Abstract
Background
Ovarian adenofibromas (AF) and cystadenofibromas (CAF) belong to the surface epithelial-stromal tumors, and may appear as solid, or solid and cystic masses mimicking ovarian cancers.
Purpose
To evaluate the capability of magnetic resonance imaging (MRI) including diffusion-weighted imaging (DWI) with apparent diffusion coefficient (ADC) measurement for the diagnosis of ovarian AF/CAF.
Material and Methods
Magnetic resonance manifestations of 13 cases of ovarian AF/CAF were retrospectively evaluated. DWI was obtained in all 13 lesions, and mean ADC values in 11 lesions were compared with those in solid portions of 27 ovarian cancers.
Results
Neither case with AF/CAF revealed high signal intensity on DWI, whereas all ovarian cancers showed high signal intensity on DWI. The ADC values in the solid portions of AF/CAF were significantly higher than those of ovarian cancers (P < 0.001). A cut-off value of 1.20 × 10–3 mm2/s for AF/CAF had a sensitivity of 82%, specificity of 93%, positive predictive value of 82%, and negative predictive value of 93%.
Conclusion
DWI with ADC measurement may be helpful in differentiating AF/CAF from ovarian cancers.
Keywords
Ovarian adenofibromas (AF) and cystadenofibromas (CAF) belong to the surface epithelial-stromal tumors, and are histologically characterized by prominent fibrous tissue components with epithelial elements, which are similar to those present in ovarian cystadenomas (1, 2). These tumors may mimic malignancy by imaging criteria because they may appear as multilocular cystic masses with solid components, or predominantly solid masses. AF/CAF may occur in women of reproductive age, and accurate preoperative diagnosis may help to avoid unnecessary surgery. The solid components of AF/CAF show very low intensity similar to the skeletal muscle on T2-weighted images (3–6). Black sponge-like appearance on T2-weighted images is one of the characteristic magnetic resonance (MR) manifestations of AF/CAF; this consists of tiny high T2 signal intensity foci within very low signal intensity solid components, reflecting scattered small cystic glandular structures within dense fibrous stromal proliferation (3). This finding is specific for AF/CAF, however, is not always observed (4–6). Low signal intensity diffuse or partial thickening of the cystic wall on T2-weighted images in multilocular cystic masses is another MR finding suggestive of ovarian CAF (6). However, AF/CAF without these characteristic MR features may be difficult to diagnose. Recently, the usefulness of diffusion-weighted imaging (DWI) in diagnosing gynecologic tumors was reported in several studies (7–10). Various malignant tumors may show high signal intensity on DWI reflecting hypercellularity and long T2 relaxation time of cancerous tissues (7). In this study we evaluated the capability of diffusion-weighted magnetic resonance imaging (MRI) with apparent diffusion coefficient (ADC) measurement to differentiate ovarian AF/CAF from ovarian cancers.
Material and Methods
Patients
The institutional review board approved this retrospective study, and informed consent was waived. We cross-referenced the database of the Department of Gynecology to identify all patients who had undergone MRI for the evaluation of gynecologic tumors between December 2005 and April 2011. Thirteen patients (mean age, 53 years, range, 27–77 years) with ovarian AF/CAF underwent MR examinations including DWI. Histological diagnosis was made by surgical resection. The 13 tumors included one benign mucinous adenofibroma, seven benign cystade-nofibromas (six serous, one mucinous), one borderline endometrioid adenofibroma, and four borderline cyst-adenofibromas (one serous, one clear cell, two mucinous). DWI of 27 consecutive surgically proven ovarian cancers exhibiting solid or complex solid and cystic masses in 25 women with a mean age of 56 years (range, 38–77 years) was also evaluated. The patients with ovarian cancers had undergone MRI including DWI between February 2005 and December 2008. The 27 ovarian cancers included four serous adenocarcinomas, one mucinous adenocarcinoma, 12 clear cell adenocarcinomas, six endometrioid adenocarcinomas, two poorly differentiated adenocarcinomas, and two mixed adenocarcinomas (endometrioid and mucinous adenocarcinomas).
MRI
Axial DWI was obtained in all patients with a spin-echo, single-shot echo planar sequence on a system with a 1.5T (Signa Excite HD; GE Healthcare, Milwaukee, WI, USA) or 3T (Signa 3T HD/HDx; GE Healthcare) superconducting MR imaging scanner using body coil transmission and an 8-channel phased-array receiver designed for cardiac studies. The parallel image-encoding techniques (array spatial sensitivity encoding technique; GE Healthcare) were used. The parameters of DWI were as follows: repetition time, 6000-6500 ms; echo time, 56.1-64 ms; b factors, 0 and 800 s/mm2; water excitation for fat signal suppression; matrix size, 128 × 192; field of view, 40 × 40 cm; 4 signals acquired; section thickness, 5-8 mm; and section gap, 0-2 mm. Among the 13 patients with ovarian AF/CAF and the 25 patients with ovarian cancers, nine and 11 patients, respectively, were evaluated at 3T, whereas four and 14 patients by 1.5T, respectively. Fast spin-echo T2-weighted images (repetition time, 4000-7000 ms; echo time, 99.3-100 ms) and fat saturated spin-echo T1-weighted images (repetition time, 600-700 ms; echo time, 7.9-9.6 ms) were obtained in all patients.
Analysis methods
The morphological appearance and internal architecture including the presence of black sponge-like appearance (BSLA), and signal intensity of the solid portion of the tumors on T2-weighted images were visually evaluated. The signal intensity was classified as low (similar to skeletal muscle) or intermediate (higher than skeletal muscle). Signal intensity of the solid portion of the tumors on DWI was visually evaluated and classified as high, intermediate, or low. Two reviewers, radiologists with 21 and 12 years of experience in gynecologic MR imaging, respectively, were blinded to histopathologic diagnosis of the ovarian tumors, and looked at all images together. Agreement between the reviewers was reached in consensus after careful individual evaluation. The mean ADC values (10–3 mm2/s) of the tumors were measured in a circular region of interest (ROI) in a representative region as large as possible within the tumor from the ADC maps on the workstation (AW4.2). The ROI was placed on the solid portion of the tumors so as not to contain cystic, hemorrhagic, or necrotic areas as much as possible by referring all MR images. However, ROI was placed within the solid portion containing tiny cysts in some AF/CAF with BSLA. In two CAF appearing as unilocular cysts with partial wall thickening, ROI was hard to place because the tiny solid areas were undetectable on the ADC maps. The Mann-Whitney U test was used to compare the mean ADC values among AF/CAF and ovarian cancers. A value of P < 0.05 was considered statistically significant. The ADC cut-off value to differentiate AF/ CAF from ovarian cancers was calculated, with its sensitivity, specificity, and positive (PPV) and negative predictive values (NPV).
Results
Both AFs appeared as single solid masses without BSLA. Six of 11 CAF appeared as multilocular, and 3/11 CAF as unilocular solid and cystic masses. The other two of the 11 CAFs appeared as solid masses containing cystic areas (Figs. 1–4). BSLA was observed in 4/13AF/CAF (31%). The solid portions of five AF/CAF showed totally low signal intensity (38%) (Fig. 1), while another five AF/CAFs showed a mixture of low to intermediate signal intensity (Figs. 2 and 4) on T2-weighted images. The solid portions of another two AF/CAFs showed intermediate signal intensity on T2-weighted images (Fig. 3). The other one CAF showed variable signal intensity on T2-weighted images as “stained-glass appearance”, and the solid portions, which were revealed on postcontrast T1-weighted images, were not clearly demonstrated on T2-weighted images. Four of 13 AF/CAF (31%) contained totally or partially intermediate signal intensity solid portions without BSLA on T2-weighted images. None of the 27 ovarian cancers had totally low signal intensity solid portions, or demonstrated BSLA on T2-weighted images.
A 71-year-old woman with right ovarian serous cystadenofibroma. (a) Axial T2-weighted fast spin-echo image shows right adnexal solid and cystic mass (arrow). Nodular solid components (arrowheads) show low signal intensity reflecting dense fibrous stroma. (b) The mean ADC value in the solid components was relatively low (1.10 × 10–3 mm2/s), however, the solid components (arrowheads) show low signal intensity on DWI probably as a consequence of the T2 blackout effect
A 77-year-old woman with left ovarian serous cystadenofibroma. (a) Axial T2-weighted fast spin-echo image shows left adnexal solid mass containing small cystic areas (arrow). The solid portion shows low to intermediate signal intensity. (b) The solid portion shows low signal intensity on DWI. The mean ADC value in the solid components was relatively high (1.43 × 10–3 mm2/s). Abundant epithelial glandular elements within solid fibrous portion are considered as the causes of signal increase on T2-weighted images and relatively high ADC value
A 68-year-old woman with right ovarian mucinous adenofibroma. (a) Axial T2-weighted fast spin-echo image shows right adnexal intermediate signal intensity solid mass (arrow). (b) The mass shows heterogeneous intermediate signal intensity on DWI (arrow). The mean ADC value in the mass was relatively high (1.36 × 10–3 mm2/s). Numerous small cysts filled with thick mucinous materials were observed throughout the mass on the cut surface of gross specimen. Thick mucinous materials are considered as the causes of signal decrease in small cystic spaces on T2-weighted images and slight signal increase on DWI
A 65-year-old woman with right ovarian borderline serous cystadeno-fibroma. (a) Axial T2-weighted fast spin-echo image shows right adnexal solid and cystic mass (arrow). The solid portions of the mass show low to intermediate signal intensity. (b) The mass shows inhomogeneous intermediate signal intensity on DWI (arrow). The mean ADC value in the mass was 1.28 × 10–3 mm2/s. Scattered relatively hypercellular areas are considered as the causes of signal increase on DWI
None of the 13 AF/CAF had high signal intensity on DWI (Figs. 1–4), whereas all 27 ovarian cancers showed homogeneous or heterogeneous high signal intensity on DWI (Fig. 5). All five borderline AF/CAF demonstrated heterogeneous intermediate signal intensity on DWI (Fig. 4). Two of eight benign AF/CAF (one mucinous adenofibroma and one mucinous cystadenofibroma) showed heterogeneous intermediate signal intensity on DWI (Fig. 3). Six of eight benign AF/CAF (six serous cystadenofibromas) had low signal intensity on DWI (Figs. 1 and 2). In two benign CAF appeared as unilocular cysts with partial wall thickening, the solid portions were too small to measure mean ADC values. The mean ADC values in solid portions of the other 11 AF/CAF (1.37 + /– 0.21; 95% confidence interval, 1.23–1.50) were significantly higher than those of the 27 ovarian cancers (1.01 + /– 0.15; 95% confidence interval, 0.95–1.07) (P < 0.001) (Fig. 6). Using a cut-off value of 1.20 for AF/CAF, a sensitivity of 82%, specificity of 93%, PPV of 82%, and NPV of 93% were found. The mean ADC values in six benign AF/CAF and those in five borderline AF/CAF were 1.43 + /– 0.25 and 1.30 + /– 0.15, respectively (P = 0.27).
A 53-year-old woman with right ovarian cancer (poorly differentiated adenocarcinoma). (a) Axial T2-weighted fast spin-echo image shows right adnexal solid mass (arrow). The mass shows intermediate signal intensity. (b) The mass shows high signal intensity on DWI (arrow). The mean ADC value in the mass was 0.81 × 10–3 mm2/s
Scatter plots of the ADC values obtained in AF/CAF (n = 11) and ovarian cancers (n = 27). The ADC values are significantly different (P < 0.001)
Discussion
Epithelial tumors of the ovaries are usually unilocular or multilocular cystic masses. The presence of solid components in the cystic masses increases the risk of malignancy (11–13). Ovarian AF/CAF are benign epithelial tumors and may appear as a solid mass or a cystic mass with solid nodular components mimicking ovarian cancers, and accurate preoperative diagnosis is required to avoid unnecessary surgery (1–3). AF/CAF belongs to the surface epithelial-stromal tumors of the ovary, and contains fibrous stroma in variable amounts in all subtypes such as serous, mucinous, endometrioid, and clear cell (1, 2). On T2-weighted images fibrous tissue has low signal intensity because of the T2 shortening effects of collagen and the decreased extracellular fluid compared with surrounding tissues (12). The solid components of AF/CAF may show characteristic “black sponge”-like appearance, or may appear as low intensity fibrous nodules or thickened cyst walls on T2-weighted images (3–6). However, abundant epithelial glandular elements, which are too small to be demonstrated as tiny cysts on imaging, may increase the signal intensity of the solid components on T2-weighted images, and the mass may mimic malignant tumor. In our series 4/13 AF/CAF (31%) had totally or partially intermediate signal intensity solid portions without BSLA on T2-weighted images, and the possibility of malignancy could not be excluded.
In our study all ovarian cancers showed high signal intensity on DWI, whereas no AF/CAF showed high signal intensity on DWI. DWI can provide excellent tissue contrast based on the molecular diffusion, and malignant tumors tend to show high signal intensity because increased cellularity of malignant tumors may restrict water diffusion and decrease the ADC value compared with benign lesions (7–10). Benign fibrous tumors may also show decreased ADC value owing to abundant collagen-producing fibro-blastic cells and a dense network of collagen fibers within the extracellular matrix, however, no signal increase is observed on DWI probably as a consequence of the T2 blackout effect (9, 10). In our series all five borderline AF/ CAF and two mucinous AF/CAF showed heterogeneous intermediate signal intensity on DWI. The causes of slight signal increase in these lesions are unclear, however, relative increased cellularity in borderline lesions, or water diffusion restriction in small cystic spaces due to thick mucinous contents may have roles. Our result and previous reports suggest that ovarian masses without high intensity solid components on DWI are considered as benign lesions (8–10). The contribution of the ADC measurement in differentiating benign and malignant ovarian tumors is controversial (8–10). Fujii et al. and Thomassin-Naggara et al. have reported that the ADC measurements in solid components did not contribute to differentiating benign from malignant ovarian lesions, whereas Takeuchi et al. reported that the ADC values in malignant ovarian tumors were significantly lower than those in benign lesions (8–10). This discrepancy is considered probably due to the pathologic variation in ovarian tumors such as the amounts of fibrous tissue, edematous change, or microscopic cystic degeneration, and a relatively small population of benign tumors in their studies (10). In our study the mean ADC values in AF/CAF were significantly higher than those in ovarian cancers. The presence of small or microscopic cystic spaces formed by epithelial elements within fibrous stroma of AF/CAF may contribute to their relative high ADC values, especially in evaluating heterogeneous lesions such as AF/CAF with BSLA. However, the relative small population, retrospective study design, and use of multiple field strength (1.5T and 3T) are limitations to our study. Further prospective studies in larger populations to support our results are needed. Especially, the number of patients with AF/CAF should be increased.
In conclusion, the benignity of ovarian AF/CAF may be suggested by lower signal intensity on DWI with higher ADC values compared with ovarian cancers. DWI with ADC measurement may be especially helpful in evaluating AF/CAF showing signal increase of solid components due to abundant epithelial glandular elements on T2-weighted images.