THE PRESURGICAL DIAGNOSIS OF DIFFUSE ADENOMYOSIS
By Helen Bickerstaff, MB, BChir Institute of Obstetrics and Gynaecology,
Royal Postgraduate Medical School, Hammersmith Hospital, London.

Introduction

Adenomyosis is a disease occurring at the interface between the endometrium and the myometrium. It is characterised by the progressive penetration of endometrial glands and stroma into the myometrium together with by smooth muscle hyperplasia and an altered local immune environment.

Adenomyosis is an important cause of dysmenorrhoea, menorrhagia and possibly subfertility. Until recently, the diagnosis could only be made after hysterectomy by which time its clinical relevance was merely academic. Based on the study of hysterectomy specimens it has become widely accepted that adenomyosis mainly affects women in the late reproductive years and perimenopause. However, recent studies using non-invasive diagnostic imaging techniques have indicated that adenomyosis may not be uncommon in symptomatic women in their twenties and thirties (1). Accurate presurgical diagnosis is also the most important prerequisite for the development of effective medical and function preserving surgical treatments (2). This review summarises the merits and demerits of available diagnostic techniques.

Histological diagnosis of adenomyosis

The definition of adenomyosis, namely the presence of endometrial foci in the myometrium, appears straightforward but the precise histopathological criteria have been the subject of considerable debate. The controversy focuses on the microscopic criteria which define superficial adenomyosis, islets within the inner third of the myometrium, which can sometimes be indistinguishable from normal indentations of the basal endometrial layer. The problem is that there is no agreement on a minimum depth of myometrial invasion, with current criteria ranging widely from the presence of glands and stroma 1 mm below the endometrial-myometrial junction to as deep as one third of the total myometrial thickness (3).

The present confusion is clearly illustrated by the findings of the Maryland WomenÕs Health Study which analysed 1114 hysterectomy reports from 15 hospitals, and 705 reports signed by 25 pathologists. The frequency of diagnosis of adenomyosis ranged from 12% to 58% among the hospitals, and 10% to 88% among the 25 pathologists. This wide variation could not be explained by differences in patient age, parity or other factors known to correlate with the incidence of adenomyosis (4).

The histological detection rate of adenomyosis is also dependent on the number and site of myometrial samples examined. This was elegantly illustrated by Bird and co-workers who found adenomyosis in 31% of routine histological sections of 200 consecutive hysterectomy specimens. However, if six extra myometrial blocks were examined, the incidence rose sharply to 61% (5).

Adenomyosis is often classified into a diffuse and focal form. Focal adenomyosis, also termed adenomyoma, refers to isolated implants surrounded by extensive but focal smooth muscle changes. It is less common than diffuse adenomyosis and not accompanied by a characteristic distortion of the inner myometrium (1). It remains to be determined whether these lesions are really related to diffuse adenomyosis or represent a different pathological process such as metaplasia of Müllerian rests.

Clinical examination

It is often stated that a progressive, uniform enlargement of the uterus accompanied by menorrhagia and dysmenorrhoea is diagnostic of uterine adenomyosis. Consequently, repeated bimanual examinations, over several months, just before and after menstruation have been recommended to detect fluctuating changes in contour, size and consistency of the uterus.

In only 10% to 20% of cases is the condition diagnosed pre-operatively. The pathological confirmation of clinically suspected cases is also low (10% to 38%) and, on basis of this low verification rate, some have argued that the clinical diagnosis of adenomyosis can no longer be considered as a reliable pre-operative indication for hysterectomy (3,6).

Endoscopy and hysterography

These techniques may sometimes indicate the presence of extensive adenomyosis. The most characteristic feature of adenomyosis on hysterography is the presence of ill defined areas of contrast intravasation extending perpendicularly from the uterine cavity into the myometrium. Unfortunately, the sensitivity of this technique is too low for clinical practice; Marshak and Eliasoph established the diagnosis on hysterography in only 38 of 150 cases with pathologically proven disease (7).

It is often claimed that hysteroscopy is an accurate technique for the diagnosis of adenomyosis but well controlled studies are still lacking. Diffuse myometrial distortion detected at the time of laparoscopy or hysteroscopy may indicate extensive adenomyosis but may also be caused by multiple small fibroids. It is unlikely that mild or moderate adenomyosis can be diagnosed visually. This was clearly demonstrated by McCausland who found adenomyosis in biopsy specimens in 33 of 50 symptomatic women who had a visually normal uterine cavity at hysteroscopy (8).

Myometrial biopsy

Myometrial needle biopsies can be taken transabdominally at the time of laparoscopy, or transvaginally under ultrasound guidance. A positive biopsy should provide evidence of ectopic endometrial islets sandwiched between strips of myometrium. Endometrial glands and stroma at the extreme end of the needle core may represent eutopic endometrium and such biopsies should be regarded as negative.

The sensitivity of random needle biopsies is low and dependent on the number of biopsies and the depth and extent of mucosal infiltration. A recent study using hysterectomy specimens reported that if four 14 gauge needle biopsies were taken at random, one or more would be positive in 70% of uteri with deep adenomyosis (9). If only superficial disease was present the sensitivity of this technique dropped to 5%. These findings are similar to a larger study by Popp et al. who took not only needle biopsies immediately after hysterectomy but also at the time of laparoscopy as well as transvaginally under ultrasound guidance (10). A single myometrial biopsy picked up only 8% to 19% of women with adenomyosis. The sensitivity of random needle biopsy is therefore too low for clinical practice.

A myometrial strip can be resected hysteroscopically. McCausland advocated this technique and demonstrated that it was possible to diagnose superficial adenomyosis histologically on biopsies measuring approximately 2 cm in length and 3 to 5 mm in depth (8). This technique is unsuitable for those women who still want to have children. Its accuracy has not yet been compared to conventional histological assessment of hysterectomy specimens.

CA 125

There is evidence that supports a role for the immune system in the pathogenesis of adenomyosis. Ota and co-workers have demonstrated that adenomyosis is associated with increased numbers of myometrial macrophages, elevated antiphospolipid auto-antibodies and CA 125 levels in peripheral blood, and deposition of IgG, C3 and C4 in ectopic foci (11). Adenomyotic uteri are also characterised by increased HLA-DR antigen expression in both eutopic and ectopic glandular cells.

Peripheral CA 125 levels are potentially useful as a serum marker for adenomyosis. Kobayashi and co-workers found that CA 125 antigens present on adenomyotic epithelial cells have a different molecular mass from those present on eutopic endometrium; the antibody binding site is however the same (12). If an antibody unique to adenomyosis could be isolated and purified then a highly specific serum screening test could be developed.

Ultrasonography

Several studies have addressed the potential role of ultrasound in the diagnosis of adenomyosis. The enhanced resolution of transvaginal ultrasound makes it superior to the transabdominal approach. Several sonographic criteria for diffuse adenomyosis have been described and are summarised in Table 1.

Table 1 ULTRASOUND CHARACTERISTICS OF ADENOMYOSIS. • ill defined hypoechoic areas • hetrogeneous myometrial echotexture • small anechioc lakes • asymetrical uterine enlargement • indistinct endometrial-myometrial border • subendometrial halo thickening

Brosens and co-workers assessed ultrasonographic details such as uterine dimensions, symmetry and echogenicity of the myometrium in order to determine which features were most predictive of adenomyosis (13). Fifty six symptomatic patients were recruited, 28 of whom were later confirmed to have adenomyosis, either by MR imaging or by histology. Although adenomyotic uteri tended to be larger and more asymmetrical than nonadenomyotic uteri this did not reach statistical significance. The disease was best predicted on basis of ill-defined heterogeneous echotexture within the myometrium.

In extensive disease, small anechoic lakes of 1 to 3 mm diameter are sometimes detected (14). Lakes of this size are at the limit of ultrasonographic resolution and may represent larger cystic implants or, more likely, focal myometrial haemorrhage. Detecting such areas of heterogeneous echogenicity within the myometrium is often complicated by the presence of various artefactual echogenic shadows and the confusion is compounded by the presence of leiomyomata and vascular calcifications.

Five further studies have compared the accuracy of pre-operative endovaginal ultrasonography with the results of the histological examination of hysterectomy specimens (14-18). These results are summarised in Table 2 and indicate that at present the diagnostic accuracy of this imaging technique remains unknown. It is important to remember that the sonographic signs are subtle, that the technique is strongly operator dependent and that different histological criteria and sampling techniques were used in these studies. We believe that in experienced hands, ultrasound is a useful tool, especially when the clinical suspicion is high.

Table 2 Accuracy of endovaginal ultrasonography in the diagnosis of diffuse adenomyosis.
	Prevelence	Sensitivity	Specificity	PPV	NPV
	(%)	(%)	(%)	(%)	(%)
Reinhold et al. (1996)	18/119 (24)	89	89	71	96
Atzori et al. (1996)	15/175 (86)	86	96.2	68.4	98
Reinhold et al. (1995)	29/100 (29)	86	86	71	94
Brosens et al. (1995)	28/56 (50)	53	75	86	77
Asher et al. (1994)	17/20 (85)	86	50	90	20
Fedele et al. (1992)	22/43 (51)	80	74	73	81
PPV: positive predictive value

MR imaging

Adenomyosis is characterised by disruption of the normal myometrial zonal anatomy which is best demonstrated on T2-weighted images where the contrast between the low signal inner myometrium (junctional zone) and the intermediate signal outer myometrium is maximal. The myometrial zonal anatomy is dependent on gonadal hormones. In premenarchal girls and postmenopausal women, it is often indistinct with a comparatively low signal intensity from both the outer myometrium and junctional zone. Ovarian suppression with gonadotrophin-releasing hormone analogues leads to an MR appearance of the uterus mimicking that of postmenopausal women; while hormone replacement therapy in postmenopausal women results in the re-appearance of myometrial zonal anatomy (19).

Diffuse adenomyosis is characterised on T2-weighted images by an irregular and diffuse thickening of the junctional zone sometimes with underlying high signal foci. Pathologically this has been shown to represent smooth muscle hyperplasia characterised by closely packed smooth muscle fibres that are poorly orientated and less vascular than the smooth muscle of the normal inner myometrium (20). It is the smooth muscle changes that are easily recognised by MR rather than foci of heterotopic glandular epithelium and stroma.

There is little doubt that MR imaging is currently the best technique for the presurgical diagnosis of adenomyosis and most studies have reported very high positive and negative predictive values (20,21). However, there is growing disagreement in the literature on what constitutes abnormal junctional zone thickening on MR imaging. Junctional zone widths between 6 and 12 mm have been quoted in the literature as diagnostic for adenomyosis (17). The current confusion has arisen because most studies have attempted to correlate the junctional zone thickness in vivo with the presence or absence of ectopic endometrial islets on histology. These studies are unlikely to be conclusive as there is as no consensus on the number of myometrial blocks that should be examined histologically or on the histological definition of superficial adenomyosis. Brosens and co-workers recently suggested that the myoproliferative changes may not always occur in concert with endometrial penetration of the myometrium and that it may be more sensible to correlate the junctional zone widths with an objective assessment of the patient's symptoms (22).

Conclusion

Adenomyosis should no longer be a retrospective diagnosis after hysterectomy. Both endovaginal ultrasonography and MR imaging are useful for the diagnosis. The use of MR imaging in routine practice has been limited by its restricted availability and cost. To some extent these factors are determined by the imaging time and, at present, most MR studies of the pelvis last for 30 minutes or longer. Advances in MR technology are likely to reduce the imaging time. This non-invasive technique is also of value in the diagnosis of ovarian, recto-vaginal and peritoneal endometriosis. Peritoneal lesions as small as 4 mm can be identified accurately with fat-saturating MR techniques (23). It is likely that this technique will replace laparoscopy as the investigation of choice for women suffering from pelvic pain and dysmenorrhoea.

The sonographic characteristics of adenomyosis are subtle and adenomyosis will remain undiagnosed if not considered. However, in experienced hands sonography can be almost as accurate as MR imaging. Finally, it should be remember that the MR or ultrasound diagnosis is based on recognising the distortion of the normal inner myometrial architecture caused by smooth muscle hyperplasia, and not on the presence of heterotopic implants.

REFERENCES

1. de Souza NM, Brosens JJ, Schweiso JE, Paraschos T, Winston RML The potential value of magnetic resonance imaging in infertility. Clin Radiology 1995; 50:75-9.
2. Wood C, Maher P, Hill D. Biopsy diagnosis and conservative treatment of adenomyosis. The Journal of the American Association of Gynecologic Laparoscpists 1994; 141:313-16
3. Azziz R. Adenomyosis: current perspectives. Obsetet Gynecol Clin North Am
4. 1989; 16: 221-35.
5. Bird CC McElin TW, Manalo-Eastrella P. The elusive adenmyosis of the uterus. Am J Obstet Gyyneacol 1972; 112: 583-93.
6. Seidman JD, Kjerulff KH. Pathological findings from the Maryland Womens Health Study - practice patterns in the diagnosis of of adenomysis. International journal of Gynecolological Pathology 1996, 15:217-221
7. Marshak RH, Eliasoph J. The roentgen findings in adenomyosis. Radiology 1955; 64:846-51
8. McCausland AM. Hysteroscopic myometrical biopsy: Its use in diagnosing adenomyosis and its clinical application. Am J Obstet Gynecol 1991; 166: 1619-28.
9. Brosens JJ, Barker FG. The role of myometrical needle biopsies in the diagnosis of adenomyosis. Fertil Steril 1995; 63 1347-9.
10. Popp LW, Schwiedessen JP, Gaetje R. Myometrial biopsy in the diagnosis of adenomyosis uteri. Am J Obstet Gynecol 1993; 169:546-9.
11. Ota H, Maki M, Shidara Y, Kodoma H, Takahashi H, Hayakawa M et al.. Effects of danazol at the immunologic level in patients with adenomoysis, with special reference to autoanyibodies: multicenter cooperative study. Am J Obstet Gynecol 1992; 167:481-6.
12. Kobayashi H, Ida W, Terao T, Kawashima Y. Molecular characteristics of the CA125 antigen produced by human endometrial epithelial cellls: comparison between eutopic and heterotopic epithelial cells. Am J Obstet Gynecol 1993; 169: 725-30.
13. Brosens JJ, de Souza NM, Barker FG, Paraschos T, Winstin RML. endovaginal ultrasonography in the diagnosos of adenomyosis uteri: identifying the predictive characteristics. Am J Obstet Gynecol 1995; 102: 471-4.
14. Fedele L, Bianchi S, Dorta M, et al.. Transvaginal ultrasonography in the diagnosis of diffuse adenomysis. Fertil Steril 1992;58:94-7.
15. Reinhold C, McCarthy S, Bret PM, Mehio A, Atri M, Zakarian R, Glaude Y, Liang LJ, Seymour RJ. Diffuse ademomyosis - comparison of endovaginal US and MR-imaging with histiopathic correlation. Radiology 1996: 199:151-8.
16. Atzori E, Tronci C, Sionis L. Transvaginal ultrasound in the diagnosis of diffus adenomyosis. Gynecologic and obstetric investigation 1996, 42:39-41.
17. Reinhold C, Atri M,Mehio A, Zakarian R, Aldis AE, Bret PM. Diffuse uterine adenomysis: morphologic criteria and diagnostic accuracy of endovaginal sonography. Radiology 1995: 197:609-14.
18. Ascher SM, Arnold LL, Patt RH, Schruefer JJ, Bagley AS, Semelka RCR, Zemen RK, Simon JA. Adenomyosis: prospective comparison of MR imaging and transvaginal sonography. Radiography 1994; 190: 803-6.
19. Hricak H, Alpers C, Crooks LE, Sheldon PE. Magnetic resonance imaging of the female pelvis: initial experience. Am J Radiol 1983; 141: 1119-28.
20. Togashi K, Ozasa H, Konishi I, Itoh H, Nishimura K, Fujisawa I, Noma S, Sagoh T, Minami S, Yamashita K, Nakano Y, Konishi J, Mori T. Enlarged uterus: differtation between adenomyosis and leiomyoma with MRI. Radiology 1989; 171:531-4.
21. Mark SA, Hricak H, Heinrichs LW, Hendrickson MR, Winkler ML, Bachica JA, Stickler JE. Adenomysis and Leiomyoma: differential diagnosis with MR imaging. Am J Radiology 1987; 163: 527-29.
22. Brosens JJ, de Souza NM, Barker FG. Uterine junctional zone: function and disease. The Lancet 1995; 356: 558-60.
23. Takahashi K, Okada S, Ozaki T, Kitao M, Sugimura K. Diagnosis of pelvic endometrosis by magnetic resonance imaging using "fat saturation" technique. Fertil.

---

Return to the Table of Contents

{FILE "e:\platform\iis\wwwroot\oblive\eago\_footers\footer.txt"}