Clinical and ultrasound characteristics of deep endometriosis affecting sacral plexus

To describe the sonomorphological changes and appearance of deep endometriosis (DE) affecting the nervous tissue of the sacral plexus (SP).


What are the novel findings of this work?
Deep endometriosis affecting the sacral plexus exhibits a distinct sonographic appearance on transvaginal ultrasound, in which lesions appear as unilateral, solid, non-uniform, hypoechogenic nodules containing hyperechogenic areas, with internal shadows, irregular spiculated contours, poor vascularization on color/ power Doppler examination and a diameter of 2 cm or more.

What are the clinical implications of this work?
Transvaginal sonography is capable of diagnosing deep endometriosis affecting the sacral plexus, which may have implications for the planning of treatment strategies.

Methods This was a retrospective study of symptomatic patients who underwent radical resection of histologically confirmed DE affecting the SP and who had undergone preoperative transvaginal sonography (TVS) between 2019 and 2023. Lesions were described based on the terms and definitions of the International Deep Endometriosis Analysis (IDEA), International Ovarian Tumor Analysis (IOTA) and Morphological Uterus Sonographic Assessment (MUSA) groups. A diagnosis
Correspondence: Dr G. Szab ó, Department of Obstetrics and Gynaecology, Faculty of Medicine, Semmelweis University, Budapest, Hungary (e-mail: szabo.gabor6@med.semmelweis-univ.hu)Accepted: 24 January 2024 of DE affecting the SP on TVS was made when the sonographic criteria of DE were visualized in conjunction with fibers of the SP and the presence of related symptoms corresponding to sacral radiculopathy.Clinical symptoms, ultrasound features and histological confirmation were analyzed for each patient included.

Results
Twenty-seven patients with DE infiltrating the SP were identified in two contributing tertiary referral centers.Median age was 37 (range, 29-45) years and all patients were symptomatic and presented one or more of the following neurological symptoms: dysesthesia in the ipsilateral lower extremity (n = 17); paresthesia in the ipsilateral lower extremity (n = 10); chronic pelvic pain radiating in the ipsilateral lower extremity (n = 9); chronic pain radiating in the pudendal region (n = 8); and motor weakness in the ipsilateral lower extremities (n = 3).All DE lesions affecting the SP were purely solid tumors in the posterior parametrium in direct contact with, or infiltrating, the S1, S2, S3 and/or S4 roots of the SP.The median of the largest diameter recorded for each of the DE nodules was 35 (range, 18-50) mm.Echogenicity was non-uniform in 23 (85%) of the DE nodules, with all but one of these nodules containing hyperechogenic areas.The shape of the lesions was irregular in 24 (89%) cases.Only one lesion exhibited a lobulated form, with all other irregular lesions showing a spiculated appearance.An acoustic shadow was produced in 20 (74%) of the nodules, all of which were internal.On color or power Doppler examination, 21 (78%) of the nodules showed no signal (color score of 1

INTRODUCTION
Deep endometriosis (DE) infiltrating the sacral plexus (SP) and SP roots is considered to be a severe form of DE in the posterior pelvic compartment 1,2 .Perineural spread from the parametrium to the SP is a possible pathological pathway.However, the progression of the disease is largely unknown and this type of endometriosis is still considered rare 3 , with the prevalence of DE infiltrating the SP not yet reported 4 .
The histological diagnosis of endometriosis is based on the presence of endometrial epithelium and/or stroma outside the uterine cavity, surrounded by fibrosis or muscular hyperplasia 5 .Endometriotic stroma contains a network of small vessels.Multiple layers of connective tissue surround the peripheral nerve fibers, with all the fascicles bound together by a thick layer called the epineurium 6 .This sheath of mesodermal origin contains fibroblasts, elastic fibers, Type-I and Type-III collagen and some adipocytes (Figure 1).
During laparoscopic surgical dissection of the SP (neurolysis), the roots of the SP can be visualized only by opening the subperitoneal spaces and the fascia covering the SP 1 .The sacral nerve roots S1-S4 can be perceived as solitary solid white bands on the posterolateral wall of the pelvis.Large DE nodules in the posterior pelvic compartment can infiltrate the parametrium and penetrate the retroperitoneal space, compressing and infiltrating the inferior hypogastric plexus and S2-S4 sacral nerve roots in the form of infracardinal lesions 3 .These lesions may involve other pelvic visceral organs (e.g.rectum, ureter, vagina).DE nodules adhering to the pelvic side wall can infiltrate the roots of the sciatic nerve or the lumbosacral trunk and are referred to as supracardinal DE lesions 3,7 .These lesions cause pain and sensory dysfunction, such as dermatome-associated dysesthesia and paresthesia of the lower extremities, and may also cause irritation of autonomous and pudendal nerve fibers arising from the S2-S4 sacral nerve roots (Figure 2).
Until now, diagnosis of DE in this location was based on clinical signs and the results of magnetic  resonance imaging (MRI) 8 .Recently, our group presented a technique to visualize the SP and its roots with transvaginal ultrasound (TVS) 9 .To the best of our

METHODS
This was a retrospective descriptive study using data obtained from two tertiary referral centers for endometriosis (Department of Obstetrics and Gynecology, Semmelweis University, Budapest, Hungary and Department of Gynecology, Center for Endometriosis, Hospital St John of God, Vienna, Austria).From these databases, we identified 27 consecutive patients with a histological diagnosis of DE infiltrating the SP following surgery for symptomatic disease, who had undergone preoperative ultrasound examination by an experienced ultrasound examiner between 2019 and 2023, of whom 19 were from Budapest and eight were from Vienna.All women had been examined by TVS according to the systematic approach of the International Deep Endometriosis Analysis (IDEA) group 10 , using high-end ultrasound equipment, namely a 2-10-MHz transvaginal probe with three-dimensional option (Hera W9 and Hera W10, Samsung, Seoul, South Korea; Voluson E10, GE Healthcare, Zipf, Austria).
Each author characterized the lesions from their own center based on patient records, ultrasound images, videoclips and ultrasound reports, according to the sonographic appearance of the lesion using the terminology of the IDEA, International Ovarian Tumor Analysis (IOTA) and Morphological Uterus Sonographic Assessment (MUSA) groups on a predefined ultrasound assessment form [10][11][12] .
Data were recorded using a predefined protocol, including clinical information, sonographic characteristics, descriptions of the location of the lesion and its relation to surrounding nerves (either extrinsic compression or intrinsic infiltration) based on intraoperative findings, and final histology.DE affecting the SP was diagnosed when features of DE were visualized in the proximity of, and/or infiltrating, the sonographic structure of the sacral root S1, S2 or S3, as reported previously 5 , in the presence of symptoms corresponding to sacral radiculopathy.Typical symptoms of sacral radiculopathy are chronic pelvic pain or dysesthesia/paresthesia radiating to the perineum (S2-S4) and the ipsilateral lower limbs and corresponding dermatome, as well as weakness in the ipsilateral lower extremities (L5-S1), depending on the actual sacral root affected by DE.
DE not only causes direct damage of the organs to a depth of at least 5 mm, but also causes secondary inflammation, reactive fibrosis and scarring.Nodules that reached the hyperechogenic outer edge of the nerve, the epineurium, and exerted traction on it were called extrinsic type.Lesions that extended through the epineurium into the fasciculi of the nerve were called intrinsic type.
IOTA terms and definitions were developed originally for the description of adnexal tumors.For the purposes of this study, we used the IOTA terms to describe the type (solid or cystic) and color score of the lesion 11 .The intralesional vascular pattern was described using a semiquantitative color score of 1-4.A color score of 1 means that no color or power Doppler signals were detected in the tumor, a score of 2 means that a minimal amount of color Doppler signal (i.e.flow) was detected, a score of 3 means that a moderate amount was detected and a score of 4 means that an abundant amount of signal was detected.Color or power Doppler ultrasound was used to visualize the vascular architecture of the DE lesion, and low pulse-repetition frequency (0.3-0.6 kHz) and low wall filter (50-100 Hz) were used to enable the detection of as many blood vessels as possible.The gain was increased to the point of significant color noise and then slowly reduced until all artifacts disappeared.
MUSA terminology was developed originally for describing myometrial lesions, including uterine smooth muscle tumors.DE is primarily a solid nodule and, therefore, MUSA terminology offered specific definitions that were used to describe the echogenicity of the solid DE lesions: uniform (hypoechogenic, isoechogenic or hyperechogenic); non-uniform with mixed echogenicity, hyperechogenic areas or cystic areas; and shadowing (edge, internal or fan-shaped) 12 .
DE nodule shape was defined as regular (oval or round) or irregular (lobulated or spiculated).Oval lesions have a dominant long axis, while round nodules have long and short axes that are similar in length.Irregular shape means that the outer contour is lobulated or spiculated.DE lesions with a lobulated shape have a smooth, wavy external contour and intact capsule.DE nodules with a spiculated contour have spiky, blurred margins 13 .
According to the dynamic ultrasonography steps of the IDEA approach, we evaluated soft markers of DE, such as site-specific tenderness and ipsilateral ovarian mobility 10 .By applying pressure between the uterus and ovary, we assessed if the ovary was fixed to the uterus medially, to the pelvic side wall laterally or to the uterosacral ligaments.To assess the status of the pouch of Douglas (POD), we used the real-time TVS-based 'sliding sign' for all patients.If the uterus was anteverted, gentle pressure was applied to the cervix using the transvaginal probe to assess whether the anterior rectum glided freely across the posterior surface of the cervix and posterior vaginal wall.If the anterior rectal wall did so, the sliding sign was considered to be positive.Then, one hand was placed over the patient's lower anterior abdominal wall in order to ballot the uterus between the palpating hand and the transvaginal probe (which was held in the other hand) to assess whether the anterior rectum glided freely over the posterior surface of the uterine fundus.If it did so, the sliding sign was also considered positive in this region.When the sliding sign was found to be positive in both of these anatomical regions (retrocervix and posterior uterine fundus), the POD was recorded as being not obliterated 10 .
Other ultrasound techniques mentioned in the IDEA approach (e.g.transvaginal strain sonoelastography), if available, were also reported 10,14,15 .Using transvaginal strain sonoelastography, the stiffness of the nodules was compared with the stiffness of the intact nerve.A strain ratio of 2.0 served as a cut-off value for the optimal distinction.
In most cases, an experienced sonographer can confidently and correctly evaluate lesions and make a differential diagnosis based on subjective evaluation of the grayscale ultrasound image with or without the added information of the color Doppler ultrasound examination results 16 .This process is called pattern recognition.In addition to the predefined protocol, the principal investigators (G.S. and G.H.) and two ultrasound examiners (I.M. and V.F.) reviewed all available digital ultrasound images and videoclips to identify possible ultrasound patterns characteristic of DE infiltrating the SP.The description agreed by the four observers based on pattern recognition was reported.
The study data were evaluated by descriptive statistical methods.Continuous data are given as mean ± SD if normally distributed and median (range) if non-normally distributed.Categorical data are given as n (%).Statistical analysis was undertaken using STATISTICA software package (version 13.5.0.17;TIBCO Software Inc., Palo Alto, CA, USA).

RESULTS
Twenty-seven patients with DE affecting the SP were identified and verified histologically following surgical resection of the DE.Their median age was 37 (range, 29-45) years.Clinical characteristics of the patients are presented in Table 1.No known comorbidities were recorded.All patients underwent laparoscopic sacral neurolysis and resection of DE nodules.The median time between the onset of symptoms and surgery was 11 (range, 4-16) months.All patients exhibited unilateral DE with extrinsic or intrinsic involvement of the SP.Sixteen lesions were on the right side and 11 lesions were on the left side.All the patients were symptomatic and presented with one or more of the following neurological symptoms: dysesthesia in the ipsilateral lower extremity (n = 17); paresthesia in the ipsilateral lower extremity (n = 10); chronic pelvic pain radiating in the ipsilateral lower extremity (n = 9); chronic pain radiating in the pudendal region (n = 8); and motor weakness in the ipsilateral lower extremities (n = 3).The sonographic characteristics of DE affecting the SP are presented in Table 2, which shows that all DE lesions affecting the SP were purely solid tumors (n = 27).The median of the largest diameter recorded for each DE nodule was 35 (range, 18-50) mm.The echogenicity was non-uniform in 23 (85%) of the DE nodules; of those, all but one contained hyperechogenic areas.The shape of the lesions was classified as irregular in 24 (89%); of those, only one lesion was lobulated and the rest were spiculated (Figure 3).In 20 (74%) of the nodules, an acoustic shadow was produced, all of which were internal (Figure 4).On color or power Doppler examination, 21 (78%) of the    nodules showed no signal (Figure 5).The remaining six (22%) lesions manifested minimal color content (color score of 2).All 17 DE lesions for which information on strain elastography was available demonstrated hard stiffness compared with their surroundings (Figure 6).
All DE nodules were located in the posterior parametrium in direct contact with the roots of the SP.In eight (30%) patients, the nodule was intrinsic, i.e. invasion and infiltration were diagnosed during surgery.The other lesions (n = 19) were extrinsic, i.e.DE affected the perineural tissue and appeared adherent to the neural structures, causing symptoms by traction and local irritation (Figure 7).The four observers evaluating the grayscale and color/power Doppler ultrasound images agreed in their description of DE affecting the SP and reached the following consensus to describe the 27 cases analyzed in this study: the most characteristic ultrasound appearance of a DE lesion affecting the SP in our series, using pattern recognition, was a purely solid, non-uniform, hypoechogenic nodule containing hyperechogenic areas, with internal shadows and irregular spiculated contours, that was poorly vascularized on color/power Doppler examination and in direct contact with the hyperechogenic epineurium of the nerves.

DISCUSSION
To date, the identification of DE nodules involving the SP on TVS has not been reported, making our study the first of its kind.The results of the present study suggest that diagnosis of DE in this rare location is feasible preoperatively using clinical signs and TVS.However, we support the additional use of MRI in women for whom surgery is considered and planned.
Ultrasonographic features of endometriosis vary.Some DE nodules have organ-specific, distinct sonographic signs, for example, the 'comet sign' in the bowel 17,18 .According to the literature summarized in the IDEA consensus statement, DE nodules tend to appear sonographically as solid, hypoechogenic to anechoic and irregular-shaped masses 10 .They may contain echogenic foci or small cystic spaces, which correspond to collections of detritus or blood.These lesions show mainly little or no blood flow on color Doppler and demonstrate high stiffness on elastography 14,15 .In a recent case report, extrapelvic DE affecting the sciatic nerve was visualized  by ultrasonography and described as a large, perineural, hypoechogenic, inhomogeneous lesion with an irregular contour engulfing the nerve 19 .The results of the present study support the sonographic characteristics described previously 19 .
An interesting feature of the ultrasonographic appearance of DE nodules affecting the SP is that most of these lesions produced an acoustic shadow and a certain number of hyperechogenic foci.These features may be due to secondary scarring of the DE nodule or the presence of an epineurium of infiltrated nerves 3 .Only in exceptional cases did cystic areas appear in these lesions.
A major difference between DE lesions and peripheral nerve sheath tumors (PNSTs) is that PNSTs originate from the nerve sheath cells.In contrast, DE nodules either adhere to the outside of the epineurium or wrap around it and may partly infiltrate the nerve.Therefore, DE nodules do not have a hyperechogenic rim.Benign PNSTs in the pelvis, such as schwannomas and neurofibromas, can be also identified with TVS 20 .However, these benign PNSTs are usually asymptomatic.The typical ultrasound appearance of a benign PNST is a solid, fusiform, hypoechogenic, non-uniform and poorly or moderately vascular tumor without acoustic shadowing.All benign PNSTs are well circumscribed by a hyperechogenic rim formed by the epineurium.Schwannomas are more inhomogeneous and show moderate vascularization.Ancient schwannomas may also show cystic areas.Neurofibromas are well-defined, homogeneously hypoechogenic and poorly vascularized tumors within hyperechogenic areas, in which nerve fibers can be seen entering and exiting the mass 21 .They demonstrate low-to-moderate stiffness on elastography (Figure 8).Malignant PNSTs are extremely rare and typically cause neurological symptoms.Their irregular spiculated borders, inhomogeneous irregular internal echogenicity and hyperechogenic rim due to the surrounding soft-tissue reaction may present a challenge for differential diagnosis.However, as they are richly vascularized 21 , color/power Doppler could be useful in the differential diagnosis between DE infiltrating the SP and malignant PNSTs.Other lesions of the SP, such as anterior sacral meningocele, present as a multilocular, thin-walled cyst on TVS 22 .
All the patients in our study were of reproductive age.As only one of the patients was under 30 years of age (29 years) and only one of the nodules had a diameter of less than 20 mm (18 mm), we suspect that DE nodules affecting the SP may take a long time to develop.In the literature, the average age of patients operated on with laparoscopy for DE infiltrating the SP is also over 30 years 1,2 .Further prospective studies should seek to clarify the natural history of DE nodules affecting the SP.
Characteristic symptoms, like pain or paresthesia radiating to the perineum and the ipsilateral lower limbs and corresponding dermatome, i.e. sacral radiculopathy, may lead to the suspicion of disease 23 .However, a diagnosis cannot be made on the basis of these symptoms alone.
A strength of our study is that it is the first of its kind to describe the ultrasound features of this rare type of DE in a clinically relevant number of patients and with a final diagnosis confirmed by radical resection and histology.A limitation of the study is its retrospective nature.However, to overcome this limitation, we used standardized terminology and a predefined ultrasound assessment form based on IOTA, MUSA and IDEA terms and definitions to describe the sonographic characteristics of the DE nodules, as well as exchanging and reviewing digital ultrasound images between four independent observers.
In conclusion, clinical signs of sacral radiculopathy and the finding on TVS of a parametrial, unilateral, solid, non-uniform, hypoechogenic nodule with hyperechogenic areas, possible internal shadowing and irregular spiculated contours, demonstrating poor vascularization on Doppler examination, and in proximity to or involving the structures of the SP, reflect symptomatic DE affecting the SP.

Figure 1
Figure 1 Microscopic images of deep endometriosis affecting a pelvic nerve (hematoxylin and eosin stained).(a) Pelvic nerve (black arrow) surrounded by loose connective tissue, with an endometrial gland (white arrow) and its stroma attached.(b) Another pelvic nerve (black arrow) with epineurium surrounded by skeletal muscle cells and vessels ( ).Endometrial gland (white arrow) and its stroma are seen infiltrating the nerve.

Figure 2
Figure 2 Laparoscopic photograph taken during sacral neurolysis, depicting dissected retroperitoneum and right S1-S2 sacral nerve roots of sacral plexus (black arrows).Deep endometriotic nodule (white arrow) is dissected from the nerve.Instrument holding a swab ( ) keeps rectum away.Clipped vessels are branches of internal iliac vein from gluteal and obturator region.Lat, lateral; Med, medial.

106
Szab ó et al. knowledge, DE affecting the SP has not yet been studied with TVS.The aim of this study was to describe the clinical and sonographic characteristics of DE infiltrating the SP.

Figure 3
Figure 3 Transvaginal ultrasound image in patient with deep endometriosis affecting left sacral plexus (SP).Transducer is lateral to cervix (C) in left lateral vaginal fornix.Deep endometriotic nodule (black arrows) with typical non-uniform appearance (hypoechogenic lesion with hyperechogenic areas) and irregular (spiculated) contour infiltrates left vaginal fornix and uterosacral ligament (S), as well as lateral and posterior wall of pelvis on patient's left side.Nerves of SP show typical 'bundle-of-straw' appearance in longitudinal section.Branches from anterior division of internal iliac vessels ( ) cross the SP.Lat, lateral; Med, medial.

Figure 5
Figure 5 Transvaginal ultrasound image with color Doppler of deep endometriotic (DE) nodule ( ) affecting sacral plexus (black arrows).DE nodule shows no signal (color score 1).Signal detected only for vessels in cervix (C) and for branches from anterior division of internal iliac vessels lateral to DE nodule.Note characteristic internal acoustic shadow (white arrow).Lat, lateral; Med, medial.

Figure 6
Figure 6 Dual-mode grayscale sonographic (a) and strain elastography (b) images of deep endometriotic nodule (black arrow) affecting sacral plexus (SP) on transvaginal ultrasound scan.Ascending-colors map is used in (b): dark blue indicates harder tissue; green and yellow indicate moderately stiff tissue; and orange and red indicate soft tissue.Lower bar in (b) indicates pressure applied.According to elastography, hypoechogenic and hard nodule with irregular spiculated contour infiltrates the SP.C, cervix; Med, medial; Lat, lateral; S, uterosacral ligament.

Figure 7
Figure 7 (a-c) Transvaginal ultrasound images of deep endometriotic (DE) nodules.(a) Extrinsic-type DE nodule ( ) affecting sacral plexus (SP) (black arrows).Nodule attached to the nerves originates from parametrium.Lesion produces internal acoustic shadow (white arrow) and has spiculated contour (yellow dashed line).(b) Intrinsic-type DE nodule.Note that hyperechogenic epineurium and hypoechogenic fascicles (black arrows) of SP pass through DE lesion (white arrow).SP runs on anterior surface of the ala of sacrum (AS) and piriformis muscle (P).(c) Intrinsic-type DE nodule at the level of the pelvic inlet.DE nodule infiltrates S1 root (black arrows) just inferior to arcuate line (AL).Note spiculated contour (yellow dashed line) of hypoechogenic lesion.(d) Image of specimen obtained from patient in (c) after dissecting nodule with laparoscopy.Compare shape of lesion ( ) with (c).Black arrow indicates pararectal fatty tissue.C, cervix; Med, medial; Lat, lateral.

Figure 8
Figure 8 Comparison of deep endometriotic (DE) nodule affecting sacral plexus (a) and neurofibroma originating from sacral plexus (b-e).(a) Transvaginal ultrasound image with color Doppler illustrating DE nodule ( ) affecting sacral plexus (black arrow).All typical characteristics are present: direct contact with hyperechogenic perineurium of nerves, spiculated contour, non-uniform echogenicity (hypoechogenic lesion with hyperechogenic areas), poor vascularization and internal acoustic shadow (white arrow).(b,c) Grayscale (b) and color-Doppler (c) transvaginal ultrasound images of neurofibroma of sacral plexus ( ).Tumor is solid and non-uniform, but is well-circumscribed by hyperechogenic rim formed by the epineurium (black arrows), and presents with no internal acoustic shadow (b) and no color Doppler signal (c).(d,e) Sonoelastographic appearance of same well-circumscribed neurofibroma ( ) as in (b,c), showing soft tissue in lesion.Dual-mode grayscale sonographic (d) and strain elastography (e) images of neurofibroma affecting sacral plexus on transvaginal scan.Ascending-colors map is used in (e): dark blue indicates harder tissue; green and yellow indicate moderately stiff tissue; and orange and red indicate soft tissue.Lower bar in (e) indicates pressure applied.A, anterior division of internal iliac artery; C, cervix; Lat, lateral; Med, medial; P, piriformis muscle.
According to pattern recognition, most DE nodules were purely solid, non-uniform, hypoechogenic nodules containing hyperechogenic areas, with internal shadows and irregular spiculated contours, and were poorly vascularized on color/power Doppler examination.

Table 1
Demographic and clinical characteristics of 27 women with deep endometriosis affecting sacral plexus Data are given as median (range), mean ± SD or n (%).

Table 2
Ultrasound findings in 27 women with deep endometriosis affecting sacral plexus *Color score: 1, no detectable blood flow; 2, minimal blood flow; 3, moderate blood flow; and 4, abundant blood flow.