Utility of risk-weighted surgical-pathological factors in early-stage cervical cancer.

BACKGROUND: Surgical-pathological risk factors were evaluated by weighting the magnitude of significance of multiple risk factors correlating to survival and treatment response in cervical cancer. METHODS: Multivariate analysis was performed for survival outcomes entering seven pathological factors obtained from 540 radical hysterectomy specimens in stage IA2-IIB cervical cancer cases. Hazard ratio (HR) in each risk factor was determined, and the sum of HR scores for the corresponding risk factors was determined per case. Survival curves and postoperative treatment response (concurrent chemoradiotherapy (CCRT) vs radiotherapy alone) were evaluated based on the extent of HR-weighted scores. RESULTS: Hazard ratios for risk factors relating to disease-free survival (DFS) was: lympho-vascular space invasion 3.95, nodal metastasis 3.88, adenocarcinoma 3.40, large tumour 2.36, positive margin 1.99, deep stromal invasion 1.29, and parametria invasion 1.21. The HR-weighted scoring method showed a high predictive value for recurrence (area-under-curve 0.836, P<0.001). Hazard ratio-weighted scores were negatively correlated to DFS, and the cases with score 12.5 showed 5-year DFS rate of 23.8%. Tumours with larger score offset the benefits of CCRT over radiotherapy alone for postoperative adjuvant treatment (P<0.001). CONCLUSION: Surgical-pathological risk factors provide valuable information for survival and management of early-stage cervical cancer when number and significance of risks are weighted.

Management of recurrent cervical cancer: a review of the literature.

OBJECTIVE: The aim of this narrative review is to update the current knowledge on the treatment of recurrent cervical cancer based on a literature review. MATERIAL AND METHODS: A web based search in Medline and CancerLit databases has been carried out on recurrent cervical cancer management and treatment. All relevant information has been collected and analyzed, prioritizing randomized clinical trials. RESULTS: Cervical cancer still represents a significant problem for public health with an annual incidence of about half a million new cases worldwide. Percentages of pelvic recurrences fluctuate from 10% to 74% depending on different risk factors. Accordingly to the literature, it is suggested that chemoradiation treatment (containing cisplatin and/or taxanes) could represent the treatment of choice for locoregional recurrences of cervical cancer after radical surgery. Pelvic exenteration is usually indicated for selected cases of central recurrence of cervical cancer after primary or adjuvant radiation and chemotherapy with bladder and/or rectum infiltration neither extended to the pelvic side walls nor showing any signs of extrapelvic spread of disease. Laterally extended endopelvic resection (LEER) for the treatment of those patients with a locally advanced disease or with a recurrence affecting the pelvic wall has been described. CONCLUSIONS: The treatment of recurrences of cervical carcinoma consists of surgery, and of radiation and chemotherapy, or the combination of different modalities taking into consideration the type of primary therapy, the site of recurrence, the disease-free interval, the patient symptoms, performance status, and the degree to which any given treatment might be beneficial.

New classification system of radical hysterectomy: emphasis on a three-dimensional anatomic template for parametrial resection.

OBJECTIVE: The international acceptance of a universal classification system for radical hysterectomy is one of the important challenges in gynecologic oncology. The recently published classification system by Querleu and Morrow is a relevant proposal that has been well received by the professional community. However, it does not include a description of parametrial resection in three dimensions, which mostly determines post-operative morbidity. METHODS: The intention of this follow-up paper was to further develop the classification system based on the four proposed types of radical hysterectomy (A-D) into a three-dimensional model using standard anatomical landmarks for definition of resection margins in longitudinal and transverse dimensions and demonstrate it on pictures. RESULTS: Resection margins were defined in longitudinal and transverse dimensions for each suggested type of radical hysterectomy on all three parts of the parametria. Besides precise description using stable anatomical landmarks, all resection lines have been shown on intra-operative photographs. CONCLUSION: Four types of radical hysteretomy can be precisely defined on a three-dimensional anatomical template, including nerve sparing procedure. Our paper should contribute to better standardization (including nomenclature) of the radical hysterectomy, enhancing harmonization of clinical practice in gynecological oncology.

Metastatic and recurrent adenocarcinoma of the uterine cervix: a long-term survival of 16 years.

PURPOSE OF INVESTIGATION: Recurrent metastatic adenocarcinoma of the cervix is associated with an extremely poor prognosis. Treatment options for recurrent disease are limited and cure is extremely rare. CASE REPORT: We report a case of a 43-year-old patient with Stage IB adenocarcinoma of the cervix. She had multiple metastatic recurrence episodes salvaged with several radical surgeries, external and intraoperative irradiation, and chemotherapy over a survival period of 16 years. CONCLUSION: We conclude that long-term multi-modal salvage treatment may achieve longer survival in rare cases with recurrent metastatic adenocarcinoma of the cervix.

Treatment of patients with ovarian carcinoma with pegylated liposomal doxorubicin: analysis of toxicities and predictors of outcome.

BACKGROUND: Pegylated liposomal doxorubicin is a new formulation with activity against epithelial ovarian carcinoma (EOC). The authors sought to determine patient characteristics that may predict for response to this treatment and favorable time to failure as well as survival. METHODS: Eight patients in a Phase I study and 44 patients in two consecutive Phase II studies who were treated with pegylated liposomal doxorubicin (40-60 mg/m2 every 3 weeks for the first two cycles and 40 mg/m2 every 4 weeks thereafter) after failing initial platinum-based chemotherapies for ovarian carcinoma were analyzed. Associations were sought for response, time to failure (TTF), and survival after the treatment and various pretreatment characteristics. RESULTS: Treatment with pegylated liposomal doxorubicin yielded 23% objective responses in measurable disease and 31% overall responses, including serum CA 125-defined responses. The median TTF was 5.2 months (95% confidence interval, 4.1-6.9 months) in all patients, and the median response duration in all responders was 13.2 months (95% confidence interval, 11.9-18.5 months). The overall median survival was 15 months (95% confidence interval, 11-40 months). The main predictive factors were tumor size and baseline hemoglobin level for TTF, and these plus Karnofsky performance status were the main predictive factors for survival. CONCLUSIONS: Pegylated liposomal doxorubicin is an effective drug when it is given as secondary therapy to patients with EOC. Lack of bulky disease is the major predictor for a favorable response, TTF, and survival. The role of this treatment in combination with other effective drugs should be explored in both previously treated and untreated patients with ovarian carcinoma.

The use of a pelvic displacement prosthesis to exclude the small intestine from the radiation field following radical hysterectomy.

OBJECTIVE: The goal of this study was to develop a device which will elevate the small intestine out of the pelvic cavity during radiation after radical surgery. METHODS: A prosthetic device of silicone plastic was designed which conforms to the pelvis. This device is filled with saline and renograffin for X-ray visualization. The capacity of the device is between 750 and 1500 cc. A small bowel contrast radiograph is performed prior to radiation to document exclusion from the radiation field. The device remains in place throughout radiation therapy and is then removed through a small incision after draining the contents of the prosthesis. RESULTS: Seven devices have been placed to date. The patients' age ranged from 35 to 65 years. All women had stage Ib1 carcinoma of the cervix and all underwent a type III radical hysterectomy with bilateral pelvic and common iliac lymphadenectomy. The indication for placement of the device was deep invasion of tumor in five patients, close margin in one patient, and positive pelvic lymph nodes in one patient. The amount of fluid instilled in the device ranged from 960 to 1200 cc. All patients had a return to normal bowel function within 3 days of surgery. All had radiologically documented exclusion of the small intestine from the radiation field prior to beginning radiation. In the postoperative period there was one major complication: a pulmonary embolism documented by pulmonary angiogram on postoperative day 2. All seven patients completed planned radiotherapy. The devices have been removed, with no adhesions to the prosthesis. CONCLUSIONS: The results of this study determine that the feasibility, safety, and efficacy of a prosthetic device in displacing the small bowel from the radiation field following radical surgery are sufficient to warrant a large-scale study. The device should be applicable to any and all tumors that require high dose pelvic radiation. It is expected that displacement of the small intestine from the radiation field will diminish overall complications and may allow delivery of radiation doses that approach colon and bladder tolerance.

A phase II study of gemcitabine and cisplatin in patients with advanced, persistent, or recurrent squamous cell carcinoma of the cervix.

OBJECTIVE: The aim of this study was to determine the response rate and toxicity of cis-platinum and gemcitabine in advanced, recurrent, or persistent squamous cell carcinoma of the cervix. METHODS: From July 1997 to January 1999, we conducted a Phase II trial in patients with advanced, persistent, or recurrent carcinoma of the cervix. The schedule employed 1250 mg/m(2) of gemcitabine on days 1 and 8 and 50 mg/m(2) of cis-platinum on day 1 in a 21-day cycle. Eligibility criteria were a GOG performance status of 0-2, adequate bone marrow reserve, serum creatinine less than 1.8 mg%, and a lesion which could be measured in two dimensions. None of the patients had received prior chemotherapy other than radiation sensitizers. Standard GOG toxicity and response criteria were used. RESULTS: Nineteen patients were enrolled into the trial. Two patients were inevaluable because of inadequate trial of drug. Seventeen patients were evaluable for response and toxicity. The median age of the patients was 47 years (range 24-72). The median number of cycles delivered was 5 (range 2-8). The incidence of grade 4 neutropenia and anemia was 2.4 and 1.2%, respectively. Two patients developed a single episode of grade 3 gastrointestinal toxicity. The overall response rate was 41% (7/17). There was 1 complete response of 14 months duration and 6 partial responses. Among those patients not previously irradiated, the response rate was 57% (4/7). Among the radiated patients, the response rate was 30% (3/10) with all responses occurring in the radiation field. CONCLUSION: This combination of cis-platinum and gemcitabine is a well-tolerated regimen which exhibits high activity in advanced, recurrent, or persistent squamous cell cervical cancer.

Cold-knife conization versus conization by the loop electrosurgical excision procedure: a randomized, prospective study.

OBJECTIVE: Our purpose was to compare the diagnostic ability and treatment efficacy of conization by the loop electrosurgical excision procedure with cold-knife conization. STUDY DESIGN: One hundred eighty women who required conization for diagnosis and treatment of cervical dysplasia or microinvasive cervical carcinoma were prospectively enrolled in a randomized clinical trial to receive either cold-knife conization or conization by the loop electrosurgical excision procedure. Conization complications, rate of lesion clearance, and therapeutic outcome were assessed for the 2 study groups. RESULTS: There were no statistically significant differences in the complication rate (P = 1.00), the rate of lesion clearance (P =.18), or the rate of disease recurrence (P =.13) between the 2 study groups. The mean follow-up was 11.2 months in the cold-knife conization group and 10.4 months in the loop-excision conization group. CONCLUSION: Cold-knife conization and loop-excision conization yield similar diagnostic and therapeutic results.

Villoglandular adenocarcinoma of the endometrium: a clinicopathologic study of 61 cases: a gynecologic oncology group study.

Papillary endometrioid or villoglandular adenocarcinoma (VGA) is a relatively common type of endometrial adenocarcinoma, but studies describing its behavior have yielded conflicting results. Patients with a component of VGA were identified in a review of 819 women entered in a Gynecology Oncology Group Study (Protocol 33) of clinical stages I and II endometrial adenocarcinoma. Cases with coexisting foci of serous or clear cell carcinoma were excluded from further consideration. Of the 61 cases that formed the study sample, there were 24 with pure villoglandular differentiation and 37 who were admixed with typical endometrioid adenocarcinoma (EA). The general clinicopathologic features of patients with pure and mixed VGA are compared with 469 patients with pure EA. The VGAs were better differentiated (grade 1 or 2--97% of VGA versus 74% EA, p = 0.001). but they were not significantly different with respect to median age, depth of invasion, or frequency of nodal spread. Six of the 61 patients with VGA died of their tumor. The disease-specific survival rate at 3 years for VGA is 94% (95% confidence interval: 0.88-0.99) compared with 88% (95% CI: 0.86-0.91) for EA. Two of the patients who died had pure villoglandular tumors and four had mixed villoglandular and endometrioid carcinoma. In view of the frequent admixture of VGA and EA and their generally similar biological characteristics, with a prognosis similar to that of typical EA, we conclude that VGA should be considered a variant of EA.

Carcinoembryonic antigen in women with isolated pelvic masses. Clinical utility?

OBJECTIVE: To determine the usefulness of serum carcinoembryonic antigen (CEA) determination in predicting the nature of an isolated pelvic mass. STUDY DESIGN: Two hundred twenty-six women with an isolated pelvic mass had a serum CEA determination preoperatively. The results were correlated with the histopathologic findings. RESULTS: CEA was elevated in 19 of the 226 women. Twelve of the 183 (7%) women with benign masses, 2 of the 17 (12%) women with tumors of low malignant potential and 5 of the 15 (33%) women with a frankly invasive epithelial ovarian cancer had elevated CEA. None of the women with a malignant germ cell or stromal tumor had elevated CEA (P = .06 for prediction of malignancy.) There were no cases of metastatic gastrointestinal malignancies in the study group. The sensitivity, specificity, positive predictive value and negative predictive value of serum CEA were 16%, 93%, 37% and 83%, respectively. The corresponding figures for serum CA-125 were 67%, 71%, 35% and 90%. There was no statistically significant correlation between elevated CEA and mucinous histology. CONCLUSION: Preoperative serum CEA determination in women with isolated pelvic masses is not useful.

Influence of quantity of lymph-vascular space invasion on the risk of nodal metastases in women with early-stage squamous cancer of the cervix.

OBJECTIVE: To determine if the quantity of lymph-vascular space invasion influences the risk of pelvic nodal metastases in women with early-stage cervical squamous carcinoma. METHODS: Between 1991 and 1997, 105 women with stages IA2, Ib, and IIa squamous carcinoma of the cervix underwent radical hysterectomy and pelvic lymphadenectomy. The histopathology slides were prospectively reviewed. If lymph-vascular space invasion (LVSI) was present, the amount was quantified in four ways: percentage of cervical histopathologic sections containing LVSI, percentage of sections with tumor containing LVSI, total number of foci of LVSI, and maximal number of foci of LVSI in 10 high-powered fields. These measures of LVSI were correlated with the risk of nodal metastases. RESULTS: Seventy-three (70%) women had tumors with LVSI. Of these, 23 had pelvic nodal metastases. All women with nodal metastases had tumors with LVSI. Using logistic regession, independent predictors of nodal metastases were depth of cervical stromal invasion (P = 0.01) and tumor size (P = 0.04). LVSI was also a significant predictor of nodal metastases based on the Mantel-Haenszel test (P = 0.01). In women whose tumors contained LVSI, logistic regression identified tumor size (P = 0.004) and LVSI in > 45% of all cervical histopathologic sections (P = 0.002) as significant predictors of nodal metastases. CONCLUSION: The quantity of LVSI, as defined by the percentage of all cervical histopathologic sections containing LVSI, correlates significantly with the risk of nodal metastases in women with early-stage squamous carcinoma of the cervix.

Risk of residual invasive disease in women with microinvasive squamous cancer in a conization specimen.

OBJECTIVE: To quantify the risk of residual invasion when cervical conization reveals microinvasive squamous carcinoma and to determine whether any factors affect this risk. METHODS: We reviewed the charts and histopathology slides of 87 women who underwent a conization that contained microinvasive squamous carcinoma, followed by either a repeat conization or hysterectomy. Depth of invasion, number of invasive foci, and status of the internal margin and post-conization endocervical curettage (ECC) were assessed. The findings were correlated with the presence of residual invasion. RESULTS: Significant predictors of residual invasion included status of the internal margin (residual invasion present in 22% of women with an involved margin versus 3% with a negative margin; P < .03) and the combined status of the internal margin and post-conization ECC (residual invasion in 4% of patients if both negative, 13% if one positive, and 33% if both positive; P < .015). Depth of invasion and number of invasive foci in the conization specimen were not significant. The power of this study to detect a 25% difference in the risk of residual invasion was 73% for depth of invasion and 75% for number of invasive foci. CONCLUSION: Women with microinvasive squamous carcinoma in a conization specimen in which both the internal conization margin and post-conization ECC are negative have a low risk of residual invasion and are candidates for follow-up or simple hysterectomy. If either the internal margin or the post-conization ECC contains dysplasia or carcinoma, the risk of residual invasion is high and warrants repeat conization before definitive treatment planning.

Phase I/II study of intraperitoneal floxuridine and platinums (cisplatin and/or carboplatin).

Previous studies have shown that intraperitoneal (i.p.) floxuridine (FUDR) is tolerated at a dose of 3 g x 3 days given in 1.5-2 L of normal saline (NS). In a randomized phase II trial by the Southwest Oncology Group, this treatment was selected for further study because of a favorable 1-year progression-free survival. We have now evaluated ip FUDR in full doses combined with i.p. cisplatin given on the third day at a dose of 60 mg/m2 in 500 mL of NS. Intraperitoneal carboplatin was partially or fully substituted for i.p. cisplatin in patients with symptomatic neuropathies. All patients also received i.p. leucovorin, as previously piloted for fluoropyrimidine modulation. Seven patients with symptomatic ascites or measurable tumors were entered, as were 11 asymptomatic patients with minimal residual (< or = 1 cm) epithelial ovarian cancer. Six cycles of the combination of i.p. FUDR + cisplatin were completed in three patients; however, the combination of FUDR with both platinums was particularly well tolerated. Intraperitoneal FUDR + carboplatin (AUC of 5) was associated with some grade 3 and 4 thrombocytopenia and neutropenia. Eight of these 11 patients are alive, and 3 have been continuously with no evidence of disease exceeding 32 months. The regimen of i.p. FUDR + i.p. cisplatin (or i.p. FUDR with both platinums) is suitable for a phase III trial testing i.p. therapy either from the outset (e.g., i.p. up front) or after achieving clinical complete responses from initial treatment without intervening relapse (i.e., i.p. consolidation) in comparison to i.p. cisplatin.

Carboplatin as a radiation sensitizer in locally advanced cervical cancer: a pilot study.

Radiation therapy is the mainstay in treatment of locally advanced cervical carcinoma. Several chemotherapeutic agents have been used as radiation sensitizers in the treatment of cervical cancer in an effort to improve local response and survival. A prospective study was designed to evaluate carboplatin as a radiosensitizer in advanced cervical cancer. Standard radiotherapy techniques were used to treat patients with Stage IIA-IIIB cervical cancer. Intravenous carboplatin was administered twice weekly concurrent with external beam radiation. Of 22 evaluable patients, there were 19 complete responders of whom 15 remain alive: 11 patients were alive and disease free at last visit for a median duration of 15 months follow-up (range, 4-43 months) and 4 patients remain alive with disease for a median duration of 17 months (range, 3-55 months). Seven have died, one of whom was without evidence of disease. There were no treatment-related deaths and no grade 4 toxicity. The most significant adverse effect was hematologic resulting in four patients with grade 3 neutropenia or anemia. There were no fistulae or late gastrointestinal or genitourinary complications. This pilot study suggests that carboplatin administered with standard radiation is safe, well-tolerated, and thus may be useful as a radiation sensitizer in the treatment of locally advanced cervical cancer.

Pelvic examination, tumor marker level, and gray-scale and Doppler sonography in the prediction of pelvic cancer.

OBJECTIVE: To determine the ability of pelvic examination, tumor marker assessment, and transvaginal ultrasonography, with selected use of Doppler ultrasonography, to predict pelvic malignancy. METHODS: Two hundred twenty-six women scheduled for operative removal of a pelvic mass were entered in the study prospectively. Each woman underwent pelvic examination, tumor marker assessment, and transvaginal ultrasonography preoperatively. Women whose gray-scale findings were suspicious for malignancy underwent Doppler ultrasonography. Suspicious findings included masses that were fixed or irregular on pelvic examination; CA 125 level greater than 35 U/mL; elevations in serum lactic dehydrogenase, alpha-fetoprotein, or hCG; and the presence of a substantial solid component on gray-scale ultrasonography. Suspicious Doppler findings included intratumoral color flow, pulsatility index less than 1.0, or resistance index 0.4 or lower. The findings were correlated with the presence of malignancy. RESULTS: If all three indicators (examination, tumor marker assessment, and gray-scale ultrasound findings) were nonsuspicious, 99% of premenopausal women and 100% of postmenopausal women had benign masses. If all three indicators were suspicious, 77% of premenopausal women and 83% of postmenopausal women had malignant tumors. Logistic regression identified ultrasound impression and tumor size to be significant predictors of malignancy in premenopausal women, whereas CA 125 level and ultrasound impression were significant in postmenopausal women. In patients with suspicious gray-scale findings, recategorization based on Doppler findings resulted in inferior diagnostic indices. CONCLUSIONS: Ultrasonographic tumor size and appearance are the best predictors of pelvic malignancy in premenopausal women, whereas CA 125 level and ultrasonographic appearance are the best predictors in postmenopausal women. Neither color nor spectral Doppler is useful in this setting.

Phase II study of liposomal doxorubicin in refractory ovarian cancer: antitumor activity and toxicity modification by liposomal encapsulation.

PURPOSE: A phase II study of liposomal doxorubicin was conducted in patients with ovarian cancer who failed to respond to platinum- and paclitaxel-based regimens. Liposomal doxorubicin was selected as a result of its superior activity against ovarian cancer xenografts relative to free doxorubicin and activity in refractory ovarian cancer patients that was noted during the phase I study. PATIENTS AND METHODS: Thirty-five consecutive patients were accrued in two institutions (22 in one and 13 in the other). All had progressive disease after either cisplatin or carboplatin and paclitaxel, or at least one platinum-based and one paclitaxel-based regimen. Patients received intravenous (I.V.) liposomal doxorubicin 50 mg/m2 every 3 weeks with a dose reduction to 40 mg/m2 in the event of grade 3 or 4 toxicities, or a lengthening of the interval to 4 weeks (and occasionally to 5 weeks) with persistence of grade 1 or 2 toxicities beyond 3 weeks. RESULTS: Nine clinical responses (one complete response [CR], eight partial responses [PRs]) were observed in 35 patients (25.7%), with seven of these having been confirmed by two consecutive computed tomographic (CT) measurements. The median progression-free survival was 5.7 months with an overall survival of 1.5 to 24+ months (median, 11 months). Although 13 patients experienced grade 3 or 4 nonhematologic skin and mucosal toxicities (either hand-foot syndrome or stomatitis), with dose modifications, the treatment was very well tolerated. Nausea that was clearly attributable to the drug, hair loss, extravasation necrosis, or decreases in ejection fraction did not occur. CONCLUSION: Liposomal doxorubicin has substantial activity against ovarian cancer refractory to platinum and paclitaxel. The responses achieved with liposomal doxorubicin were durable and maintained with minimal toxicity. This liposomal formulation should be evaluated further in combination with other drugs in less refractory patients.

Residual disease in the uterus after preoperative radiotherapy and hysterectomy in stage IB cervical carcinoma.

Sixty-one patients with FIGO IB cervical cancer treated with planned preoperative radiotherapy (dose to point A: 52-93 Gy, mean 73 Gy) and hysterectomy from 1969 to 1993 were retrospectively reviewed. Patient characteristics and treatment parameters and their association with residual tumor in the hysterectomy specimen were analyzed. Glandular (adenocarcinoma and adenosquamous) tumors were smaller than squamous tumors: 6/11 (55%) were < 6 cm in diameter, versus 12/50 (24%) squamous tumors (p = 0.03). Glandular tumors had a higher incidence of residual disease: 10/11 (91%) versus 24/50 (48%) (p = 0.01). There was no association between presence of pathologic residual disease in the hysterectomy specimen and tumor size, morphology (endophytic vs. exophytic), patient age, dose to point A, time to deliver radiotherapy, or interval between radiotherapy and hysterectomy. Overall 34/61 (56%) patients had residual disease in their hysterectomy specimens after planned preoperative radiotherapy. There were significantly more glandular tumors than squamous tumors with residual disease, even though glandular tumors were a group of smaller tumors.

Pathologic models to predict outcome for women with endometrial adenocarcinoma: the importance of the distinction between surgical stage and clinical stage--a Gynecologic Oncology Group study.

BACKGROUND: Numerous pathologic factors have been identified as important in predicting outcome for women with endometrial adenocarcinoma. However, most patients have a mixture of good and bad factors. For these women, the prognosis is uncertain, and it is often unclear whether postoperative therapy is indicated. METHODS: Using univariate and multivariate analysis, we investigated the pathologic factors commonly reported to be of prognostic significance, using data from 819 patients with clinical Stages I and II endometrial adenocarcinoma from a Gynecologic Oncology Group study. Since the clinical stage frequently underestimated the surgical stage, models that designate the relative risk associated with each of the variables were created for both clinical and surgical Stage I and 11 patients. RESULTS: We confirmed the importance of age, depth of myometrial invasion, and to a lesser degree, histologic grade, and cell type, as independent prognostic variables. CONCLUSIONS: The relative risk of death can be determined using a simple multiplicative calculation, and the absolute risk can be estimated by inspection of the accompanying figures. These data can be used to provide patients with prognostic information and to help determine the need for postoperative adjuvant therapy.

The role of endocervical curettage at cervical conization for high-grade dysplasia.

OBJECTIVE: To quantify the risk of invasive cancer above the location where the conization specimen was taken in patients with an endocervical curettage (ECC) positive for dysplasia at conization for high-grade cervical intraepithelial neoplasia (CIN), and to determine if any pathologic features may influence this risk. METHODS: The charts of 104 patients who underwent cervical conization for high-grade dysplasia followed by repeat conization or hysterectomy at Los Angeles County + University of Southern California Women's Hospital between January 1986 and December 1992 were reviewed retrospectively. Patients with invasive cancer or glandular dysplasia on the initial conization were excluded. The ECC performed immediately after conization biopsy (conization ECC) was benign in 63 patients and contained dysplasia in 41. All available conization ECC specimens that contained dysplasia were evaluated for volume of dysplasia and degree of cytologic atypia. Fisher exact test was used for statistical comparison between and within groups. RESULTS: Invasive cancer was not present in any patients in the benign ECC group but was present in nine (22%) patients in the dysplasia group (P < .0001); five of these patients had microinvasion (no more than 3 mm of stromal invasion and no lymph-vascular space involvement) and four had frank invasion. Comparison of patients with involved endocervical margins revealed that none of 37 patients in the benign ECC group versus eight of 27 patients in the dysplasia group had invasive cancer (P < .0005). All patients with invasion were 35 years or older and all patients with frank invasion were 50 years or older. Neither volume nor cytologic grade of dysplasia in the ECC was predictive of invasion in the residual cervix. CONCLUSIONS: An ECC at conization positive for dysplasia is an important predictor of invasion in the residual cervix of patients whose conization reveals high-grade intraepithelial neoplasia and should be routinely performed. Women 50 years or older with both a positive endocervical margin and conization ECC should undergo repeat conization before further therapy. Women under 50 years of age should undergo repeat conization if fertility is not desired; otherwise, close follow-up is necessary to exclude the presence of an invasive lesion in the residual cervix.