Ultrasound has been largely used for identifying malignancy in adnexal masses. Most studies have focused on the discrimination between benign and malignant lesions.1 On the contrary, only few studies have focused on the characterization of different types of ovarian malignancies by ultrasound.2–5
Recent morphologic, immunohistochemical, and molecular genetics studies suggest that there are 2 types of epithelial ovarian cancer (EOC), namely, EOC type I and EOC type II.6 Type I tumors comprise low-grade serous, low-grade endometrioid, clear cell, mucinous, and transitional cell carcinomas, whereas type II tumors comprise high-grade serous, high-grade endometrioid, malignant mixed mesodermal tumors, and undifferentiated carcinomas.6
According to Kurman and Shih,6 these 2 types of EOC are characterized by different clinical behaviors as well as different morphological, immunohistochemical, and molecular genetics features.
In addition, there is some evidence that type I cancers arise from well-characterized precursor lesions such as borderline tumors and endometriosis.7
To the best of our knowledge, no study has assessed the ultrasound features of type I and type II EOC. Our null hypothesis is that there are no differences in ultrasound features between these 2 types of EOC.
In the present study, we aimed to assess if our hypothesis is correct or, alternatively, there are differences in ultrasound features of type I and type II EOC.
This is a retrospective observational study. Eligible subjects were every woman diagnosed and treated for epithelial invasive ovarian cancer at our institution between January 1995 and June 2012.
Clinical, ultrasound, and histopathologic records were reviewed. Patients were identified through hospital database, and records retrieved for review.
Inclusion criterion was histologically proven diagnosis of epithelial invasive ovarian cancer. Exclusion criteria were as follows: surgery performed outside our institution, incomplete clinical or ultrasound records, and ultrasound not performed preoperatively.
Clinical data evaluated were as follows: patient’s age, patient’s symptoms (asymptomatic, abdominal swelling, constipation, abdominal discomfort or pain, bloating), menopausal status (defined as 1 year of absence of menstruation in women >45 years old; a woman was considered as postmenopausal if >50 years), if hysterectomy had been performed, preoperative serum CA-125 levels, histological diagnosis, and tumor stage.
Ultrasound data analyzed in the present study were tumor volume (expressed in milliliters), bilaterality, tumor features according to the classification of Granberg et al8 (unilocular, unilocular solid, multilocular, multilocular solid, and solid) and tumor color score (absent, scanty, moderate, and abundant).9 Transvaginal ultrasound evaluation was performed according to a predetermined protocol.9 In cases of large tumors, fluid in Douglas pouch and suspicion of pelvic implants transabdominal ultrasound were also performed. In cases with bilateral tumors, we selected the largest one for analysis. Tumor volume was calculated measuring the 3 orthogonal planes and applying the prolate ellipsoid formula (volume = length × height × width × 0.5233).
Tumors were classified as EOC type I or type II according to Kurman and Shih.6 Type I tumors comprise low-grade serous, low-grade endometrioid, clear cell, mucinous, and transitional cell carcinomas, whereas type II tumors comprise high-grade serous, high-grade endometrioid, malignant mixed mesodermal tumors, and undifferentiated carcinomas. We used only morphological criteria because immunohistohemical and molecular genetics analyses were not available for most cases.
Tumors were staged according to International Federation of Gynecology and Obstetrics classification.10 Categorical variables are presented as raw number and percentage.
Kolmogorov-Smirnov test was used to determine data distribution in continuous variables. Continuous variables are presented as mean and SD or as median and range, depending on data distribution.
Categorical variables were compared using χ2 test. Continuous variables were compared using 1-way analysis of variance with Bonferroni post hoc test or Mann-Whitney U or Kruskal-Wallis tests, depending on data distribution.
Power analysis was not performed. P < 0.05 was considered as statistically significant. All statistical analyses were performed using the SPSS 15.0 statistical package (SPSS Inc, Chicago, IL). Institutional review board approval was obtained for this study.
Three hundred twenty women were eligible for this study. Seventy-six were excluded because ultrasound was not performed preoperatively (n = 32), surgery was not performed at our institution (n = 28), and because of incomplete records (n = 16).
Two hundred forty-four women were ultimately included. Patients’ mean age was 55.2 (SD, 12.8) years, ranging from 19 to 84 years old.
Sixty-seven women (27.5%) had type I EOC, and 177 (72.5%) had type II EOC. Histologic diagnoses are shown in Table 1.
We observed that women with type I EOC were younger, presented asymptomatic at diagnosis more frequently, and had lower CA-125 levels and lower tumor stage than women with type II EOC (Table 2).
Ultrasound features are shown in Table 3. Type II EOCs were more frequently identified as a solid mass and were smaller lesions than type I EOC. Amount of flow within the tumor was not statistically different among groups.
Ultrasound features according to specific histologic type are shown in Table 4. Interestingly, high-grade serous carcinomas and undifferentiated carcinomas presented more frequently as highly vascularized small solid tumors as compared with all other histotypes.
In the present study, we have assessed the clinical and sonographic features of primary EOC according to the current classification regarding its pathogenesis. According to our data, type II EOC presents in older, postmenopausal women. Most of them are symptomatic and have advanced disease at diagnosis. Sonographically, these tumors are generally small, highly vascularized solid tumors. On the contrary, type I EOC presents in younger women and generally asymptomatic at diagnosis, and usually tumors are diagnosed at earlier stage.
The strength of our study is that, to the best of our knowledge, it is the first to report specifically on the features of EOC according to the current paradigm of EOC carcinogenesis.6
Our study has some limitations. First, it is a retrospective analysis, so not all cases could be included because data were incomplete in a significant number of them. Second, molecular genetics and immunohistochemical analysis were not available for most cases, and we classified the tumors just based on morphological data. This could lead to some cases to be misclassified as type I or type II.
The potential clinical relevance of our results is that our findings could provide some clues to clinicians faced with a diagnosis of ovarian cancer about the type of EOC. Regarding the generalizability of our results, it is difficult to determine because this is a retrospective study, and similar information in the literature is scanty. In a recent review article by Lalwani et al,5 the authors reported some data that are in agreement with our findings.
According to these authors, high-grade serous carcinomas constitute more than 90% of all serous carcinomas (94.5% in our series); most of these tumors manifest at advanced stage (89.7% in our series), and at imaging, they usually present as a unilocular solid, multilocular solid, or a solid mass (98.7% of our cases presented as unilocular solid, multilocular solid, or solid mass). Lalwani et al5 stated that 84% of these tumors are bilateral. However, we found only 38% of bilateral cases at ultrasound diagnosis.
Low-grade serous carcinomas account less than 10% of serous carcinomas (5.5% in our series). At imaging, these tumors appear as large unilocular solid or multilocular solid masses. In our series, 66.6% of low-grade serous carcinomas presented as large unilocular solid or multilocular solid masses. Bilaterality was not a frequent finding (22.2%). This latter finding is in contrast with data reported by Lalwani et al5 (bilaterality rate for low-grade serous carcinomas as high as 74%–77%).
According to Lalwani et al,5 endometrioid carcinomas generally present as unilateral large multilocular cystic masses. In our series, about a half of cases had these ultrasound patterns. However, high-grade endometrioid carcinomas presented at advanced stage in 70% as compared with 21.4% in low-grade endometrioid carcinomas.
Clear cell carcinomas constituted 5.4% of all epithelial neoplasms in our series. At imaging, these tumors usually present as large unilocular solid or multilocular solid lesions (91.6% in our series).5 Mucinous carcinomas had a similar appearance than clear cell carcinomas.
In conclusion, we report on the ultrasound and clinical features of EOC according to histotype. We have found that some differences exist between type I and type II EOC. Although the clinical significance of these findings is still to be determined, this information could provide some clues to clinicians faced with a diagnosis of ovarian cancer.