- Email: [email protected]
Classification of Microhysteroscopic Images and their Correlation with Histologic Diagnoses
May 2003, Vol. 10, No. 2
The Journal of the American Association of Gynecologic Laparoscopists
Classification of Microhysteroscopic Images and their Correlation with Histologic Diagnoses Jorge E. Dotto, M.D., Baltasar Lema, M.D., Jorge E. Dotto, Jr., M.D., and Jacques Hamou, M.D. Abstract
(J Am Assoc Gynecol Laparosc 10(2):233–246, 2003)
Study Objective. To confirm that images observed at hysteroscopy correlate with histopathologic diagnoses. Design. Double-blind study (Canadian Task Force classification II-2). Setting. Gynecologic cancer center, private institute. Subjects. One thousand four hundred thirty-six uterine cavities. Intervention. Hysteroscopy. Measurements and Main Results. Images were classified as normal hysteroscopy, benign lesion, low-risk hyperplasia, highrisk hyperplasia, and carcinoma. A hysteroscopic diagnosis was made and biopsy specimens were obtained with Kevorkiantype curettes or Sims curettes. Tissues were studied by a pathologist, after which we compared endoscopic diagnoses with anatopathologic diagnoses. Significant correlation was found between suspicion based on images and histologic confirmation (p = 0.001). The Cramer V coefficient, which measures the relationship between both methods, was high: 0.925. The Cramer V coefficient takes values ranging from zero, to indicate lack of correlation, to 1, to indicate perfect correlation. A λ symmetry coefficient of 0.96 is interpreted as probable improvement in the prediction of histologic diagnosis based on images. Conclusion. This classification system can be useful for a systematic approach to hysteroscopic findings and to improve communication among specialists. It is based on the degree of hysteroscopic suspicion aimed at early detection of endometrial cancer and its precursor lesions.
Development of the microhysteroscope has made hysteroscopy a safe, efficient, atraumatic technique. Not only does it provide satisfactory assessment of the uterine cavity in the office, it contributes to early detection of benign and premalignant lesions as well as endometrial cancer. The major advantage of this type of endoscopy is the high-quality vision that allows direct observation of endometrium.1–5 Experience acquired with this technique coupled with endometrial biopsies in all cases led to correlation of hysteroscopic images with the diagnosis in every case. Given this fact, we decided to classify hysteroscopic images with histologic diagnoses based on degree of suspicion.
tive age, and according to indication in the menopausal group. We did not use anesthesia or preoperative drugs. After viewing the lesion, we took a tissue sample that was sent to the pathologist labeled “endometrial biopsy.” Neither hysteroscopists nor pathologists had other information regarding the patients. Patients were divided in two groups: 285 women made up the asymptomatic high-risk group6 (obesity, hypertension, diabetes mellitus, nulliparity, family cancer, personal cancer, e.g., colon, breast); the second group consisted of 1151 patients with spotting or postmenopausal and perimenopausal metrorrhagia. All of them underwent microhysteroscopy with a Hamou no. 1, 4-mm, rigid hysteroscope (Karl Storz, Tuttlingen, Germany).7 This instrument has an endoscope 4 mm in diameter, 25 cm long, with 90-degree vision, and foreobliquity of 30 degrees. It has two oculars: one allows direct vision, and the other is lateral and mobile. Four magnifications are possible: 1, 20, 60, and 150 × (lateral ocular). The optic system offers up to 20 magnifications in panoramic vision and 150 in contact vision; we have found magnifications of 1 and 20 useful in studying endometrium. For uterine distention we used the Hamou microhysteroflator (CO2) with an outflow of 40 ml/minute.
Materials and Methods The evaluated population in this double-blind study consisted of 1508 women (mean age 50 yrs, range 35–85 yrs). Of these, 72 (4.8%) were withdrawn from the study due to cervical stenosis or discomfort during the examination. This brought the number of uterine cavities examined to 1436. Hysteroscopies were done after menstruation on day 6 to day 14 of the cycle in women who were of reproduc-
From Departments of Hysteroscopy (Dr. Dotto) and Pathology (Dr. Lema), Instituto Argentino de Diagnóstico y Tratamiento, Asociación Argentina del Cáncer (Dr. Dotto, Jr.), Buenos Aires, Argentina; and Universitaire d’Hysteroscopie, Maternité, Antoine Béclere, University of Paris, Paris, France (Dr. Hamou). Address reprint requests to Jorge E. Dotto, M.D., Instituto Argentino de Diagnóstico y Tratamiento, Amenábar 2879 - C1428CQU, Buenos Aires, Cap. Fed., Argentina; fax 54 11 4780 0730. Submitted April 29, 2002. Accepted for publication December 3, 2002. Reprinted from the JOURNAL OF THE AMERICAN ASSOCIATION OF GYNECOLOGIC LAPAROSCOPISTS, May 2003, Vol. 10 No. 2 © 2003 The American Association of Gynecologic Laparoscopists. All rights reserved. This work may not be reproduced in any form or by any means without written permission from the AAGL. This includes but is not limited to, the posting of electronic files on the Internet, transferring electronic files to other persons, distributing printed output, and photocopying. To order multiple reprints of an individual article or request authorization to make photocopies, please contact the AAGL.
233
Correlation of Microhysteroscopic Images with Histologic Diagnoses Dotto et al
Documentation was obtained with a Printer Sony U 1200. A video camera was used. Since biopsy specimens obtained with the hysteroscope were often inadequate, we used Kevorkian-type or Sims curettes. After identifying the biopsy site with the hysteroscope, the microcurette is placed for a biopsy; the hysteroscope is reintroduced to confirm the adequacy of the site. Histologic classification of hyperplasia and precursor lesions was simple, complex, and atypical hyperplasia.8
Toward the luteal phase of the cycle the surface becomes more irregular; greater congestion changes the pale mucosal lining to a brighter color. In this second half of the cycle less CO2 pressure is required to distend the cavity. The endometrial surface is rough and yellowish red. With the distal tip of the microhysteroscope one can appreciate the depression it generates, which is useful to estimate the thickness of the lining (Figure 2). At higher magnification vessels can be seen more clearly. Patients must be examined with care since greater mucosal congestion is likely to result in tearing and hemorrhage.
Classification of Hysteroscopic Images We used the following classification of hysteroscopic images based on our experience and on the hysteroscopic nomenclature of Italian authors9–13: normal hysteroscopy, benign lesion, low-risk hyperplasia, high-risk hyperplasia, and carcinoma.
Postmenopausal Women: Atrophic Endometrium The cavity is easily identified. The surface is whitish, no gland openings are seen, and few vessels are present due to atrophy. Petechiae can be seen on the surface due to submucosal bleeding. This shows how thin and friable the epithelium is.14
Normal Hysteroscopy
Benign Lesions
Menstruating Women: Proliferative or Secretory Endometrium During the proliferative phase of a normal cycle, the mucosal lining of the uterine cavity has a whitish pink appearance. No surface vessels are seen; glandular openings are shallow depressions that have no evidence of vascularization (Figure 1). The thin mucosal lining is visualized as a compact image.
Endometrial Changes due to Hormone Therapy Different patterns are observed depending on agents prescribed for dysfunctional metrorrhagia or combined estrogen-progesterone hormone replacement therapy (HRT), and are interpreted as a hyperplasic change. Suspicious changes may be evident in a mucosa with cysts,
FIGURE 1. Proliferative endometrium, absence of surface vessels, whitish pink color, compact appearance.
234
May 2003, Vol. 10, No. 2
The Journal of the American Association of Gynecologic Laparoscopists
FIGURE 2. Secretory endometrium, irregular surface, more intense color. Vessels can be seen due to congestion. Endoscope leaves indentation.
pseudopolyps, and vascular proliferation, all of which may lead to a mistaken presumptive diagnosis of hyperplasia. This pseudohyperplastic pattern is associated with unopposed estrogen therapy.
Endometritis The reddish color with “starry sky” white configuration is similar in appearance to diffuse colpitis seen with a colposcope (Figure 4) and can also appear as whitish plaques that are friable and quick to bleed (Figure 5). Specimens may be taken for microbial identification.
Myomas, Submucous and Intramural The submucous myoma is perceived as a hemispheric structure with a smooth surface and endometrial lining. The wide base is solid in appearance, and when probed with the tip of the endoscope, pediculate types are sometimes seen. Some myomas are covered with large superficial vessels. The intramural myoma protrudes into the cavity and can be identified by diagnostic hysteroscopy. Some of these myomata are partially submucous. Others can be transformed into submucous lesions by negative pressure known as hydromassage during operative hysteroscopy.
Synechiae The constitution, location, and extension of these scarring processes can be evaluated. Multiple synechiae tend to block the view of the uterine cavity and can lead to an unsatisfactory diagnosis. It may be necessary to examine them under general anesthesia. Synechiae may be formed by endometrium, muscle fibers, and connective tissue. Endometrial synechiae are fine and fragile. They are located in the periphery of the cavity and are easily detached (Figure 6). Myofibrous synechiae are thicker and tend to be central or close to tubal openings (Figure 7). Connective tissue synechiae are large, thick, or irregular, and may be multiple. Cavity distortion is evident and introduction of the scope is difficult (Figure 8).
Basal Adenomas, Endometrial Polyps Due to their typical aspect these lesions are easy to diagnose. When other disorders such as hyperplasia or necrobiotic myomas are also present, however, they can be difficult to diagnose and may be considered suspicious. Functional polyps can be multiple and small, and have the same transformations as endometrium. Some are red, tend to be solitary and pedunculated, and can have a whitish hue and be vascularized. Endometrial polyps have a soft consistency (Figure 3).5
Foreign Bodies The position of an intrauterine device (IUD) within the cavity and in relation to the walls of the uterus can be assessed. If one of the tips is embedded in the endometrial lining or in the orifice of one of the tubes, removal under direct visual control is possible.
235
Correlation of Microhysteroscopic Images with Histologic Diagnoses Dotto et al
FIGURE 3. Whitish endometrial polyp-like formation. On the left, an irregular development of endometrial mucosa is seen extending into the isthmus. Microhysteroscopic diagnosis: endometrial polyp + low-risk hyperplasia. Histopathologic diagnosis: endometrial polyp + simple hyperplasia.
FIGURE 4. Endometritis close to left ostium visualized as reddish with abundant punctation.
236
May 2003, Vol. 10, No. 2
The Journal of the American Association of Gynecologic Laparoscopists
FIGURE 5. Endometritis throughout the cavity, white plaques that bleed readily.
FIGURE 6. Fine and fragile endometrial synechiae, easily ruptured with tip of endoscope.
237
Correlation of Microhysteroscopic Images with Histologic Diagnoses Dotto et al
FIGURE 7. Myofibrous synechiae are thicker in this case, located close to tubal ostium. Remaining endometrium is atrophic.
FIGURE 8. Connective tissue synechiae with partial debridement.
238
May 2003, Vol. 10, No. 2
The Journal of the American Association of Gynecologic Laparoscopists
Uterine Malformations Cases of uterus didelphys or uterine septa show characteristic hysteroscopic images (Figure 9).
This yields several different images (Figure 11). When multiple cystic atrophies are associated with other disorders such as necrobiotic myomas, the microhysteroscopic image can be suspicious. Histopathology is required to dispel doubt. Confusion can arise in the case of images of multiple polypoid adenomas associated with cystic atrophy. We considered these to be suspicious of high-risk hyperplasia and they were later diagnosed by anatomopathology as simple hyperplasias. In polypoid hyperplasia, various polyps of different sizes are observed. They show marked vascularization and interpapillary bridges.
Placental Remnants, Placental Polyps These lesions are dark red and smooth, or appear in a necrotic white pattern. They tend to form mucosal bridges, and the darker color of some of these adhesions correlates with active vascularization. Low-Risk Hyperplasia Endometrial mucosa has a nonhomogeneous appearance with projections and marked vascularization. The present histologic classification divides hyperplasias as simple, which includes simple hyperplasia and glandular cystic hyperplasia of the old classification system; complex, previously named adenomatous hyperplasia; and atypical.8 Correlation exists between this classification and an array of microhysteroscopic aspects. Endometrial hyperplasia is defined as exuberant development of endometrial mucosa; this irregular growth can be focal or generalized. The appearance of simple hyperplasia is similar to that of normal endometrium, with normal glandular distribution but of greater thickness. This can be measured by means of the depression (>7 mm) generated by the distal tip of the endoscope (Figure 10). In classic glandular cystic hyperplasia (GCH) the opening of the glands forms a relief with small cysts that is sometimes associated with cystic atrophy (this is apparently due to involution or regression that some CGH features).
High-Risk Hyperplasia The endometrial surface shows polypoid proliferation as well as interpapillary bridges that can be evident on a hemorrhagic background. This disorder can be distinguished only histologically. Adenomatous hyperplasia has an irregular mammilated surface, and an imprint is left by the scope on the surface (Figure 12). Carcinoma The microhysteroscopic appearance is multiple and varied, but if the image can be visualized it is characteristic. The lesion may be circumscribed or diffused. The focal lesion may be a whitish exophytic polypoid proliferation with atypical vascularization and hemorrhagic areas, or papillary with a brainlike aspect and necrotic areas (Figure 13).15,16
FIGURE 9. Uterine malformation, septum in fundus.
239
Correlation of Microhysteroscopic Images with Histologic Diagnoses Dotto et al
FIGURE 10. Small, whitish, elevated, well-defined area in right horn ostium. Microhysteroscopic diagnosis: low-risk hyperplasia. Histopathologic diagnosis: simple hyperplasia.
FIGURE 11. Patient with an IUD with exuberant endometrial proliferation and formation of small cysts. Microhysteroscopic diagnosis: IUD + low-risk hyperplasia. Histopathologic diagnosis: IUD + simple hyperplasia (glandular-cystic).
240
May 2003, Vol. 10, No. 2
The Journal of the American Association of Gynecologic Laparoscopists
FIGURE 12. Hysteroscopic image showing mucosal proliferation with greater vascularity. Hysteroscopic diagnosis: high-risk hyperplasia. Histopathologic diagnosis: atypical hyperplasia.
FIGURE 13. Uterine cavity occupied by exophytic hemorrhagic proliferation with necrotic zones and whitish brainlike areas. Microhysteroscopic diagnosis: endometrial adenocarcinoma. Histopathologic diagnosis: endometrial adenocarcinoma.
241
Correlation of Microhysteroscopic Images with Histologic Diagnoses Dotto et al
An architectural disarrangement with partial necrotic or hemorrhagic areas is generally observed. Vascularization is totally disorganized. On the other hand, it is possible to find endometrial carcinomas in atrophic endometrium (Figure 14) or combined pathologies (Figure 15). This observation is similar to published reports.17,18 At this point we should state that the classification system allows us to suspect but not to confirm a diagnosis. With hysteroscopy, our efforts are aimed at identifying pathologic foci to guide biopsies under microhysteroscopic control. Direct macroscopic observation will detail the extent of the affected area, topography, and extension to the isthmus and endocervix.
two specimens were not included in the final analysis because our aim was to study only endometrial pathology. Eight complex and three atypical hyperplasias were labeled high-risk hyperplasias by microhysteroscopy, with overdiagnosis in four patients with mixed disease that distorted the microhysteroscopic images, and two simple hyperplasias (glandular, cystic) associated with necrobiotic myomas and multiple adenomas associated with cystic atrophies. The 70 low-risk hyperplasias diagnosed by microhysteroscopy corresponded to 60 simple hyperplasias; the remaining 10 endometrial modifications stemmed from HRT. It is sometimes difficult in simple hyperplasias to distinguish dysfunctional forms of endometrial modifications produced by HRT, and the same is true in the differential diagnosis between high-risk hyperplasia and highly differentiated adenocarcinoma. A case in point is one woman with endometrial carcinoma who came to the clinic accompanying her mother who was scheduled for a study. Since she belonged to a high-risk group, it was suggested that she be examined as well, and endometrial carcinoma was diagnosed. Another example is a patient with an endocervical polyp. She was examined because of a family history of endometrial and breast carcinoma (mother, sister) and was diagnosed with complex hyperplasia. Correlation was significant (p <0.001, Cramer V correlation coefficient 0.925, λ coefficient 0.96).
Results Of 1436 patients, the study failed to be completed in 24 (1.7%) due to synechiae that prevented complete visualization of the cavity (Table 1). We classified these as relative failures. It was not possible to arrive at a microhysteroscopic diagnosis of suspicion of carcinoma in 2 of 21 endometrial carcinomas since these women had multiple synechiae and the whole cavity could not be visualized. However, biopsy specimens confirmed the diagnosis. Microhysteroscopic suspicion of two endocervical carcinomas was confirmed by histopathology. However, these
FIGURE 14. Adenocarcinoma in proximity of left ostium confirmed by histopathology. The uterine cavity has characteristic whitish atrophic endometrium with scarce vessels and petechiae.
242
May 2003, Vol. 10, No. 2
The Journal of the American Association of Gynecologic Laparoscopists
FIGURE 15. Submucosal myoma in fundus of uterine cavity on left wall in front of left ostium. Microhysteroscopic image of a small carcinoma. Microhysteroscopic diagnosis: small carcinoma. Histopathologic diagnosis: endometrial carcinoma.
TABLE 1. Diagnosis by Microhysteroscopy and Biopsy No. of Microhysteroscopy Cases
Diagnosis
19 2 15
Endometrial adenocarcinoma Endocervical carcinoma High-risk hyperplasia
70
Low-risk hyperplasia
98 41 603 91 437 60 72 Totals
Endometrial adenoma Isthmic polyps Atrophic endometria Submucosal myomas Normal endometria Synechiae Examination not performed 1508 patients
No. of Histopathology Cases 21 2 3 8 2 2 60 10 98 41 603 91 437 58 72
243
Diagnosis Endometrial adenocarcinoma Endocervical squamous cell carcinoma Atypical adenomatous hyperplasia Complex hyperplasi Simple hyperplasia Cystic atrophy Simple hyperplasia Hormone therapy dysfunction Endometrial adenoma Isthmic polyps Atrophic endometria Submucosal myomas Normal endometria Synechiae Examination not performed 1508 patients
Correlation of Microhysteroscopic Images with Histologic Diagnoses Dotto et al
Table 2 correlates microhysteroscopic images and corresponding histologic confirmation in all 1436 patients. This includes 24 women for whom no presumptive diagnosis of suspicion was arrived at despite microhysteroscopic study due to the presence of synechiae (relative failures). The highlighted diagonal of the table shows the number of cases for which images leading to suspicion and histology coincided. Cases where disease was overdiagnosed are found under the main diagonal. Ten cases that were diagnosed by microhysteroscopy as low-risk hyperplasia were considered benign by histology; these were patients who had been receiving HRT, and the relevant visual characteristics had been overdiagnosed. Four cases were over diagnosed as high-risk hyperplasia and were in fact simple hyperplasias with associated mixed disorders: two simple hyperplasias (GCH) plus necrobiotic myomas (Figure 16), and multiple adenomas with cystic atrophy. Cases of underdiagnosis are found in the upper part of the table, above the highlighted diagonal. Two carcinomas were confirmed by histology in which the complete cavity could not be visualized due to the presence of multiple synechiae. These cases were included in the group of benign lesions. Table 3 shows significant correlation between suspicion based on images and histologic confirmation (p = 0.001). The Cramer V coefficient, which measures the relationship between both methods, was high: 0.925. (The Cramer V coefficient takes values between zero to indicate
TABLE 2. Correlation between Microhysteroscopy Images and Histologic Confirmation Histology Hysteroscopy NH BL LRH HRH CA Totals
NH
BL
SH
CH
CA
Total Images
1040 0 0 0 0 1040
0 290 10 0 0 300
0 0 60 4 0 64
0 0 0 11 0 11
0 2 0 0 19 21
1040 292 70 15 19 1436
NHI = normal hysteroscopic image; BL = benign lesion LRH = low-risk hyperplasia; HRH = high-risk hyperplasia; CA = carcinoma; NH = normal histology; BH = benign histology; SH = simple hyperplasia; CH = complex hyperplasia; CA = carcinoma.
lack of correlation and 1 to indicate perfect correlation.) A λ symmetry coefficient of 0.96 is interpreted as probable improvement in prediction of histologic diagnosis based on images (SAS, Cary, NC). To determine the method’s general sensitivity and specificity, we defined normal and benign pathology as negative and the rest as positive (simple hyperplasia, complex hyperplasia, carcinoma). Of 96 women who had some type of confirmed positive pathology, 94 were correctly diagnosed (4 were overdiagnosed as
FIGURE 16. Suspected microhysteroscopic diagnosis: high-risk hyperplasia due to associated disorders (overdiagnosis). Histopathologic diagnosis: polypoid adenoma, simple hyperplasia (GCH) + necrobiotic myoma.
244
May 2003, Vol. 10, No. 2
The Journal of the American Association of Gynecologic Laparoscopists
search for malignant or potentially malignant lesions. This diagnostic exploration offers guidance in the case of suspicious lesions or negative biopsies: further investigation of these lesions can help avoid mistaken conclusions.7–10,17,18 We conclude that the diagnostic method is highly specific and very sensitive at detecting pathology. Predictive values calculated from specimens indicate that it is useful to predict endometrial lesions.
TABLE 3. Sensibility, Specificity, and Positive and Negative Predictive Values of Microhysteroscopy Pathology Method
Abnormal (+)
Normal (-)
Total Images
94 2 96
10 1330 1340
104 1332 1436
Abnormal (+) Normal (-) Total hystopathology confirmation Sensitivity Specificity Positive predictive value Negative predictive value
References 1. Carabias J, Bonilla Musoles F, et al: Atlas de Histeroscopia. Barcelona, Jims, 1985
94/96, 97.9% 1330/1340, 99.3% 94/104, 90.4% 1330/1332, 99.8%
2. Dotto J: Early endometrial cancer detection and its precursors in high-risk patients using cytology and microhysteroscopy. Doctoral thesis, Facultad de Medicina, University of Buenos Aires, Argentina, 1989
having mixed pathology). Two patients with carcinoma were correctly classified with benign pathologies by diagnosis (synechiae). The method had 97.9% sensitivity. Of 1340 patients with benign pathologies, 10 were considered by microhysteroscopic diagnosis to have other types of pathologies. The method had 99.3% specificity. Similarly, the proportion of patients with positive images who were correctly diagnosed showed a positive predictive value of 90.4%. The proportion of patients with negative images that were correctly diagnosed showed a negative predictive value of 99.8%.
3. Arrighi A, Testa R, Ormane S, et al: La pesquisa del carcinoma endometrial. Rev Soc Obstet Gynecol 73:3–7, 1994 4. Scarselli G, Mencaglia L, Tantini C, et al: Attualita e propettive de una nuova tecnica endoscopica nella routine ginecologica: La microcolpohisteroscopia. Patol Clin Obstet Ginecol 11:343–374, 1983 5. Hamou J: Adenocarcinoma of the endometrium. In Hysteroscopy and Microhysteroscopy. Text and Atlas. Edited by J Hamou. Paris, Appleton & Lange, p 99, 1984 6. Dotto J, Ghinelli C, Novelli J, et al: Correlación citohistopatológica en patología endometrial. Rev Soc Obstet Gynecol 59:237, 1980
Discussion Although hysteroscopy has been performed increasingly since the 1980s, we identified a need for a classification of images such as those that exist for other endoscopic procedures. Our systematization of hysteroscopic images is based on over 18 years of experience. We studied our first 1436 hysteroscopies with a histologic control and have used this classification ever since, observing strong correlation between images and histology. The two false negative results that turned out to be carcinomas were not misinterpretations of the images since direct visualization of the lesions was not possible due to the presence of synechiae. The diagnosis was made based on histologic evaluation of specimens obtained with a microcurette in areas not subjected to inspection. The four cases of hysteroscopic overdiagnosis were solved with biopsies. Thus although this method has a high correlation with final histologic results, it should be used only for diagnostic exploration. The final diagnosis of precursor and malignant lesions (low- and high-risk hyperplasias, carcinomas) should be made by histopathology. Within this framework, we believe this classification is useful for a systematic approach to endoscopic findings and can serve as a communication tool between the endoscopist and the treating gynecologist when the two specialists manage a patient. The former can report findings to the latter. It can also be useful for reports between endoscopist and pathologist; the endoscopist’s diagnostic impression can be briefly stated, thus guiding the pathologist’s
7. Hamou JE: Mycrohysteroscopie, une novelle technique en endoscopie, ses applications. Acta Endosc 10:415–422, 1980 8. Kurman R, Kaminsky P, Norris H: The behavior of endometrial hyperplasia. A long term study of untreated hyperplasia in 170 patients. Cancer 56:403–412, 1985 9. Mencaglia L, Tantini C, Pappalardo S, et al: Classificazzione e dati epidemiologici sulle iperplasia endometriali. Presented at the Congresso Nazionale de Endocrinologia Ginecologia, Madonna di Campligio, Italy, March 18–25, 1984 10. Mencaglia L, Scarselli G: Etats precancereux et cancereux de l’endometre. In Hysteroscopie et Microcolpohysteroscopie. Atlas et Traite. Edited by J Hamou. Palermo, Italy, Cofese, pp 145–163, 1984 11. Mencaglia L, Scarselli G, Tantini C, et al: Programma di screening per il carcinoma dell´endometrio. Bol Assoc Ital Endosc Ginecol 11:123–126, 1985 12. Mencaglia L, Perino A, Hamou J: Hysteroscopy in perimenopausal and postmenopausal women with abnormal uterine bleeding. J Reprod Med 32:577–582, 1987 13. Mencaglia L, Valle R, Perino A, et al: Endometrial carcinoma and its precursors: Early detection and treatment. Int J Gynaecol Obstet 31:107–116, 1990 14. Van Herendael B, Stevens M, Flakiewicz-Kula CH, et al: Dating of the endometrium by microhysteroscopy. Gynecol Obstet Invest 24:114–118, 1987
245
Correlation of Microhysteroscopic Images with Histologic Diagnoses Dotto et al
15. Mencarelli R, Mainenti M, Ottaviani A, et al: Hysteroscopy in the early diagnosis of endometrial neoplasia. Personal experience with 200 consecutive cases. Minerva Ginecol 44:429–432, 1992
17. Conoscenti G, Meir Y, Fischer-Tamaro L, et al: The diagnostic capacities of transvaginal echography and hysteroscopy in the characterization of endometrial pathology. Minerva Ginecol 47:293–300, 1995
16. Scarselli G, Pesci R, Tantini C, et al: Screening per l’adenocarcinoma endometriale. Patol Clin Obstet Ginecol 13:1–4, 1985
18. Bokhman JV: Two pathogenic types of endometrial carcinoma. Gynecol Oncol 15:10–17, 1983
246
The Journal of the American Association of Gynecologic Laparoscopists
Classification of Microhysteroscopic Images and their Correlation with Histologic Diagnoses Jorge E. Dotto, M.D., Baltasar Lema, M.D., Jorge E. Dotto, Jr., M.D., and Jacques Hamou, M.D. Abstract
(J Am Assoc Gynecol Laparosc 10(2):233–246, 2003)
Study Objective. To confirm that images observed at hysteroscopy correlate with histopathologic diagnoses. Design. Double-blind study (Canadian Task Force classification II-2). Setting. Gynecologic cancer center, private institute. Subjects. One thousand four hundred thirty-six uterine cavities. Intervention. Hysteroscopy. Measurements and Main Results. Images were classified as normal hysteroscopy, benign lesion, low-risk hyperplasia, highrisk hyperplasia, and carcinoma. A hysteroscopic diagnosis was made and biopsy specimens were obtained with Kevorkiantype curettes or Sims curettes. Tissues were studied by a pathologist, after which we compared endoscopic diagnoses with anatopathologic diagnoses. Significant correlation was found between suspicion based on images and histologic confirmation (p = 0.001). The Cramer V coefficient, which measures the relationship between both methods, was high: 0.925. The Cramer V coefficient takes values ranging from zero, to indicate lack of correlation, to 1, to indicate perfect correlation. A λ symmetry coefficient of 0.96 is interpreted as probable improvement in the prediction of histologic diagnosis based on images. Conclusion. This classification system can be useful for a systematic approach to hysteroscopic findings and to improve communication among specialists. It is based on the degree of hysteroscopic suspicion aimed at early detection of endometrial cancer and its precursor lesions.
Development of the microhysteroscope has made hysteroscopy a safe, efficient, atraumatic technique. Not only does it provide satisfactory assessment of the uterine cavity in the office, it contributes to early detection of benign and premalignant lesions as well as endometrial cancer. The major advantage of this type of endoscopy is the high-quality vision that allows direct observation of endometrium.1–5 Experience acquired with this technique coupled with endometrial biopsies in all cases led to correlation of hysteroscopic images with the diagnosis in every case. Given this fact, we decided to classify hysteroscopic images with histologic diagnoses based on degree of suspicion.
tive age, and according to indication in the menopausal group. We did not use anesthesia or preoperative drugs. After viewing the lesion, we took a tissue sample that was sent to the pathologist labeled “endometrial biopsy.” Neither hysteroscopists nor pathologists had other information regarding the patients. Patients were divided in two groups: 285 women made up the asymptomatic high-risk group6 (obesity, hypertension, diabetes mellitus, nulliparity, family cancer, personal cancer, e.g., colon, breast); the second group consisted of 1151 patients with spotting or postmenopausal and perimenopausal metrorrhagia. All of them underwent microhysteroscopy with a Hamou no. 1, 4-mm, rigid hysteroscope (Karl Storz, Tuttlingen, Germany).7 This instrument has an endoscope 4 mm in diameter, 25 cm long, with 90-degree vision, and foreobliquity of 30 degrees. It has two oculars: one allows direct vision, and the other is lateral and mobile. Four magnifications are possible: 1, 20, 60, and 150 × (lateral ocular). The optic system offers up to 20 magnifications in panoramic vision and 150 in contact vision; we have found magnifications of 1 and 20 useful in studying endometrium. For uterine distention we used the Hamou microhysteroflator (CO2) with an outflow of 40 ml/minute.
Materials and Methods The evaluated population in this double-blind study consisted of 1508 women (mean age 50 yrs, range 35–85 yrs). Of these, 72 (4.8%) were withdrawn from the study due to cervical stenosis or discomfort during the examination. This brought the number of uterine cavities examined to 1436. Hysteroscopies were done after menstruation on day 6 to day 14 of the cycle in women who were of reproduc-
From Departments of Hysteroscopy (Dr. Dotto) and Pathology (Dr. Lema), Instituto Argentino de Diagnóstico y Tratamiento, Asociación Argentina del Cáncer (Dr. Dotto, Jr.), Buenos Aires, Argentina; and Universitaire d’Hysteroscopie, Maternité, Antoine Béclere, University of Paris, Paris, France (Dr. Hamou). Address reprint requests to Jorge E. Dotto, M.D., Instituto Argentino de Diagnóstico y Tratamiento, Amenábar 2879 - C1428CQU, Buenos Aires, Cap. Fed., Argentina; fax 54 11 4780 0730. Submitted April 29, 2002. Accepted for publication December 3, 2002. Reprinted from the JOURNAL OF THE AMERICAN ASSOCIATION OF GYNECOLOGIC LAPAROSCOPISTS, May 2003, Vol. 10 No. 2 © 2003 The American Association of Gynecologic Laparoscopists. All rights reserved. This work may not be reproduced in any form or by any means without written permission from the AAGL. This includes but is not limited to, the posting of electronic files on the Internet, transferring electronic files to other persons, distributing printed output, and photocopying. To order multiple reprints of an individual article or request authorization to make photocopies, please contact the AAGL.
233
Correlation of Microhysteroscopic Images with Histologic Diagnoses Dotto et al
Documentation was obtained with a Printer Sony U 1200. A video camera was used. Since biopsy specimens obtained with the hysteroscope were often inadequate, we used Kevorkian-type or Sims curettes. After identifying the biopsy site with the hysteroscope, the microcurette is placed for a biopsy; the hysteroscope is reintroduced to confirm the adequacy of the site. Histologic classification of hyperplasia and precursor lesions was simple, complex, and atypical hyperplasia.8
Toward the luteal phase of the cycle the surface becomes more irregular; greater congestion changes the pale mucosal lining to a brighter color. In this second half of the cycle less CO2 pressure is required to distend the cavity. The endometrial surface is rough and yellowish red. With the distal tip of the microhysteroscope one can appreciate the depression it generates, which is useful to estimate the thickness of the lining (Figure 2). At higher magnification vessels can be seen more clearly. Patients must be examined with care since greater mucosal congestion is likely to result in tearing and hemorrhage.
Classification of Hysteroscopic Images We used the following classification of hysteroscopic images based on our experience and on the hysteroscopic nomenclature of Italian authors9–13: normal hysteroscopy, benign lesion, low-risk hyperplasia, high-risk hyperplasia, and carcinoma.
Postmenopausal Women: Atrophic Endometrium The cavity is easily identified. The surface is whitish, no gland openings are seen, and few vessels are present due to atrophy. Petechiae can be seen on the surface due to submucosal bleeding. This shows how thin and friable the epithelium is.14
Normal Hysteroscopy
Benign Lesions
Menstruating Women: Proliferative or Secretory Endometrium During the proliferative phase of a normal cycle, the mucosal lining of the uterine cavity has a whitish pink appearance. No surface vessels are seen; glandular openings are shallow depressions that have no evidence of vascularization (Figure 1). The thin mucosal lining is visualized as a compact image.
Endometrial Changes due to Hormone Therapy Different patterns are observed depending on agents prescribed for dysfunctional metrorrhagia or combined estrogen-progesterone hormone replacement therapy (HRT), and are interpreted as a hyperplasic change. Suspicious changes may be evident in a mucosa with cysts,
FIGURE 1. Proliferative endometrium, absence of surface vessels, whitish pink color, compact appearance.
234
May 2003, Vol. 10, No. 2
The Journal of the American Association of Gynecologic Laparoscopists
FIGURE 2. Secretory endometrium, irregular surface, more intense color. Vessels can be seen due to congestion. Endoscope leaves indentation.
pseudopolyps, and vascular proliferation, all of which may lead to a mistaken presumptive diagnosis of hyperplasia. This pseudohyperplastic pattern is associated with unopposed estrogen therapy.
Endometritis The reddish color with “starry sky” white configuration is similar in appearance to diffuse colpitis seen with a colposcope (Figure 4) and can also appear as whitish plaques that are friable and quick to bleed (Figure 5). Specimens may be taken for microbial identification.
Myomas, Submucous and Intramural The submucous myoma is perceived as a hemispheric structure with a smooth surface and endometrial lining. The wide base is solid in appearance, and when probed with the tip of the endoscope, pediculate types are sometimes seen. Some myomas are covered with large superficial vessels. The intramural myoma protrudes into the cavity and can be identified by diagnostic hysteroscopy. Some of these myomata are partially submucous. Others can be transformed into submucous lesions by negative pressure known as hydromassage during operative hysteroscopy.
Synechiae The constitution, location, and extension of these scarring processes can be evaluated. Multiple synechiae tend to block the view of the uterine cavity and can lead to an unsatisfactory diagnosis. It may be necessary to examine them under general anesthesia. Synechiae may be formed by endometrium, muscle fibers, and connective tissue. Endometrial synechiae are fine and fragile. They are located in the periphery of the cavity and are easily detached (Figure 6). Myofibrous synechiae are thicker and tend to be central or close to tubal openings (Figure 7). Connective tissue synechiae are large, thick, or irregular, and may be multiple. Cavity distortion is evident and introduction of the scope is difficult (Figure 8).
Basal Adenomas, Endometrial Polyps Due to their typical aspect these lesions are easy to diagnose. When other disorders such as hyperplasia or necrobiotic myomas are also present, however, they can be difficult to diagnose and may be considered suspicious. Functional polyps can be multiple and small, and have the same transformations as endometrium. Some are red, tend to be solitary and pedunculated, and can have a whitish hue and be vascularized. Endometrial polyps have a soft consistency (Figure 3).5
Foreign Bodies The position of an intrauterine device (IUD) within the cavity and in relation to the walls of the uterus can be assessed. If one of the tips is embedded in the endometrial lining or in the orifice of one of the tubes, removal under direct visual control is possible.
235
Correlation of Microhysteroscopic Images with Histologic Diagnoses Dotto et al
FIGURE 3. Whitish endometrial polyp-like formation. On the left, an irregular development of endometrial mucosa is seen extending into the isthmus. Microhysteroscopic diagnosis: endometrial polyp + low-risk hyperplasia. Histopathologic diagnosis: endometrial polyp + simple hyperplasia.
FIGURE 4. Endometritis close to left ostium visualized as reddish with abundant punctation.
236
May 2003, Vol. 10, No. 2
The Journal of the American Association of Gynecologic Laparoscopists
FIGURE 5. Endometritis throughout the cavity, white plaques that bleed readily.
FIGURE 6. Fine and fragile endometrial synechiae, easily ruptured with tip of endoscope.
237
Correlation of Microhysteroscopic Images with Histologic Diagnoses Dotto et al
FIGURE 7. Myofibrous synechiae are thicker in this case, located close to tubal ostium. Remaining endometrium is atrophic.
FIGURE 8. Connective tissue synechiae with partial debridement.
238
May 2003, Vol. 10, No. 2
The Journal of the American Association of Gynecologic Laparoscopists
Uterine Malformations Cases of uterus didelphys or uterine septa show characteristic hysteroscopic images (Figure 9).
This yields several different images (Figure 11). When multiple cystic atrophies are associated with other disorders such as necrobiotic myomas, the microhysteroscopic image can be suspicious. Histopathology is required to dispel doubt. Confusion can arise in the case of images of multiple polypoid adenomas associated with cystic atrophy. We considered these to be suspicious of high-risk hyperplasia and they were later diagnosed by anatomopathology as simple hyperplasias. In polypoid hyperplasia, various polyps of different sizes are observed. They show marked vascularization and interpapillary bridges.
Placental Remnants, Placental Polyps These lesions are dark red and smooth, or appear in a necrotic white pattern. They tend to form mucosal bridges, and the darker color of some of these adhesions correlates with active vascularization. Low-Risk Hyperplasia Endometrial mucosa has a nonhomogeneous appearance with projections and marked vascularization. The present histologic classification divides hyperplasias as simple, which includes simple hyperplasia and glandular cystic hyperplasia of the old classification system; complex, previously named adenomatous hyperplasia; and atypical.8 Correlation exists between this classification and an array of microhysteroscopic aspects. Endometrial hyperplasia is defined as exuberant development of endometrial mucosa; this irregular growth can be focal or generalized. The appearance of simple hyperplasia is similar to that of normal endometrium, with normal glandular distribution but of greater thickness. This can be measured by means of the depression (>7 mm) generated by the distal tip of the endoscope (Figure 10). In classic glandular cystic hyperplasia (GCH) the opening of the glands forms a relief with small cysts that is sometimes associated with cystic atrophy (this is apparently due to involution or regression that some CGH features).
High-Risk Hyperplasia The endometrial surface shows polypoid proliferation as well as interpapillary bridges that can be evident on a hemorrhagic background. This disorder can be distinguished only histologically. Adenomatous hyperplasia has an irregular mammilated surface, and an imprint is left by the scope on the surface (Figure 12). Carcinoma The microhysteroscopic appearance is multiple and varied, but if the image can be visualized it is characteristic. The lesion may be circumscribed or diffused. The focal lesion may be a whitish exophytic polypoid proliferation with atypical vascularization and hemorrhagic areas, or papillary with a brainlike aspect and necrotic areas (Figure 13).15,16
FIGURE 9. Uterine malformation, septum in fundus.
239
Correlation of Microhysteroscopic Images with Histologic Diagnoses Dotto et al
FIGURE 10. Small, whitish, elevated, well-defined area in right horn ostium. Microhysteroscopic diagnosis: low-risk hyperplasia. Histopathologic diagnosis: simple hyperplasia.
FIGURE 11. Patient with an IUD with exuberant endometrial proliferation and formation of small cysts. Microhysteroscopic diagnosis: IUD + low-risk hyperplasia. Histopathologic diagnosis: IUD + simple hyperplasia (glandular-cystic).
240
May 2003, Vol. 10, No. 2
The Journal of the American Association of Gynecologic Laparoscopists
FIGURE 12. Hysteroscopic image showing mucosal proliferation with greater vascularity. Hysteroscopic diagnosis: high-risk hyperplasia. Histopathologic diagnosis: atypical hyperplasia.
FIGURE 13. Uterine cavity occupied by exophytic hemorrhagic proliferation with necrotic zones and whitish brainlike areas. Microhysteroscopic diagnosis: endometrial adenocarcinoma. Histopathologic diagnosis: endometrial adenocarcinoma.
241
Correlation of Microhysteroscopic Images with Histologic Diagnoses Dotto et al
An architectural disarrangement with partial necrotic or hemorrhagic areas is generally observed. Vascularization is totally disorganized. On the other hand, it is possible to find endometrial carcinomas in atrophic endometrium (Figure 14) or combined pathologies (Figure 15). This observation is similar to published reports.17,18 At this point we should state that the classification system allows us to suspect but not to confirm a diagnosis. With hysteroscopy, our efforts are aimed at identifying pathologic foci to guide biopsies under microhysteroscopic control. Direct macroscopic observation will detail the extent of the affected area, topography, and extension to the isthmus and endocervix.
two specimens were not included in the final analysis because our aim was to study only endometrial pathology. Eight complex and three atypical hyperplasias were labeled high-risk hyperplasias by microhysteroscopy, with overdiagnosis in four patients with mixed disease that distorted the microhysteroscopic images, and two simple hyperplasias (glandular, cystic) associated with necrobiotic myomas and multiple adenomas associated with cystic atrophies. The 70 low-risk hyperplasias diagnosed by microhysteroscopy corresponded to 60 simple hyperplasias; the remaining 10 endometrial modifications stemmed from HRT. It is sometimes difficult in simple hyperplasias to distinguish dysfunctional forms of endometrial modifications produced by HRT, and the same is true in the differential diagnosis between high-risk hyperplasia and highly differentiated adenocarcinoma. A case in point is one woman with endometrial carcinoma who came to the clinic accompanying her mother who was scheduled for a study. Since she belonged to a high-risk group, it was suggested that she be examined as well, and endometrial carcinoma was diagnosed. Another example is a patient with an endocervical polyp. She was examined because of a family history of endometrial and breast carcinoma (mother, sister) and was diagnosed with complex hyperplasia. Correlation was significant (p <0.001, Cramer V correlation coefficient 0.925, λ coefficient 0.96).
Results Of 1436 patients, the study failed to be completed in 24 (1.7%) due to synechiae that prevented complete visualization of the cavity (Table 1). We classified these as relative failures. It was not possible to arrive at a microhysteroscopic diagnosis of suspicion of carcinoma in 2 of 21 endometrial carcinomas since these women had multiple synechiae and the whole cavity could not be visualized. However, biopsy specimens confirmed the diagnosis. Microhysteroscopic suspicion of two endocervical carcinomas was confirmed by histopathology. However, these
FIGURE 14. Adenocarcinoma in proximity of left ostium confirmed by histopathology. The uterine cavity has characteristic whitish atrophic endometrium with scarce vessels and petechiae.
242
May 2003, Vol. 10, No. 2
The Journal of the American Association of Gynecologic Laparoscopists
FIGURE 15. Submucosal myoma in fundus of uterine cavity on left wall in front of left ostium. Microhysteroscopic image of a small carcinoma. Microhysteroscopic diagnosis: small carcinoma. Histopathologic diagnosis: endometrial carcinoma.
TABLE 1. Diagnosis by Microhysteroscopy and Biopsy No. of Microhysteroscopy Cases
Diagnosis
19 2 15
Endometrial adenocarcinoma Endocervical carcinoma High-risk hyperplasia
70
Low-risk hyperplasia
98 41 603 91 437 60 72 Totals
Endometrial adenoma Isthmic polyps Atrophic endometria Submucosal myomas Normal endometria Synechiae Examination not performed 1508 patients
No. of Histopathology Cases 21 2 3 8 2 2 60 10 98 41 603 91 437 58 72
243
Diagnosis Endometrial adenocarcinoma Endocervical squamous cell carcinoma Atypical adenomatous hyperplasia Complex hyperplasi Simple hyperplasia Cystic atrophy Simple hyperplasia Hormone therapy dysfunction Endometrial adenoma Isthmic polyps Atrophic endometria Submucosal myomas Normal endometria Synechiae Examination not performed 1508 patients
Correlation of Microhysteroscopic Images with Histologic Diagnoses Dotto et al
Table 2 correlates microhysteroscopic images and corresponding histologic confirmation in all 1436 patients. This includes 24 women for whom no presumptive diagnosis of suspicion was arrived at despite microhysteroscopic study due to the presence of synechiae (relative failures). The highlighted diagonal of the table shows the number of cases for which images leading to suspicion and histology coincided. Cases where disease was overdiagnosed are found under the main diagonal. Ten cases that were diagnosed by microhysteroscopy as low-risk hyperplasia were considered benign by histology; these were patients who had been receiving HRT, and the relevant visual characteristics had been overdiagnosed. Four cases were over diagnosed as high-risk hyperplasia and were in fact simple hyperplasias with associated mixed disorders: two simple hyperplasias (GCH) plus necrobiotic myomas (Figure 16), and multiple adenomas with cystic atrophy. Cases of underdiagnosis are found in the upper part of the table, above the highlighted diagonal. Two carcinomas were confirmed by histology in which the complete cavity could not be visualized due to the presence of multiple synechiae. These cases were included in the group of benign lesions. Table 3 shows significant correlation between suspicion based on images and histologic confirmation (p = 0.001). The Cramer V coefficient, which measures the relationship between both methods, was high: 0.925. (The Cramer V coefficient takes values between zero to indicate
TABLE 2. Correlation between Microhysteroscopy Images and Histologic Confirmation Histology Hysteroscopy NH BL LRH HRH CA Totals
NH
BL
SH
CH
CA
Total Images
1040 0 0 0 0 1040
0 290 10 0 0 300
0 0 60 4 0 64
0 0 0 11 0 11
0 2 0 0 19 21
1040 292 70 15 19 1436
NHI = normal hysteroscopic image; BL = benign lesion LRH = low-risk hyperplasia; HRH = high-risk hyperplasia; CA = carcinoma; NH = normal histology; BH = benign histology; SH = simple hyperplasia; CH = complex hyperplasia; CA = carcinoma.
lack of correlation and 1 to indicate perfect correlation.) A λ symmetry coefficient of 0.96 is interpreted as probable improvement in prediction of histologic diagnosis based on images (SAS, Cary, NC). To determine the method’s general sensitivity and specificity, we defined normal and benign pathology as negative and the rest as positive (simple hyperplasia, complex hyperplasia, carcinoma). Of 96 women who had some type of confirmed positive pathology, 94 were correctly diagnosed (4 were overdiagnosed as
FIGURE 16. Suspected microhysteroscopic diagnosis: high-risk hyperplasia due to associated disorders (overdiagnosis). Histopathologic diagnosis: polypoid adenoma, simple hyperplasia (GCH) + necrobiotic myoma.
244
May 2003, Vol. 10, No. 2
The Journal of the American Association of Gynecologic Laparoscopists
search for malignant or potentially malignant lesions. This diagnostic exploration offers guidance in the case of suspicious lesions or negative biopsies: further investigation of these lesions can help avoid mistaken conclusions.7–10,17,18 We conclude that the diagnostic method is highly specific and very sensitive at detecting pathology. Predictive values calculated from specimens indicate that it is useful to predict endometrial lesions.
TABLE 3. Sensibility, Specificity, and Positive and Negative Predictive Values of Microhysteroscopy Pathology Method
Abnormal (+)
Normal (-)
Total Images
94 2 96
10 1330 1340
104 1332 1436
Abnormal (+) Normal (-) Total hystopathology confirmation Sensitivity Specificity Positive predictive value Negative predictive value
References 1. Carabias J, Bonilla Musoles F, et al: Atlas de Histeroscopia. Barcelona, Jims, 1985
94/96, 97.9% 1330/1340, 99.3% 94/104, 90.4% 1330/1332, 99.8%
2. Dotto J: Early endometrial cancer detection and its precursors in high-risk patients using cytology and microhysteroscopy. Doctoral thesis, Facultad de Medicina, University of Buenos Aires, Argentina, 1989
having mixed pathology). Two patients with carcinoma were correctly classified with benign pathologies by diagnosis (synechiae). The method had 97.9% sensitivity. Of 1340 patients with benign pathologies, 10 were considered by microhysteroscopic diagnosis to have other types of pathologies. The method had 99.3% specificity. Similarly, the proportion of patients with positive images who were correctly diagnosed showed a positive predictive value of 90.4%. The proportion of patients with negative images that were correctly diagnosed showed a negative predictive value of 99.8%.
3. Arrighi A, Testa R, Ormane S, et al: La pesquisa del carcinoma endometrial. Rev Soc Obstet Gynecol 73:3–7, 1994 4. Scarselli G, Mencaglia L, Tantini C, et al: Attualita e propettive de una nuova tecnica endoscopica nella routine ginecologica: La microcolpohisteroscopia. Patol Clin Obstet Ginecol 11:343–374, 1983 5. Hamou J: Adenocarcinoma of the endometrium. In Hysteroscopy and Microhysteroscopy. Text and Atlas. Edited by J Hamou. Paris, Appleton & Lange, p 99, 1984 6. Dotto J, Ghinelli C, Novelli J, et al: Correlación citohistopatológica en patología endometrial. Rev Soc Obstet Gynecol 59:237, 1980
Discussion Although hysteroscopy has been performed increasingly since the 1980s, we identified a need for a classification of images such as those that exist for other endoscopic procedures. Our systematization of hysteroscopic images is based on over 18 years of experience. We studied our first 1436 hysteroscopies with a histologic control and have used this classification ever since, observing strong correlation between images and histology. The two false negative results that turned out to be carcinomas were not misinterpretations of the images since direct visualization of the lesions was not possible due to the presence of synechiae. The diagnosis was made based on histologic evaluation of specimens obtained with a microcurette in areas not subjected to inspection. The four cases of hysteroscopic overdiagnosis were solved with biopsies. Thus although this method has a high correlation with final histologic results, it should be used only for diagnostic exploration. The final diagnosis of precursor and malignant lesions (low- and high-risk hyperplasias, carcinomas) should be made by histopathology. Within this framework, we believe this classification is useful for a systematic approach to endoscopic findings and can serve as a communication tool between the endoscopist and the treating gynecologist when the two specialists manage a patient. The former can report findings to the latter. It can also be useful for reports between endoscopist and pathologist; the endoscopist’s diagnostic impression can be briefly stated, thus guiding the pathologist’s
7. Hamou JE: Mycrohysteroscopie, une novelle technique en endoscopie, ses applications. Acta Endosc 10:415–422, 1980 8. Kurman R, Kaminsky P, Norris H: The behavior of endometrial hyperplasia. A long term study of untreated hyperplasia in 170 patients. Cancer 56:403–412, 1985 9. Mencaglia L, Tantini C, Pappalardo S, et al: Classificazzione e dati epidemiologici sulle iperplasia endometriali. Presented at the Congresso Nazionale de Endocrinologia Ginecologia, Madonna di Campligio, Italy, March 18–25, 1984 10. Mencaglia L, Scarselli G: Etats precancereux et cancereux de l’endometre. In Hysteroscopie et Microcolpohysteroscopie. Atlas et Traite. Edited by J Hamou. Palermo, Italy, Cofese, pp 145–163, 1984 11. Mencaglia L, Scarselli G, Tantini C, et al: Programma di screening per il carcinoma dell´endometrio. Bol Assoc Ital Endosc Ginecol 11:123–126, 1985 12. Mencaglia L, Perino A, Hamou J: Hysteroscopy in perimenopausal and postmenopausal women with abnormal uterine bleeding. J Reprod Med 32:577–582, 1987 13. Mencaglia L, Valle R, Perino A, et al: Endometrial carcinoma and its precursors: Early detection and treatment. Int J Gynaecol Obstet 31:107–116, 1990 14. Van Herendael B, Stevens M, Flakiewicz-Kula CH, et al: Dating of the endometrium by microhysteroscopy. Gynecol Obstet Invest 24:114–118, 1987
245
Correlation of Microhysteroscopic Images with Histologic Diagnoses Dotto et al
15. Mencarelli R, Mainenti M, Ottaviani A, et al: Hysteroscopy in the early diagnosis of endometrial neoplasia. Personal experience with 200 consecutive cases. Minerva Ginecol 44:429–432, 1992
17. Conoscenti G, Meir Y, Fischer-Tamaro L, et al: The diagnostic capacities of transvaginal echography and hysteroscopy in the characterization of endometrial pathology. Minerva Ginecol 47:293–300, 1995
16. Scarselli G, Pesci R, Tantini C, et al: Screening per l’adenocarcinoma endometriale. Patol Clin Obstet Ginecol 13:1–4, 1985
18. Bokhman JV: Two pathogenic types of endometrial carcinoma. Gynecol Oncol 15:10–17, 1983
246