Differences in the incidence of obstetric complications depending on the extent and location of adenomyosis lesions

Abstract Objectives Although adenomyosis is reportedly associated with adverse pregnancy outcomes, clinical factors related to the high risk of obstetric complications are unclear. This study aimed to elucidate the characteristics of adenomyosis lesions associated with the increased incidence of obstetric complications based on imaging findings. Methods This was a retrospective, observational cohort study conducted in a tertiary perinatal care center. Eighty-eight singleton pregnant women with adenomyosis were included in the study. Based on magnetic resonance imaging or ultrasonography before and/or during pregnancy, patients were classified according to three types of image characteristics: the extent of adenomyosis lesion (focal type or diffuse type), location of the lesion (extrinsic type, intrinsic type, or indeterminate type), the positional relationship between the lesion and the placenta (placenta distant from adenomyosis or placenta over adenomyosis), and the incidence of obstetric complications were examined. Results Patients with diffuse type adenomyosis are significantly more likely to have spontaneous second-trimester miscarriage (diffuse type vs. focal type: 16.7 vs. 0%, p < .01), preterm premature rupture of membranes (19.4 vs. 1.9%, p < .01), and preeclampsia (25.0 vs. 7.7%, p = .02), as compared to those with focal type adenomyosis. In a comparison of the three location types, the incidence of placental malposition was higher in patients with the extrinsic type adenomyosis (extrinsic type vs. intrinsic type vs. indeterminate type: 20.0 vs. 6.7 vs. 2.3%, p = .03). Comparisons between the types of the placenta over or distant from adenomyosis lesion displayed no significant differences in the frequencies of obstetric complications. Conclusions We demonstrated that the frequency of obstetric complications related to adenomyosis varies depending on the extent and location of the lesion; patients with diffuse type adenomyosis have an increased risk of spontaneous second-trimester miscarriage, preterm premature rupture of membranes, and preeclampsia, while patients with extrinsic type adenomyosis have an increased risk of placental malposition. Imaging evaluation of adenomyosis prior to conception or early in pregnancy may be useful for the obstetrical risk assessment among patients with adenomyosis.


Introduction
Adenomyosis, which is characterized by the presence of endometrial glands in the myometrium, is associated with pelvic pain, metrorrhagia, and infertility.With the progress in infertility treatment and advanced maternal age, pregnancies complicated by adenomyosis are on the rise.Recent studies have reported that adenomyosis is associated with an increased incidence of many obstetric complications: such as spontaneous second-trimester miscarriage, preterm delivery, placental malposition, hypertensive disorders of pregnancy (HDP), cesarean delivery, postpartum hemorrhage (PPH), small-for-gestational-age (SGA) infants [1][2][3][4][5][6][7][8].However, clinical factors which lead to the development of these obstetric complications in pregnant women with adenomyosis remain unknown.
Adenomyosis can be assessed by the extent or the location of the lesion and is divided into several subtypes based on the imaging findings of ultrasonography and magnetic resonance imaging (MRI).Extension of the lesion throughout the entire myometrium is classified as the diffuse type, whereas a lesion surrounded by normal tissue is classified as the focal type [9].In non-gravid women, the diffuse type causes greater menstrual bleeding than the focal type [10].Adenomyosis can also be classified by the location of the lesion into the extrinsic type: the lesion located on the serosal side of the uterus, and the intrinsic type: the lesion located on the luminal side.The extrinsic type is associated with a high likelihood of infertility [10], while in the intrinsic type, hormone therapy may trigger serious uterine bleeding [11].There have been serial reports indicating that the extent or the location of adenomyosis lesions is associated with distinct clinical symptoms in a non-pregnant state.However, whether these classifications of adenomyosis lesions are related to adverse pregnancy outcomes remains unclear.
Therefore, based on the imaging findings of adenomyosis according to extent and location of the lesion, the present study aimed to clarify the characteristics of adenomyosis, which contribute to the development of obstetric complications, to provide the evidence to identify the high-risk pregnant women with adenomyosis in perinatal management.

Materials and methods
Under the approval by the Institutional Review Board of The University of Tokyo (3053-1), clinical information was retrospectively obtained from the medical records of pregnant women with adenomyosis who were perinatally managed and delivered after 12 weeks of gestation at the University of Tokyo Hospital from January 2010 to February 2022.Women with multiple pregnancies, fetal abnormalities, uterine deformities, or previous uterine surgery were excluded from this study.
Pregnant women who had MRI and/or transvaginal ultrasonography (TVUS) findings before and/or during the first trimester of pregnancy that met the following criteria were considered cases with adenomyosis.On MRI, adenomyosis was diagnosed when one of the following two diagnostic criteria was met: (1) a myometrial mass with indistinct margins with primarily low signal intensity, or (2) diffuse or focal thickening of the junctional zone forming an ill-defined area of low signal intensity on T2-weighted images [12].On TVUS, adenomyosis was diagnosed based on the following morphological ultrasonographic features [13]: myometrial cyst, hyperechogenic islands, or echogenic subendometrial lines and buds as direct features of adenomyosis, which are considered typical of adenomyosis; and asymmetrical thickening, fan-shaped shadowing, globular uterus, translesional vascularity, irregular junctional zone, or interrupted junctional zone as indirect features of adenomyosis, which are considered a consequence of ectopic endometrium in the myometrium.The diagnosis of adenomyosis by TVUS is confirmed by meeting at least one of the direct features with reference to the presence of indirect features.
Patients who met the above diagnostic features were divided based on the imaging findings of their lesions into three groups for each of the following parameters: (1) extent of the lesion (focal type or diffuse type), (2) lesion location (extrinsic type, intrinsic type, or indeterminate type), (3) positional relationship between the placenta and the lesion (placenta distant from adenomyosis or placenta over adenomyosis).These classifications are illustrated in Figure 1.

Extent of lesions was classified based on a study
by Van den Bosch et al. as the diffuse type or focal type [14].If >25% of the circumference of the lesion was surrounded by the normal myometrium, it was defined as the "focal type"; whereas when < 25% of the lesion was surrounded by the normal myometrium, it was classified as the "diffuse type".2. With regard to the location of the lesions, adenomyosis was classified as the "extrinsic type" when it resided adjacent to the serosal side of the uterus without affecting the inner structures, and as "intrinsic type" when it resided adjacent to the uterine inner layer without affecting the serosal side [15]."Indeterminate type" is the one that does not match either of these two criteria.This classification includes lesions extending throughout the uterine myometrium, which can be mostly regarded as the diffuse type, and multiple lesions.3. Regarding the positional relationship between the placenta and the adenomyosis lesion, adenomyosis was classified as "placenta distant from adenomyosis" when the adenomyosis lesion was not present in the myometrium which was in contact with the placenta, whereas adenomyosis was classified as "placenta over adenomyosis" when any part of the adenomyosis lesion was included in the myometrium in contact with the placenta.
The positional relationship between the placenta and the adenomyosis lesion was assessed based on ultrasonography images taken up to 20 weeks of gestation.Other classifications were assessed using MRI and/or TVUS before or during the first trimester of pregnancy.Ultrasonography was performed using Voluson P8, S8, E10 or S10 (GE Healthcare, Japan).The transvaginal probes used in this study were IC9-RS, RIC5-9 A-RS and E8C-RS probes and transabdominal probes were 4 C-RS, C1-5-RS, C1-6-D, RM6C, C2-9-RS and RAB6-RS probes, respectively.Ultrasonography was performed by obstetrician-gynaecologists, with images saved as Digital Imaging and Communications in Medicine data in the image filing system Claio.Ultrasonography images of the uterus and the placenta saved in Claio were retrospectively analyzed by two Japan Society of Ultrasonics in Medicine-certified ultrasound specialists, who classified the images according to the three indicators described above and were blinded from the patients' perinatal outcomes.
The following perinatal outcomes were examined: spontaneous second-trimester miscarriage, preterm premature rupture of membranes (pPROM), preterm delivery, spontaneous preterm delivery, placental malposition, HDP, preeclampsia, cesarean delivery, PPH, and light-for-date (LFD) infants.Spontaneous secondtrimester miscarriage was defined as spontaneous miscarriage delivered between 14 and 21 weeks of gestation.Spontaneous preterm delivery was defined as preterm delivery excluding iatrogenic preterm delivery due to reasons such as preeclampsia or placental malposition.In accordance with the diagnostic criteria stated by the International Society for the Study of Hypertension in Pregnancy, the diagnosis of HDP and its classification including preeclampsia were made [16].Placental malposition included placenta previa and placenta previa or low-lying placenta that required cesarean delivery.PPH was defined as blood loss !500 g for vaginal delivery and !1000 g for cesarean delivery.An LFD infant was defined as an infant with a birth weight below the 10th percentile.
Statistical analyses were performed using Stata software version 14.0 (Stata Corp LLC, USA).Frequencies of obstetric complications and other categorical variables were analyzed using Fisher's exact test, while blood loss and other continuous variables were analyzed using the Mann-Whitney U test.Fisher's exact The lesion was classified as the "extrinsic type" if it presents on the serosal side of the uterus, and as "intrinsic type" if it presents in the uterine inner layer."Indeterminate type" includes lesions extending through all layers of uterine muscle and multiple lesions.(c) Positional relationship between the lesion and the placenta: the lesion was classified as "placenta distant from adenomyosis" when the adenomyosis lesion was not present in the myometrium which was in contact with the placenta, whereas it was classified as "placenta over adenomyosis" when the adenomyosis lesion was included in the myometrium in contact with the placenta.
test and Bonferroni method were used for comparison between the three groups.For these tests, p < .05 was considered to indicate a significant difference.For significant differences, odds ratios (ORs) with 95% confidence intervals (CI) were calculated.

Results
The present study included 88 patients with adenomyosis, who were classified based on three types of image characteristics: the extent of adenomyosis lesions, lesion location, and the positional relationship between the lesion and the placenta, as shown in Supplementary Figure 1.Of the 88 patients with adenomyosis, 66 patients (75.0%) were diagnosed by MRI before pregnancy, while 22 patients (25.0%) were diagnosed by TVUS alone.Of the 22 patients diagnosed by TVUS alone, 12 patients were diagnosed before pregnancy, while 10 patients were diagnosed in the first trimester of pregnancy.In the classification based on the extent of the lesions, 52 patients (59.0%) were classified as the focal type, while 36 patients (40.9%) were classified as the diffuse type.
Table 1 shows the comparison of patient characteristics and obstetric complications between the focal type and the diffuse type.Patient characteristics (maternal age, percentage of primipara, and frequency of pregnancies achieved with assisted reproductive technology) did not differ significantly between the two groups.Spontaneous second-trimester miscarriage was significantly more frequent in the diffuse type than in the focal type (diffuse type vs. focal type: 16.7 vs. 0%, p < .01).Diffuse type adenomyosis was also associated with significantly higher frequencies of pPROM (19.4 vs. 1.9%,OR ¼ 17.0, 95% CI: 2.0-141, p < .01)and preeclampsia (25.0 vs. 7.7%, OR ¼ 4.0, 95% CI: 1.1-14.2,p ¼ .02),compared with the focal type.A significant difference was not observed in the incidence of preterm delivery including spontaneous preterm delivery, placental malposition, HDP, cesarean delivery, PPH, or LFD infants between the two groups.
In the classification based on the lesion location, 30 patients were of the extrinsic type (34.0%) and 15 patients were of the intrinsic type (17.0%), while 43 patients were of the indeterminate type (48.8%) (Supplementary Figure 1).Representative images of TVUS and MRI are shown in Figure 2.Among 52 patients categorized as the focal type, 30, 15, and 7 patients were classified into extrinsic type, intrinsic type, and indeterminate type, respectively.All 36 patients categorized as the diffuse type were classified into the indeterminate type, and 43 indeterminate type patients consisted of 36 diffuse type and 7 focal type patients.In the comparison between the three types of the lesion location, there were significant differences in the frequencies of second-trimester miscarriage (extrinsic type vs. intrinsic type vs. indeterminate type: 0 vs. 0 vs. 14.0%,p ¼ .03),pPROM (3.3 vs. 0 vs. 20.9%,p ¼ .02),and placental malposition (20.0 vs. 6.7 vs. 2.3%, p ¼ .03)(Table 2).Post hoc test showed a significant difference between the extrinsic type and indeterminate type groups with respect to the frequency of placental malposition (extrinsic type vs. indeterminate type: 20.0 vs. 2.3%, p ¼ .02).Post hoc test applied to the results of second-trimester miscarriage and pPROM showed no statistically significant differences between any of the types.There were no significant differences in other obstetric complications between the three groups.
Based on the positional relationship between the placenta and the adenomyosis lesion, 29 patients (33.0%) were classified as "placenta distant from Unless stated otherwise, data are given as n (%).IQR: interquartile range; ART: assisted reproductive technology; pPROM: preterm premature rupture of membranes; HDP: hypertensive disorders of pregnancy; PPH: postpartum hemorrhage; LFD: light-for-date; OR: odds ratio; CI: confidence interval.
adenomyosis," while 59 patients (67.0%) were classified as "placenta over adenomyosis."Comparisons between the "over adenomyosis" and "distant from adenomyosis" lesions displayed no significant difference in the incidence of obstetric complications (Table 3).

Discussion
The current study demonstrated that the frequency of obstetric complications varies with the extent and location of the adenomyosis lesions.Imaging evaluation of adenomyosis before and/or during pregnancy may aid in detecting adenomyosis in pregnant women who are at high risk for obstetric complications.
This study aimed to clarify which patient with adenomyosis warrants particularly careful perinatal management in actual clinical settings based on the imaging findings.In previous studies that compared patients with adenomyosis to patients without adenomyosis, patient characteristics such as age and history of infertility treatment not consistent, and the  Unless stated otherwise, data are given as n (%).IQR: interquartile range; ART: assisted reproductive technology; pPROM: preterm premature rupture of membranes; HDP: hypertensive disorders of pregnancy; PPH: postpartum hemorrhage; LFD: light-for-date; OR: odds ratio; CI: confidence interval.
confounding factors associated with perinatal outcomes were not eliminated.Due to the strong association between adenomyosis and infertility, patients with adenomyosis are an extremely unbalanced population, that is, majority of the patients are older, primiparous, and conceive with assisted reproductive technology.These factors alone might put patients with adenomyosis at high risk for obstetric complications, including preeclampsia.The present study compared patients with adenomyosis among whom these underlying characteristics were similar.Consequently, patients with diffuse type adenomyosis in which the lesions extend throughout a wider area of the uterine myometrium demonstrated higher incidences of spontaneous second-trimester miscarriage, pPROM, and preeclampsia, which indicates that the presence of adenomyosis itself adversely affects pregnancy outcomes.
Classification of the adenomyosis lesions into the diffuse or the focal type is often used in clinical practice.However, in many cases, this classification is applied subjectively, and few studies have employed clear criteria.Focusing on objectivity and reproducibility in classification, the present study measured the percentage of the lesion circumference surrounded by the normal myometrium based on the study by Van den Bosch et al. to classify lesions into the diffuse or the focal type [14].This classification was performed by two ultrasound specialists blinded to the perinatal outcomes.Only one study by Tamura et al. has reported the obstetric complications in the diffuse type adenomyosis.This study found that preeclampsia was more frequent in the diffuse type [8].However, this study had several limitations: it was a retrospective multicenter questionnaire study, criteria for imaging classifications were unclear, and the timing of the imaging findings used for the classification was unclear.The findings in the present study were obtained from diffuse/focal classification based on consistent criteria at a single center and are therefore likely to be more reliable.Although the current study limited the timing of the image classification to prepregnancy or the first trimester, our results show that classifying adenomyosis as diffuse or focal can differentiate patients regarding perinatal risks early in pregnancy, which is valuable for actual perinatal management.However, even in the focal type, obstetric complications occur more frequently than in the general population of women of the same age.In the focal type, preterm delivery, placental malposition, and cesarean delivery occurred in 32.7, 13.5, and 55.8%, respectively.Thus, although not as much as those with the diffuse type, careful perinatal management is also required for pregnant women with the focal type.
We also assessed adenomyosis by the lesion location according to commonly used location-based classification method as follows: intrinsic type, in which adenomyosis is located on the uterine cavity side; extrinsic type, in which adenomyosis is on the serosal side; and indeterminate type, in which adenomyosis is far advanced and diffusely located in the myometrium [3,4].Comparison of these categories revealed differences in the frequency of spontaneous secondtrimester miscarriage, pPROM, and placental malposition.As all the cases with the diffuse type adenomyosis were classified as the indeterminate type, the trend toward increased frequency of spontaneous second-trimester miscarriage and pPROM in the indeterminate type is likely due to its shared incidence with the diffuse type.Interestingly, the frequency of placental malposition was found to be significantly increased in the extrinsic type.Previous studies have shown that the intrinsic type develops due to factors such as trauma to the endometrium [15,17,18], while the extrinsic type is considered to develop due to deep infiltrating endometriosis (DIE) in the pelvis [15,19,20].Placental malposition is known to occur more frequently in endometriosis, particularly in DIE [21].Our result suggests that the increased incidence of placental malposition in the extrinsic type might be associated with underlying DIE in these patients.
Our analysis of the classification based on the positional relationship of the adenomyosis to the placenta was based on the hypothesis that if the adenomyosis and the placenta are close to each other, placentation and placental function would be more directly affected by the presence of adenomyosis, thereby making obstetric complications more likely.In a previous retrospective cohort study conducted on 20 patients, the incidence of SGA infants was increased when the placenta was located over the adenomyosis [22], but the small sample size clearly warranted further investigation.In our study, there was no difference in the frequency of complications.Since our sample size is limited, and it was difficult to assess the amount of the lesion overlapping the placenta, further studies are required to conclude whether the positional relationship between the placenta and the adenomyosis affects obstetric complications.
The strength of our study is that this study includes the largest number of cases compared to previous reports that analyzed the pregnancy outcome of women with adenomyosis based on detailed information from individual medical records.However, the sample size is not large enough to clearly identify the differences in the incidence of complications among the groups.A possible limitation is the misdiagnosis of adenomyosis, which may cause potential bias.However, since MRI and TVUS have high sensitivity and specificity in detecting adenomyosis (MRI: sensitivity 70-89%, specificity 86-92.5%,TVUS: 72-86%, 81-92.8%,respectively) [23,24,25], the impact of this limitation is likely to be small.However, especially for TVUS, although the characteristic ultrasound sings of adenomyosis have been identified, it is reported that those signs have a certain degree of variable reliability in detecting adenomyosis [13].In this study, according to the recent consensus on diagnostic criteria [13] , the diagnosis of adenomyosis was confirmed by the presence of direct signs considered typical of adenomyosis by two board-certified ultrasound specialists.
In conclusion, we found that patients with diffuse type adenomyosis are at high risk for spontaneous second-trimester miscarriage, pPROM, and preeclampsia.Moreover, the frequency of placental malposition was increased in patients with extrinsic type adenomatosis.Our results suggest that assessment of adenomyosis by MRI or ultrasonography before or during the first trimester of pregnancy may enable the stratification of high-risk groups for obstetric complications in pregnant women with adenomyosis.

Figure 1 .
Figure 1.Schematic representation of the classification of adenomyosis based on imaging findings.(a)Extent of lesions: the lesion >25% of the circumference was surrounded by normal myometrium was defined as the "focal type"; whereas that <25% was surrounded by normal myometrium was classified as the "diffuse type".Circumference of the lesion surrounded by the normal myometrium, as is illustrated indicated by the red line.(b) Location of the lesion: The lesion was classified as the "extrinsic type" if it presents on the serosal side of the uterus, and as "intrinsic type" if it presents in the uterine inner layer."Indeterminate type" includes lesions extending through all layers of uterine muscle and multiple lesions.(c) Positional relationship between the lesion and the placenta: the lesion was classified as "placenta distant from adenomyosis" when the adenomyosis lesion was not present in the myometrium which was in contact with the placenta, whereas it was classified as "placenta over adenomyosis" when the adenomyosis lesion was included in the myometrium in contact with the placenta.

Figure 2 .
Figure 2. Representative images of the classification of adenomyosis based on the location of the lesion by TVUS and MRI.The typical ultrasound images for the extrinsic type (a), the intrinsic type (b), and the indeterminate type (c) are shown.MRI images (T2 weighted images) of the corresponding patients are also shown below them.Yellow arrows indicate the circumference of the adenomyosis lesion.The white arrow indicates the endometrium of the uterus.TVUS: transvaginal ultrasonography; MRI: magnetic resonance imaging.

Table 1 .
Extent of adenomyosis lesions and obstetric complications.

Table 2 .
Location of adenomyosis lesions and obstetric complications.< .05,p values calculated by Fisher's exact test among three groups are shown.

Table 3 .
Positional relationship between the placenta and the adenomyosis lesions and obstetric complications.Values were calculated by Fisher's exact test.Unless stated otherwise, data are given as n (%).IQR: interquartile range; ART: assisted reproductive technology; pPROM: preterm premature rupture of membranes; HDP: hypertensive disorders of pregnancy; PPH: postpartum haemorrhage; LFD: light-for-date; OR: odds ratio; CI: confidence interval.