Frozen section analysis is widely used in the intraoperative evaluation of adnexal masses, and it is often useful to determine the appropriate surgical strategy. The discriminative preoperative evaluation of adnexal masses is rather complex. Various diagnostic models, such as the risk of malignancy index,1–3 are used in the preoperative work-up. Despite all efforts, the interpretation of the nature of the adnexal mass is often inaccurate. The prospective, randomized, controlled study by Yazbek et al4 shows that the quality of gynecologic ultrasonography has a significant influence on the management of patients with suspected ovarian cancer. Experienced sonographers diagnosed benign adnexal pathology more accurately.
Intraoperative pathologic examination aids in making an informed decision for determining the extent of surgery and helps prevent both undertreatment and overtreatment. This is especially important for young women who may be managed conservatively with preservation of fertility.
Frozen section analysis is generally accepted as a reliable method for determining the nature of an adnexal mass. The accuracy in detecting invasive malignancies has been assessed in previous studies. A systematic review by Geomini et al5 shows high levels of sensitivity and specificity (71–100% and 98–100%, respectively) for frozen section analysis. In contrast, frozen section analysis of borderline adnexal masses appears to be less accurate. Tempfer et al6 present a pooled analysis of four studies that included 317 women with borderline adnexal masses. The overall sensitivity was 71.1% and the positive predictive value was 84.3%.
It might be wise to analyze frozen sections from all patients with adnexal masses to obtain as much information as possible for determining the optimal surgical procedure. However, analyzing frozen sections extends the operating time and thus the duration of anesthesia. Furthermore, frozen section analysis increases costs and implies a heavier workload for pathologists.
Most gynecologists decide before surgery whether frozen section analysis is needed, and they may alter the decision during surgery, depending on the intraoperative findings. Various factors influence the surgeon's use of frozen section analysis, depending on the suspicion of malignancy.7 The objective of this study was to determine the factors that influence the use of frozen section analysis in adnexal masses and the factors that predict malignancy.
MATERIALS AND METHODS
This retrospective cohort study was conducted in the Radboud University Nijmegen Medical Centre and 10 cooperating referral hospitals in the east of the Netherlands between January 2005 and September 2009. The study was approved by the medical ethics committee of the Radboud University Nijmegen Medical Centre. Women who were admitted for surgical treatment of an adnexal mass with unknown histology were included in the study. A subgroup (n=548) of the study population has been described in our previous publication about the risk of malignancy index.3 The patients for whom frozen section analysis was cancelled were excluded from participation in this study if there was clear evidence of malignancy during the surgical procedure, such as pleural effusions and evidence of distal organ involvement.
Menopausal status was registered. Ultrasonography for assessing the locularity of the tumor, laterality, and presence of solid areas, and the determination of the serum level of CA 125 were all parts of the routine preoperative evaluation. All ultrasonography was performed by experienced sonographers. Ultrasonography was performed transvaginally and was combined with abdominal ultrasonography if necessary. Doppler flow studies were not performed. No ultrasound morphologic grading system was used. The presence of adhesions and the diameter of the tumor were retrospectively obtained from the surgery and pathology reports. The use of frozen section analysis was registered, as was the corresponding histopathologic outcome. The final diagnosis of the adnexal mass was based on the full histopathologic examination of all surgical specimens removed. A total of six pathology units covered the pathology activities of the 11 hospitals that participated in this study. One of these units covered four hospitals, two units each covered two hospitals, and three units each covered one hospital.
The histopathologic diagnosis of the frozen section analysis was compared with the final histopathology. The sensitivity, specificity, and the positive and negative predictive values of frozen section analysis for malignant tumors were calculated.
Factors that potentially influenced the use of frozen section analysis and potentially predicted the presence of malignancy, including menopausal status, CA 125 level, ultrasound characteristics, presence of adhesions, and tumor size, were studied.
We used univariable logistic regression to study the ability of the variables to predict the use of frozen section analysis and to predict invasive malignancy in patients with adnexal masses. We used multivariable logistic regression with stepwise selection procedures to identify variables that contributed independently to the use of frozen section analysis and to the presence of invasive malignancy. All variables that reached the level of significance at 0.10 in the univariable logistic regression were valid for entry in the selection procedure, and we used P=.05 for staying in the model. We used SPSS 16.0.1 for Windows for all statistical analyses.
During the study period, 722 women underwent surgery for an adnexal mass, and 670 (93%) were included in the study. The 52 women (70%) for whom frozen section analysis was cancelled because of clear evidence of malignancy during the surgical procedure were excluded from the study. Of the 670 participants, 531 (80%) had benign gynecologic conditions diagnosed (ie, final histopathologic diagnosis), 70 (10%) had borderline malignancy diagnosed, and 69 (10%) had malignant disease diagnosed. General gynecologists performed the surgery for 503 women (75%), gynecologic oncologists performed the surgery for 53 women (8%), and either a general gynecologist or a gynecologic oncologist performed the surgery for 114 women (17%).
Frozen sections were analyzed for 323 patients (48%); 206 of these sections (64%) showed benign ovarian cysts, 55 (17%) showed borderline malignancies, and 62 (19%) showed malignant adnexal masses. Frozen sections were analyzed for 39% of the benign ovarian cysts, for 79% of the borderline malignancies, and for 90% of the malignant adnexal masses. Table 1 presents the characteristics and final histologic types of all the cases in relation to frozen section analysis. The adnexal masses of seven patients without frozen section analysis were identified as malignant only after the final histopathologic examination (10%). Five of these patients were postmenopausal and two were premenopausal. Three of the seven patients presented with a multilocular mass and solid areas on ultrasonography. One of them had a CA 125 level of 60 units/mL, and another presented with adhesions during surgery. Two patients presented with multilocularity only on ultrasonography. One patient presented with solid areas only on ultrasonography and one presented with only a CA 125 level of 40 units/mL.
Table 2 shows the comparison of frozen section analysis with the final histopathologic diagnosis. The frozen section analysis was concordant with the final pathology findings for 292 patients (90%). The sensitivity, specificity, and positive and negative predictive values of frozen section analysis were 84% (95% confidence interval [CI] 75–93%), 99% (95% CI 98–100%), 95% (95% CI 89–100%), and 96% (95% CI 94–98%), respectively, for malignant tumors. The accuracy was 96% (95% CI 94–98%).
Table 3 shows factors that potentially influence the use of frozen section analysis or predict the presence of malignancy. In a univariable analysis, all factors tested were significant predictors of the use of frozen section analysis. Predictive factors for the presence of malignancy were the CA 125 level, locularity of the tumor, bilaterality, presence of solid areas, presence of adhesions, and diameter of the tumor. Table 4 shows the multivariable regression analyses used to identify independent predictive factors for frozen section analysis use and for the presence of malignancy. The CA 125 level, locularity of the tumor, and the presence of solid areas were predictors of both the use of frozen section analysis and the presence of malignancy. The presence of adhesions predicted malignancy, but not the use of frozen section analysis. In contrast, menopausal status and tumor size were predictors of the use of frozen section analysis, but not of malignancy.
Intraoperative frozen section analysis is an important and reliable tool in the clinical management of patients with adnexal masses. The main problem in this management is the risk of malignancy, which is why adnexal masses have to be carefully assessed before surgery. As reported in previous publications,5,8,9 frozen section analysis is a reliable method for detecting invasive malignancies during the operative procedure. In our study, we found high levels of sensitivity (84%) and positive predictive value (95%), which are comparable with those of previous studies.5,8,9 Little information has been published concerning the factors that influence the decision to analyze a frozen section during adnexal mass surgery. We therefore have investigated which factors influence the use of frozen section analysis in adnexal masses and which factors predict malignancy.
In the recent study by Brun et al,7 patient age older than 50 years, tumor size larger than 10 cm, and preoperative evidence of malignancy were associated with more use of frozen section analysis. We also have identified tumor size as an independent predictive factor for using frozen section analysis. Other factors influencing the use of frozen section analysis in our study were the CA 125 level, locularity of the tumor, and presence of solid areas. Tumor size predicted the use of frozen section analysis, but not the presence of malignancy. This is compatible with the literature, because tumor size is generally not considered an independent predictor of malignancy.10
In our cohort, frozen sections were more often analyzed for postmenopausal women, but menopausal status was not identified as an independent predictor of malignancy. This conflicts with data on the risk of malignancy index that identified postmenopausal status as an independent risk factor for malignancy.1 Women for whom frozen section analysis was cancelled because of clear evidence of malignancy during the surgical procedure were excluded from the study. These women, however, were more often postmenopausal than the women in our study group (83% compared with 58%). This might have caused a selection bias. Furthermore, it might be possible that menopausal status would have been identified as an independent predictor of malignancy in a larger sample size. Adhesions revealed during surgery were associated with malignancy. However, adhesions did not lead to more use of frozen section analysis. The presence of adhesions is the only variable that cannot be assessed reliably preoperatively; it has to be assessed during surgery. Although we do not have all the necessary information, the data suggest that the decision to analyze a frozen section is more often based on the preoperative findings than intraoperative ones, or that the presence of adhesions is an underestimated predictor of malignancy.
Unfortunately, whether ascites were present was unknown for 68% of the patients. Therefore, we were unable to determine whether the presence of ascites is a predictor of analyzing frozen sections to determine malignancy. The presence of ascites could indicate malignancy and probably has an effect on the management of these cases.
The malignancies of seven patients (which represent 10% of all malignancies) were not found intraoperatively. Six of these patients were wrongly diagnosed with benign diseases; therefore, they were not properly staged. Analyzing frozen sections would have been of great value in these cases. The surgeon suspected malignancy in one case but decided that the patient would not be a suitable candidate for radical surgery. Therefore, minimally invasive surgery was performed and several biopsies were performed. In this case, frozen section analysis would not have changed the management approach.
Frozen section analysis has its limitations. It is not accurate in all cases. A report by Geomini et al11 shows that tumor size has an effect on the accuracy of frozen section analysis. For masses with a diameter of 10 cm or larger, a benign result of the frozen section analysis was less reliable than for masses with a diameter of less than 10 cm. In the group with masses of 10 cm or more, 11% of the women for whom frozen section analysis indicated a benign cyst turned out to have a malignant or borderline tumor according to the final pathology. Not only tumor size but also the mucinous histologic type limits the accuracy of frozen section analysis.5,9,12,13
Frozen section analysis is not always useful. If the probability of malignancy is low before surgery, then it is unlikely that frozen section analysis will change the clinical management. This is also true when the malignancy becomes evident before or during surgery. It is important that the use of frozen section analysis is considered carefully in each case to avoid superfluous testing, but it also should be used when necessary to determine the extent of the surgical procedure.
In conclusion, menopausal status and tumor size are associated with more use of frozen section analysis, but they have not been identified as factors associated with malignancy. The frozen section analysis is useful when the CA 125 levels are greater than 35 units/mL and when there are multilocular tumors, solid areas on ultrasonography, and adhesions revealed during surgery.
1. Jacobs I, Oram D, Fairbanks J, Turner J, Frost C, Grudzinskas JG. A risk of malignancy index incorporating CA 125, ultrasound and menopausal status for the accurate preoperative diagnosis of ovarian cancer. Br J Obstet Gynaecol 1990;97:922–9.
2. Tingulstad S, Hagen B, Skjeldestad FE, Halvorsen T, Nustad K, Onsrud M. The risk-of-malignancy index to evaluate potential ovarian cancers in local hospitals. Obstet Gynecol 1999;93:448–52.
3. van den Akker PA, Aalders AL, Snijders MP, Kluivers KB, Samlal RA, Vollebergh JH, et al. Evaluation of the Risk of Malignancy Index in daily clinical management of adnexal masses. Gynecol Oncol 2010;116:384–8.
4. Yazbek J, Raju SK, Ben-Nagi J, Holland TK, Hillaby K, Jurkovic D. Effect of quality of gynaecological ultrasonography on management of patients with suspected ovarian cancer: a randomised controlled trial. Lancet Oncol 2008;9:124–31.
5. Geomini P, Bremer G, Kruitwagen R, Mol BW. Diagnostic accuracy of frozen section diagnosis of the adnexal mass: a metaanalysis. Gynecol Oncol 2005;96:1–9.
6. Tempfer CB, Polterauer S, Bentz EK, Reinthaller A, Hefler LA. Accuracy of intraoperative frozen section analysis in borderline tumors of the ovary: a retrospective analysis of 96 cases and review of the literature. Gynecol Oncol 2007;107:248–52.
7. Brun JL, Cortez A, Rouzier R, Callard P, Bazot M, Uzan S, et al. Factors influencing the use and accuracy of frozen section diagnosis of epithelial ovarian tumors. Am J Obstet Gynecol 2008;199:244.e1–7.
8. Medeiros LR, Rosa DD, Edelweiss MI, Stein AT, Bozzetti MC, Zelmanowicz A, et al. Accuracy of frozen-section analysis in the diagnosis of ovarian tumors: a systematic quantitative review. Int J Gynecol Cancer 2005;15:192–202.
9. Ilvan S, Ramazanoglu R, Ulker Akyildiz E, Calay Z, Bese T, Oruc N. The accuracy of frozen section (intraoperative consultation) in the diagnosis of ovarian masses. Gynecol Oncol 2005;97:395–9.
10. McDonald JM, Doran S, DeSimone CP, Ueland FR, DePriest PD, Ware RA, et al. Predicting risk of malignancy in adnexal masses. Obstet Gynecol 2010;115:687–94.
11. Geomini PM, Zuurendonk LD, Bremer GL, de Graaff J, Kruitwagen RF, Mol BW. The impact of size of the adnexal mass on the accuracy of frozen section diagnosis. Gynecol Oncol 2005;99:362–6.
12. Rose PG, Rubin RB, Nelson BE, Hunter RE, Reale FR. Accuracy of frozen-section (intraoperative consultation) diagnosis of ovarian tumors. Am J Obstet Gynecol 1994;171:823–6.
© 2011 The American College of Obstetricians and Gynecologists
13. Puls L, Heidtman E, Hunter JE, Crane M, Stafford J. The accuracy of frozen section by tumor weight for ovarian epithelial neoplasms. Gynecol Oncol 1997;67:16–9.