Evaluating the development of endometriosis and adenomyosis lesions over time: An ultrasound study of symptomatic women

All authors contributed to the study design. Ligita Jokubkiene and Povilas Sladkevicius performed initial ultrasound examinations, and Ligita Jokubkiene performed follow‐up ultrasound examinations. All authors interpreted data. Sofie Orlov performed statistical analyses and drafted the manuscript with supervision from Povilas Sladkevicius and Ligita Jokubkiene. All authors contributed to reviewing and editing the manuscript. All authors approved the final submitted version of the manuscript.

The study received approval from the Regional Ethical Review Board of Lund, Sweden, on September 11, 2018, with reference number 555/2018. Women were given written and verbal information about the study. Written consent to participate was obtained and could be withdrawn at any point.

The recruitment of women is illustrated in a flowchart (Figure  1 ). A total of 178 women with symptoms suggestive of endometriosis or adenomyosis and normal initial ultrasound examination in 2014–2017 were eligible for this study. Among them, 78 women were excluded: 30 declined participation without providing a reason, 23 did not respond to invitation, 16 expressed interest but were unable to participate due to various reasons, 5 had undergone hysterectomy, 3 were pregnant above 12 weeks of gestational age, and 1 woman withdrew her consent. Thus, 100 women were included in this study. Demographic data for the included women at follow‐up are presented in Table  1 . Flowchart of the recruitment of women to the study. Demographic data of the included women at follow‐up ultrasound examination. Note : Data are given as n (%) or median (interquartile range). Abbreviations: COC, combined oral contraceptives; GnRH, gonadotropin‐releasing hormone; LNG‐IUS, levonorgestrel‐releasing intrauterine system. p ‐values are given from comparison of women without ( n  = 87) and with endometriosis or adenomyosis lesions at ultrasound ( n  = 13) at follow‐up, using Mann–Whitney U test for continuous variables and Chi‐square test or Fisher's exact test for categorical variables. (1) In women without endometriosis or adenomyosis lesions, LNG‐IUS in combination with systemic progestin (three women), COC (one woman), or GnRH analog (one woman), and (2) in women with endometriosis or adenomyosis lesions, combination of LNG‐IUS with systemic progestin (one woman). Menopausal hormone therapy in one woman and bioidentical progesterone in one woman. The median age of women with endometriosis or adenomyosis lesions at follow‐up ultrasound examination was higher compared to women without lesions (42 years [IQR 37–45] vs. 31 years [IQR 27–37], p  < 0.001). No significant differences were observed in other demographic variables between these groups. Two women were pregnant, one woman in gestational week 5, and one in gestational week 7, at the time of the follow‐up ultrasound examination. None of the included women underwent surgery for ovarian or deep endometriosis during follow‐up. Ultrasound findings of endometriosis or adenomyosis lesions are presented in Table  2 and Figure  2 . At follow‐up ultrasound examination, endometriosis or adenomyosis lesions were found in 13/100 women, constituting a prevalence of 13% (95% CI 7.1–21.2). Endometriosis lesions were found in 8/100 women (8% [95% CI 3.5–15.2]). Endometrioma was found in 7/100 women (7% [95% CI 2.9–13.9]), with a median maximal diameter of 10 mm (range 6.8–21.0). Deep endometriosis in 4/100 women (4% [95% CI 1.1–10.0]): 1 woman had isolated deep endometriosis in the vaginal wall with maximal diameter of 9 mm, 3 women had unilateral deep endometriosis in the uterosacral ligament with a median maximal diameter of 12 mm (range 3.0–17.0), and of these, 1 woman had an additional lesion in rectum with a maximal diameter of 16 mm. Adenomyosis was found in 6/100 women, representing a prevalence of 6% (95% CI 2.2–12.6). Direct and indirect features of adenomyosis are shown in Table  3 . A total of nine women (9% [95% CI 4.2–16.4]) had endometriosis or adenomyosis lesions in only one site. Fluid in pouch of Douglas was observed in 38/100 women (38% [95% CI 28.5–48.3]), with a median diameter of 12.1 mm (IQR 6.9–17.5). No superficial endometriosis lesions were found. Ultrasound findings of endometriosis and/or adenomyosis lesions at follow‐up ( n  = 100). Endometrioma and deep endometriosis in the uterosacral ligament in two women and endometrioma, deep endometriosis in the uterosacral ligament and in the rectum in one woman. Deep endometriosis in the vaginal wall. One woman had concomitant endometrioma. Development of endometrioma over time in the left ovary: (A) normal ovary at initial ultrasound examination and (B) ovary with endometrioma at follow‐up 7 years later. Ultrasound features of adenomyosis according to revised definitions by The Morphological Uterus Sonographic Assessment group at follow‐up. Note : Data are given as n (%) and calculated from women included in this study ( n  = 100). During the follow‐up examination, no endometriosis or adenomyosis lesions were found in 87/100 women (87% [95% CI 78.8–92.9]), 77/100 (77% [95% CI 67.5–84.8]) had normal ultrasound findings, 6/100 (6% [95% CI 2.2–12.6]) had only indirect features of adenomyosis, and 4/100 (4% [95% CI 1.1–10.0]) had other pathological findings, including polycystic ovaries ( n  = 2), uterine fibroid ( n  = 1), and ovarian cystadenoma ( n  = 1). Of 13 women with endometriosis or adenomyosis lesions at follow‐up, 12 women (92%) were older than 35 years. Seven women (54%) were on hormonal treatment and six women (46%) were not on hormonal treatment at both examinations. There was no difference in hormonal treatment at both initial and follow‐up examinations between women with endometriosis or adenomyosis lesions compared to those without lesions (7/13 [54%] vs. 52/87 [60%] and 7/13 [54%] vs. 43/87 [50%], p  = 0.685 and p  = 0.766, respectively). In total, 38/100 women (38%) were on hormonal treatment during both examinations (duration, median: 36 months [IQR10–75]), 29/100 women (29%) were not on hormonal treatment at both examinations, 12/100 women (12%) began hormonal treatment during the follow‐up period (duration, median: 21 months [IQR 3–56]), and 21/100 women (21%) discontinued hormonal treatment during the follow‐up period. Table  4 shows the presence and intensity of symptoms at initial and follow‐up ultrasound examinations. At initial ultrasound examination, there were no differences in the presence or intensity of symptoms between the two groups of women who subsequently developed visible endometriosis or adenomyosis lesions at follow‐up ultrasound and those who did not. Symptoms at initial and follow‐up ultrasound examination and alteration over time. Note : Data are given as n (%) and median (interquartile range). Abbreviation: VAS, visual analog scale. p ‐values are given for comparison between women without ( n  = 87) and with endometriosis or adenomyosis lesions at ultrasound ( n  = 13) at initial (a) and follow‐up (b) examinations, using Mann–Whitney U test for continuous variables and Chi‐square test for categorical variables. If the expected count in any group was less than five, either Fisher's exact test or Fisher–Freeman–Halton exact test was used for categorical variables. Severity was assessed using VAS and categorized as mild (0–30 mm), moderate (31–70 mm), and severe (71–100 mm). p ‐values are given for comparison for the severity categories (severe, moderate, and mild) and for severe vs non‐severe and mild vs non‐mild, respectively. Alteration of symptoms was described using minimal clinical important difference (MCID) of 10 mm between initial and follow‐up examination. Alterations of 10 mm or more were described as decreased or increased intensity and alterations below 10 mm were described as unchanged intensity. Heavy menstrual bleeding was assessed only at follow‐up and considered present if the woman experienced at least one of the following: bleeding more than 7 days, bleeding with clots, needing to change sanitary protection at least every other hour more than once, or needing to use double sanitary protection. At follow‐up, women with endometriosis or adenomyosis lesions reported lower intensity of dysmenorrhea and chronic pelvic pain compared to women without lesions (48 mm [IQR 16–79] vs. 73 mm [IQR 46–85] and 45 mm [IQR 26–57] vs. 57 mm [IQR 36–75], p  = 0.087 and p  = 0.026, respectively). Additionally, severe dysmenorrhea tended to be less prevalent in women with endometriosis or adenomyosis lesions compared to women without lesions (4/13 women [31%] vs. 46/87 women [53%], p  = 0.137). None of the women with endometriosis or adenomyosis lesions at follow‐up reported severe chronic pelvic pain or severe dyschezia, whereas 21/87 women (24%) and 20/87 women (23%) without endometriosis or adenomyosis lesions did ( p  = 0.064 and p  = 0.065). No differences were observed in other symptoms between women with and without endometriosis or adenomyosis lesions at follow‐up. In women with endometriosis or adenomyosis lesions, the intensity of dysmenorrhea was lower at follow‐up compared to initial examination (48 mm [IQR 16–79] vs. 65 mm [IQR 33–88], p  = 0.002, respectively). No significant differences were observed in the intensity and presence of other symptoms. In women without endometriosis or adenomyosis lesions, the presence of dysuria and dyschezia was lower at follow‐up compared to the initial examination (28/87 women [32%] vs. 51/87 women [59%], p  < 0.001, and 56/87 women [64%] vs. 68/87 women [78%], p  = 0.017, respectively). There were no differences in the presence or intensity of other symptoms between the two examinations in this group of women. Alterations of symptoms over time are shown in Table  4 . Alterations of the intensity of dysmenorrhea over time were significantly different between women with endometriosis or adenomyosis lesions and women without lesions (in women with lesions: intensity decreased in 7/13 women [54%], increased in 0/13 women [0%], and was unchanged in 6/13 women [46%] vs. in women without lesions: intensity decreased in 35/86 women [40%], increased in 32/86 women [37%], and was unchanged in 19/86 women [22%], p  = 0.019). None of the women with endometriosis or adenomyosis lesions reported an increase in the intensity of dysmenorrhea between the initial and follow‐up examinations, in contrast to 32/86 women (37%) without lesions ( p  = 0.008). Increased intensity of chronic pelvic pain between initial and follow‐up examinations tended to be less common in women with endometriosis or adenomyosis lesions compared to women without lesions (3/13 women [23%] vs. 35/86 women [41%], p  = 0.223). At follow‐up, the use of analgesics was less frequent in women with endometriosis or adenomyosis lesions (8/13 women [62%] vs. 11/13 women [85%], p  = 0.250) and women without lesions (63/87 women [72%] vs. 76/87 women [87%], p  = 0.007) compared to the initial examination. Neither pregnancy nor delivery between the ultrasound examinations was associated with differences in the presence, severity, intensity, or alterations of symptoms over time, except for the intensity of dyschezia which was higher in women who were pregnant compared to those who were not pregnant between the ultrasound examinations (64 mm [IQR 52–87] vs. 49 mm [IQR 22–75], p  = 0.031). Women with hormonal treatment at follow‐up reported lower intensity of dysmenorrhea (52 mm [IQR 23–77] vs. 81 mm [IQR 68–88], p  < 0.001), chronic pelvic pain (47 mm [IQR 21–61] vs. 63 mm [IQR 49–82], p  < 0.001), and dyschezia (48 mm [IQR 15–69] vs. 68 mm [IQR 51–86], p  = 0.011) compared to women without hormonal treatment. Severe dysmenorrhea at follow‐up was less common in women with hormonal treatment compared to women without (16/50 women [32%] vs. 34/50 women [68%], p  < 0.001).

We have found that 13% of women with symptoms suggestive of endometriosis or adenomyosis and normal initial ultrasound examination develop visible endometriosis or adenomyosis lesions during long‐term follow‐up. At follow‐up ultrasound examination, 8% of women had endometriosis and 6% of women had adenomyosis. Presence and intensity of symptoms at initial examination were similar between women who subsequently developed endometriosis or adenomyosis lesions and those who did not. At follow‐up, women who developed endometriosis or adenomyosis lesions reported lower intensity of dysmenorrhea and chronic pelvic pain compared to those without lesions. This study is the first prospective long‐term follow‐up study of women with symptoms suggestive of endometriosis, being examined by an experienced ultrasound examiner twice with at least 5 years intervals to assess possible development of visible endometriosis and adenomyosis lesions over time. The study population consisted of a selected group of symptomatic women, who had normal initial ultrasound examinations excluding the presence of visible endometrioma and deep endometriosis in a previous study. 13 Age seems to be an important factor in the development of visible lesions. Almost all women who developed endometriosis or adenomyosis lesions were older than 35 years, indicating a delay in the development of visible lesions at ultrasound examination. Age can also be an important factor in the young population; higher prevalence of endometrioma, deep endometriosis, and adenomyosis being reported in young women aged 21–25 years compared to those aged 12–16 years. 24 Age has been demonstrated to correlate with the prevalence of adenomyosis. 25 Endometriosis is estimated to affect 5%–10% of women of reproductive age. 26 However, there are no screening studies on the general population, and the reported prevalences are those in selected populations. Previously, we reported a prevalence of endometriosis being 25% among symptomatic women attending our Ultrasound Unit, 13 a finding consistent with similar prevalence observed in women with subfertility undergoing their first in vitro fertilization treatment. 27 In this study, we focused on symptomatic women without previously detectable endometriosis lesions. The observed finding of 13% of women with endometriosis or adenomyosis is in agreement with the prevalence in general population. However, due to the specific patient cohort selection, direct comparisons are challenging. It is difficult to compare ultrasound‐based and surgery‐based studies due to limitations in both modalities for detecting different types of endometriotic lesions. In surgery‐based studies, the prevalence of deep endometriosis can be underestimated if the lesions are small or because of severe adhesions in the pelvis. 7 , 10 Since there are no ultrasound‐based studies evaluating development of endometriosis or adenomyosis lesions over time in non‐pregnant women without abnormal initial findings, comparisons can be made only with surgery‐based studies. In a study by Knox et al., 28 19% of women with dysmenorrhea during adolescence were diagnosed with mild superficial endometriosis during laparoscopy as adults. These women were more likely to experience severe dysmenorrhea, intermenstrual pain, dyspareunia, and dyschezia compared to women who did not undergo laparoscopy during the follow‐up period of 4–14 years (mean 10.24 years). The differences in prevalence and symptoms observed in comparison to our study may be attributed to the detection of different phenotypes of endometriosis (superficial vs. endometrioma, deep endometriosis, or adenomyosis) due to different diagnostic methods, as well as differences in the assessment of symptom severity. Contrary to our study, Sasamoto et al. 29 found that severe dysmenorrhea and acyclic pelvic pain were more common in women with endometriosis confirmed at surgery compared to women without known endometriosis, both at a baseline and follow‐up. Differences might be explained by phenotypes of endometriosis and their possible effect on symptom severity. In the study by Sasamoto et al., 29 women had mainly superficial endometriosis observed at surgery, while women in our study had endometrioma, deep endometriosis, or adenomyosis at ultrasound. However, in agreement with our study, Sasamoto et al. 29 found no increase in severe dysmenorrhea over time in women with endometriosis. Differences in symptomatology could be important for the clinical management and assessment of women presenting with symptoms suggestive of endometriosis. Symptoms suggestive of endometriosis lack specificity and may overlap with those of other conditions. It is possible that some women may have had endometriosis that was not detectable during the initial ultrasound examination. Alternatively, these symptoms could be related to other conditions or simply be indicative of primary dysmenorrhea. Nevertheless, it is important to consider the possibility of endometriosis in women with dysmenorrhea, pelvic pain, dyspareunia, dysuria, and dyschezia to promote early detection and hormonal treatment. 9 However, the symptom intensity should not be the leading factor for follow‐up. An important strength of this study is the systematic ultrasound examinations conducted and reported in accordance with international guidelines 11 , 12 , 21 and being performed by experienced ultrasound examiners with expertise in endometriosis and adenomyosis. This study is the first to assess development of endometriosis or adenomyosis lesions in symptomatic women at ultrasound, thereby contributing novel insights to the field. A limitation of this study is relatively high rate of loss to follow‐up. Women could not participate in follow‐up due to various reasons, potentially introducing selection bias. It remains uncertain whether women who declined participation had more or less severe symptoms, underwent surgical or hormonal treatment for endometriosis, were diagnosed with other diseases causing symptoms, or declined participation due to unwillingness to undergo additional ultrasound examination without initial abnormal findings. Additionally, some women relocated from the area. Furthermore, lack of information about surgery, other than hysterectomy, oophorectomy, or for advanced endometriosis, performed during the follow‐up period is a limitation. However, none of the included women underwent surgery for ovarian or deep endometriosis during follow‐up. Another limitation is the possible effect on the development of endometriosis or adenomyosis during pregnancy and hormonal treatment during follow‐up. However, we did not observe any differences in hormonal treatment and pregnancies or deliveries between women who developed endometriosis or adenomyosis lesions and those who did not. The perfect circumstances to study the natural development of endometriosis would be in women without any hormonal treatment or pregnancy during the study period. However, it would be neither possible nor ethically justifiable to prohibit women from using hormonal treatment or becoming pregnant during follow‐up. Furthermore, it would be interesting to study disease development in asymptomatic women. The limitation of ultrasound in detecting superficial endometriosis 9 could result in the underestimation of endometriosis presence during both ultrasound examinations. Unfortunately, performing laparoscopy solely for research purposes in these women is not ethically justifiable, although would aid in confirming or excluding the presence of superficial endometriosis. Some studies suggest that presence of fluid in the pouch of Douglas can facilitate assessment of superficial endometriosis 30 and ultrasound features have recently been proposed for this purpose. 31 Our study population is relatively small, and the results have to be validated in larger studies. However, this study can serve as a basis for sample calculations in further investigations.

Our findings suggest that one in eight symptomatic women with normal ultrasound examination may develop endometriosis and adenomyosis lesions visible at ultrasound over time. However, majority of women remain having normal ultrasound examinations despite symptoms. Symptom intensity at initial ultrasound examination and exacerbation of dysmenorrhea or chronic pelvic pain during follow‐up were not associated with the development of endometriosis or adenomyosis lesions visible at ultrasound, suggesting that even women with less severe symptoms might benefit from a follow‐up ultrasound when indicated.

The natural course of endometriosis is complex and not fully understood. The diagnostic delay of endometriosis is approximately 7–12 years, where women with earlier onset of symptoms and with more severe symptoms, experience longer diagnostic delays. 1 , 2 , 3 , 4 Early detection and treatment of endometriosis, especially in adolescents and young women, is important to reduce the risk of disease progression and diminished quality of life. 4 , 5 Transvaginal ultrasound is the first‐choice diagnostic method for endometriosis with high accuracy for endometrioma and deep endometriosis when performed by an experienced examiner. 6 , 7 , 8 , 9 Deep endometriosis lesions can be missed during laparoscopy, particularly in cases of small lesions, or obliterated the pouch of Douglas. 7 , 10 The International Deep Endometriosis Analysis group consensus protocol provides recommendations on how to perform and describe ultrasound examinations for endometriosis in the pelvis. 11 The revised consensus by The Morphological Uterus Sonographic Assessment group describes direct and indirect ultrasound features of adenomyosis. 12 In a previous study, 13 we found that a substantial amount (48%) of women with symptoms suggestive of endometriosis had normal ultrasound findings despite being examined by experienced ultrasound examiners in endometriosis. To the best of our knowledge, there are no published studies on the development of visible endometriosis or adenomyosis lesions at ultrasound examination over time in women with symptoms suggestive of endometriosis and normal initial ultrasound examination. In adolescents and/or young women, endometriosis and/or adenomyosis lesions on ultrasound examination and magnetic resonance imaging are more common with increasing age. 14 , 15 The primary objective of this study was to investigate if symptomatic women with a previous normal ultrasound examination develop visible endometriosis or adenomyosis lesions during follow‐up ultrasound examination after at least 5 years. The secondary objective was to assess if the presence and intensity of symptoms changed over time and if they were related to ultrasound findings at follow‐up.

The authors have no conflict of interest to declare.

This prospective observational cohort study was conducted at the Ultrasound Unit, Department of Obstetrics and Gynecology, Skane University Hospital, Malmo, Sweden, a tertiary care center. Eligible women were those referred to our unit by regional hospitals or private practitioners in the period between 2014 and 2017 with symptoms suggestive of endometriosis or adenomyosis, at least dysmenorrhea, but who had normal findings at first ultrasound examination. These women were identified in our previous study 13 and invited to participate in the current study via an invitation letter containing information about the follow‐up study. Follow‐up ultrasound examinations were performed between February 2022 and December 2022. Exclusion criteria included a history of hysterectomy and/or oophorectomy, and pregnancy above 12 weeks of gestational age at the time of ultrasound examination. Demographic data, including age, parity, pregnancy, and/or delivery during the follow‐up period, history and duration of infertility, current and previous hormonal treatment, duration of hormonal treatment, and analgesic usage, were collected at both ultrasound examinations. Prior to ultrasound examination, women reported the presence and intensity of symptoms suggestive of endometriosis or adenomyosis, including dysmenorrhea, chronic pelvic pain, dyspareunia, dysuria, and dyschezia, with and without hormonal treatment using a questionnaire developed and routinely used at our Ultrasound Unit. The intensity of symptoms was assessed using a visual analog scale (VAS) ranging from 0 to 100 mm, categorized as mild (0–30 mm), moderate (31–70 mm), and severe (71–100 mm). 16 Self‐assessment of heavy menstrual bleeding was reported only at follow‐up and was considered to be present if at least one of the following was experienced: bleeding lasting more than 7 days, presence of clots in the bleeding, necessity to change sanitary pads at least every other hour more than once, or need to use double sanitary pads. 17 , 18 , 19 Alterations of symptoms were assessed by comparing the presence and the intensity of symptoms assessed using VAS and described using the minimal clinically important difference (MCID) of 10 mm between initial and follow‐up examination. 20 Alterations of 10 mm or more were described as decreased or increased intensity and alterations below 10 mm were described as unchanged intensity. The initial ultrasound examinations were performed between 2014 and 2017 by one of three examiners with expertise in advanced ultrasound examination of endometriosis. All follow‐up ultrasound examinations were performed by one (LJ) of these examiners. Ultrasound examinations were performed systematically as described in our previous publication 13 and in accordance with the International Deep Endometriosis Analysis group consensus protocol 11 using Voluson E8 (2014–2015) and E10 ultrasound equipment (GE, Milwaukee, WI, USA) with volumetric vaginal transducer of 5–9 MHz. Abnormal findings were measured in three orthogonal planes. Ovarian cysts, including endometrioma, were described according to the terminology proposed by the International Ovarian Tumor Analysis group. 21 The typical appearance of an endometrioma was described as a unilocular cyst with ground‐glass echogenicity and no‐to‐moderate vascularization. 22 Deep endometriosis lesions were described according to the International Deep Endometriosis Analysis consensus. 11 The peritoneum in the pouch of Douglas was examined for superficial endometriosis, using pattern recognition to identify hypoechoic lesions that disrupt the normal appearance of the peritoneum. 23 Adenomyosis was assessed using both 2D and 3D ultrasound and described according to the revised definitions of The Morphological Uterus Sonographic Assessment features of adenomyosis. 12 Direct features of adenomyosis included myometrial cysts, hyperechoic islands, and echogenic subendometrial lines and buds. Indirect features included a globular uterus, asymmetrical myometrial thickening, fan‐shaped shadowing, translesional vascularity, irregular junctional zone, and interrupted junctional zone. 12 We defined adenomyosis as the presence of at least one direct feature. Women with only indirect features of adenomyosis were assigned to the group of women without endometriosis or adenomyosis lesions in statistical analysis. The primary outcome of this study was the presence of visible endometriosis or adenomyosis lesions at follow‐up ultrasound examination. Secondary outcomes were symptom intensity and changes in symptoms over time in women who developed visible endometriosis or adenomyosis lesions and those who did not. Statistical analyses were conducted using IBM SPSS Statistics version 27 (IBM, Armonk, NY, USA). Prevalence at follow‐up was reported as counts and percentages with 95% confidence interval (CI). Given the non‐normally distribution of continuous variable, descriptive statistics were expressed as medians and interquartile ranges (IQR). Comparisons between women with and without endometriosis or adenomyosis lesions at ultrasound were performed to analyze the presence and severity of symptoms both at initial and follow‐up examination, as well as to assess alterations of symptoms over time. Symptom alterations between the initial and follow‐up examinations within each group were reported separately. Differences in non‐related continuous variables were evaluated using the Mann–Whitney U test. Categorical variables were compared using the Chi‐squared test; if the expected count in any group was less than five, either Fisher's exact test or Fisher–Freeman–Halton exact test was used. For related samples, categorical variables were assessed using the McNemar test, while continuous variables were assessed using Wilcoxon signed‐rank test. p ‐values <0.05 were considered statistically significant.