Diagnostic and therapeutic efficacy of cervical conization J. R. VAN NAGELL, JR., M.D.
J.
C. PARKER, jR., M.D.
L. P. HICKS, M.D. R. CONRAD, B.S. G. ENGLAND, B.S.
Lexington, Kentucky
Cervical conization was performed on 756 patients at the University of Kentucky Medical Center from july I, 1964, to january 1, 1973. Sixty-six patients were pregnant at the time of conization. Eighty-six per cent of patients with cytologic findings of carcinoma in situ had hi5tologic verification of carcinoma in situ or sel!ere dysplasia, and there was absolute correlation between cytology and histology in 75 per cent of patients with occult inl!asive cancer. Cervical biopsies without colposcopic direction predicted either severe dysplasia or carcinoma in situ in 77 per cent of cases but were accurate in only seven of 24 patients with occult invasive cancer. Carcinoma in situ was present in 30 per cent of hysterectomy specimens following conization but recurred in only seven per cent of patients followed without hysterectomy. Recurrent carcinoma in situ following hysterectomy was more common in patients with residual intraepithelial cancer in the uterus but was independent of the size of the vaginal cuff removed. Major postconization complications requiring hospitalization occurred in 3.4 per cent of nonpregnant patients and in 7.5 per cent of pregnant patients.
SINCE ITS INITIAL description by Lisfranc 1 in 1815, the technique of cervical conization has undergone many changes. Sims, 2 in 1861, dosed the cervix with silver wire sutures following conization and was the first to obtain healing by primary intention. Likewise, Sturmdorf,3 in 1916, described a method inverting cervicovaginal epithelium to cover the conization site which is still used in many institutions today. More recently, several investigators4 • 5 have described methods of injecting vasoconstrictive solutions into the cervix at the time of conization in order to minimize blood loss. Despite these technical advances, cervical
conization remams a major operative procedure associated with a five to ten per cent incidence of postoperative complications.4• 6 Perhaps even more important than changes in conization technique have been changes in the indications for its performance. Whereas conization was initially performed for invasive cervical cancer and the repair of cervical lacerations/ it has been used during the past three decades for the diagnosis and treatment of cervical epithelial abnormalities. With the utilization of colposcopy to direct cervical biopsies, it has been suggested by some investigators that cervical conization is being performed too frequently. 8 • 9 The present study was undertaken to determine the diagnostic and therapeutic efficacy of cervical conization as a method in the evaluation of patients with cervical neoplasia.
From the Departments of Obstetrics and Gynecology and Pathology, University of Kent!.I.Cky Medical Center. Supported lry the American Cancer Society, Kentucky Division. Received for publicatwn September 25, 1974.
Methods and results
Revised january 31, 1975.
From July 1, 1964, to January I. 1973, cervical conization was performed on 756 patients at the University of Kentucky Medical Center. Sixty-six of these patients were pregnant at the time of conization.
Accepted February 13, 1975. Reprint requests: Dr.]. R. van Nagell, Jr., Department of Obstetrics and Gynecology, University of Kentucky Medical Center, Lexington, Kent!J.Cky 40506.
134
Efficacy of cervical conization
Volume 124 Number 2
Conization was carried out as a diagnostic method in patients with cytologic or biopsy findings of cervical epithelial abnormalities. Cytologic and histologic abnormalities were graded according to the following classification: mild dysplasia, moderate dysplasia, severe dysplasia, atypical cellular changes, carcinoma in situ, mitroinvasive carcinoma, and invasive carcinoma. Cervical conization was performed in the absence of biopsy-proved invasive cancer within two months on all patients with cytologic or biopsy evidence of persistent moderate dysplasia, severe dysplasia, carcinoma in citu, or microinvasive carcinoma. Persistent moderate dysplasia was defined as moderate dysplasia on two occasions separated by at least three months. Microinvasive cancer was defined as the presence of stromal invasion to a depth of less than 5 mm. Conization was performed according to the method of Parker and colleagues. 10 Twelve to sixteen sections were made of each conization specimen, and additional step sections were cut when an area indicative of epithelial abnormality was noted. Areas of cervical epithelial abnormality were described as involving squamous epithelium, both glandular and squamous epithelium, or glandular epithelium alone. Basic duiracteristics of the patients studied are presented in Table I. The mean age, gravidity, and location of cervical epithelial abnormality was essentially the same in those patients who were pregnant and in those who were not. Seven hundred and seven patients had cervical conizations at this institution within eight weeks of cytologic examination. Conization findings in these patients are presented in Table II. Cytology and histology specimens were reviewed in all cases in which there was not absolute cytology-histology correlation. Eighty-six per cent of patients with cervical cytologic findings of carcinoma in situ had histologic verification of severe dysplasia or intraepithelial carcinoma on conization. Seventeen patients with normal or mildly dysplastic cytologic findings had severe dysplasia or carcinoma in situ on conization. However, no patient whose cervical cytology was normal or showed mild or moderate dysplasia had histologic evidence of invasive cancer in the conization specimen. Twenty-eight patients (four per cent) with no visible cervical lesions had invasive cancer at conization. Twenty-one of these patients had cytologic diagnoses of invasive cancer, and the remainder had cytologic abnormalities ranging from atypical cellular changes to microinvasive cancer. Age as an independent variable had no direct effect on the cytology-histology correlation. Thirty-eight patients (5.4 per cent) had cytologic findings which were
135
Table I. Basic characteristics of patients undergoing cervical conization Characteristics
Nonpregnant
Pregnant
No. of patients Mean age (yr.) Mean gravidity Location of lesion Squamous epithelium alone(%) Squamous and glandular epithelium (%) Glandular epithelium alone(%)
690 37.0 4.6
66 33.3 5.2
62.6
57.8
34.5
35.9
2.9
6.3
not severe enough to warrant conization. However, cervical biopsy abnormalities were more advanced and indicated the need for conization. Cervical biopsies were performed on 53 7 patients prior to conization. Biopsies were taken from areas in the cervix which failed to take up Schiller's stain, since colposcopy was not performed at this institution during the time of investigation. Endocervical curettage was carried out only in those patients with visible endocervical lesions. The histologic correlation between punch biopsies and conization is presented in Table III. Twenty-four patients with invasive cancer and no visible cervical lesions had punch biopsies prior to conization. Invasive cancer was confined to the squamous epithelium in eleven cases and involved both squamous and glandular epithelium in three cases. Punch biopsies showed microinvasive or invasive cancer in nine patients (37.5 per cent) and carcinoma in situ in II patients (45.8 per cent). Thirteen patients with suspected invasive cancer on punch biopsy had cervical conization in order to determine the extent of invasion. Seven of these patients had invasive cancer and six had intraepithelial or microinvasive cancer. The relationship between location of the epithelial abnormality and the diagnostic accuracy of cervical biopsy is presented in Tables IV and V. Cervical biopsies showed either severe dysplasia or carcinoma in situ in I45 of I93 patients (75.I per cent) with carcinoma in situ limited to the squamous epithelium and in 119 of 151 patients (78. 7 per cent) whose carcinoma in situ involved both squamous and glandular epithelium. Diagnostic accuracy of punch biopsies was significantly reduced in those lesions confined to the endocervical glands. Eleven patients had either carcinoma in situ or invasive adenocarcinoma of the endocervix, and only six of them (54.5 per cent) were diagnosed accurately on punch biopsy.
January 15, 1976 Obstet. Gvnecol.
136 Van Nagell et al.
J.
Am.
Table II. Cytology-histology correlation in patients undergoing cervical conization Conization findings Mild dysplasia or less Cervical cytologic findings
No. of patients
No.
Normal Mild dysplasia Moderate dysplasia Severe dysplasia Atypical cells Carcinoma in situ Microinvasive carcinoma Invasive carcinoma
9 29 103 163 19 345 7 32
4 II 9 13 3 12 0 0
I
Moderate dysplasia
%
No.
44.4 37.9 8.6 8.0 15.8 3.5
0 6 32 25 2
I
6
Severe dysplasia
%
No.
20.7 30.8 2.'i.3 10.5 1.7
2 9 36 2 22 0 0
()
0
I
Microinvasive carcinoma
Carcinoma in situ
%
No.
11.2 6.9 8.6 22.1 10.5 6.4
4 10 49 76 11 274 4 7
%
No.
44.4 34.5 47.1 46.6 57.9 79.4 57.1 21.9
0 0 4 12 0 28 I
I
I
%
Invasive carcinoma No.
I
%
0 0
4
3.9 7.4
()
8.1 14.3 12.5
3 2 21
0.6 5.3 0.9 28.6 65.6
I I
Table III. Biopsy-conization correlation in patients with cervical epithelial abnormalities Conization findings Mild dysplasia or less Biopsy findings
No.
Normal Atypical cellular changes Mild dysplasia Moderate dysplasia Severe dysplasia Carcinoma in situ Microinvasive carcinoma Invasive carcinoma No. of patients
3 0 8 5 4 2 0 0 22
I
Moderate dysplasia
%
No.
13.6
12 I 10 23 13 I 0 0
36.4 22.7 18.2 9.1
I
%
No.
20.0 1.7 16.7 38.2 21.7 1.7
6 I 6 15 19 12 0 0
60
I
Microinvasive carcmoma
Carcinoma in situ
Severe dysplasia
I
%
No.
10.2 1.7 10.2 25.4 32.2 20.3
27 6 8 32 55 198 3 5
59
%
No.
8.1 1.8 2.4 16.5 16.5 59.4 0.9 1.5
2 0 1 3 3 25 3 I
I
Invasive carcinoma
%
No.
5.3
0 0 0 3 1 11 2 7
2.6 7.9 7.9 65.8 7.9 2.6
%
12.5 4.2 45.8 8.3 29.2
24
38
334
I
Table IV. Punch biopsy-conization correlation according to location of lesion in patients with carcinoma in situ Biopsy findings Severe dysplasia Location of carcinoma in situ on conization Squamous epithelium only Squamous and glandular epithelium Glandular epithelium only
Patients
No.
193 151
34 21
Sixty-six patients were pregnant at the time of conization and sixty-three of these patients had conization within six weeks of cytologic examination. Thirtyseven (82.2 per cent) of the 45 patients with carcinoma in situ on conization had cytologic evidence of severe dysplasia or carcinoma in situ. The one pregnant patient with occult invasive cervical cancer had
0
I
Other
Carcinoma in situ
%
No.
17.6 13.8
Ill 98
I
%
No.
57.5 64.9
48 32
100
0
I
% 24.9 21.3
cytologic evidence of malignancy. Fifty-three of these patients had cervical biopsies prior to conization. Cervical biopsies were accurate in 78.9 per cent of pregnant patients with carcinoma in situ. However, only one of the three patients with microinvasive cancer was diagnosed on cervical biopsy. The one pregnant patient with invasive cancer was not biopsied
Efficacy of cervical conization 137
Volume 124 Number 2
Table V. Punch biopsy-conization correlation according to location of lesion in patients with occult invasive cervical cancer BiosfrY findings Carcinoma in situ Location of invasive carcinoma on conization
No. of patients
Squamous epithelium only Squamous and glandular epithelium Glandular epithelium only
11 3
6 1
54.5 33.3
10
4
40.0
No.
l
%
prior to conization. Hysterectomy was performed with cesarean section or in the immediate postpartum period in 43 patients with carcinoma in situ. Twelve patients (27.9 per cent) had residiual carcinoma in situ and one had microinvasive carcinoma in the hysterectomy specimen. Complications of conization are presented in Table VI. Sixteen patients required readmission for excessive cervical bleeding. This bleeding was successfully treated with resuture, vaginal packing, or cryotherapy in thirteen patients, and emergency hysterectomy was required in three patients. Bleeding occurred five to ten days after conization in all patients. Six patients required blood transfusions following conization related bleeding. Uterine perforation at the time of conization occurred in two patients, and three patients experienced spontaneous abortion within one week of conization. The incidence of major conization complications, excluding the 161 patients whose hysterectomy immediately followed frozen section examination of the cone specimen, was 3.4 per cent in nonpregnant patients and 7.5 per cent in pregnant patients. Three hundred and seventy four patients with carcinoma in situ or microinvasive carcinoma who desired no further pregnancies underwent hysterectomy following conization. Vaginal hysterectomy was performed in 299 patients and abdominal hysterectomy in the remainder. A wide vaginal cuff was excised with the hysterectomy specimen in 226 cases. All hysterectomy specimens were examined microscopically to determine the incidence of residual epithelial abnormality. Sections were routinely taken through the anterior, posterior, and lateral walls of the cervix as well as additional sections in areas of histologic abnormality. No patient had undiagnosed invasive cancer in the hysterectomy specimens following conization. Carcinoma in situ was present in 112 hysterectomy specimens (29.9 per cent) and microinvasive
Microinvasive carcinoma No.
I
Invasive carcinoma
%·
No.
9.1 33.3
2
%
No.
18.2
2
0
5
0
I
Other
I
% 18.2 33.3
1
50.0
10.0
Table VI. Complications of cervical conization Nonpregnant Complications
No.
No. of patients Complication Cervical bleeding requiring hospitalization Uterine perforation Pelvic infection Pregnancy loss Total
529
I
%
Pregnant No.
I
%
66
14
2.6
2
3.0
2 2
0.4 0.4
0 0 3
4.5
3.4
5
7.5
0
18
cancer in six (1.6 per cent). Following conization, patients were seen in the gynecologic oncology clinic every three months for two years and every six months thereafter. The duration of follow-up ranged from 12 to 110 months (mean 50 months). Recurrent vaginal cancer following hysterectomy occurred in five patients ( 1.1 per cent), and there was residual carcinoma in situ in the hysterectomy specimen of four of these patients. Vaginectomy was performed on three of these patients and radiation therapy in two. All of these patients are alive with no evidence of recurrent disease three to 60 months following therapy. One hundred and nine patients with histologically proved carcinoma in situ who desired further pregnancies have been followed without hysterectomy. Seven of these patients (6.4 per cent) have required a second conization and hysterectomy for recurrent carcinoma in situ.
Comment The findings of this study indicate that cervical conization is an effective method in the diagnosis of cervical epithelial abnormalities. However, it is a major operative procedure associated with a significant incidence of complications. For this reason, conization should be performed only in those instances in which it
138 Van Nagel! et al.
can provide definite diagnostic or therapeutic advantages to less radical procedures. Cervical cytologic examination is an excellent screening procedure and suggests the need for further histologic confirmation by biopsy or conization. Furthermore, cytologic examination and cervical biopsies are often complementary in the diagnosis of cervical epithelial abnormalities. This is exemplified by those patients (5.4 per cent) from this institution with normal or mildly dysplastic cytologic findings whose conizations were performed because of biopsy evidence of more advanced epithelial abnormalities. The most effective method for reducing the number of cervical conizations is to increase the accuracy of cervical biopsies. In the present series, only seven of 24 patients with invasive carcinoma in the absence of a cervical lesion had the correct diagnosis made by cervical biopsy. Similarly, only 76.8 per cent of the carcinoma in situ or severe dysplasia was diagnosed by biopsy alone. Stafl and Mattingly 11 reported colposcopic prediction of histopathology to be clinically accurate in 85 per cent of patients with cervical epithelial abnormalities. Likewise, Townsend and associates12 noted an absolute correlation between colposcopically directed biopsies and cervical conization of 84.5 per cent when the lesion did not involve the endocervical canal. Clearly, the accuracy of colposcopy depends upon the skill and experience of the colposcopist. However, it does provide an additional diagnostic alternative to conization if the lesion can be completely defined and the squamocolumnar jum:tion fully visualized. Conization during pregnancy was directly responsible for a five per cent incidence of fetal loss. In addition, two patients experienced significant postconization hemorrhage. Ortiz and Newton, 13 in a study of 47 pregnant patients with abnormal cervical cytologic findings, found that colposcopically directed biopsy was an effective means to histologically evaluate the cervix and was unlikely to miss invasive cancer because of the physiologic eversion of the cervix during pregnancy. Similarly, Stafl and Mattingly 11 found that no unsatisfactory or incomplete colposcopic examinations occurred in 89 pregnant patients, and
REFERENCES I. Lisfranc, J.: Gaz. Med. France 2: 835, 1815. 2. Sims, J. M.: Trans. Med. Soc. N. Y. 137: 367, 1861. 3. Sturmdorf, A.: Surg. Gynecol. Obstet. 22: 93, 1916. 4. Davis, R. M., Cook, J. K., and Kirk, R. F.: Obstet. Gynecol. 40: 23, 1972. 5. DeLee, S. T.: Int. Surg. 50: 549, 1968.
January I!\. 1976 Am. J. Obstet. Gynecol.
conization was not required in any case. Conization during pregnancy is therefore most dearly indicated when the squamocolumnar junction cannot be visualized or when there is biopsy evidence of microinvasive cancer. The observation that the recurrence rate of carcinoma in situ following conization is significantly less than the incidence of residual intraepithelial carcinoma in the hysterectomy specimen is not new. Creasman and Rutledge 1\ in a study of 861 patients treated for carcinoma in situ at the M. D. Anderson HospitaL noted residual carcinoma in situ in 22 per cent of hysterectomy specimens, whereas recurrent carcinoma in situ following conization was found in only seven per cent. In a previous investigation from this institution, van Nagell and colleagues 15 found that the incidence of residual carcinoma in situ following conization was inversely related to the amount of inflammation present in the hysterectomy specimen. Perhaps the inflammatory response of the cervix following conization destroys residual carcinoma in situ or makes epithelial abnormalities difficult to interpret. Recurrent vaginal carcinoma in situ following hysterecrom} occurred most commonly in those patients who had residual carcinoma in situ in the hysterectomy specimen but was independent of the size of the vaginal euff excised. There is little doubt that the definitive therapy for cervical intraepithelial carcinoma in the woman who has completed her family is hysterectomy. However, for the patient who desires further pregnancy, directed biopsy and conization provide satisfactory diagnostic and, in certain cases, therapeutic alternatives to more radical procedures. In this regard, conization is particularly useful in the diagnosis of lesions extending up the endocervical canal in which the squamocolumnar junction cannot be visualized. Many advances in the management of malignant cervical disease have come about through greater individualization of therapy. It is not unreasonable to suspect that similar advantages will result from the application of this philosophy to patients with premalignant disease of the cervix.
6. McCann, S. W., Micka!, A., and Crapanzano, J. T.. Obstet. Gynecol. 33: 470, 1969. 7. Emmet, T. A.: AM. J. OasTET. GYNECOL. 7: 442, 1874. 8. Hollyock, V. E., and Chanen, W.: AM. J. 0BSTET. GYNECOL. 114: 185, 1972. 9. Ostergard, D. R., and Gondos, B.: AM. J. OasTET. GYNECOL. 115: 783, 1973.
Volume 124 Number 2
10. Parker, R. T., Cuyler, W.K., Kaufman, L.A., Carter, B., Thomas, W. L., Creadick, E. N., Turner, V. H., Peete, C. H., and Cherny, W. B.: AM.J. OssTET. GYNECOL. 80:693. 1960. 11. Staflfi A., and Mattingly, R. F.: Obstet. Gynecol. 41: 168, 1973. 12. Townsend, D. E., Ostergard, D. R., and Mishell, D. R.: AM. J. 0BSTET. GYNECOL. 108: 429, 1970.
Efficacy of cervical conization
139
13. Ortiz, R., and Newton, M.: AM.j. OBSTET. GYNECOL. 109: 46, 1971. 14. Creasman, W. T., and Rutledge, F.: Obstet. Gynecol. 39: 373, 1971. 15. van Nagell, J. R., Jr., Roddick, J. W., Jr., Cooper, R. M., and Triplett, H. B.: AM. J. 0BSTET. GYNECOL. 113: 948, 1972.
J.
C. PARKER, jR., M.D.
L. P. HICKS, M.D. R. CONRAD, B.S. G. ENGLAND, B.S.
Lexington, Kentucky
Cervical conization was performed on 756 patients at the University of Kentucky Medical Center from july I, 1964, to january 1, 1973. Sixty-six patients were pregnant at the time of conization. Eighty-six per cent of patients with cytologic findings of carcinoma in situ had hi5tologic verification of carcinoma in situ or sel!ere dysplasia, and there was absolute correlation between cytology and histology in 75 per cent of patients with occult inl!asive cancer. Cervical biopsies without colposcopic direction predicted either severe dysplasia or carcinoma in situ in 77 per cent of cases but were accurate in only seven of 24 patients with occult invasive cancer. Carcinoma in situ was present in 30 per cent of hysterectomy specimens following conization but recurred in only seven per cent of patients followed without hysterectomy. Recurrent carcinoma in situ following hysterectomy was more common in patients with residual intraepithelial cancer in the uterus but was independent of the size of the vaginal cuff removed. Major postconization complications requiring hospitalization occurred in 3.4 per cent of nonpregnant patients and in 7.5 per cent of pregnant patients.
SINCE ITS INITIAL description by Lisfranc 1 in 1815, the technique of cervical conization has undergone many changes. Sims, 2 in 1861, dosed the cervix with silver wire sutures following conization and was the first to obtain healing by primary intention. Likewise, Sturmdorf,3 in 1916, described a method inverting cervicovaginal epithelium to cover the conization site which is still used in many institutions today. More recently, several investigators4 • 5 have described methods of injecting vasoconstrictive solutions into the cervix at the time of conization in order to minimize blood loss. Despite these technical advances, cervical
conization remams a major operative procedure associated with a five to ten per cent incidence of postoperative complications.4• 6 Perhaps even more important than changes in conization technique have been changes in the indications for its performance. Whereas conization was initially performed for invasive cervical cancer and the repair of cervical lacerations/ it has been used during the past three decades for the diagnosis and treatment of cervical epithelial abnormalities. With the utilization of colposcopy to direct cervical biopsies, it has been suggested by some investigators that cervical conization is being performed too frequently. 8 • 9 The present study was undertaken to determine the diagnostic and therapeutic efficacy of cervical conization as a method in the evaluation of patients with cervical neoplasia.
From the Departments of Obstetrics and Gynecology and Pathology, University of Kent!.I.Cky Medical Center. Supported lry the American Cancer Society, Kentucky Division. Received for publicatwn September 25, 1974.
Methods and results
Revised january 31, 1975.
From July 1, 1964, to January I. 1973, cervical conization was performed on 756 patients at the University of Kentucky Medical Center. Sixty-six of these patients were pregnant at the time of conization.
Accepted February 13, 1975. Reprint requests: Dr.]. R. van Nagell, Jr., Department of Obstetrics and Gynecology, University of Kentucky Medical Center, Lexington, Kent!J.Cky 40506.
134
Efficacy of cervical conization
Volume 124 Number 2
Conization was carried out as a diagnostic method in patients with cytologic or biopsy findings of cervical epithelial abnormalities. Cytologic and histologic abnormalities were graded according to the following classification: mild dysplasia, moderate dysplasia, severe dysplasia, atypical cellular changes, carcinoma in situ, mitroinvasive carcinoma, and invasive carcinoma. Cervical conization was performed in the absence of biopsy-proved invasive cancer within two months on all patients with cytologic or biopsy evidence of persistent moderate dysplasia, severe dysplasia, carcinoma in citu, or microinvasive carcinoma. Persistent moderate dysplasia was defined as moderate dysplasia on two occasions separated by at least three months. Microinvasive cancer was defined as the presence of stromal invasion to a depth of less than 5 mm. Conization was performed according to the method of Parker and colleagues. 10 Twelve to sixteen sections were made of each conization specimen, and additional step sections were cut when an area indicative of epithelial abnormality was noted. Areas of cervical epithelial abnormality were described as involving squamous epithelium, both glandular and squamous epithelium, or glandular epithelium alone. Basic duiracteristics of the patients studied are presented in Table I. The mean age, gravidity, and location of cervical epithelial abnormality was essentially the same in those patients who were pregnant and in those who were not. Seven hundred and seven patients had cervical conizations at this institution within eight weeks of cytologic examination. Conization findings in these patients are presented in Table II. Cytology and histology specimens were reviewed in all cases in which there was not absolute cytology-histology correlation. Eighty-six per cent of patients with cervical cytologic findings of carcinoma in situ had histologic verification of severe dysplasia or intraepithelial carcinoma on conization. Seventeen patients with normal or mildly dysplastic cytologic findings had severe dysplasia or carcinoma in situ on conization. However, no patient whose cervical cytology was normal or showed mild or moderate dysplasia had histologic evidence of invasive cancer in the conization specimen. Twenty-eight patients (four per cent) with no visible cervical lesions had invasive cancer at conization. Twenty-one of these patients had cytologic diagnoses of invasive cancer, and the remainder had cytologic abnormalities ranging from atypical cellular changes to microinvasive cancer. Age as an independent variable had no direct effect on the cytology-histology correlation. Thirty-eight patients (5.4 per cent) had cytologic findings which were
135
Table I. Basic characteristics of patients undergoing cervical conization Characteristics
Nonpregnant
Pregnant
No. of patients Mean age (yr.) Mean gravidity Location of lesion Squamous epithelium alone(%) Squamous and glandular epithelium (%) Glandular epithelium alone(%)
690 37.0 4.6
66 33.3 5.2
62.6
57.8
34.5
35.9
2.9
6.3
not severe enough to warrant conization. However, cervical biopsy abnormalities were more advanced and indicated the need for conization. Cervical biopsies were performed on 53 7 patients prior to conization. Biopsies were taken from areas in the cervix which failed to take up Schiller's stain, since colposcopy was not performed at this institution during the time of investigation. Endocervical curettage was carried out only in those patients with visible endocervical lesions. The histologic correlation between punch biopsies and conization is presented in Table III. Twenty-four patients with invasive cancer and no visible cervical lesions had punch biopsies prior to conization. Invasive cancer was confined to the squamous epithelium in eleven cases and involved both squamous and glandular epithelium in three cases. Punch biopsies showed microinvasive or invasive cancer in nine patients (37.5 per cent) and carcinoma in situ in II patients (45.8 per cent). Thirteen patients with suspected invasive cancer on punch biopsy had cervical conization in order to determine the extent of invasion. Seven of these patients had invasive cancer and six had intraepithelial or microinvasive cancer. The relationship between location of the epithelial abnormality and the diagnostic accuracy of cervical biopsy is presented in Tables IV and V. Cervical biopsies showed either severe dysplasia or carcinoma in situ in I45 of I93 patients (75.I per cent) with carcinoma in situ limited to the squamous epithelium and in 119 of 151 patients (78. 7 per cent) whose carcinoma in situ involved both squamous and glandular epithelium. Diagnostic accuracy of punch biopsies was significantly reduced in those lesions confined to the endocervical glands. Eleven patients had either carcinoma in situ or invasive adenocarcinoma of the endocervix, and only six of them (54.5 per cent) were diagnosed accurately on punch biopsy.
January 15, 1976 Obstet. Gvnecol.
136 Van Nagell et al.
J.
Am.
Table II. Cytology-histology correlation in patients undergoing cervical conization Conization findings Mild dysplasia or less Cervical cytologic findings
No. of patients
No.
Normal Mild dysplasia Moderate dysplasia Severe dysplasia Atypical cells Carcinoma in situ Microinvasive carcinoma Invasive carcinoma
9 29 103 163 19 345 7 32
4 II 9 13 3 12 0 0
I
Moderate dysplasia
%
No.
44.4 37.9 8.6 8.0 15.8 3.5
0 6 32 25 2
I
6
Severe dysplasia
%
No.
20.7 30.8 2.'i.3 10.5 1.7
2 9 36 2 22 0 0
()
0
I
Microinvasive carcinoma
Carcinoma in situ
%
No.
11.2 6.9 8.6 22.1 10.5 6.4
4 10 49 76 11 274 4 7
%
No.
44.4 34.5 47.1 46.6 57.9 79.4 57.1 21.9
0 0 4 12 0 28 I
I
I
%
Invasive carcinoma No.
I
%
0 0
4
3.9 7.4
()
8.1 14.3 12.5
3 2 21
0.6 5.3 0.9 28.6 65.6
I I
Table III. Biopsy-conization correlation in patients with cervical epithelial abnormalities Conization findings Mild dysplasia or less Biopsy findings
No.
Normal Atypical cellular changes Mild dysplasia Moderate dysplasia Severe dysplasia Carcinoma in situ Microinvasive carcinoma Invasive carcinoma No. of patients
3 0 8 5 4 2 0 0 22
I
Moderate dysplasia
%
No.
13.6
12 I 10 23 13 I 0 0
36.4 22.7 18.2 9.1
I
%
No.
20.0 1.7 16.7 38.2 21.7 1.7
6 I 6 15 19 12 0 0
60
I
Microinvasive carcmoma
Carcinoma in situ
Severe dysplasia
I
%
No.
10.2 1.7 10.2 25.4 32.2 20.3
27 6 8 32 55 198 3 5
59
%
No.
8.1 1.8 2.4 16.5 16.5 59.4 0.9 1.5
2 0 1 3 3 25 3 I
I
Invasive carcinoma
%
No.
5.3
0 0 0 3 1 11 2 7
2.6 7.9 7.9 65.8 7.9 2.6
%
12.5 4.2 45.8 8.3 29.2
24
38
334
I
Table IV. Punch biopsy-conization correlation according to location of lesion in patients with carcinoma in situ Biopsy findings Severe dysplasia Location of carcinoma in situ on conization Squamous epithelium only Squamous and glandular epithelium Glandular epithelium only
Patients
No.
193 151
34 21
Sixty-six patients were pregnant at the time of conization and sixty-three of these patients had conization within six weeks of cytologic examination. Thirtyseven (82.2 per cent) of the 45 patients with carcinoma in situ on conization had cytologic evidence of severe dysplasia or carcinoma in situ. The one pregnant patient with occult invasive cervical cancer had
0
I
Other
Carcinoma in situ
%
No.
17.6 13.8
Ill 98
I
%
No.
57.5 64.9
48 32
100
0
I
% 24.9 21.3
cytologic evidence of malignancy. Fifty-three of these patients had cervical biopsies prior to conization. Cervical biopsies were accurate in 78.9 per cent of pregnant patients with carcinoma in situ. However, only one of the three patients with microinvasive cancer was diagnosed on cervical biopsy. The one pregnant patient with invasive cancer was not biopsied
Efficacy of cervical conization 137
Volume 124 Number 2
Table V. Punch biopsy-conization correlation according to location of lesion in patients with occult invasive cervical cancer BiosfrY findings Carcinoma in situ Location of invasive carcinoma on conization
No. of patients
Squamous epithelium only Squamous and glandular epithelium Glandular epithelium only
11 3
6 1
54.5 33.3
10
4
40.0
No.
l
%
prior to conization. Hysterectomy was performed with cesarean section or in the immediate postpartum period in 43 patients with carcinoma in situ. Twelve patients (27.9 per cent) had residiual carcinoma in situ and one had microinvasive carcinoma in the hysterectomy specimen. Complications of conization are presented in Table VI. Sixteen patients required readmission for excessive cervical bleeding. This bleeding was successfully treated with resuture, vaginal packing, or cryotherapy in thirteen patients, and emergency hysterectomy was required in three patients. Bleeding occurred five to ten days after conization in all patients. Six patients required blood transfusions following conization related bleeding. Uterine perforation at the time of conization occurred in two patients, and three patients experienced spontaneous abortion within one week of conization. The incidence of major conization complications, excluding the 161 patients whose hysterectomy immediately followed frozen section examination of the cone specimen, was 3.4 per cent in nonpregnant patients and 7.5 per cent in pregnant patients. Three hundred and seventy four patients with carcinoma in situ or microinvasive carcinoma who desired no further pregnancies underwent hysterectomy following conization. Vaginal hysterectomy was performed in 299 patients and abdominal hysterectomy in the remainder. A wide vaginal cuff was excised with the hysterectomy specimen in 226 cases. All hysterectomy specimens were examined microscopically to determine the incidence of residual epithelial abnormality. Sections were routinely taken through the anterior, posterior, and lateral walls of the cervix as well as additional sections in areas of histologic abnormality. No patient had undiagnosed invasive cancer in the hysterectomy specimens following conization. Carcinoma in situ was present in 112 hysterectomy specimens (29.9 per cent) and microinvasive
Microinvasive carcinoma No.
I
Invasive carcinoma
%·
No.
9.1 33.3
2
%
No.
18.2
2
0
5
0
I
Other
I
% 18.2 33.3
1
50.0
10.0
Table VI. Complications of cervical conization Nonpregnant Complications
No.
No. of patients Complication Cervical bleeding requiring hospitalization Uterine perforation Pelvic infection Pregnancy loss Total
529
I
%
Pregnant No.
I
%
66
14
2.6
2
3.0
2 2
0.4 0.4
0 0 3
4.5
3.4
5
7.5
0
18
cancer in six (1.6 per cent). Following conization, patients were seen in the gynecologic oncology clinic every three months for two years and every six months thereafter. The duration of follow-up ranged from 12 to 110 months (mean 50 months). Recurrent vaginal cancer following hysterectomy occurred in five patients ( 1.1 per cent), and there was residual carcinoma in situ in the hysterectomy specimen of four of these patients. Vaginectomy was performed on three of these patients and radiation therapy in two. All of these patients are alive with no evidence of recurrent disease three to 60 months following therapy. One hundred and nine patients with histologically proved carcinoma in situ who desired further pregnancies have been followed without hysterectomy. Seven of these patients (6.4 per cent) have required a second conization and hysterectomy for recurrent carcinoma in situ.
Comment The findings of this study indicate that cervical conization is an effective method in the diagnosis of cervical epithelial abnormalities. However, it is a major operative procedure associated with a significant incidence of complications. For this reason, conization should be performed only in those instances in which it
138 Van Nagel! et al.
can provide definite diagnostic or therapeutic advantages to less radical procedures. Cervical cytologic examination is an excellent screening procedure and suggests the need for further histologic confirmation by biopsy or conization. Furthermore, cytologic examination and cervical biopsies are often complementary in the diagnosis of cervical epithelial abnormalities. This is exemplified by those patients (5.4 per cent) from this institution with normal or mildly dysplastic cytologic findings whose conizations were performed because of biopsy evidence of more advanced epithelial abnormalities. The most effective method for reducing the number of cervical conizations is to increase the accuracy of cervical biopsies. In the present series, only seven of 24 patients with invasive carcinoma in the absence of a cervical lesion had the correct diagnosis made by cervical biopsy. Similarly, only 76.8 per cent of the carcinoma in situ or severe dysplasia was diagnosed by biopsy alone. Stafl and Mattingly 11 reported colposcopic prediction of histopathology to be clinically accurate in 85 per cent of patients with cervical epithelial abnormalities. Likewise, Townsend and associates12 noted an absolute correlation between colposcopically directed biopsies and cervical conization of 84.5 per cent when the lesion did not involve the endocervical canal. Clearly, the accuracy of colposcopy depends upon the skill and experience of the colposcopist. However, it does provide an additional diagnostic alternative to conization if the lesion can be completely defined and the squamocolumnar jum:tion fully visualized. Conization during pregnancy was directly responsible for a five per cent incidence of fetal loss. In addition, two patients experienced significant postconization hemorrhage. Ortiz and Newton, 13 in a study of 47 pregnant patients with abnormal cervical cytologic findings, found that colposcopically directed biopsy was an effective means to histologically evaluate the cervix and was unlikely to miss invasive cancer because of the physiologic eversion of the cervix during pregnancy. Similarly, Stafl and Mattingly 11 found that no unsatisfactory or incomplete colposcopic examinations occurred in 89 pregnant patients, and
REFERENCES I. Lisfranc, J.: Gaz. Med. France 2: 835, 1815. 2. Sims, J. M.: Trans. Med. Soc. N. Y. 137: 367, 1861. 3. Sturmdorf, A.: Surg. Gynecol. Obstet. 22: 93, 1916. 4. Davis, R. M., Cook, J. K., and Kirk, R. F.: Obstet. Gynecol. 40: 23, 1972. 5. DeLee, S. T.: Int. Surg. 50: 549, 1968.
January I!\. 1976 Am. J. Obstet. Gynecol.
conization was not required in any case. Conization during pregnancy is therefore most dearly indicated when the squamocolumnar junction cannot be visualized or when there is biopsy evidence of microinvasive cancer. The observation that the recurrence rate of carcinoma in situ following conization is significantly less than the incidence of residual intraepithelial carcinoma in the hysterectomy specimen is not new. Creasman and Rutledge 1\ in a study of 861 patients treated for carcinoma in situ at the M. D. Anderson HospitaL noted residual carcinoma in situ in 22 per cent of hysterectomy specimens, whereas recurrent carcinoma in situ following conization was found in only seven per cent. In a previous investigation from this institution, van Nagell and colleagues 15 found that the incidence of residual carcinoma in situ following conization was inversely related to the amount of inflammation present in the hysterectomy specimen. Perhaps the inflammatory response of the cervix following conization destroys residual carcinoma in situ or makes epithelial abnormalities difficult to interpret. Recurrent vaginal carcinoma in situ following hysterecrom} occurred most commonly in those patients who had residual carcinoma in situ in the hysterectomy specimen but was independent of the size of the vaginal euff excised. There is little doubt that the definitive therapy for cervical intraepithelial carcinoma in the woman who has completed her family is hysterectomy. However, for the patient who desires further pregnancy, directed biopsy and conization provide satisfactory diagnostic and, in certain cases, therapeutic alternatives to more radical procedures. In this regard, conization is particularly useful in the diagnosis of lesions extending up the endocervical canal in which the squamocolumnar junction cannot be visualized. Many advances in the management of malignant cervical disease have come about through greater individualization of therapy. It is not unreasonable to suspect that similar advantages will result from the application of this philosophy to patients with premalignant disease of the cervix.
6. McCann, S. W., Micka!, A., and Crapanzano, J. T.. Obstet. Gynecol. 33: 470, 1969. 7. Emmet, T. A.: AM. J. OasTET. GYNECOL. 7: 442, 1874. 8. Hollyock, V. E., and Chanen, W.: AM. J. 0BSTET. GYNECOL. 114: 185, 1972. 9. Ostergard, D. R., and Gondos, B.: AM. J. OasTET. GYNECOL. 115: 783, 1973.
Volume 124 Number 2
10. Parker, R. T., Cuyler, W.K., Kaufman, L.A., Carter, B., Thomas, W. L., Creadick, E. N., Turner, V. H., Peete, C. H., and Cherny, W. B.: AM.J. OssTET. GYNECOL. 80:693. 1960. 11. Staflfi A., and Mattingly, R. F.: Obstet. Gynecol. 41: 168, 1973. 12. Townsend, D. E., Ostergard, D. R., and Mishell, D. R.: AM. J. 0BSTET. GYNECOL. 108: 429, 1970.
Efficacy of cervical conization
139
13. Ortiz, R., and Newton, M.: AM.j. OBSTET. GYNECOL. 109: 46, 1971. 14. Creasman, W. T., and Rutledge, F.: Obstet. Gynecol. 39: 373, 1971. 15. van Nagell, J. R., Jr., Roddick, J. W., Jr., Cooper, R. M., and Triplett, H. B.: AM. J. 0BSTET. GYNECOL. 113: 948, 1972.
