GYNECOLOGIC
ONCOLOGY
46,
182-185 (1992)
The Significance of Cone Biopsy Resection Margins S. PATERSON-BROWN, FRCS, MRCOG, 0. A. CHAPPATTE, FRCS, MRCOG, S. K. CLARK, M.B.B.CHIR., A. WRIGHT, MBBS, P. MAXWELL, MRCP, N. A. TAUB, M.Sc., AND K. S. RAJU, M.D., MRCOG St. Thomas’ Hospital, Lambeth Palace Road, London, SE1 7EH, United Kingdom Received September 19, 1991
This retrospective study was conducted to assessthe
This 1Zyear retrospective study examines the significance of relationship between margin involvement of a cone biopsy margin involvement with dysplasia at cone biopsy in relation to with CIN to subsequent abnormal cytology and histology follow-up. Of 300 cone biopsies, 123 (41%) had margin involveand to see whether further treatment or close follow-up ment. These cases of margin involvement were associated with more severe dysplasia (P < 0.0901) and a higher chance of sub- are more appropriate for this group of women. sequent abnormal cytological follow-up (P < 0.0001) and residual MATERIALS AND METHODS dysplasia at subsequent surgery (P < 0.0001). Involvement of the endocervical margin at the initial cone biopsy was a sensitive Three hundred sixty-three women who underwent knife predictor of future abnormality, with an incidence of subsequent abnormal cytology of 29% and residual disease of 82% in those cone biopsies between 1975 and 1987 at St. Thomas’ Hosundergoing further surgery. In those cases where the margins pital were studied. The ages ranged from 17 to 70 years with a mean of 37, and of these, 79 (22%) were nullihad been clear, the incidence of abnormal cytological follow-up was 8%, with 12% residual diseasein those undergoing subsequent parous and 38 were postmenopausal. Referral was made surgery. Margin involvement was a better predictor of residual following abnormal routine cervical cytology in 262 (72%) diseaseat repeat surgery than abnormal follow-up cytology (pos- patients; 59 (167o) were referred for abnormal bleeding itive predictive value, 79% vs 60%, respectively). o 1992ACAIICC and 42 (12%) for other gynecological problems. Press,Inc.
The indication for cone biopsy was abnormal cervical smear cytology (between 1975 and 1980) and unsatisfactory colposcopy (between 1980 and 1987). Preoperative INTRODUCTION cervical assessment was performed with Lugol’s iodine Controversy exists regarding the follow-up manage- and/or colposcopy, and cone biopsy was performed using ment of women treated by cone biopsy for cervical in- the cold knife technique only. The depth of the cone and traepithelial neoplasia (CIN) when the resection margins the histology were recorded with particular reference to are involved. Further surgery has been advocated in the the involvement or otherwise of the endo- and ectocervpast while more recent opinions suggest that careful cy- ical resection margins. Dysplastic lesions were classified tological follow-up [1,15,5] and endocervical curettage according to the system of Richart, and they were referred [lo] are safe and are better predictors of residual disease to as CIN, CINl (mild), CIN2 (moderate), and CIN3 (severe dysplasia or carcinoma in situ), microinvasive carthan margin involvement alone. Abnormal follow-up cytology, after cone margin in- cinoma, and invasive cancer [20]. Follow-up was performed regularly with cervical smears volvement with CIN, has been variably reported to occur taken at 3, 6, and 12 months and then annually. Ninetyin 28-100% of cases [1,4,13,8]. When further surgery is performed for CIN diagnosed nine women had further surgery; seven of these were for at cone biopsy, residual disease has been associated with invasive cancer. Results were analyzed on computer using a Minitab previous cone margin involvement and variably reported in 34-69% of cases[21,2,15,10]. Likewise the relationship package (Minitab Inc., State College, PA). Data were between endocervical margin involvement and residual compared by the x2 test with Yates’ correction for condysplasia is unclear, with some reporting a direct corre- tinuity, Fisher’s exact test, or the trend test, as appropriate . lation [ 15,10] and others no correlation [17]. 182 OC90-8258/92 $4.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
183
CONE BIOPSY RESECTION MARGINS
TABLE 1 Cone Biopsy Margin Clearance Related to Histology
CIN 1 CIN 2 CIN 3 Microinvasive and invasive Total
Margins clear
Margins involved
Total
21 31 117
1 8 101
22 39 218
8
13
21
177
123
300
Note. Margin involvement correlated with degree of dysplasia: x2 = 19.9, DF = 3, P < 0.0002.
RESULTS Of 363 cone biopsies, 33 had no evidence of dysplasia and were excluded from the study. Of the 330 remaining, 303 (92%) had cervical intraepithelial neoplasia (CIN), of which 26 had CINl, 47 had CIN2, and 230 had CIN3. Of the 27 remaining, 17 had microinvasive disease and 10 had frank invasion. Of the 330 cones, the excision margins were unsuitable for comment in 30 casesand these are therefore excluded from further analysis. Of the 300 cone biopsies with histologically known margins, 177 (59%) had clear margins and 123 (41%) had involved margins. In 123 cones with margin involvement, the endocervical margin alone was involved in 65 (53%), the ectocervical margin alone was involved in 18 (15%), both margins were involved in 15 (12%), and in 25 (20%) the margin involved was unspecified. Tables 1 and 2 show that margin involvement was related to the severity of the dysplasia (P < 0.0002) and this was most marked at the endocervical margin (P < 0.0001). Further surgery in the form of repeat cone biopsy (12) or hysterectomy (71) was performed in 83 patients. Of these, 32 had clear initial cone biopsy margins and surgery was indicated due to microinvasive or invasive disease (5)) recurrent dysplasia (ll), or unrelated gynecological
symptoms (16). Fifty-one patients who had previously had involved cone biopsy margins had surgery performed either due to suspected incomplete excision or due to recurrent dysplasia within 6 months (26). All surgery was performed within 6 months of initial cone biopsy. Residual disease was found in the surgical specimens of 44 patients. Four of these had clear initial cone biopsy margins, while in 40 the margins had been involved (Table 3). Cytological follow-up was performed in all patients not undergoing immediate definitive treatment except for 12 patients who were lost to follow-up. Of the others the first 3-month smear was abnormal in 27% of caseswhere either margin of the cone biopsy had been involved, while it was 9% if it had been clear of dysplasia; this difference was statistically significant (P < 0.001). Subsequent abnormal cytology within 3 years occurred in 16 and 5%, respectively, where margins had been involved or clear of disease. The likelihood of requiring further surgery for residual CIN was related to margin involvement (16% when clear and 47% when involved; P < 0.0001); this was especially notable at the endocervical margin. Thirty-six of the 80 (65 endocervical margin alone, 15 both endo- and ectocervical margins involved) patients with endocervical margin involvement underwent further surgery (45%), while 41 of the 195 (177 clear margins, 18 ectocervical alone involvement) patients with clear endocervical margin had further surgery (21%) (P < 0.0001). Conversely clearance at the ectocervical margin did not correlate significantly with further surgery (36% when involved, 27% when clear), which occurred in 12 of 33 (18 ectocervical alone, 15 both margins) with an involved ectocervical margin as opposed to 65 of 242 (177 clear, 65 endocervical margin alone) with a clear ectocervical margin. The value of cone biopsy margin involvement in predicting residual disease in those undergoing repeat surgery was calculated. Involvement of any margin correctly predicted residual disease in 79% (positive predictive value), while clear margins correctly predicted no residual disease
TABLE 2 Margin Involvement Related to Histology (Total of 123 Cases)
CIN 1 CIN 2 CIN 3 Microinvasive and invasive Total
Endocervical margin alone
Ectocervical margin alone
Both margins
Margin not specified
0 4 54
1 3 12
0 1 13
0 0 22
7
2
1
3
65
18
15
25
Note. Endocervical margin involvement correlated with dysplasia: x2 = 15.3, DF = 1, P < 0.0001. Ectocervical margin involvement did not correlate with dysplasia: P < 0.1.
184
PATERSON-BROWN
ET AL.
TABLE 3 Residual Disease at Subsequent Surgery Related to Cone Margin Involvement
Total number Residual disease Percentage
Involved margins (n = 51)
Clear margins (n = 32)
Endo
Ecto
Both
Unspec
32 4 12
29 22 76
6 4 67
5 4 80
11 10 90
Note. Endo, endocervical margin; ecto, ectocervical margin; both, both margins; unspec, margin unspecified. Residual disease correlated with cone margin involvement: x2 = 33.2, P < O.OOQl.Residual disease correlated with endocervical margin involvement: x2 = 24.5, P < 0.0001. Residual disease did not correlate with ectocervical margin involvement: P > 0.1.
in 88% (negative predictive value). If the endocervical margin was involved the positive predictive value increased to 82%. The specificity of margin clearance describes those cases with no residual disease which had clear margins: with both margins clear this was 72%) while it was 84% with endocervical margin clearance. The sensitivity describes those cases of residual disease where there had been margin involvement; this occured in 91% of those with either margin involved and 77% with endocervical margin involvement. The value of dysplasia on cytological follow-up alone after cone biopsy in predicting residual disease at repeat surgery was also calculated. Of 38 patients with negative smears, 17 (45%) had residual disease, while 21 of 35 (60%) with dysplastic smears had residual disease. Follow-up cytology had a negative predictive value of 55%, a positive predictive value of 60%, a specificity of 60%, and a sensitivity of 55%. By combining normal cytology with margin clearance in predicting no residual disease the specificity rose to 97%, but the sensitivity of the converse (i.e., abnormal cytology and margin involvement associated with residual disease) fell to 43%.
[15,10]. Margin involvement was a better predictoi of residual dysplasia at repeat surgery than abnormal cytological follow-up with a positive predictive value of 79%, a negative predictive value of 88%, a sensitivity of 91%, and a specificity of 72%. This is contrary to the findings of Buxton et al. [5], who found abnormal followup cytology to be a better indicator of residual dysplasia than margin involvement, although using positive cytology to predict residual disease they had a false positive rate of 43%. Using margin involvement as a predictor of residual disease we had a false positive rate of 21%. With this false positive rate, clearly margin involvement alone does not indicate that further surgery should be the best option, but with a significant risk of residual disease being present, close follow-up is mandatory, and if abnormal cytology persists, this study would suggest that the risk of residual dysplasia is further increased to 97%. Conversely complete excision of CIN disease, although still associated with dysplasia in subsequent surgical specimens, is much less likely, with a negative predictive value of 88%; again this is a better indicator than normal cytological follow-up, which, when assessed alone, has a negative predictive value of 55%. This study has investigated cold knife cone biopsy, DISCUSSION which is used increasingly rarely due to the more popular Local ablation of cervical dysplasia, which depends and less destructive loop diathermy technique. “Large upon an accurate colposcopic diagnosis, is being increas- loop excision of the transformation zone” (LLETZ) proingly criticized following the recent reports questioning vides a simple and cheap means of excision biopsy with the reliability of colposcopically directed punch biopsies minimal morbidity [19]. Whether the interpretations from [6,7,9,12,18,22,23]. The trend is thus going back toward this study can be extrapolated to the LLETZ biopsy reexcision biopsy to secure a definitive histological diagnosis main to be seen, but until this information is available, and to plan future management. Evaluation of the re- we must continue to pay respect to the histological insected margins of the cervical biopsy is important in this volvement of the endocervical margin with CIN disease. context in identifying those patients with incomplete exREFERENCES cision, who may be at increased risk of abnormal followup and further surgery. M., Ingemarsson, I., Lindberg, L. G., and Nordqvist, Margin involvement of the cervical cone biopsy was 1. Ahlgren, S. R. B. Conisation as treatment of carcinoma in situ of the uterine related to both subsequent dysplastic smears and residual cervix, Obstet. Gynecol. 46(2), 135-140 (1975). dysplasia at repeat surgery. This was particularly true of 2. Benedet, J. L., Anderson, G. H., Simpson, M. L., and Shaw, D. the endocervical margin, and this is in agreement with Colposcopy, conization, and hysterectomy practices: A current perspective, Obstet. Gynecol. 60(5), 539-545 (1982). previous reports which refer to it as the critical margin
CONE BIOPSY RESECTION MARGINS 3. Bevan, J. R., Attwood, M. E., Jordan, J. A., Lucas, A., and Newton, J. R. Treatment of preinvasive disease of the cervix by cone biopsy, Br. J. Obstet. Gynaecol. 88, 1140-1144 (1981). 4. Bjerre, B., Eliasson, G., Linell, F., Soderberg, H., and Sjoberg, N. Conisation as only treatment of carcinoma in situ of the uterine cervix, Am. J. Obstet. Gynecol. X25(2), 143-152 (1976). 5. Buxton, E. J., Luesley, D. M., Wade-Evans, T., and Jordan, J. A. Residual disease after cone biopsy: Completeness of excision and follow-up cytology as predictive factors, Obstet. Gynecol. 70(4), 529-532 (1987). 6. Carter, P. G., Grant-Harris, V., and Wilson, P. 0. G. Br. J. Obstet. Gynaecol. 98, 232-233 (1991). 7. Chappatte, 0. A., Byrne, D., Raju, K. S., Nayagam, M., and Kenney, A. Histological differences between colposcopic directed biopsy and loop excision of the transformation zone (LLETZ): A cause for concern, Gynecol. Oncol. 43(l), 46-50 (1991). 8. Holdt, D. G., Jacobs, A. J., Scott, J. C., and Adam, G. M. Diagnostic significance and sequelae of cone biopsy, Am. J. Obstet. Gynecol. 143(3), 312-318 (1982). 9. Howe, D. T., and Vincenti, A. C. Is large loop excision of the transformation zone (LLETZ) more accurate than colposcopically directed punch biopsy in the diagnosis of cervical intraepithelial neoplasia? Br. J. Obstet. Gynaecol. 98, 588-591 (1991). 10. Husseinzadeh, N., Shbaro, I., and Wesseler, T. Predictive value of cone margins and post-cone endocervical curettage with residual disease in subsequent hysterectomy, Gynecol. Oncol. 33, 198-200 (1989). 11. Jones, H. W. Cone biopsy in the management of cervical intraepithelial neoplasia, Clin. Obstet. Gynecol. 26, 968-979 (1983). 12. Jones, H. W., and Buller, R. E. The treatment of cervical intraepithelial neoplasia by cone biopsy, Am. J. Obstet. Gynecol. 137, 882-886 (1980).
185
13. Larsson, G. Conization for cervical dysplasia and carcinoma in situ: Long term follow-up of 1013 women, Ann. Chir. Gynaecol. 70,7985 (1981). 14. Lopes, A., Pearson, S. E., Mor-Yosef, S., Ireland, D., and Monaghan, J. M. Is it time for a reconsideration of the criteria for cone biopsy? Br. J. Obstet. Gynaecol. 96, 1345-1347 (1989). 15. Lubicz, S., Ezekweche, C., Allen, A., and Schiffer, M. Significance of cone biopsy margins in the management of patients with cervical neoplasia, J. Reprod. Med. 29, 179-185 (1984). 16. Mor-Yosef, S., Lopes, A., Pearson, S., and Monaghan, J. M. Loop diathermy cone biopsy, Obstet. Gynecol. 75(5), 884-886 (1990). 17. Ostergard, D. R. Prediction of clearance of cervical intraepithelial neoplasia by conisation, Obstet. Gynecol. 56(l), 77-80 (1980). 18. Phipps, J. H., Gunasekara, P. C., and Lewis, B. V. Occult cervical carcinoma revealed by large loop diathermy, Br. Med. J. II 8660, 453-456 (1989). 19. Prendiville, W., Cullimore, J., and Norman, S. Large loop excisian of the transformation zone (LETZ). A new method of management of women with cervical intraepithelial neoplasia. Br. J. Obstet. Gynecol. 96, 1054-1060 (1989). 20. Richart, R. M. Cervical intraepithelial neoplasia, Puthol. Ann. 8, 301-328 (1973). 21. Schiffer, M. A., Greene, H. J., Pomerance, W., and Moltz, A. Cervical conization for diagnosis and treatment of carcinoma in situ, Am. J. Obstet. Gynecol. 93(6), 889-895 (1965). 22. Skehan, M., Soutter, W. P., Lim, K., Krausz, and Pryse-Davies, J. Reliability of colposcopy and directed punch biopsy, Br. J. Obstet. Gynecol. 97, 811-816 (1990). 23. Yarnoz, M. C., Cortes, J., Llompart, M., Torrecabota, J., Rossello, J. J., and Amengual, E. The colposcopy and the cone biopsy in the diagnosis, treatment and follow-up of 81 cases of cervical intraepithelial neoplasia, Eur. J. Gynaecol. Oncol. 4,345-349 (1988).
ONCOLOGY
46,
182-185 (1992)
The Significance of Cone Biopsy Resection Margins S. PATERSON-BROWN, FRCS, MRCOG, 0. A. CHAPPATTE, FRCS, MRCOG, S. K. CLARK, M.B.B.CHIR., A. WRIGHT, MBBS, P. MAXWELL, MRCP, N. A. TAUB, M.Sc., AND K. S. RAJU, M.D., MRCOG St. Thomas’ Hospital, Lambeth Palace Road, London, SE1 7EH, United Kingdom Received September 19, 1991
This retrospective study was conducted to assessthe
This 1Zyear retrospective study examines the significance of relationship between margin involvement of a cone biopsy margin involvement with dysplasia at cone biopsy in relation to with CIN to subsequent abnormal cytology and histology follow-up. Of 300 cone biopsies, 123 (41%) had margin involveand to see whether further treatment or close follow-up ment. These cases of margin involvement were associated with more severe dysplasia (P < 0.0901) and a higher chance of sub- are more appropriate for this group of women. sequent abnormal cytological follow-up (P < 0.0001) and residual MATERIALS AND METHODS dysplasia at subsequent surgery (P < 0.0001). Involvement of the endocervical margin at the initial cone biopsy was a sensitive Three hundred sixty-three women who underwent knife predictor of future abnormality, with an incidence of subsequent abnormal cytology of 29% and residual disease of 82% in those cone biopsies between 1975 and 1987 at St. Thomas’ Hosundergoing further surgery. In those cases where the margins pital were studied. The ages ranged from 17 to 70 years with a mean of 37, and of these, 79 (22%) were nullihad been clear, the incidence of abnormal cytological follow-up was 8%, with 12% residual diseasein those undergoing subsequent parous and 38 were postmenopausal. Referral was made surgery. Margin involvement was a better predictor of residual following abnormal routine cervical cytology in 262 (72%) diseaseat repeat surgery than abnormal follow-up cytology (pos- patients; 59 (167o) were referred for abnormal bleeding itive predictive value, 79% vs 60%, respectively). o 1992ACAIICC and 42 (12%) for other gynecological problems. Press,Inc.
The indication for cone biopsy was abnormal cervical smear cytology (between 1975 and 1980) and unsatisfactory colposcopy (between 1980 and 1987). Preoperative INTRODUCTION cervical assessment was performed with Lugol’s iodine Controversy exists regarding the follow-up manage- and/or colposcopy, and cone biopsy was performed using ment of women treated by cone biopsy for cervical in- the cold knife technique only. The depth of the cone and traepithelial neoplasia (CIN) when the resection margins the histology were recorded with particular reference to are involved. Further surgery has been advocated in the the involvement or otherwise of the endo- and ectocervpast while more recent opinions suggest that careful cy- ical resection margins. Dysplastic lesions were classified tological follow-up [1,15,5] and endocervical curettage according to the system of Richart, and they were referred [lo] are safe and are better predictors of residual disease to as CIN, CINl (mild), CIN2 (moderate), and CIN3 (severe dysplasia or carcinoma in situ), microinvasive carthan margin involvement alone. Abnormal follow-up cytology, after cone margin in- cinoma, and invasive cancer [20]. Follow-up was performed regularly with cervical smears volvement with CIN, has been variably reported to occur taken at 3, 6, and 12 months and then annually. Ninetyin 28-100% of cases [1,4,13,8]. When further surgery is performed for CIN diagnosed nine women had further surgery; seven of these were for at cone biopsy, residual disease has been associated with invasive cancer. Results were analyzed on computer using a Minitab previous cone margin involvement and variably reported in 34-69% of cases[21,2,15,10]. Likewise the relationship package (Minitab Inc., State College, PA). Data were between endocervical margin involvement and residual compared by the x2 test with Yates’ correction for condysplasia is unclear, with some reporting a direct corre- tinuity, Fisher’s exact test, or the trend test, as appropriate . lation [ 15,10] and others no correlation [17]. 182 OC90-8258/92 $4.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
183
CONE BIOPSY RESECTION MARGINS
TABLE 1 Cone Biopsy Margin Clearance Related to Histology
CIN 1 CIN 2 CIN 3 Microinvasive and invasive Total
Margins clear
Margins involved
Total
21 31 117
1 8 101
22 39 218
8
13
21
177
123
300
Note. Margin involvement correlated with degree of dysplasia: x2 = 19.9, DF = 3, P < 0.0002.
RESULTS Of 363 cone biopsies, 33 had no evidence of dysplasia and were excluded from the study. Of the 330 remaining, 303 (92%) had cervical intraepithelial neoplasia (CIN), of which 26 had CINl, 47 had CIN2, and 230 had CIN3. Of the 27 remaining, 17 had microinvasive disease and 10 had frank invasion. Of the 330 cones, the excision margins were unsuitable for comment in 30 casesand these are therefore excluded from further analysis. Of the 300 cone biopsies with histologically known margins, 177 (59%) had clear margins and 123 (41%) had involved margins. In 123 cones with margin involvement, the endocervical margin alone was involved in 65 (53%), the ectocervical margin alone was involved in 18 (15%), both margins were involved in 15 (12%), and in 25 (20%) the margin involved was unspecified. Tables 1 and 2 show that margin involvement was related to the severity of the dysplasia (P < 0.0002) and this was most marked at the endocervical margin (P < 0.0001). Further surgery in the form of repeat cone biopsy (12) or hysterectomy (71) was performed in 83 patients. Of these, 32 had clear initial cone biopsy margins and surgery was indicated due to microinvasive or invasive disease (5)) recurrent dysplasia (ll), or unrelated gynecological
symptoms (16). Fifty-one patients who had previously had involved cone biopsy margins had surgery performed either due to suspected incomplete excision or due to recurrent dysplasia within 6 months (26). All surgery was performed within 6 months of initial cone biopsy. Residual disease was found in the surgical specimens of 44 patients. Four of these had clear initial cone biopsy margins, while in 40 the margins had been involved (Table 3). Cytological follow-up was performed in all patients not undergoing immediate definitive treatment except for 12 patients who were lost to follow-up. Of the others the first 3-month smear was abnormal in 27% of caseswhere either margin of the cone biopsy had been involved, while it was 9% if it had been clear of dysplasia; this difference was statistically significant (P < 0.001). Subsequent abnormal cytology within 3 years occurred in 16 and 5%, respectively, where margins had been involved or clear of disease. The likelihood of requiring further surgery for residual CIN was related to margin involvement (16% when clear and 47% when involved; P < 0.0001); this was especially notable at the endocervical margin. Thirty-six of the 80 (65 endocervical margin alone, 15 both endo- and ectocervical margins involved) patients with endocervical margin involvement underwent further surgery (45%), while 41 of the 195 (177 clear margins, 18 ectocervical alone involvement) patients with clear endocervical margin had further surgery (21%) (P < 0.0001). Conversely clearance at the ectocervical margin did not correlate significantly with further surgery (36% when involved, 27% when clear), which occurred in 12 of 33 (18 ectocervical alone, 15 both margins) with an involved ectocervical margin as opposed to 65 of 242 (177 clear, 65 endocervical margin alone) with a clear ectocervical margin. The value of cone biopsy margin involvement in predicting residual disease in those undergoing repeat surgery was calculated. Involvement of any margin correctly predicted residual disease in 79% (positive predictive value), while clear margins correctly predicted no residual disease
TABLE 2 Margin Involvement Related to Histology (Total of 123 Cases)
CIN 1 CIN 2 CIN 3 Microinvasive and invasive Total
Endocervical margin alone
Ectocervical margin alone
Both margins
Margin not specified
0 4 54
1 3 12
0 1 13
0 0 22
7
2
1
3
65
18
15
25
Note. Endocervical margin involvement correlated with dysplasia: x2 = 15.3, DF = 1, P < 0.0001. Ectocervical margin involvement did not correlate with dysplasia: P < 0.1.
184
PATERSON-BROWN
ET AL.
TABLE 3 Residual Disease at Subsequent Surgery Related to Cone Margin Involvement
Total number Residual disease Percentage
Involved margins (n = 51)
Clear margins (n = 32)
Endo
Ecto
Both
Unspec
32 4 12
29 22 76
6 4 67
5 4 80
11 10 90
Note. Endo, endocervical margin; ecto, ectocervical margin; both, both margins; unspec, margin unspecified. Residual disease correlated with cone margin involvement: x2 = 33.2, P < O.OOQl.Residual disease correlated with endocervical margin involvement: x2 = 24.5, P < 0.0001. Residual disease did not correlate with ectocervical margin involvement: P > 0.1.
in 88% (negative predictive value). If the endocervical margin was involved the positive predictive value increased to 82%. The specificity of margin clearance describes those cases with no residual disease which had clear margins: with both margins clear this was 72%) while it was 84% with endocervical margin clearance. The sensitivity describes those cases of residual disease where there had been margin involvement; this occured in 91% of those with either margin involved and 77% with endocervical margin involvement. The value of dysplasia on cytological follow-up alone after cone biopsy in predicting residual disease at repeat surgery was also calculated. Of 38 patients with negative smears, 17 (45%) had residual disease, while 21 of 35 (60%) with dysplastic smears had residual disease. Follow-up cytology had a negative predictive value of 55%, a positive predictive value of 60%, a specificity of 60%, and a sensitivity of 55%. By combining normal cytology with margin clearance in predicting no residual disease the specificity rose to 97%, but the sensitivity of the converse (i.e., abnormal cytology and margin involvement associated with residual disease) fell to 43%.
[15,10]. Margin involvement was a better predictoi of residual dysplasia at repeat surgery than abnormal cytological follow-up with a positive predictive value of 79%, a negative predictive value of 88%, a sensitivity of 91%, and a specificity of 72%. This is contrary to the findings of Buxton et al. [5], who found abnormal followup cytology to be a better indicator of residual dysplasia than margin involvement, although using positive cytology to predict residual disease they had a false positive rate of 43%. Using margin involvement as a predictor of residual disease we had a false positive rate of 21%. With this false positive rate, clearly margin involvement alone does not indicate that further surgery should be the best option, but with a significant risk of residual disease being present, close follow-up is mandatory, and if abnormal cytology persists, this study would suggest that the risk of residual dysplasia is further increased to 97%. Conversely complete excision of CIN disease, although still associated with dysplasia in subsequent surgical specimens, is much less likely, with a negative predictive value of 88%; again this is a better indicator than normal cytological follow-up, which, when assessed alone, has a negative predictive value of 55%. This study has investigated cold knife cone biopsy, DISCUSSION which is used increasingly rarely due to the more popular Local ablation of cervical dysplasia, which depends and less destructive loop diathermy technique. “Large upon an accurate colposcopic diagnosis, is being increas- loop excision of the transformation zone” (LLETZ) proingly criticized following the recent reports questioning vides a simple and cheap means of excision biopsy with the reliability of colposcopically directed punch biopsies minimal morbidity [19]. Whether the interpretations from [6,7,9,12,18,22,23]. The trend is thus going back toward this study can be extrapolated to the LLETZ biopsy reexcision biopsy to secure a definitive histological diagnosis main to be seen, but until this information is available, and to plan future management. Evaluation of the re- we must continue to pay respect to the histological insected margins of the cervical biopsy is important in this volvement of the endocervical margin with CIN disease. context in identifying those patients with incomplete exREFERENCES cision, who may be at increased risk of abnormal followup and further surgery. M., Ingemarsson, I., Lindberg, L. G., and Nordqvist, Margin involvement of the cervical cone biopsy was 1. Ahlgren, S. R. B. Conisation as treatment of carcinoma in situ of the uterine related to both subsequent dysplastic smears and residual cervix, Obstet. Gynecol. 46(2), 135-140 (1975). dysplasia at repeat surgery. This was particularly true of 2. Benedet, J. L., Anderson, G. H., Simpson, M. L., and Shaw, D. the endocervical margin, and this is in agreement with Colposcopy, conization, and hysterectomy practices: A current perspective, Obstet. Gynecol. 60(5), 539-545 (1982). previous reports which refer to it as the critical margin
CONE BIOPSY RESECTION MARGINS 3. Bevan, J. R., Attwood, M. E., Jordan, J. A., Lucas, A., and Newton, J. R. Treatment of preinvasive disease of the cervix by cone biopsy, Br. J. Obstet. Gynaecol. 88, 1140-1144 (1981). 4. Bjerre, B., Eliasson, G., Linell, F., Soderberg, H., and Sjoberg, N. Conisation as only treatment of carcinoma in situ of the uterine cervix, Am. J. Obstet. Gynecol. X25(2), 143-152 (1976). 5. Buxton, E. J., Luesley, D. M., Wade-Evans, T., and Jordan, J. A. Residual disease after cone biopsy: Completeness of excision and follow-up cytology as predictive factors, Obstet. Gynecol. 70(4), 529-532 (1987). 6. Carter, P. G., Grant-Harris, V., and Wilson, P. 0. G. Br. J. Obstet. Gynaecol. 98, 232-233 (1991). 7. Chappatte, 0. A., Byrne, D., Raju, K. S., Nayagam, M., and Kenney, A. Histological differences between colposcopic directed biopsy and loop excision of the transformation zone (LLETZ): A cause for concern, Gynecol. Oncol. 43(l), 46-50 (1991). 8. Holdt, D. G., Jacobs, A. J., Scott, J. C., and Adam, G. M. Diagnostic significance and sequelae of cone biopsy, Am. J. Obstet. Gynecol. 143(3), 312-318 (1982). 9. Howe, D. T., and Vincenti, A. C. Is large loop excision of the transformation zone (LLETZ) more accurate than colposcopically directed punch biopsy in the diagnosis of cervical intraepithelial neoplasia? Br. J. Obstet. Gynaecol. 98, 588-591 (1991). 10. Husseinzadeh, N., Shbaro, I., and Wesseler, T. Predictive value of cone margins and post-cone endocervical curettage with residual disease in subsequent hysterectomy, Gynecol. Oncol. 33, 198-200 (1989). 11. Jones, H. W. Cone biopsy in the management of cervical intraepithelial neoplasia, Clin. Obstet. Gynecol. 26, 968-979 (1983). 12. Jones, H. W., and Buller, R. E. The treatment of cervical intraepithelial neoplasia by cone biopsy, Am. J. Obstet. Gynecol. 137, 882-886 (1980).
185
13. Larsson, G. Conization for cervical dysplasia and carcinoma in situ: Long term follow-up of 1013 women, Ann. Chir. Gynaecol. 70,7985 (1981). 14. Lopes, A., Pearson, S. E., Mor-Yosef, S., Ireland, D., and Monaghan, J. M. Is it time for a reconsideration of the criteria for cone biopsy? Br. J. Obstet. Gynaecol. 96, 1345-1347 (1989). 15. Lubicz, S., Ezekweche, C., Allen, A., and Schiffer, M. Significance of cone biopsy margins in the management of patients with cervical neoplasia, J. Reprod. Med. 29, 179-185 (1984). 16. Mor-Yosef, S., Lopes, A., Pearson, S., and Monaghan, J. M. Loop diathermy cone biopsy, Obstet. Gynecol. 75(5), 884-886 (1990). 17. Ostergard, D. R. Prediction of clearance of cervical intraepithelial neoplasia by conisation, Obstet. Gynecol. 56(l), 77-80 (1980). 18. Phipps, J. H., Gunasekara, P. C., and Lewis, B. V. Occult cervical carcinoma revealed by large loop diathermy, Br. Med. J. II 8660, 453-456 (1989). 19. Prendiville, W., Cullimore, J., and Norman, S. Large loop excisian of the transformation zone (LETZ). A new method of management of women with cervical intraepithelial neoplasia. Br. J. Obstet. Gynecol. 96, 1054-1060 (1989). 20. Richart, R. M. Cervical intraepithelial neoplasia, Puthol. Ann. 8, 301-328 (1973). 21. Schiffer, M. A., Greene, H. J., Pomerance, W., and Moltz, A. Cervical conization for diagnosis and treatment of carcinoma in situ, Am. J. Obstet. Gynecol. 93(6), 889-895 (1965). 22. Skehan, M., Soutter, W. P., Lim, K., Krausz, and Pryse-Davies, J. Reliability of colposcopy and directed punch biopsy, Br. J. Obstet. Gynecol. 97, 811-816 (1990). 23. Yarnoz, M. C., Cortes, J., Llompart, M., Torrecabota, J., Rossello, J. J., and Amengual, E. The colposcopy and the cone biopsy in the diagnosis, treatment and follow-up of 81 cases of cervical intraepithelial neoplasia, Eur. J. Gynaecol. Oncol. 4,345-349 (1988).
