Heterogeneity of clinical symptoms and therapeutic strategies for different subtypes of adenomyosis: An initial single- center study in China

Objective: To investigate the relationship between the magnetic resonance imaging (MRI) classification of different clinical symptoms and corresponding therapeutic efficacy in adenomyosis patients. Methods: From January 2015 to October 2020, a total of 468 patients diagnosed with adenomyosis through MRI examination at Peking University Third Hospital were included in this retrospective cohort study. Totals of 184 (39.3%), 208 (44.4%), 17 (3.6%)


| INTRODUC TI ON
Adenomyosis is a common gynecologic disease, and its main manifestations are dysmenorrhea, menorrhagia, and infertility. Adenomyosis is difficult to cure, and seriously affects the life quality of patients. 1,2 Due to varied imaging diagnostic criteria, non-standardized histologic definitions, and heterogeneous patient populations, it is difficult to determine the true prevalence of adenomyosis. The prevalence of adenomyosis ranges between 5% and 70%, but most prospective studies have shown that the prevalence is approximately 20%. 3,4 Patients have heterogeneous clinical characteristics 5 : 15%-30% have mainly dysmenorrhea, while 40%-60% have mainly menorrhagia. Previous studies have found that the infertility rate in adenomyosis patients is up to 27%. 6 The adenomyosis negatively affects in vitro fertilization, pregnancy, and the live birth rate, as well as increasing the risk of miscarriage. 7,8 The lesion location, distribution, and size, and the range of myometrial invasion and accompanying diseases are different among different patient populations. Therefore, adenomyosis is a heterogeneous disease that must be rationally classified. Accurate classification of adenomyosis and the formulation of treatment strategies are both urgent problems to be solved in clinical practice.
Although there is no consensus on the classification of adenomyosis, magnetic resonance imaging (MRI) might be the most accurate diagnostic and classification method for adenomyosis. 9,10 MRI classification is mainly based on the distribution and location of lesions according to the method proposed by Bazot and Daraï 11 and by Kishi et al. 12 The relatively acceptable classifications are listed as follows: subtype I (intrinsic) adenomyosis: lesions are located in the uterine inner layer without affecting its outer structures; subtype II (extrinsic) adenomyosis: the lesions are located in the uterine outer layer without affecting its inner structures; subtype III (intramural) adenomyosis: lesions exist independently and are not related to the outer and inner structures; and subtype IV (indeterminate) adenomyosis: adenomyosis not conforming to one of the previous types. Different subtypes may represent heterogeneous etiologies and pathogeneses. Subtype I may be caused by direct invasion of the endometrium into the inner and medial myometrium; subtype II may be caused by direct invasion of the extrauterine endometriosis into the serosa; subtype III may be caused by metaplasia or epithelial-mesenchymal transformation, and subtype IV may be caused by a heterogeneous mixture of advanced disease. 13,14 Drug therapy is the preferred treatment for adenomyosis. Poor drug control may lead to further surgery according to the patient's specific situation. However, we have found that different patients have different responses to drug therapy in clinical practice.
Previous studies have found that only half of patients can benefit from progesterone therapy. [15][16][17][18][19] Adenomyosis may lead to a high proportion of infertility. With the opening of China's family planning policy, the increasing female population has fertility requirements.
Improving the fertility of patients with adenomyosis is an urgent problem to be solved. Therefore, exploring the therapeutic efficacy for different adenomyosis subtypes can provide guidance for better treatment selection. The classification method adopted in our study is improved on the basis of the classification method proposed by Kishi et al.,12 adopting intrinsic, extrinsic, and intramural subtypes, and defining the indeterminate subtype as penetrating subtype according to their MRI characteristics.
To our knowledge, this is the largest study on adenomyosis classification, and is the first classification study in China. A new classification of adenomyosis is proposed in this study on the basis of the nomenclature method. The aim of this study is to investigate the relationship between the MRI classification of different clinical symptoms and corresponding therapeutic efficacy in these adenomyosis patients, to improve the fertility of young adenomyosis patients, and provide a sufficient theoretical basis for treatment.

| Inclusion and exclusion criteria
This is a retrospective cohort study and was approved by the Peking University Third Hospital Medical Science Research ethics committee (M2022056). From January 2015 to October 2020, a total of 1220 patients were diagnosed with adenomyosis and underwent pelvic MRI examination at the Peking University Third Hospital. MRI was performed when adenomyosis was first diagnosed and before treatment. Patients with gynecologic malignant tumors, menopause, pregnancy, giant uterine myomas (single diameter >5 cm), and uterine malformations were excluded. Two experienced radiologists reviewed the MRI results and selected 468 patients with typical lesions that could be clearly and uniformly classified.

| Clinical characteristics of adenomyosis questionnaire
We designed the clinical characteristics of the adenomyosis questionnaire using the Delphi Method referring to literature on clinical characteristics of adenomyosis. The MRI scans were reviewed by radiologists first, then patients were identified, and the questionnaires were filled out. From January to March 2021, the questionnaires were completed through outpatient service, medical records review, and telephone follow up. All 468 patients signed informed consent forms and completed questionnaires by doctors. The quantitative K E Y W O R D S adenomyosis, clinical symptoms, infertility, magnetic resonance imaging classification, therapeutic efficacy standard of dysmenorrhea degree was the visual analog scale (VAS) score. Menorrhagia referred to five or more wet pieces of overnight sanitary napkins daily. Infertility referred to regular sexual life without contraception and no pregnancy for more than 1 year, and exclusion of male factors. Anemia referred to hemoglobin of 110 g/L or less. Effective treatment referred to a decrease in VAS score of three or more points and/or a decrease in menstrual volume by at least half.

| MRI classification
MRI was used for classification. Subtype I (intrinsic) (Figure 1a

| Statistical analysis
Statistical analysis was performed using SPSS software version 24.0 (IBM, Armonk, NY, USA). Continuous variables conforming to a normal distribution are expressed as the mean ± standard deviation, and a t test or analysis of variance was used to compare them. Those not conforming to a normal distribution are represented by the median, and the rank sum test was used for comparison. Categorical variables are expressed as numbers or percentages, and the χ 2 test or Fisher exact test was used. The threshold for statistical significance was set as P < 0.05.

| Clinical characteristics of patients with different subtypes
The clinical characteristics of the different subtypes are shown in Table 2. There was no significant difference in terms of age among intrinsic, intramural, and penetrating subtypes. However, their ages were significantly older than that in the extrinsic subtype cohort. The proportion of patients with dysmenorrhea was increasing in intrinsic, extrinsic, intramural, and penetrating subtypes, but there was no significant difference in terms of VAS max among the four subtypes. The proportion of patients with menorrhagia in the intrinsic subtype was higher than that in the extrinsic and intramural subtypes, and there was no significant difference comparing with the penetrating subtype. The incidence of anemia in the intrinsic subtype was higher than that in the extrinsic subtype, similar to that in the intramural subtype, and lower than that in the penetrating subtype. However, there was no significant difference in terms of moderate to severe anemia (hemoglobin ≤90 g/L) between intrinsic and penetrating subtypes (57/184 [31.0%] vs 26/59 [44.1%]; P = 0.067). There was no significant difference in the incidence of deep infiltrating endometriosis (DIE) or endometriotic cysts between intrinsic and intramural subtypes, and their rates were significantly lower than those of extrinsic and penetrating subtypes. The proportion of patients with extrinsic disease that was complicated by DIE and endometriotic cysts was significantly higher than that among patients in the penetrating subtype.

| Drug treatment of patients with different subtypes
Drug treatments of adenomyosis include oral contraceptives (OCs), gonadotropin-releasing hormone agonists (GnRHa), the levonorgestrel-releasing intrauterine device (LNG-IUD), denorgestrel, and nonsteroidal anti-inflammatory drugs. The drug treatments of the four subtypes are shown in Table 3 intrinsic and penetrating subtypes were higher than those in the extrinsic subtype (P = 0.117).
Progesterone (which acts as a summary of LNG-IUD, OCs, and (70.6%), respectively. The total effective rate of progesterone treatment in intrinsic subtype was significantly lower than that in the extrinsic subtype (P < 0.001). The effective rate of penetrating subtype was higher than that of intrinsic subtype.

| DISCUSS ION
Adenomyosis is a common gynecologic disease in the population of childbearing age. The main clinical manifestations are severe pain, abnormal uterine bleeding, and infertility. This progressive disease can also affect the birth outcome, lead to complications during pregnancy, and affects women's quality of life. 20 the time before childbirth. Kishi et al. 12 found no significant difference in symptoms such as dysmenorrhea and menorrhagia among patients with different subtypes. Mean while Bourdon et al. 22 found that women in the intrinsic subtype were more likely to develop menorrhagia than those in the extrinsic subtype, and the difference was correlated with a history of uterine surgery. However, in terms of pain symptoms (dysmenorrhea, dyspareunia, chronic pelvic pain), both the presence and intensity were similar between the two subtypes. We found that the proportion of dysmenorrhea in the intrinsic subtype was lower than that in the extrinsic subtype, and that the proportion in the extrinsic subtype was lower than those in the intramural and penetrating subtypes. We speculate that this is because the extrinsic subtype has a higher rate of merging with DIE and endometriotic cysts. DIE and endometriotic cysts is generally associated with dysmenorrhea. The growth of the lesion in the intramural subtype was limited to the middle of the myometrium, expanding outward, and causing abnormal contraction of normal muscular tissue. We also found that the proportion of menorrhagia in the intrinsic subtype was higher than that in the extrinsic subtype. These findings agree with those of Bourdon et al. 22 However, the proportion of patients with menorrhagia in the intrinsic subtype was not significantly different from that in the penetrating subtype.
If penetrating subtype developed from intrinsic subtype, the proportion of patients with menorrhagia should be higher than that in the intrinsic subtype. However, there was an insignificant difference between them. Therefore, we propose that the penetrating subtype may develop from the extrinsic subtype rather than the intrinsic subtype. The patients included in the above two studies were all surgically treated patients with relatively serious conditions. Further studies are necessary. Therefore, we enlarged the sample size and added non-surgical patients for further analysis.
Previous studies did not focus on differences in infertility among different subtypes. The fertility status of women worldwide is low and the demand for fertility is increasing. Many patients have to accept multiple rounds of assisted reproduction. However, adenomyosis will cause a high infertility rate, and a low success rate of assisted reproduction. Therefore, closer attention should be paid to the fertility of patients with adenomyosis. To find a better fertility guidance method, we should analyze the infertility problems of adenomyosis, especially the comparison of infertility problems of different subtypes. Iwasawa et al. 23 found that the extrinsic subtype of infertile patients who underwent in vitro fertilization/intracytoplasmic sperm injection had fewer pregnancy losses and more live births than the advanced group (same as our penetrating subtype).
In our study, we found that the intrinsic subtype feature, mainly secondary infertility, which is related to endometrial damage caused by repeated curettage and fertility, would reduce the receptivity of the endometrium. Therefore, although the infertility rate was low, the success rate of assisted reproduction may also be affected. Primary infertility was predominant in extrinsic subtype, and it was mainly related to DIE, which might influence the microenvironment of the pelvic cavity, and produce toxic effects in the tubes carrying eggs and sperm. These adverse effects would cause infertility without endometrial damage. Although the infertility rate was relatively high, the success rate of assisted reproduction was also higher. Therefore, we should pay more attention to intrinsic infertility patients in clinical practice. We should use some methods to change endometrial receptivity to improve the success rate of pregnancy in the process of assisted reproduction.
In clinical practice, we find that the effectiveness of drug therapy varies between patients with different adenomyosis subtypes.
GnRHa is effective for almost all patients with adenomyosis.
However, it cannot be used for a long-term course because of its strong adverse effects and poor adherence. Other common drugs, such as OCs, LNG-IUD, and denorgestrel, are all progesterone preparations and can be used as long-term maintenance therapy.
However, there are still a considerable number of patients in whom progesterone has poor therapeutic efficacy. Hitherto, there is no good method to predict the efficacy of drug therapy beforehand.
At present, there are only two studies on the clinical therapeutic efficacy in different subtypes of adenomyosis. Matsubara et al. 24 found that intrinsic subtype patients were prone to progesteronerelated serious unpredictable bleeding. On the other hand, extrinsic subtype patients responded well to progesterone therapy.
The life quality after progesterone treatment in intrinsic patients was poorer than that in extrinsic patients, suggesting differences in the efficacy and safety of progesterone treatment. We found that the effective rate of progesterone treatment in the intrinsic subtype was significantly lower than that in the extrinsic subtype.
Meanwhile, the penetrating subtype had a rate between that of intrinsic and extrinsic subtypes. Chen et al. 18  was the worst (37, 57.7%) and the shedding rate the highest (32, 50%) in indeterminate subtype (n = 64). We found that the effective rates of LNG-IUD in intrinsic and penetrating subtypes were lower than those in extrinsic subtype. The descending/shedding rate of LNG-IUD was higher than that in the extrinsic subtype.
Our findings on the efficacy of LNG-IUD of different subtypes are not completely consistent with the previous study. These discrepancies may be ascribed to the small sample size, recall bias of retrospective studies, and the short time-course of follow up.
However, based on the theory that the expression of progesterone receptors in intrinsic patients is lower, 14 the intrinsic subtype should be less sensitive to LNG-IUD than the extrinsic subtype. They found that in the glandular and stromal cells of intrinsic subtype patients, progesterone receptor-immunostained cells were significantly fewer than the estrogen receptor-stained cells (P < 0.05). On the other hand, progesterone and estrogen receptor expression was not different in the extrinsic subtype. Therefore, the intrinsic subtype has lower progesterone receptor expression.
In our study, the progesterone therapeutic efficacy in the intrinsic subtype was worse than that in the extrinsic subtype. These find-  . Further study is necessary to explore the molecular mechanism of the pathogenesis of adenomyosis, and to find potential therapeutic targets for new targeted drugs for those patients in whom progestogen is ineffective.
In conclusion, there are significant differences in age, dysmenorrhea, menorrhagia, and infertility among patients with different subtypes of adenomyosis. It is speculated that most penetrating adenomyosis has developed from the extrinsic subtype. The effective rate of progesterone in the intrinsic subtype is significantly lower than that in the extrinsic subtype. Current treatment methods cannot produce satisfactory results in women planning pregnancy. Therefore, MRI can be used to predict the clinical symptoms, disease severity, and therapeutic efficacy. MRI is helpful for selecting an individualized therapeutic plan for patients with adenomyosis.

AUTH O R CO NTR I B UTI O N S
Xiaotong Han and Xinran Gao contributed to the conception and design of the study, acquisition of data, analysis and interpretation of data, and drafting the article; Feng Wang, Zhongyu Liu, and Chunliang Shang contributed to acquisition of data; and Hongyan Guo contributed to the conception and design of the study. All authors revised the article critically for important intellectual content, gave final approval of the version to be submitted, and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

ACK N OWLED G M ENTS
We thank the Beijing Association for Science and Technology-Young Elite Scientist Sponsorship Program By Bast for the sources of funding (BYESS2022031). We are grateful to Yan Zhou for reviewing the MRI scans and to Molin Wang and Lu Liu for collecting medical records.

CO N FLI C T O F I NTE R E S T
The authors have no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
Research data are not shared.