- Email: [email protected]
Stage IA2 squamous carcinoma of the cervix: Difficult diagnosis and therapeutic dilemma
Stage IA2 squamous carcinoma of the cervix: Difficult diagnosis and therapeutic dilemma Benjamin E. Greer, MD, David C. Figge, MD, Hisham K. Tamimi, MD, Joanna M. Cain, MD, and Roger B. Lee, MD Seattle, Washington The International Federation of Gynecology and Obstetrics instituted a change in the classification for carcinoma of the cervix with a new substage IA2. The criteria for this substage exceed the generally accepted criteria for microinvasion. Fifty patients with early invasive squamous cell carcinoma of the cervix were treated from 1976 through 1983 with a cone biopsy followed by a radical hysterectomy and pelvic lymph node dissection. These patients were reviewed to evaluate the ability to make the histologic diagnosis and to examine the natural history of the disease with maximal treatment. Histologically positive margins were found at the time of cone biopsy in 66% (33/50) of the patients. Negative margins at the time of cone biopsy were identified in 34% (17/50) of the patients. Residual invasive disease at the time of radical hysterectomy was found in 24% (4/17) of the patients with negative margins. Two of the 50 patients had positive lymph nodes. Three patients had recurrent metastatic disease. This study demonstrates that a preoperative diagnosis of stage IA2 invasive squamous cell carcinoma of the cervix is a difficult diagnosis to establish and creates a therapeutic dilemma regarding treatment. (AM J OSSTET GYNECOL 1990;162:1406-11.)
Key words: Stage IA2 squamous carcinoma of cervix, microinvasion The concept of microinvasive carcinoma of the cervix as a distinct clinical entity was introduced in 1947. J The International Federation of Gynecology and Obstetrics (FIGO) in 1961 defined stage IA carcinoma of the cervix as early stromal invasion. In the early 1970s FIGO defined stage IA as microinvasive carcinoma but failed to provide guidelines for establishing the histologic diagnosis. In 1974 the Society of Gynecologic Oncology (SGO) provided the working definition: a micro invasive lesion is one in which neoplastic epithelium invades the stroma in one or more places to a depth of 3 mm or less below the base of the epithelium, and in which lymphatic or blood vessel involvement is not demonstrated. These histologic criteria have been generally accepted in the United States. The Oncology Committtee of the FIGO recently mandated changes in the definition of stage IA carcinoma of the cervix. 2 Stage IA carcinoma of the cervix is subdivided into stage IAJ and stage IA 2. Stage IAJ is defined as minimally microscopically evident stromal invasion. This includes only those patients with minute
From the Department of Obstetrics and Gynecology, University of Washington Medical Center. Presented at the Fifty-sixth Annual Meeting of the Pacific Coast Obstetrical and Gynecological Society, Coronado, California, September 17-21,1989. Reprint requests: Benjamin E. Greer, MD, Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, RH-20, University of Washington Medical Center, Seattle, WA 98195. 6/6/19908
1406
foci of invasion visible only microscopically and not penetrating more than 1 mm below the basement membrane. Stage IA2 is defined as lesions detected microscopically that can be measured. The upper limit of measurement must not exceed a depth of invasion of more than 5 mm taken from the basement membrane, either surface or glandular, from where it originates; and a second dimension, the horizontal spread, must not exceed 7 mm. Vascular involvement, either venous or lymphatic, should not alter the staging. The committee suggested that the diagnosis should be made on microscopic examination of removed tissue, preferably obtained by conization of the cervix, which should include the entire lesion. The criteria for the newly defined stage IA2 therefore clearly exceed the criteria for microinvasion as defined by the SGO. Current practice emphasizes the use of colposcopy in the diagnosis of preinvasive and invasive carcinomas. Conization of the cervix is reserved for distinguishing between microinvasion and true invasive carcinoma or those cases with positive endocervical curettage, discrepancy between cytology and biopsy, or unsatisfactory colposcopy results. In certain cases it may serve as a therapeutic modality. We have two major reservations concerning the new definition estalished by the FIGO. One is that it may be a difficult diagnosis to establish even with conization of the cervix. Our second concern is that since the stage IA2 clearly exceeds the generally accepted definition as established by the SGO, there must be two separate treatment regimens for patients
Volume 162 Number 6
with a diagnosis of stage IA2 cervical carcinoma. Traditionally stage IA carcinoma of the cervix has implied conservative surgery, such as cone biopsy or simple extrafascial hysterectomy. Therapy decisions regarding stage IA2 without a clear understanding of the fine points of the definition may result in undertreatment of patients because the scope of the definition clearly exceeds the currently accepted criteria for conservative surgery. This study was conducted to evaluate the ability to establish the histologic diagnosis of stage IA2 by conization of the cervix and to evaluate the natural history and results of a group of patients with early invasive squamous cell carcinoma of the cervix with maximal treatment. Material and methods
From 1976 through 1983 the Division of Gynecologic Oncology at the University of Washington Medical Center treated 50 patients with early invasive squamous cell carcinoma of the cervix in which a radical hysterectomy and pelvic lymph node dissection had been preceded by conization biopsy of the cervix. The clincial record and pathologic reports were reviewed on each patient. The patient'S age, race, parity, initial cytologic results, presence or absence of gross lesions, whether the patient had colposcopy, and if so the outcome and adequacy of the examination, and the diagnosis on directed biopsies were recorded. The pathologic condition of the cone biopsy specimens was evaluated for positive or negative margins, depth of invasion, capillary-like space invasion, multifocal disease, and confluence. The horizontal spread of the tumors was not measured. The pathologic condition of each radical hysterectomy specimen was examined for residual disease after conization, extension beyond the cervix, and site and number of metastatic lymph nodes. A period of at least 5 years has elapsed since the treatment of each patient. The disease status, site of recurrence, subsequent treatments, and outcomes were recorded. Results
The mean age of the 50 patients was 37 years (range, 22 to 72 years). The mean parity was 2.8. Forty-two of the patients were white and two each of native American, Oriental, black, and Mexican-American background. Routine periodic screening was responsible for alerting the physician to the diagnosis in all but one instance. This patient had postmenopausal bleeding and an endocervical curettage demonstrated carcinoma in situ. In review, six other patients had a history of vaginal spotting. Forty-nine of the patients had cytologic studies performed at the initial examination. The Papanicolaou smears suggested carcinoma in situ in 22 pa-
Stage lA2 squamous carcinoma of cervix
1407
tients and invasive cancer in 16. The cytologic condition was reported as merely "atypical" in 10 patients, 20% of the total group. One patient had a negative smear but a gross lesion. No visible lesion was present in 94% (47/50) of the patients. In the three patients with gross lesions, two had cone biopsy and the remaining patient had colposcopy and a biopsy consistent with carcinoma in situ but with questionable invasion. Colposcopy was performed in 70% (35/50) of trre women. Twenty-three of the examinations were considered adequate. The indications to proceed with conization of the cervix were based on a question of invasion in 15 patients, carcinoma in situ in six patients, and cervical intraepithelial neoplasia with a positive endocervical curettage in two patients. Twelve of the :35 patients undergoing colposcopy were considered to have inadequate examinations. Eight of those 12 patients had a biopsy diagnosis of carcinoma in situ. Four of the patients did not have biopsies and had conization of the cervix. Conization of the cervix without colposcopy was performed in 30% (15/50) of patients. The cone biopsies were all performed by referring physicians, and the pathologic reports were reviewed at the University of Washington Medical Center. Po,itive margins for neoplasia on the cone were present in 66% (33/50) with invasive disease. Twenty-seven of those patients had residual carcinoma at the tine of radical hysterectomy and pelvic lymph node dissection. Six of those 27 patients had only residual carcinoma in situ. There were six patients with positive margins at the time of conization of the cervix that revealed no evidence of residual disease at time of radical operation. Negative margins on the conization specimen were confirmed in 17 of the 50 cases (34%). This represents a group of patients who would qualify as stage 1.'\2 carcinoma. Four of the 17 (24%) had residual invasive disease at the time of radical hysterectomy despite the negative margins on the cone biopsy specimen. Two additional patients had residual carcinoma in situ. Capillary-like space invasion was demonstrated in 32% (16/50) of the conization specimens, multifocal invasion in 12% (6/50), and confluence was evident in 48% (24/50). Of the three patients with recurrent invasive carcinoma, one had all three other factors present, one had only confluence, and one was without any factors. No conclusion can be drawn from these findings. Two patients had evidence of positive metastatic diSease in the lymph nodes. One patient with a positive right common iliac node had a 2.2 mm invasive lesion with confluence on cone biopsy with negative margins and no residual cancer at the time of radical hysterectomy. The patient was treated postoperatively with radiation therapy of 5000 cGy to the whole pelvis and
1408 Greer et al.
June 1990 Am J Obstet Gynecol
Fig. 2. The maximal horizontal spread of this cancer would not be demonstrated as a result of the radial histologic sections.
Fig. 1. This figure demonstrates the altedtionin the depth of invasion after a cervical biopsy. '
4500 cGy to the paraaortic lymph nodes to an upper border of T12. She is alive and free of disease at 87 months. The other patient is dead of disease despite postoperative radiation therapy. Ninety-two percent (46/50) of the patients had follow-up. Forty-three of these 46 patients were followed up for more than 5 years. The range was from 8 to 123 months (mean follow-up, 68.3 months). There were four patients lost to follow-up in the immediate postoperative period when the patients returned to their referring community. Four recurrences developed after treatment. One patient had carcinoma in situ at the apex and was treated successfully with an upper vaginectomy. Three patients recurred with invasive disease. One patient had recurrence both in the pelvis and liver at 25 months. The patient underwent palliative radiation therapy and chemotherapy and survived an additional 7 months. One of the patients had a positive left external iliac node and received 5000 cGy to the whole pelvis postoperatively and disease recurred 3 months later. She survived 3 months with additional treatment. One patient had recurrence of disease in the pelvis at 61 months after radical surgery. The patient underwent radiation therapy for the recurrence with whole pelvic external radiation and intraoperative electron beam therapy. Currently she is alive with disease. Postoperative radiation therapy was administered to four patients. Two patients had positive lymph nodes; one of these patients is clinically without evidence of disease and one is dead of disease. Two patients had
extension beyond the cervix on final pathologic review and received adjuvant whole pelvis radiation therapy. Both are clinically without evidence of disease. Two additional patients received radiation therapy for recurrent disease.
Comment The new FIGO definition of stage IA2 disease presumably evolved from the volumetric concepts of microinvasive disease. 3 These volumetric concepts have merit in a reserach setting but are impractical for clinical practice. The planar definition of the 5 mm depth of invasion and 7 mm of horizontal spread as adapted by FIGO was made without data to support its utilization. A presumptive diagnosis of stage IA2 carcinoma by cone biopsy of the cervix was possible in only 34% (17/50) of cases with early invasive carcinoma in this series of patients. Twenty-four percent (4/17) of these cases had residual cancer in the surgical specimen at the time of radial hysterectomy despite apparently negative margins at the time of the cone biopsy. One patient with 2.2 mm of invasion with negative margins had a positive common iliac node. The diagnosis of stage IA2 disease becomes untenable if the lesion extends to or beyond the limits of the cone biopsy. This occurred in two thirds (33/50) of the cases of early carcinoma in this study. In such instances, it is customary to proceed with radical surgery because of the unknown depth of residual carcinoma. Although the diagnosis in each instance was established by cone biopsy, only 26% (13/50) of the cases of early carcinoma in this series could be appropriately considered as stage IA2 lesions. The validity of such a staging classification must be open to challenge.
Volume 162 Number 6
We can perceive two other mcUor pitfalls in reaching the diagnosis of stage IA. carcinoma. Colposcopy and directed biopsy are integral steps in establishing the diagnosis of invasive cancer, and failure of the clinician to avail himselfor herself of this approach would be widely criticized. Col~oscopic biopsies were performed in 64% of the patients in this series. In each instance, the biopsy presumably removed part of the lesion and therefore may have altered the depth of invasion as seen on the final cone biopsy. This is illustrated in Fig. 1. The validity of a diagnosis of stage IA2 carcinoma in any case when directed biopsies had encroached on the lesion in question could be justifiably rejected. A second criticism involves the accepted techniques of pathologic study. Cone biopsy specimens, after initial fixation in formalin, are divided into quadrants and step-sectioned. The maximal extent of horizontal spread could be underestimated by this procedure. Fig. 2 demonstrates a narrow, long circumferential lesion in which the maximal horizontal spread would not be reflected by the histologic condition of the cone. The increasing use of las'er conization of the cervix, with its attendant coagulation artifact that may obscure margins and depth of invasion, can only add to the difficulty of making a diagnosis. The 4% incidence of metastatic lymph nodes and the 6% recurrence rate of invasive carcinoma in this series of cases with early invasive cancer emphasize the necessity of establishing guidelines for therapy. Unfortunately, there are too few patients in this series who "fit" the criteria for either the FIGO or SGO definition of microinvasion to make any conclusions or statements. Recent articles support the SGO definition for microinvasion,,· 7 Rarely patients with metastatic lesions have been reported with <3 mm of invasive cancer. s This was demonstrated in this series. Recently, the argument was advanced that guidelines advanced by the SGO may result in unnecessary radical surgery.9 A more logical perspective is to consider the potential number of failures in the patients that exceed the SGO definition of microinvasion and still meet the criteria for stage IA2 by the new FIGO definition. A recent paper summarizing reports in the literature concluded that the risk of metastatic disease in the lymph nodes in cases with <3 mm invasion does not exceed 1%.5 Lesions with a depth of invasion between 3.1 and 5 mm have a nodal metastasis rate of 8%. The difference is highly statistically significant (p < 0.001). It would be unconscionable to accept this potential failure rate merely to reduce the number of radical hysterectomies performed for this disease. The literature does not provide clear direction concerning the issues of patterns of invasion, capillary-like space invasion, or horizontal spread. A clear understanding of the problems in making a
Stage IA2 squamous carcinoma of cervix
1409
diagnosis and the implications of the histologic findings will permit experienced clinicians to make appropriate choices of treatment for their patients. Clinical staging has two purposes: it describes firm criteria that permit comparison of data and it establishes guidelines for treatment. The FIGO classification for stage IA2 lesions would appear to adequately serve neither purpose. By contrast, the SGO definition for microinvasion serves a practical purpose, and conservative treatment in cases so defined will result in a cure in virtually all instances. Radical hysterectomy and pelvic lymph node dissection should be mandated for patients whose histologic findings exceed the SGO definition for microinvasion. This approach provides the maximum opportunity for cure of invasive cancer. At the next FIGO meeting the current definition of stage IA. should be reevaluated and modified to meet the requirements of clinical practice. REFERENCES I. Mestwerdt G. Die Fruhdiagnose des Kollumkarzinoms. Zentralbl Gynakol 1947;69:198-202. 2. Pettersson F. Annual report on results of treatment in carcinoma of the uterus, vagina and ovary and trophoblastic disease. Stockholm: International Federation of Gynecology and Obstetrics, 1988:20. 3. Burghardt E, Holzer E. Diagnosis and treatment of microinvasive carcinoma of the cervix uteri. Obstet Gynecol 1977;46:641-53. 4. van NagellJRJr, Greenwell N, Powell DF, Donaldson ES, Hanson MB, Gay EC. Microinvasive carcinoma of the cervix. AMJ OBSTET GYNECOL 1983;145:981-91. 5. Simon NL, Gore H, Shingleton HM, Soong SJ, Orr JW Jr, Hatch KD. Study of superficially invasive carcinoma of the cervix. Obstet Gynecol 1986;68: 19-24. 6. Maiman MA, Fruchter RG, DiMaio TM, Boyce JG. Superficially invasive squamous cell carcinoma of the cervix. Obstet Gynecol 1988;72:399-403. 7. Tsukamoto N, Kaku T, Matsukuma K, et al. The problem of stage Ia (FIGO, 1985) carcinoma of the uterine cervix. Gynecol Oncol 1989;34:1-6. 8. Hasumi K, Sakamoto A, Sugano H. Microinvasive carcinoma of the uterine cervix. Cancer 1980;45:928-31. 9. Kolstad P. Follow-up study of 232 patients with stage Ia2 and 411 patients with stage Ia2 squamous cell carcinoma of the cervix (microinvasive carcinoma). Gynecol Oncol 199;33:265-72.
Editors' note: This manuscript was revised after these
discussions were presented.
Discussion DR. CONLEY G.
LACEY, La Jolla, California. The authors retrospectively studied 50 patients with early squamous cell carcinoma of the cervix to determine the clinical feasibility of the FIGO cervical cancer classification system as it applies to stage IA2 disease. All information was extracted from the patients' records. A measurement of the lesion diameter or breadth, a critical dimension of the FIGO system, was not included in the analysis. Considerable useful information alerting us to the perils involved in the diagnosis and management of
1410 Greer et al.
early cervical cancer was obtained from this study. For example, the impracticality of the FICO system is suggested by the fact that 33 of the 50 patients (66%) were excluded from FICO substage IA2 by virtue of positive cone margins. This is not as serious an indictment as it may sound, however, since as a practical matter, most patients with positive margins would be excluded from conservative treatment even with use of the SCO criteria, particularly if the margin was involved by invasive disease. The remaining 17 patients in the study had negative cone margins and otherwise conformed to the author's criteria for FICO stage IA2. It is these 17 patients who should form the nucleus for this study and the analysis of these 17 patients must provide the evidence to refute the FICO system. Six of these 17 patients had residual disease on the hysterectomy specimen, two with carcinoma in situ and four with microinvasive or invasive carcinoma. We do not know the outcome of these patients. It would also be helpful to know the exact millimeter-by-millimeter depth of invasion, the breadth of invasion, the presence of node or other metastases, the presence of other risk factors such as vascular space involvement, the recurrence rate, and the survival for this group of 17 patients. Whereas some of this information may be present in this study, it is not segregated from the 33 patients excluded on the basis of positive cone margins and therefore cannot be identified and used to refute the FICO criteria for stage IA 2 • Howe~er, the authors do appropriately point out that the danger in use of FICO stage IA2 criteria for patient management is clearly documented in the literature, in which the reported incidence of pelvic lymph node metastases in patients with invasion between 3 and 5 mm is 8%. Dr. Creer, how many of the patients in this study had between 3 and 5 mm of invasion and what was the outcome of these patients? This article is a modest impeachment of FICO substage IA 2. It would be greatly strengthened by a careful pathologic review with inclusion of all FICO criteria and other known risk factors. Results should be segregated according to those that do or do not conform to the study parameters. Thus, while I sympathize with the authors' opposition to FICO substage IA 2, I cannot bring myself to join their indictment on the basis of the evidence presented in this paper. I have one additional question. What w.ere the characteristics of the four patients whose disease characteristics conformed to the FICO criteria, who had negative cone margins but had residual invasive disease in the hysterectomy specimen? DR. PHILLIP DISAIA, Irvine, California. I wish to defend the "Fathers of FICO" at least in part. Until the new staging, we had a system that separated everything into microinvasion or stage lB. Stage IB encompassed everything from early stromal invasion to a 6 cm tumor. FICO, in an attempt to define what Europeans term "microcarcinoma," studied the natural history of small cancers of the cervix. What is disappointing about this new staging system is that we do not have something to pass on to the practicing clinician. The FICO staging
June 1990 Am J Obstet Gyneco1
system should not dictate therapy, and if we allow it to, we are misinformed. My recommendation is that clinicians apply the guidelines of the SCO, which were designed to help the clinican determine therapy. It has been the experience of most gynecologic oncologists that it is safe to apply the SCO guideline that simple hysterectomy is adequate for patients with <3 mm of invasion, with no vascular involvement, and no areas of confluence. Dr. Creer, would you comment on that part of the SCO definition of microinvasion that has been controversial: confluence? Should it be included in the criteria? Finally, what is your recommendation to the practicing clinician about simple hysterectomy for patients with these lesions? DR. FRANK R. GAMBERDELLA, Santa Barbara, California. I was interested in the 17 cases with negative margins. I think you said four did not have negative results. You commented on the possible inadequacy of the carbon dioxide laser for conization. Should the clinician concerned about the possiblity of invasive lesion use a carbon dioxide laser cone or a traditional coldknife cone for definitive diagnosis and better evaluation of the margins? DR. GREER (Closing). Dr. Disaia's question is the essence of this paper. As a gynecologic oncologist, I understand the implications of the FICO classification. I have concerns about the interpretation of the FICO classification in areas where tumor boards do not understand the implications of the definition of stage IA 2. They could recommend conservative management for a patient who needs radical surgery. As I pointed out in my conclusions, the experienced clinician will not fall into this trap. The SCO definition suitably fits the definition of microinvasion and the patient who fulfills these criteria can be conservatively managed. As presented in the text of the paper, no conclusions can be made regarding the risk factors of capillary-like space invasion, multifocal invasion, or confluence. Twelve of the 17 patients with negative margins had invasions of::;3 mm, and five patients had invasion of 3.1 to 5 mm. Residual invasive carcinoma at the time of radical operation was found in two patients in each measurement group. The four patients with residual invasive carcinoma are alive and free of disease after maximal treatment with radical operation. A patient with 1.8 mm of invasive carcinoma with capillary-like space invasion, confluence, multi focal invasion, and negative margins at the time of conization had recurrence at 61 months. She is alive with disease despite additional treatment. Laser conization of the cervix with its coagulation artifact may obscure margins and depth of invasion. Any patient requiring a conization with a suggestion of invasion on the Papanicolaou smear, colposcopy. or biopsy should have cold-knife conization. This provides the most accurate information to make appropriate treatment decisions. I have one additional comment. A recent article suggested that lesions invading between 3 and 5 mm could be adequately treated with a total abdominal hyster-
Volume 162 Number 6
ectomy and a pelvic lymphadenectomy.l A 22.5% incidence of metastatic disease in parametrial lymph nodes in radical hysterectomy specimens has been recently reported. 2 Patients with stage IB invasive carcinoma of the cervix had an 11.4% incidence of positive parametrial nodes. These data are important to remember as we continue to evolve our guidelines for management of early invasive carcinoma of the cervix.
Stage IA2 squamous carcinoma of cervix
1411
REFERENCES 1. Simon NL, Gore H, Shingleton HM, Soong SJ, Orr JW Jr, Hatch KD. Study of superficially invasive carcinoma of the cervix. Obstet Gynecol 1986;68: 19-24. 2. Girardi F, Lichtenegger MD, Tamussino K, Haas J. The importance of parametrial lymph nodes in the treatment of cervical cancer. Gynecol Oncol 1989;34:206-11.
Key words: Stage IA2 squamous carcinoma of cervix, microinvasion The concept of microinvasive carcinoma of the cervix as a distinct clinical entity was introduced in 1947. J The International Federation of Gynecology and Obstetrics (FIGO) in 1961 defined stage IA carcinoma of the cervix as early stromal invasion. In the early 1970s FIGO defined stage IA as microinvasive carcinoma but failed to provide guidelines for establishing the histologic diagnosis. In 1974 the Society of Gynecologic Oncology (SGO) provided the working definition: a micro invasive lesion is one in which neoplastic epithelium invades the stroma in one or more places to a depth of 3 mm or less below the base of the epithelium, and in which lymphatic or blood vessel involvement is not demonstrated. These histologic criteria have been generally accepted in the United States. The Oncology Committtee of the FIGO recently mandated changes in the definition of stage IA carcinoma of the cervix. 2 Stage IA carcinoma of the cervix is subdivided into stage IAJ and stage IA 2. Stage IAJ is defined as minimally microscopically evident stromal invasion. This includes only those patients with minute
From the Department of Obstetrics and Gynecology, University of Washington Medical Center. Presented at the Fifty-sixth Annual Meeting of the Pacific Coast Obstetrical and Gynecological Society, Coronado, California, September 17-21,1989. Reprint requests: Benjamin E. Greer, MD, Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, RH-20, University of Washington Medical Center, Seattle, WA 98195. 6/6/19908
1406
foci of invasion visible only microscopically and not penetrating more than 1 mm below the basement membrane. Stage IA2 is defined as lesions detected microscopically that can be measured. The upper limit of measurement must not exceed a depth of invasion of more than 5 mm taken from the basement membrane, either surface or glandular, from where it originates; and a second dimension, the horizontal spread, must not exceed 7 mm. Vascular involvement, either venous or lymphatic, should not alter the staging. The committee suggested that the diagnosis should be made on microscopic examination of removed tissue, preferably obtained by conization of the cervix, which should include the entire lesion. The criteria for the newly defined stage IA2 therefore clearly exceed the criteria for microinvasion as defined by the SGO. Current practice emphasizes the use of colposcopy in the diagnosis of preinvasive and invasive carcinomas. Conization of the cervix is reserved for distinguishing between microinvasion and true invasive carcinoma or those cases with positive endocervical curettage, discrepancy between cytology and biopsy, or unsatisfactory colposcopy results. In certain cases it may serve as a therapeutic modality. We have two major reservations concerning the new definition estalished by the FIGO. One is that it may be a difficult diagnosis to establish even with conization of the cervix. Our second concern is that since the stage IA2 clearly exceeds the generally accepted definition as established by the SGO, there must be two separate treatment regimens for patients
Volume 162 Number 6
with a diagnosis of stage IA2 cervical carcinoma. Traditionally stage IA carcinoma of the cervix has implied conservative surgery, such as cone biopsy or simple extrafascial hysterectomy. Therapy decisions regarding stage IA2 without a clear understanding of the fine points of the definition may result in undertreatment of patients because the scope of the definition clearly exceeds the currently accepted criteria for conservative surgery. This study was conducted to evaluate the ability to establish the histologic diagnosis of stage IA2 by conization of the cervix and to evaluate the natural history and results of a group of patients with early invasive squamous cell carcinoma of the cervix with maximal treatment. Material and methods
From 1976 through 1983 the Division of Gynecologic Oncology at the University of Washington Medical Center treated 50 patients with early invasive squamous cell carcinoma of the cervix in which a radical hysterectomy and pelvic lymph node dissection had been preceded by conization biopsy of the cervix. The clincial record and pathologic reports were reviewed on each patient. The patient'S age, race, parity, initial cytologic results, presence or absence of gross lesions, whether the patient had colposcopy, and if so the outcome and adequacy of the examination, and the diagnosis on directed biopsies were recorded. The pathologic condition of the cone biopsy specimens was evaluated for positive or negative margins, depth of invasion, capillary-like space invasion, multifocal disease, and confluence. The horizontal spread of the tumors was not measured. The pathologic condition of each radical hysterectomy specimen was examined for residual disease after conization, extension beyond the cervix, and site and number of metastatic lymph nodes. A period of at least 5 years has elapsed since the treatment of each patient. The disease status, site of recurrence, subsequent treatments, and outcomes were recorded. Results
The mean age of the 50 patients was 37 years (range, 22 to 72 years). The mean parity was 2.8. Forty-two of the patients were white and two each of native American, Oriental, black, and Mexican-American background. Routine periodic screening was responsible for alerting the physician to the diagnosis in all but one instance. This patient had postmenopausal bleeding and an endocervical curettage demonstrated carcinoma in situ. In review, six other patients had a history of vaginal spotting. Forty-nine of the patients had cytologic studies performed at the initial examination. The Papanicolaou smears suggested carcinoma in situ in 22 pa-
Stage lA2 squamous carcinoma of cervix
1407
tients and invasive cancer in 16. The cytologic condition was reported as merely "atypical" in 10 patients, 20% of the total group. One patient had a negative smear but a gross lesion. No visible lesion was present in 94% (47/50) of the patients. In the three patients with gross lesions, two had cone biopsy and the remaining patient had colposcopy and a biopsy consistent with carcinoma in situ but with questionable invasion. Colposcopy was performed in 70% (35/50) of trre women. Twenty-three of the examinations were considered adequate. The indications to proceed with conization of the cervix were based on a question of invasion in 15 patients, carcinoma in situ in six patients, and cervical intraepithelial neoplasia with a positive endocervical curettage in two patients. Twelve of the :35 patients undergoing colposcopy were considered to have inadequate examinations. Eight of those 12 patients had a biopsy diagnosis of carcinoma in situ. Four of the patients did not have biopsies and had conization of the cervix. Conization of the cervix without colposcopy was performed in 30% (15/50) of patients. The cone biopsies were all performed by referring physicians, and the pathologic reports were reviewed at the University of Washington Medical Center. Po,itive margins for neoplasia on the cone were present in 66% (33/50) with invasive disease. Twenty-seven of those patients had residual carcinoma at the tine of radical hysterectomy and pelvic lymph node dissection. Six of those 27 patients had only residual carcinoma in situ. There were six patients with positive margins at the time of conization of the cervix that revealed no evidence of residual disease at time of radical operation. Negative margins on the conization specimen were confirmed in 17 of the 50 cases (34%). This represents a group of patients who would qualify as stage 1.'\2 carcinoma. Four of the 17 (24%) had residual invasive disease at the time of radical hysterectomy despite the negative margins on the cone biopsy specimen. Two additional patients had residual carcinoma in situ. Capillary-like space invasion was demonstrated in 32% (16/50) of the conization specimens, multifocal invasion in 12% (6/50), and confluence was evident in 48% (24/50). Of the three patients with recurrent invasive carcinoma, one had all three other factors present, one had only confluence, and one was without any factors. No conclusion can be drawn from these findings. Two patients had evidence of positive metastatic diSease in the lymph nodes. One patient with a positive right common iliac node had a 2.2 mm invasive lesion with confluence on cone biopsy with negative margins and no residual cancer at the time of radical hysterectomy. The patient was treated postoperatively with radiation therapy of 5000 cGy to the whole pelvis and
1408 Greer et al.
June 1990 Am J Obstet Gynecol
Fig. 2. The maximal horizontal spread of this cancer would not be demonstrated as a result of the radial histologic sections.
Fig. 1. This figure demonstrates the altedtionin the depth of invasion after a cervical biopsy. '
4500 cGy to the paraaortic lymph nodes to an upper border of T12. She is alive and free of disease at 87 months. The other patient is dead of disease despite postoperative radiation therapy. Ninety-two percent (46/50) of the patients had follow-up. Forty-three of these 46 patients were followed up for more than 5 years. The range was from 8 to 123 months (mean follow-up, 68.3 months). There were four patients lost to follow-up in the immediate postoperative period when the patients returned to their referring community. Four recurrences developed after treatment. One patient had carcinoma in situ at the apex and was treated successfully with an upper vaginectomy. Three patients recurred with invasive disease. One patient had recurrence both in the pelvis and liver at 25 months. The patient underwent palliative radiation therapy and chemotherapy and survived an additional 7 months. One of the patients had a positive left external iliac node and received 5000 cGy to the whole pelvis postoperatively and disease recurred 3 months later. She survived 3 months with additional treatment. One patient had recurrence of disease in the pelvis at 61 months after radical surgery. The patient underwent radiation therapy for the recurrence with whole pelvic external radiation and intraoperative electron beam therapy. Currently she is alive with disease. Postoperative radiation therapy was administered to four patients. Two patients had positive lymph nodes; one of these patients is clinically without evidence of disease and one is dead of disease. Two patients had
extension beyond the cervix on final pathologic review and received adjuvant whole pelvis radiation therapy. Both are clinically without evidence of disease. Two additional patients received radiation therapy for recurrent disease.
Comment The new FIGO definition of stage IA2 disease presumably evolved from the volumetric concepts of microinvasive disease. 3 These volumetric concepts have merit in a reserach setting but are impractical for clinical practice. The planar definition of the 5 mm depth of invasion and 7 mm of horizontal spread as adapted by FIGO was made without data to support its utilization. A presumptive diagnosis of stage IA2 carcinoma by cone biopsy of the cervix was possible in only 34% (17/50) of cases with early invasive carcinoma in this series of patients. Twenty-four percent (4/17) of these cases had residual cancer in the surgical specimen at the time of radial hysterectomy despite apparently negative margins at the time of the cone biopsy. One patient with 2.2 mm of invasion with negative margins had a positive common iliac node. The diagnosis of stage IA2 disease becomes untenable if the lesion extends to or beyond the limits of the cone biopsy. This occurred in two thirds (33/50) of the cases of early carcinoma in this study. In such instances, it is customary to proceed with radical surgery because of the unknown depth of residual carcinoma. Although the diagnosis in each instance was established by cone biopsy, only 26% (13/50) of the cases of early carcinoma in this series could be appropriately considered as stage IA2 lesions. The validity of such a staging classification must be open to challenge.
Volume 162 Number 6
We can perceive two other mcUor pitfalls in reaching the diagnosis of stage IA. carcinoma. Colposcopy and directed biopsy are integral steps in establishing the diagnosis of invasive cancer, and failure of the clinician to avail himselfor herself of this approach would be widely criticized. Col~oscopic biopsies were performed in 64% of the patients in this series. In each instance, the biopsy presumably removed part of the lesion and therefore may have altered the depth of invasion as seen on the final cone biopsy. This is illustrated in Fig. 1. The validity of a diagnosis of stage IA2 carcinoma in any case when directed biopsies had encroached on the lesion in question could be justifiably rejected. A second criticism involves the accepted techniques of pathologic study. Cone biopsy specimens, after initial fixation in formalin, are divided into quadrants and step-sectioned. The maximal extent of horizontal spread could be underestimated by this procedure. Fig. 2 demonstrates a narrow, long circumferential lesion in which the maximal horizontal spread would not be reflected by the histologic condition of the cone. The increasing use of las'er conization of the cervix, with its attendant coagulation artifact that may obscure margins and depth of invasion, can only add to the difficulty of making a diagnosis. The 4% incidence of metastatic lymph nodes and the 6% recurrence rate of invasive carcinoma in this series of cases with early invasive cancer emphasize the necessity of establishing guidelines for therapy. Unfortunately, there are too few patients in this series who "fit" the criteria for either the FIGO or SGO definition of microinvasion to make any conclusions or statements. Recent articles support the SGO definition for microinvasion,,· 7 Rarely patients with metastatic lesions have been reported with <3 mm of invasive cancer. s This was demonstrated in this series. Recently, the argument was advanced that guidelines advanced by the SGO may result in unnecessary radical surgery.9 A more logical perspective is to consider the potential number of failures in the patients that exceed the SGO definition of microinvasion and still meet the criteria for stage IA2 by the new FIGO definition. A recent paper summarizing reports in the literature concluded that the risk of metastatic disease in the lymph nodes in cases with <3 mm invasion does not exceed 1%.5 Lesions with a depth of invasion between 3.1 and 5 mm have a nodal metastasis rate of 8%. The difference is highly statistically significant (p < 0.001). It would be unconscionable to accept this potential failure rate merely to reduce the number of radical hysterectomies performed for this disease. The literature does not provide clear direction concerning the issues of patterns of invasion, capillary-like space invasion, or horizontal spread. A clear understanding of the problems in making a
Stage IA2 squamous carcinoma of cervix
1409
diagnosis and the implications of the histologic findings will permit experienced clinicians to make appropriate choices of treatment for their patients. Clinical staging has two purposes: it describes firm criteria that permit comparison of data and it establishes guidelines for treatment. The FIGO classification for stage IA2 lesions would appear to adequately serve neither purpose. By contrast, the SGO definition for microinvasion serves a practical purpose, and conservative treatment in cases so defined will result in a cure in virtually all instances. Radical hysterectomy and pelvic lymph node dissection should be mandated for patients whose histologic findings exceed the SGO definition for microinvasion. This approach provides the maximum opportunity for cure of invasive cancer. At the next FIGO meeting the current definition of stage IA. should be reevaluated and modified to meet the requirements of clinical practice. REFERENCES I. Mestwerdt G. Die Fruhdiagnose des Kollumkarzinoms. Zentralbl Gynakol 1947;69:198-202. 2. Pettersson F. Annual report on results of treatment in carcinoma of the uterus, vagina and ovary and trophoblastic disease. Stockholm: International Federation of Gynecology and Obstetrics, 1988:20. 3. Burghardt E, Holzer E. Diagnosis and treatment of microinvasive carcinoma of the cervix uteri. Obstet Gynecol 1977;46:641-53. 4. van NagellJRJr, Greenwell N, Powell DF, Donaldson ES, Hanson MB, Gay EC. Microinvasive carcinoma of the cervix. AMJ OBSTET GYNECOL 1983;145:981-91. 5. Simon NL, Gore H, Shingleton HM, Soong SJ, Orr JW Jr, Hatch KD. Study of superficially invasive carcinoma of the cervix. Obstet Gynecol 1986;68: 19-24. 6. Maiman MA, Fruchter RG, DiMaio TM, Boyce JG. Superficially invasive squamous cell carcinoma of the cervix. Obstet Gynecol 1988;72:399-403. 7. Tsukamoto N, Kaku T, Matsukuma K, et al. The problem of stage Ia (FIGO, 1985) carcinoma of the uterine cervix. Gynecol Oncol 1989;34:1-6. 8. Hasumi K, Sakamoto A, Sugano H. Microinvasive carcinoma of the uterine cervix. Cancer 1980;45:928-31. 9. Kolstad P. Follow-up study of 232 patients with stage Ia2 and 411 patients with stage Ia2 squamous cell carcinoma of the cervix (microinvasive carcinoma). Gynecol Oncol 199;33:265-72.
Editors' note: This manuscript was revised after these
discussions were presented.
Discussion DR. CONLEY G.
LACEY, La Jolla, California. The authors retrospectively studied 50 patients with early squamous cell carcinoma of the cervix to determine the clinical feasibility of the FIGO cervical cancer classification system as it applies to stage IA2 disease. All information was extracted from the patients' records. A measurement of the lesion diameter or breadth, a critical dimension of the FIGO system, was not included in the analysis. Considerable useful information alerting us to the perils involved in the diagnosis and management of
1410 Greer et al.
early cervical cancer was obtained from this study. For example, the impracticality of the FICO system is suggested by the fact that 33 of the 50 patients (66%) were excluded from FICO substage IA2 by virtue of positive cone margins. This is not as serious an indictment as it may sound, however, since as a practical matter, most patients with positive margins would be excluded from conservative treatment even with use of the SCO criteria, particularly if the margin was involved by invasive disease. The remaining 17 patients in the study had negative cone margins and otherwise conformed to the author's criteria for FICO stage IA2. It is these 17 patients who should form the nucleus for this study and the analysis of these 17 patients must provide the evidence to refute the FICO system. Six of these 17 patients had residual disease on the hysterectomy specimen, two with carcinoma in situ and four with microinvasive or invasive carcinoma. We do not know the outcome of these patients. It would also be helpful to know the exact millimeter-by-millimeter depth of invasion, the breadth of invasion, the presence of node or other metastases, the presence of other risk factors such as vascular space involvement, the recurrence rate, and the survival for this group of 17 patients. Whereas some of this information may be present in this study, it is not segregated from the 33 patients excluded on the basis of positive cone margins and therefore cannot be identified and used to refute the FICO criteria for stage IA 2 • Howe~er, the authors do appropriately point out that the danger in use of FICO stage IA2 criteria for patient management is clearly documented in the literature, in which the reported incidence of pelvic lymph node metastases in patients with invasion between 3 and 5 mm is 8%. Dr. Creer, how many of the patients in this study had between 3 and 5 mm of invasion and what was the outcome of these patients? This article is a modest impeachment of FICO substage IA 2. It would be greatly strengthened by a careful pathologic review with inclusion of all FICO criteria and other known risk factors. Results should be segregated according to those that do or do not conform to the study parameters. Thus, while I sympathize with the authors' opposition to FICO substage IA 2, I cannot bring myself to join their indictment on the basis of the evidence presented in this paper. I have one additional question. What w.ere the characteristics of the four patients whose disease characteristics conformed to the FICO criteria, who had negative cone margins but had residual invasive disease in the hysterectomy specimen? DR. PHILLIP DISAIA, Irvine, California. I wish to defend the "Fathers of FICO" at least in part. Until the new staging, we had a system that separated everything into microinvasion or stage lB. Stage IB encompassed everything from early stromal invasion to a 6 cm tumor. FICO, in an attempt to define what Europeans term "microcarcinoma," studied the natural history of small cancers of the cervix. What is disappointing about this new staging system is that we do not have something to pass on to the practicing clinician. The FICO staging
June 1990 Am J Obstet Gyneco1
system should not dictate therapy, and if we allow it to, we are misinformed. My recommendation is that clinicians apply the guidelines of the SCO, which were designed to help the clinican determine therapy. It has been the experience of most gynecologic oncologists that it is safe to apply the SCO guideline that simple hysterectomy is adequate for patients with <3 mm of invasion, with no vascular involvement, and no areas of confluence. Dr. Creer, would you comment on that part of the SCO definition of microinvasion that has been controversial: confluence? Should it be included in the criteria? Finally, what is your recommendation to the practicing clinician about simple hysterectomy for patients with these lesions? DR. FRANK R. GAMBERDELLA, Santa Barbara, California. I was interested in the 17 cases with negative margins. I think you said four did not have negative results. You commented on the possible inadequacy of the carbon dioxide laser for conization. Should the clinician concerned about the possiblity of invasive lesion use a carbon dioxide laser cone or a traditional coldknife cone for definitive diagnosis and better evaluation of the margins? DR. GREER (Closing). Dr. Disaia's question is the essence of this paper. As a gynecologic oncologist, I understand the implications of the FICO classification. I have concerns about the interpretation of the FICO classification in areas where tumor boards do not understand the implications of the definition of stage IA 2. They could recommend conservative management for a patient who needs radical surgery. As I pointed out in my conclusions, the experienced clinician will not fall into this trap. The SCO definition suitably fits the definition of microinvasion and the patient who fulfills these criteria can be conservatively managed. As presented in the text of the paper, no conclusions can be made regarding the risk factors of capillary-like space invasion, multifocal invasion, or confluence. Twelve of the 17 patients with negative margins had invasions of::;3 mm, and five patients had invasion of 3.1 to 5 mm. Residual invasive carcinoma at the time of radical operation was found in two patients in each measurement group. The four patients with residual invasive carcinoma are alive and free of disease after maximal treatment with radical operation. A patient with 1.8 mm of invasive carcinoma with capillary-like space invasion, confluence, multi focal invasion, and negative margins at the time of conization had recurrence at 61 months. She is alive with disease despite additional treatment. Laser conization of the cervix with its coagulation artifact may obscure margins and depth of invasion. Any patient requiring a conization with a suggestion of invasion on the Papanicolaou smear, colposcopy. or biopsy should have cold-knife conization. This provides the most accurate information to make appropriate treatment decisions. I have one additional comment. A recent article suggested that lesions invading between 3 and 5 mm could be adequately treated with a total abdominal hyster-
Volume 162 Number 6
ectomy and a pelvic lymphadenectomy.l A 22.5% incidence of metastatic disease in parametrial lymph nodes in radical hysterectomy specimens has been recently reported. 2 Patients with stage IB invasive carcinoma of the cervix had an 11.4% incidence of positive parametrial nodes. These data are important to remember as we continue to evolve our guidelines for management of early invasive carcinoma of the cervix.
Stage IA2 squamous carcinoma of cervix
1411
REFERENCES 1. Simon NL, Gore H, Shingleton HM, Soong SJ, Orr JW Jr, Hatch KD. Study of superficially invasive carcinoma of the cervix. Obstet Gynecol 1986;68: 19-24. 2. Girardi F, Lichtenegger MD, Tamussino K, Haas J. The importance of parametrial lymph nodes in the treatment of cervical cancer. Gynecol Oncol 1989;34:206-11.