigmoid nodules was 27.23 ± 8,974 mm, with a median of 27 mm.

• The mean distance of intestinal lesions from the external anal sphincter was 123.4 ± 17.75 mm, with a median of 120 mm.

The data in Table 3 represent the description of the pathologies observed intraoperatively. The following are observed:

• Most patients had a normal appearance of the uterus (65.7%), 15.2% of the patients had fibroids and 19.2% of the patients had adenomyosis.

• 58.6% of the patients had anterior sac lesions, more frequently with a black appearance (50.5%), with an average size of 25.18 ± 6.746 mm.

• 41.4% of the patients had right parametric appearance, 37.4%  with black appearance, and 51.5% of the patients had left parametric lesions, more frequently class B2 (28.3%) and with black appearance (44.4%); 24.2% of the pa­tients had straight parametric lesions, more frequently class B2.

• 21.2% of the patients had lesions of the rectovaginal sep­tum, more frequently class A2 (11.1%).

• 24.2% of the patients had right USL lesions, more frequently with a black appearance (18.2%) and with an average size of 14.32 ± 6.611 mm, and 26.3% of the patients had left USL lesions, more frequently with a black appearance (20.2%) and with an average size of 14.07 ± 7.532 mm.

• 50.5% of the patients had intestinal lesions, more frequently class C3 (22.2%), 36.4% had rectal lesions, more frequently single (33.3%), with an average size of 22.61 ± 8.771 mm, 19.2% had sigmoid lesions, more frequently single (15.2%), with an average size of 25.11 ± 8.704 mm, and 3% of the patients had ileal lesions, with an average size of 16.67 ± 5.774 mm;

• 21.2% of the patients had bladder invasion and 1% of the patients had diaphragmatic or appendicular invasion.

The data from Table 4 and Figures 1-3 represent the comparison between MRI and intraoperative endometriosis diagnoses. The results show that:

The Associations of MRI diagnoses versus intra­ope­ra­tively are statistically significant, according to Fisher tests (p<0.05), noting that:

• In the case of parametric lesions, the specificity was 87.1% and the sensitivity was 37.3%.

• In the case of USL lesions, the specificity was 70.5% and the sensitivity was 70.3%.

• In the case of rectovaginal septal lesions, the specificity was 81.8% and the sensitivity was 63.2%.

• In the case of rectal nodules, the specificity was 96.8% and the sensitivity was 77.8%.

• In the case of sigmoid nodules, the specificity was 96.2% and the sensitivity was 55.6%.

• In the case of bladder invasion, the specificity was 100% and the sensitivity was 66.7%.

The dimensions of the rectal nodules had a non­pa­rametric distribution (p<0.05) and the dimensions of the sigmoid nodules had a normal distribution (p>0.05) according to the Shapiro-Wilk test.

The correlation between the dimensions of the observed rectal nodules MRI versus intraoperatively is significant and of high degree (p=0.001, R=0.599).

The correlation between the dimensions of the sig­moid nodules observed by MRI versus intraoperatively is significant and of very high degree (p=0.005, R=0.809), the similarity between dimensions being much higher.

Discussion

An accurate preoperative assessment of disease extension is required for planning the complete surgical ex­ci­sion, but such assessment is difficult with physical exa­mi­nation. Various sonographic approaches (trans­va­gi­nal, transrectal, endoscopic transrectal) have been used for this purpose, but they do not allow a panoramic eva­lua­tion. Furthermore, exploratory laparoscopy has li­mi­ta­tions in demonstrating deep endometriotic lesions hid­den by adhesions or located in the subperitoneal space.

Solid deep lesions have low to intermediate signal intensity with punctate regions of high signal intensity on T1-weighted images, show uniform low signal in­ten­sity on T2-weighted images, and can demonstrate en­hance­ment on contrast-enhanced images. Magnetic re­sonance imaging is a useful tool. The MR imaging features depend on the type of lesions: infiltrating small implants, solid deep lesions mainly located in the posterior cul-de-sac and involving the uterosacral ligaments and torus uterinus, or visceral endometriosis involving the bladder and rectal wall⁽⁸⁾, being an additional method to clinical examination and transvaginal or transrectal sonography in the evaluation of patients with deep infiltrating endometriosis⁽⁹⁾.

Conclusions

The vast majority of patients had changes observed at the clinical examination. There is a correlation between the dimensions of the rectosigmoid nodules observed on MRI versus intraoperatively, this association being highly statistically significant.

MRI had a clearly superior specificity regarding parametric invasion, rectovaginal septum, sigmoid invasion and of the rectum.

If the ultrasound is useful in suspected cases of en­do­metriosis, MRI plays an essential role in guiding the diagnosis of endometriosis⁽⁴⁾. These two provide dif­ferent and complementary information. Performing these investigations should be discussed depending on the type of endometriosis suspected, the proposed treat­ment strategy and the information provided to the patient.

MRI, as a noninvasive diagnostic tool, offers essential advantages regarding the classification and therapy planning for patients with DIE⁽⁵⁾.

None of the evaluated imaging modalities were able to detect overall pelvic endometriosis with enough accuracy that they would suggest to replace surgery. The laparoscopy remains the gold standard for the diagnosis and treatment of endometriosis along with histopathological exam^(2,6).

MRI features have a potential diagnostic utility in iden­ti­fying the need for colorectal surgery in patients with DIE⁽⁷⁾. Despite some limitations, magnetic re­so­nance imaging is able to directly demonstrate deep pelvic en­dometriosis.   

Conflict of interests: The authors declare no con­flict of interests.