The Status and Distance of Cone Biopsy Margins as a Predictor of Excision Adequacy for Endocervical Adenocarcinoma In Situ

Transcription

1 ANATOMIC PATHOLOGY The Status and Distance of Cone Biopsy Margins as a Predictor of Excision Adequacy for Endocervical Adenocarcinoma In Situ NEAL S. GOLDSTEIN, MD, AND ANIU MANI, MD Cervical cone biopsy has become an important surgical procedure for endocervical adenocarcinoma in situ (AIS), especially for patients who desire to retain their fertility. Establishing the usefulness of the endocervical margin status in cone biopsy specimens as a predictor of residual AIS is paramount. We examined the status of the endocervical margin in the cone biopsy specimen, the distance between the most proximal AIS and the endocervical margin in the cone biopsy specimen, and the endocervical curettage (ECC) specimen performed at the time of cone biopsy and residual AIS in the hysterectomy specimens of 61 patients with specimens accessioned from 1968 through 1997; 43 (30%) of patients with a negative endocervical margin had residual AIS in the hysterectomy specimen. Conversely, 10 of 18 (56%) patients with a positive endocervical margin in the cone biopsy specimen had no AIS in the hysterectomy specimen. All 6 patients with AIS in the ECC specimen had residual AIS. No patient with an endocervical margin in the cone biopsy specimen greater than 10 mm had residual AIS. Patients with distances less than 10 mm had equal percentages of residual AIS. In general, more patients with a negative endocervical margin in the cone biopsy specimen had no residual AIS in the hysterectomy specimen than those with a positive endocervical margin in the cone biopsy specimen. However, the status of this margin is not useful for predicting the presence of residual AIS. Pathologists should report the distance between the endocervical cone biopsy margin and the closest AIS. (Key words: Cervix; Endocervix; Adenocarcinoma in situ; Cone biopsy; Margins) Am J Clin Pathol 1998,109: Endocervical adenocarcinoma in situ (AIS) is an acknowledged precursor lesion of invasive endocervical adenocarcinoma. 1-3 has been the traditional surgical procedure for AIS. However, its routine use has been questioned because many patients with AIS are in their childbearing years and desire to retain their fertility. 4-8 In this context, the cervical cone biopsy has become an important surgical procedure for AIS. The major concern of using the cone biopsy as a definitive surgical procedure is its ability to completely excise the AIS, because any AIS that is left in the uterus has the potential to progress to invasive adenocarcinoma. Predicting which patients have a high likelihood of having residual AIS in the uterus after cone biopsy is paramount. The From the Department of Anatomic Pathology, William Beaumont Hospital, Royal Oak, Michigan. Manuscript received July 14, 1997; revision accepted August 4,1997. Address reprint requests to Dr Goldstein: Department of Anatomic Pathology, William Beaumont Hospital, 3601 West Thirteen Mile Rd, Royal Oak, MI status of the margins in the cone biopsy specimen is a histologic feature that would seem to be predictive of residual AIS in the uterus, but some studies have suggested that this may not be the case. Furthermore, the optimum minimal distance between the endocervical margin in a cone biopsy specimen and AIS that would markedly reduce the likelihood of residual AIS has yet to be established. We retrospectively studied the specimens of patients with AIS who had undergone a cervical cone biopsy followed by hysterectomy to determine the relationship between margins in the cone biopsy specimen and residual AIS in the hysterectomy specimen. MATERIALS AND METHODS Seventy-seven patients with AIS who had a cone biopsy shortly followed by hysterectomy were identified using the William Beaumont Hospital surgical pathology computer system from January 1, 1968 through June 1, Patients with small invasive adenocarcinomas in the cone biopsy specimen were excluded Twelve patients had coexistent squamous dysplasia. Sixteen of the 77 patients were 727

2 728 ANATOMIC PATHOLOGY excluded from the study; 5 had undergone large loop excision of the transformation zone biopsies that left the endocervical margin difficult to assess. In the other 11 excluded patients, the hysterectomy specimens were devoid of AIS but not all of the cervix was histologically examined. Thirteen of the hysterectomy specimens that were included in the study had residual AIS but had a subtotally examined cervix. All of the previous or concomitant cervical or uterine specimens of the remaining 61 study patients were examined to exclude invasive adenocarcinoma. All of the cone biopsy specimens in the study were obtained by cold-knife or laser surgical procedures. All specimens were oriented by an ectocervical-placed suture that marked the 12-o'clock position. They were inked and totally embedded. The mean number of blocks per cone biopsy specimen was 5.1 (range, 3-9 blocks). Fifty cone biopsy specimens had serially sectioned blocks with at least three slide levels per block. The mean number of slides per block was 3.8 (range, 3-12 slides). Forty eight patients (79%) had undergone an endocervical curettage (ECC) that was performed during the same operation, immediately following the cone biopsy. Fifty-two (85%) of the cone biopsy specimens had been sectioned longitudinally, allowing the distance between any AIS and the inked endocervical margin to be measured. The status of the endocervical (proximal) inked margin in these specimens was recorded as negative (no AIS at the endocervical margin) or positive (AIS present at the endocervical margin). The distance from the closest AIS to the inked endocervical margin was measured in the specimens with negative endocervical margins and categorized as less than 2 mm, 2 to 5 mm, 5.1 to 10 mm, or more than 10 mm. The number of quadrants with positive margins was recorded. The other 9 cone biopsy specimens (15%) had been sectioned by initially transversely amputating the proximal, endocervical end of the cone, producing a doughnut-shaped section. The remaining cone biopsy specimen was then sectioned in the usual longitudinal manner. The endocervical margin in these specimens was recorded as negative (no AIS was present in the transverse-oriented section) or positive (AIS present in the section). A distance between the AIS and the proximal end of the specimen could not be measured in this group of specimens. The status of the deep stromal surgical margin from all the cone biopsy specimen margins was recorded as negative (AIS not present at the inked margin) or positive (AIS present at the inked margin). The presence or absence of AIS in the ECC was recorded. The mean number of cervical tissue blocks from the hysterectomy specimens was 15 (range, 8-37). The presence or absence of AIS or invasive adenocarcinoma in each specimen was recorded. The mean time between the cone biopsy and hysterectomy was 25 days (range, 4-73 days). The mean age of the patients at diagnosis was 43 years (range, years). RESULTS No AIS was present at the endocervical margin of the cone biopsy specimen in 43 patients (Table 1). Thirty of the 43 patients (70%) had no AIS in the specimen, and 13 (30%) had residual AIS in the hysterectomy specimen. Of the 18 patients with AIS at the endocervical margin, 10 (56%) had no AIS in the specimen, and 8 (44%) had AIS in the hysterectomy specimen. One of the 8 patients also had a coexistent small invasive adenocarcinoma located in the proximal endocervix. Of the 61 patients, 52 had a longitudinally sectioned endocervical margin in the cone biopsy specimen. The endocervical margin was negative for 36 of these patients, 11 (31%) of whom had AIS in the hysterectomy specimen. The distance between the inked endocervical margin and the most proximal AIS was less than 2 mm in the specimens of 15 patients. Of these 15, 6 (40%) had AIS in the hysterectomy specimen. Similarly, 3 (30%) of 10 patients in whom the endocervical margins were negative in the cone biopsy specimen for a distance of 2 to 5 mm and 2 (33%) of 6 patients in whom the endocervical margins were negative for a distance of 5.1 to 10 mm had AIS in the hysterectomy specimen. All 5 patients who had more than 10 mm of normal mucosa situated between the most proximal AIS and the endocervical margin of the cone biopsy specimen had no AIS in their hysterectomy specimens. Sixteen patients had a positive endocervical margin in a longitudinally sectioned cone biopsy specimen (see Table 1). Ten of these patients (62%) had no AIS in the hysterectomy specimen. There was no apparent relationship between the number of endocervical margin quadrants in the cone biopsy specimen that contained AIS and the presence of residual AIS in the hysterectomy specimen. Seven of the ten patients (70%) with AIS in one quadrant and all 3 patients (100%) with AIS in two quadrants of the endocervical margin of the cone biopsy specimen had no residual AIS. All 3 patients with AIS in three AJCP June 1998

3 GOLDSTEIN AND MANI 729 Cone Biopsy Margins as a Predictor of Excision Adequacy TABLE 1. STATUS OF THE ENDOCERVICAL MARGINS IN THE CONE BIOPSY SPECIMEN AND RESIDUAL ADENOCARCINOMA IN SITU (AIS) IN 61 PATIENTS* Specimens Status of Endocervical MarginlNo. of Cone Specimens No AIS Present AIS Present All patients Negative (n = 43) Positive (n = 18) Longitudinally sectioned endocervical margin (n = 52) Negative (n = 36) Distance of AIS to margin (mm) < 2 (n = 15) 2-5 (n = 10) (n = 6) >10 (n = 5) No. of positive quadrants (n = 16) 1 (n = 10) 2 (n = 3) 3 (n = 3) 4(n = 0) Transverse sectioned endocervical margin (n = 9) Negative (n = 7) Positive (n = 2) *Data are given as number (percentage). + One patient had a small coexistent invasive adenocarcinoma in the hysterectomy specimen. 30 (70) 10 (56) 25 (69) 9(60) 7(70) 4(67) 5 (100) 10 (63) 7(70) 5 (71.4) 0 13(30) 8 + (44) 11 (31) 6(40) 3(30) 2(33) 6(38) 3(30) 2 (28.6) 2 (100) TABLE 2. STATUS OF DEEP STROMAL MARGIN IN CONE BIOPSY SPECIMENS AND ENDOCERVICAL CURETTAGE AND RESIDUAL ADENOCARCINOMA IN SITU (AIS)* Status Cone Biopsy No AIS Present (n - = 40) AIS Present (n = 21) Deep stromal margin in cone biopsy specimen Negative Positive Endocervical curettage Negative Positive 58 (100) 42 (100) 6 (100) 40 (69) 32 (76) 18(31) 3(100) 10 (24) 6 (100) Data are given as number (percentage). quadrants of the endocervical margin of the cone biopsy specimen had residual AIS. Nine patients had a transversely sectioned endocervical margin in the cone biopsy specimen. Seven of these patients had a negative endocervical margin, of whom 2 (28%) had residual AIS in the hysterectomy specimen. Five of the 7 patients had AIS that was within several millimeters of the endocervical end of the amputated cone biopsy specimen. Both patients with AIS in the transversely sectioned endocervical margin had residual AIS. There was no relationship between the status of the deep stromal margin and residual AIS (Table 2). All 3 patients with a positive deep stromal margin in the cone biopsy specimen had AIS in the hysterectomy specimen; these patients also had positive endocervical margins in the cone biopsy specimen. Eighteen of the 58 patients (31%) with a negative deep stromal margin in the cone biopsy specimen had residual AIS. Ten (24%) of 42 patients with a negative ECC specimen had residual AIS in the hysterectomy specimen (see Table 2). All 6 patients who had AIS in the ECC specimen had residual AIS in the hysterectomy specimen. Three of these 6 patients had a negative endocervical margin in the cone biopsy specimen, and all 3 had negative deep stromal margins. Vol. 109 No. 6

4 730 ANATOMIC PATHOLOGY DISCUSSION Cervical conization has become an established surgical procedure for the treatment of AIS. 4,8,13-21 Most authors believe that a patient with AIS does not require a second cone biopsy or hysterectomy if the margins are free of AIS in the initial cone biopsy specimen. The results of our study generally support this recommendation but they also highlight the context in which the recommendation needs to be made. Fewer patients with a negative endocervical margin in the cone biopsy specimen had residual AIS in the hysterectomy specimen than patients with a positive margin. The status of the endocervical margin in the cone biopsy specimen was not a useful predictor of residual AIS for an individual patient because 30% of patients with a negative endocervical margin in the cone biopsy specimen had residual AIS, and conversely, 56% of patients with a positive endocervical margin in the cone biopsy specimen had no residual AIS. Our results are similar to those of other studies (Tables 3 and 4). A tabulation of the results from these studies, TABLE 3. RESIDUAL ADENOCARCINOMA IN SITU (AIS) FOLLOWING CONE BIOPSY WITH A NEGATIVE ENDOCERVICAL MARGIN IN THE CONE BIOPSY SPECIMEN* or Second Cone Biopsy Source No. of Negative Endocervical Margins No AIS Present AIS + Present Im et al 8 Wolf et all 8 Nicklin et al 15 Poyner et al 1 Muntz et al 20 Hopkins et al 16 Ostor et al 13 Bertrand et al 4 Andersen and Arffmann 19 Qizilbash 10 Luesley et al 15 Widrich et al 21 Present study Total (55) 14 (67) 9(82) 6(60) 11 (92) 6(86) 4 (100) 4 (100) 1 1(50) 30 (70) 109 (77) 4(45) 7(33) 2(18) 4(40) 1(8) 1(14) 1(50) 13 (30) 33 (23) *Data are given as number (percentage). + Some AIS specimens also had coexistent invasive adenocarcinoma. TABLE 4. RESIDUAL ADENOCARCINOMA IN SITU (AIS) FOLLOWING CONE BIOPSY WITH A POSITIVE ENDOCERVICAL MARGIN IN THE CONE BIOPSY SPECIMEN * Source No. of Positive Endocervical Margins No AIS Present or Second Cone Biopsy AIS Present Im et al 8 Wolf et al 18 Nicklin et al 15 Poyner et al 1 Muntz et al 20 Hopkins et al 16 Ostor et al 13 Bertrand et al 4 Andersen and Arffmann 19 Luesley et al 5 Widrich et al 21 Present study Total (33) 9(47) 6(54) 3(30) 1(20) 2(33) 1 (100) 2(50) 5(36) 10 (56) 49(44) 4(67) 10 (53) 5(45) 7(70) 4(80) 4(67) 2(50) 9(64) 8(44) 61 (56) *Data are given as number (percentage). AJCP June 1998

5 GOLDSTEIN AND MANI 731 Cone Biopsy Margins as a Predictor of Excision Adequacy including ours, shows that 23% of the patients with a negative endocervical margin in the cone biopsy specimen had residual AIS (see Table 3) compared with 56% of the patients with a positive endocervical margin in the cone biopsy specimen (see Table 4). There are two possible reasons that the status of the endocervical margin in the cone biopsy specimen has no relationship with the presence of residual AIS in the hysterectomy specimen. One theory is that AIS can be multifocal, with foci of AIS developing at different levels of the endocervical canal. Residual AIS in the upper endocervical canal could be a discontinuous, independently proliferating focus that was not removed by the cone biopsy. However, there is little evidence for AIS multicentricity being the major cause of residual AIS after cone biopsy. Adenocarcinoma in situ is almost always a unicentric disease that usually originates at the squamocolumnar junction or within the transformation zone and extends proximally in a contiguous manner without skip lesions. 4 ' 5 ' 11,12 ' 14,19 ' In addition, the frequencies of AIS multicentricity and residual AIS after cone biopsy are dissimilar. Two studies that defined multicentric AIS as several separate foci of AIS involving different portions of the endocervical mucosa and separated by normal mucosa documented multicentricity in 13% and 15% of patients. 4 ' 13 These values are substantially less than the 24% overall incidence in the literature (30.3% in this study) of residual AIS found in uteri after cone biopsy with negative endocervical margins. The second, more plausible, theory is that AIS extended contiguously beyond the proximal end of the endocervical margin of the cone biopsy. Fluhmann, 28 in his richly illustrated monograph, showed that normal endocervical mucosal anatomy consists of longitudinal, elongated, and sometimes twisted grooves, invaginations, and infoldings that split into smaller clefts between the plicae palmatae. These clefts and grooves are commonly, but incorrectly, referred to as endocervical glands. A longitudinal section of the endocervical canal that tangentially cuts across an AlS-containing small cleft would give the appearance of a multifocal neoplastic process. This anatomy is conducive to the proximal spread of AIS below the endocervical mucosal surface. As the cone biopsy scalpel turns toward the endocervical lumen, deep infoldings of the endocervical mucosa that are in continuity with more distal mucosa are left in the wall. Adenocarcinoma in situ could be present in the deep mucosal infoldings, while the overlying surface mucosa and superficial mucosal clefts that constitute the endocervical margin in the cone biopsy Transversely sectioned proximal endocervix at the level of the endocervical margin of a cone biopsy. The darker shaded lines represent adenocarcinoma in situ. The inner circle represents the outer edge of the cone biopsy specimen. Adenocarcinoma in situ that is situated peripheral to the edge of the endocervical margin would be left in the cervix after cone biopsy. specimen are devoid of AIS (Figure). The implication is that the preferred shape of the cone biopsy is a cylinder, with the proximal end of the specimen removed with as close to 90 degrees as possible. 4 The larger the surface area of the proximal margin, the greater the likelihood of identifying any AIS that is in the deep mucosal infoldings. The percentage of patients with residual AIS in the hysterectomy specimen was similar when the endocervical margin in the cone biopsy specimen was free of AIS by less than 2 mm, 2 to 5 mm, or 5.1 to 10 mm. Only when the'distance between the most proximal AIS and the endocervical margin exceeded 10 mm did all the uteri become free of AIS. This group of patients most likely had short lengths of AIS or long cone biopsy specimens. A distance of 10 mm probably provides an adequate length of endocervix so that any irregular tongues of AIS growth around the circumference of the endocervix are encompassed within the cone biopsy specimen. It also probably accounts for the lack of measurement precision. The proximal edge of the cone biopsy specimen is irregular and jagged, and ink can course down into the infoldings. The results of this study support the opinions of other authors; when there is a considerable length of resection margin, the Vol. 109 No. 6

6 732 ANATOMIC PATHOLOGY likelihood of additional AIS in the upper endocervical canal is small, enabling the decision to opt for uterine conservation. 15 This result also suggests that pathologists should report the distance between the most proximal AIS and the endocervical margin in the cone biopsy specimen. 15 Although short lengths between the most proximal AIS and the endocervical margin in the cone biopsy specimen probably have no implications for the likelihood of complete excision, longer lengths probably have merit. A strong association was observed between the presence of AIS in the ECC performed at the time of cone biopsy and residual AIS. All of the patients with AIS in the ECC specimen had residual AIS in the uterus. The predictive value of this procedure when the specimen is devoid of AIS is limited. Adenocarcinoma in situ can be present in the endocervical mucosa infoldings and invaginations within the wall that are not exposed to the curetting instrument. The cone biopsy has become a standard surgical procedure for AIS. Although the status of the endocervical margin in the cone biopsy specimen is a poor indicator of the likelihood of residual AIS, it and the ECC are the only available markers of the likelihood of residual AIS. Acknowledgments: We thank Donna Sloat for her medical illustration skills and John Watts, MD, for his editorial suggestions. REFERENCES 1. Poyner EA, Barakat RR, Hoskins WJ. Management and followup of patients with adenocarcinoma in situ of the uterine cervix. Gynecol Oncol. 1995;57: Kashimura M, Shinohara M, Oikawa K, et al. An adenocarcinoma in situ of the uterine cervix that developed into invasive adenocarcinoma after 5 years. Gynecol Oncol. 1990;36: Kennedy AW, El Tabbakh GH, Biscotti CV, et al. Invasive adenocarcinoma of the cervix following LLETZ (large loop excision of the transformation zone) for adenocarcinoma in situ. Gynecol Oncol. 1995;58: Bertrand M, Lickrish GM, Colgan TR. The anatomic distribution of cervical adenocarcinoma in situ: implications for treatment. Am / Obstet Gynecol. 1987;157: Luesley DM, Jordan JA, Woodman CBJ, et al. A retrospective review of adenocarcinoma-in-situ and glandular atypia of the uterine cervix. Br J Obstet Gynecol. 1987;94: Boon ME, Baak JPA, Kurver PJH, et al. Adenocarcinoma in situ of the cervix: an underdiagnosed lesion. Cancer. 1981;48: Bousfield L, Pacey F, Young Q, et al. Expanded cytologic criteria for the diagnosis of adenocarcinoma in situ of the cervix and related lesions. Acta Cytol. 1980;24: Im DD, Duska LR, Rosenshein NB. Adequacy of conization margins in adenocarcinoma in situ of the cervix as a predictor of residual disease. Gynecol Oncol. 1995;59: Friedell GH, Hertig AT, Younge PA. The problem of early stromal invasion in carcinoma in situ of the uterine cervix. Arch Pathol. 1958;66: Qizilbash AH. In-situ and microinvasive adenocarcinoma of the uterine cervix: a clinical, cytologic, and histologic study of 14 cases. Am J Clin Pathol. 1975;64: Burghardt E. In situ and microinvasive adenocarcinoma. In: Friedman EA, ed. Early Histological Diagnosis of Cervical Cancer. Philadelphia, Pa: Saunders; 1973: Jaworski RC. Endocervical glandular dysplasia, adenocarcinoma in situ, and early invasive (microinvasive) adenocarcinoma of the uterine cervix. Semin Diagn Pathol. 1990;7: Ostor AG, Pagano R, Davoren RAM, et al. Adenocarcinoma in situ of the cervix, hit / Gynecol Pathol. 1984;3: Fu YS, Berek JS, Hilborne LH. Diagnostic problems of in situ and invasive adenocarcinomas of the uterine cervix. Appl Pathol. 1987;5: Nicklin JL, Wright RG, Bell JR, et al. A clinicopathological study of adenocarcinoma in situ of the cervix: the influence of cervical HPV infection and other factors, and the role of conservative surgery. Aust NZJ Obstet Gynaecol. 1991;31: Hopkins MP, Roberts JA, Schmidt RW. Cervical adenocarcinoma in situ. Obstet Gynecol. 1988;71: Cullimore JE, Luesley DM, Rollason TP, et al. A prospective study of conization of the cervix in the management of cervical intraepithelial glandular neoplasia (CIGN): a preliminary report. Br J Obstet Gynecol. 1992;99: Wolf JK, Levenback C, Malpica A, et al. Adenocarcinoma in situ of the cervix: significance of cone biopsy margins. Obstet Gynecol. 1996;88: Andersen ES, Arffmann E. Adenocarcinoma in situ of the uterine cervix: a clinicopathologic study of 36 cases. Gynecol Oncol. 1989;35: Muntz HG, Bell DA, Lage JM, et al. Adenocarcinoma in situ of the uterine cervix. Obstet Gynecol. 1992;80: Widrich T, Kennedy AW, Myers TM, et al. Adenocarcinoma in situ of the uterine cervix: management and outcome. Gynecol Oncol. 1996;61: Christopherson WM, Nealon N, Gray LA. Noninvasive precursor lesions of adenocarcinoma and mixed adenosquamous carcinoma of the cervix uteri. Cancer. 1979;44: Boon ME, Kirk RS, Rietveld-Scheffers PEM. The morphogenesis of adenocarcinoma of the cervix: a complex pathological entity. Histopathology. 1981;5: Teshima S, Shimosato Y, Kishi K, et al. Early stage adenocarcinoma of the uterine cervix: histopathologic analysis with consideration of histogenesis. Cancer. 1982;56: Noda K, Kimura K, Ikeda M, Teshima K. Studies on the histogenesis of cervical adenocarcinoma. Int ] Gynecol Pathol. 1983;1: Betsill WG Jr, Clark AH. Early endocervical glandular neoplasia, I: histomorphometry and cytomorphology. Acta Cytol. 1986;30: Jaworski RC, Pacey F, Greenberg MA, et al. The histologic diagnosis of adenocarcinoma in situ and related lesions of the cervix uteri: adenocarcinoma in situ. Cancer. 1988;61: Fluhmann CF. The Cervix Uteri and Its Diseases. Philadelphia, Pa: Saunders; 1961: AJCP June 1998