CME JOURNAL OF MAGNETIC RESONANCE IMAGING 00:00–00 (2013)
Technical Note
MRI Appearances of Mucinous Borderline Ovarian Tumors: Pathological Correlation Feng Hua Ma, MD,1 Shu Hui Zhao, MD,1 Jin Wei Qiang, MD, PhD,1* Guo Fu Zhang, MD, PhD,2y Xue Zhen Wang, MD,2 and Li Wang, MD3 MUCINOUS TUMOR IS a common epithelial tumor of the ovary, which can be classified as benign, borderline, and malignant tumors according to their pathological features. Mucinous borderline ovarian tumors (MBOT) account for 17% of mucinous ovarian tumors and 32% of all borderline ovarian tumors respectively (1). The first comprehensive clinicopathologic report on MBOT appeared in 1973 (2). MBOT are characterized as tumors with low malignant potential, mitotic activity and nuclear atypia, and without obvious invasion of the stroma (2). However, MBOT may present with penitoneal implants, lymph node metastases, and a high recurrence rate after the resection (1,3). As they tend to occur in young women of childbearing age and have a good prognosis, the current management guidelines for MBOT are different from that of benign or malignant tumors, with a trend toward fertility-sparing surgery. Gynecologists often determine the surgical approach based on the results of intraoperative frozen section diagnosis; nevertheless, because of their often large size and internal heterogeneity, the accuracy of frozen section diagnosis is only 60–75% and MBOT are more likely to be misdiagnosed (4–7). Therefore, the accurate characterization of MBOT with the preoperative imaging would be extremely helpful in surgical planning. Studies have suggested that MRI can detect and characterize ovarian morphological features and is a useful tool for distinguishing benign from malignant ovarian tumors (8–11). Recently, the novel techniques such as perfusion- and diffusion-weighted imaging (PWI and DWI) have achieved the promising results in the characterization of ovarian masses (12,13). However, little data are available on the imaging feature of MBOT (6,9–11). This study describes the MRI morphological features of MBOT and correlated them with pathological findings to improve the accuracy of preoperative diagnosis.
Purpose: The purpose of this study is to evaluate the MRI features of mucinous borderline ovarian tumors (MBOT). Materials and Methods: MRI morphology of 30 MBOT proven MBOT by surgery and pathology was retrospectively studied and correlated with the histopathological findings. On MRI, tumors were classified into three morphological categories: (i) unilocular cyst in five (17%) tumors. (ii) multilocular cyst in 23 (76%) tumors. (iii) solid mass in 2 (7%) tumors. MRI features of tumors were identified including the multilocularity (23/30, 77%), honeycomb loculi (15/30, 50%), signal discrepancy (different signal intensity on T1WI and T2WI) (19/30, 63%), thickened wall or septa (>3 mm) (16/30, 53%). Results: Intestinal type and endocervical type of MBOT, two distinctly histologic subtypes, were found in 20 (67%) and 10 (33%) tumors respectively. There were a higher prevalence of multilocularity (P ¼ 0.026), honeycomb loculi (P ¼ 0.025), and signal discrepancy (P ¼ 0.024) in intestinal type than endocervical type of MBOT. Conclusion: Typical MRI features of MBOT are large multilocular tumors with honeycomb loculi, heterogeneous signal intensity of the loculi, and thickened wall or septa. Key Words: ovary; mucinous tumors; borderline; magnetic resonance imaging J. Magn. Reson. Imaging 2013;00:000–000. C 2013 Wiley Periodicals, Inc. V
1 Department of Radiology, Jinshan Hospital, Shanghai Medical College, Fudan University, Shanghai, China. 2 Department of Radiology, Obsterics and Gynecology Hospital, Shanghai Medical College, Fudan University, Shanghai, China. 3 Department of Pathology, Jinshan Hospital, Shanghai Medical College, Fudan University, Shanghai, China. Contract grant sponsor: Shanghai Municipal Science and Technology Commission; Contract grant number: 124119a3300; Contract grant sponsor: Shanghai Municipal Health Bureau; Contract grant number: 2008-196. *Address reprint requests to: J.W.Q., Department of Radiology, Jinshan Hospital, Shanghai Medical College, Fudan University, 1508 Longhang Road, Jinshan District, Shanghai 201508, China. E-mail: [email protected] y Co-corresponding atuhor: G.F.Z., Department of Radiology, Obsterics and Gynecology Hospital, Shanghai Medical College, Fudan University, 419 Fangxie Road, Huangpu District, Shanghai 200011, China. E-mail: [email protected] Received June 14, 2013; Accepted August 21, 2013. DOI 10.1002/jmri.24408 View this article online at wileyonlinelibrary.com. C 2013 Wiley Periodicals, Inc. V
MATERIALS AND METHODS Subjects The institutional review boards of our hospitals approved this retrospective study and informed consent was obtained from each patient for academic use 1
2
of their clinical data. Patients with the suspected ovarian tumors (by gynecologic examination, biomarkers, ultrasound, or computed tomography) were enrolled in a prospective MRI study project of ovarian tumor from February 2008 to November 2012 at Jinshan Hospital and Obstetrics and Gynaecology Hospital of Fudan University. Among 530 cases proven by surgery and histology, we found a total of 28 patients with MBOT (8 cases from J.S.H and 20 cases from O.G.H) and retrospectively studied their MRI features. The age of the patients ranged from 16 to 75 years (mean age, 39 years). Twenty-one (75%) of them were premenopausal and 7 (25%) postmenopausal. All patients underwent surgery (laparoscopy in 11 cases, laparotomy in 17 cases) within the 1 week after completing the MRI scan. MRI Scanning MRI was performed with a 1.5 Tesla (T) MR unit (Symphony or Avanto, Siemens, Erlangen, Germany) with a phased-array pelvic coil. The patient lay in the supine position and breathed freely. The following sequences were obtained: axial T1weighted spin-echo imaging (T1WI) (repetition time/ echo time [TR/TE] ¼ 340/10 ms); T1WI flash twodimensional (2D) with fat saturation (TR/TE ¼196/ 2.9 ms); T2-weighted turbo spin-echo imaging (T2WI) with and without fat saturation (TR/TE ¼ 8000/83 ms and TR/TE ¼ 4000/98 ms), and sagittal and coronal T2WI turbo spin-echo (TR/TE ¼ 8000/98 ms). The contrast-enhanced T1WI flash 2D with fat saturation (TR/TE ¼ 196/2.9 ms) was performed in the axial, sagittal and coronal planes after the intravenous administration of 0.1 mmol.kg1 Gadopentetate dimeglumine (Gd-DTPA, Magnevist; Bayer Schering, Guangzhou, China) injected at a rate of 2–3 mL.s1. The scanning parameters were as follows: slice thickness ¼ 5 mm; gap ¼ 1.5 mm; matrix ¼ 256 256; field of view ¼ 20–25 cm 34 cm; excitations ¼ 4. The scanning range was from the inferior public symphysis to the renal hilum and was extended beyond the dome of tumor in the cases with huge masses. Image Analysis The MR images were reviewed independently by two radiologists (F.H.M. and J.W.Q.), with 10 and 25 years of experience in gynecological imaging. Discrepancies were resolved in consensus. The following potential MR features (i) the laterality, shape, and size; (ii) the presence and number of loculus; (iii) the thickness of the septa and wall; (iv) the presence and types of solid component (mural nodule, papillary projection, solid area, or mass); (v) the signal intensity of cystic and solid components; (vi) the enhancement pattern; and (vii) the presence of pelvic ascites and associated ovarian lesions, were assessed. Tumors were classified as purely cystic, predominantly cystic, mixed cystic-solid, or solid. Cystic tumors were classified as unilocular and multilocular (2 loculi). Signal discrepancy was defined as the different signal intensity between the adjacent loculus on both T1WI and T2WI.
Ma et al.
The thickness of wall or septa greater than 3 mm was defined as thickening. The size of tumor, wall, septa and solid component were measured on the contrastenhanced T1WI. Histopathological Examination The histopathological evaluations were performed by a pathologist who specialized in gynecological pathology. A radiologist who specialized in gynecological imaging worked together with the pathologist to evaluate the specimen grossly. After determining the shape and size of the tumor, the fresh specimen was dissected into 1- to 5-cm sections. Grossly, the tumor was assessed for cystic, solid or mixed cystic-solid; the presence of loculus; the nature of the intracystic content, including the color and the viscosity; the thickness of the wall and septa, and the size, shape, architecture, and distribution of the solid component. Microscopically, the tumor was classified into one of two distinct subtypes according its histologic characteristic: intestinal type and endocervical type of MBOT (I-MBOT and E-MBOT). The staging of tumor was performed according to 2002 International Federation of Gynecology and Obstetrics (FIGO). Statistical Analysis Statistical analysis was performed with SPSS 17.0 for Windows (SPSS Inc., Chicago, IL). Continuous parametric variables were compared using Student’s t-test between I-MBOT and E-MBOT, and categorical variables were compared using the Fisher’s exact test. All tests were two-sided. A P-value less than 0.05 were considered statistically significant. RESULTS Clinical and Pathological Findings A total of 30 tumors in 28 patients were found. The tumor size ranged from 4.2 cm to 28.3 cm (mean, 18.3 6 5.6 cm). There were bilateral tumors in two patients and unilateral tumors in 26 patients. Histologically, 20 (67%) tumors were I-MBOT and 10 (33%) tumors were E-MBOT. There were five associated lesions, including three endometriomas, one serous cystadenoma, and one mature teratoma. All cases were at stage Ia (FIGO). MRI Features According to the presence and number of septa or solid component, three MRI morphological patterns were identified. (i) Unilocular cyst: 5 (17%) tumors were unilocular with distinct border and oval shape. The cystic component was pathologically confirmed clear mucus, and its signal intensity was similar to that of urine on both T1WI and T2WI. The cystic wall was thin and uniform with a thickness less than 3 mm. There was one or more papillary projection from the cyst wall, varied from 0.6 cm to 2.5 cm and demonstrated enhancement after the administration of
MRI Appearances and Pathology of Ovarian Tumors
3
Figure 1. A 39-year-old woman with a MBOT. The tumor shows a unilocular cyst with a large intracystic papillary projection (arrow). The signal intensity of the cystic component is homogenously urine-like. The papillary projection is moderate signal intensity on axial T2-weighted fat-saturated image (a) and avidly enhanced on contrast-enhanced T1-weighted fat-saturated image (b).
Gd-DTPA (Fig. 1). (ii) Multilocular cyst: 23 (76%) tumors appeared this pattern consisting of multilocular (over 7 loculi) cystic mass. Honeycomb loculi, which composed of grouped innumerable little loculi (3 mm), which enhanced moderately or vigorously, was seen in 16 (53%) tumors (Fig. 4). A solid component was found in 5 (17%) tumors with the largest diameter ranged from 1.5 cm to 4.5 cm and moderate enhancement. Gross pathological sections showed that the innumerable tiny loculi were found within the solid component. (iii) Solid tumor: 2 of 30 tumors (7%) appeared as a solid
Figure 2. A 33-year-old woman with a MBOT. The tumor shows a multilocular cyst with honeycomb loculi (arrows) which demonstrates a mix of low, moderate and high signal intensities on T2-weighted fat-saturated image (a). The low signal on T2-weighted fat-saturated image appears as a solid area (arrowheads) on contrast-enhanced image (b) and a fine honeycomb structure in the dissected specimen. Histology reveals numerous loculi (yellow arrowheads) in this intestinal type of MBOT (c) (H&E 100).
4
Ma et al.
Figure 3. A 26-year-old woman with a MBOT. Axial T1-weighted image (a), T2-weighted fat-saturated image (b), and coronal T2-weighted image (c) demonstrate a multilocular cyst with various signal intensities between adjacent loculi. The loculi show no obvious enhancement on contrast-enhanced T1-weighted fat-saturated image (d).
lobulated mass with hypointense on T1WI and a slightly hyperintense on T2WI. The tumors were enhanced moderately and heterogeneously, and pathologically compact papillary architecture (Fig. 5). The MRI features of the subtypes of MBOT are summarized in Table 1. There was a significant difference
between I-MBOT and E-MBOT in the maximum diameter (P ¼ 0.028), with the I-MBOT tending to be larger. Multilocularity was present in 23 of 30 (76%) MBOT, a higher incidence being found in the I-MBOT (18/20, 90%) than in E-MBOT (5/10, 50%) (P ¼ 0.026). Honeycomb loculi were present in 15 of 30 (50%), a higher
Figure 4. A 43-year-old woman with a MBOT. Sagittal T2-weighted fat-saturated image (a) and contrast-enhanced T1weighted fat-saturated image (b) demonstrate a unilocular cyst with an incomplete septa, thickened wall (arrow) and avidly enhanced intracystic multiple papillary projections.
MRI Appearances and Pathology of Ovarian Tumors
5
Figure 5. A 60-year-old woman with a MBOT. The tumor appears as a solid lobulated mass which is heterogeneously hyperintense on T2-weighted image (a) and moderately enhanced on contrast-enhanced T1-weighted fat-saturated image (b).
incidence in the I-MBOT (13/20, 63%) than in the E-MBOT (2/10, 20%) (P ¼ 0.025). Signal discrepancy was present in 18 of 30 (60%) MBOT, a higher incidence in I-MBOT (15/20, 75%) than in E-MBOT (3/ 10, 30%) (P ¼ 0.024). Correlation Between Pathology and MRI Findings Gross pathology confirmed MRI findings and four patterns of mucus were noted: (i) clear mucus, which was similar to signal intensity of urine, was found in 17 tumors; (ii) yellowish and viscous mucus in 8 tumors, seen frequently in the larger louculi, appeared a moderately intense on T1WI and a hyperintense on T2WI; (iii) clear jelly-like mucus in 8 tumors, seen commonly in the smaller loculi, was hypointense on T1WI and moderately intense on T2WI; (iv) white colloid in 9 tumors, seen in the honeycomb loculi, was hyperintense on T1WI and hypointense on T2WI. DISCUSSION MBOT account for 30–50% of BOT and happen in patients ranging in age from 9 to 88 years, but almost half of women were young ranging from 35 to 47 years (1,14,15). MBOT present at earlier stages than invasive tumors, with approximately 90% at stage I. In our series, the mean age of patients was 39 years, and all cases were at stage Ia. Owing to a favorable prognosis of MBOT, the conservative surgery of unilateral salpingo-oophorectomy is often performed to
preserve patient’s fertility (4,7,16). However, a consequently high recurrence rate calls for the need of complete surgical staging, which consists of total abdominal hysterectomy and bilateral salpingooophorectomy, cytological analyses of the ascites or the peritoneal lavage fluid, a multiple peritoneal biopsy, and an appendectomy for I-MBOT (4,7,16). Although studies have suggested that MRI is useful in characterizing complex adnexal masses (8–11) and quantitative multiparametric MRI (PWI, DWI,T2 map, etc.) have been confirmed having value in distinguishing malignant from benign ovarian tumors (12,13), the comprehensive MRI features specific to MBOT have not been studied.. Bent et al reviewed the MRI appearances of BOT and classified them into four morphological categories: unilocular cyst (19%); minimally septate cyst with papillary projections (19%); markedly septate lesion with plaque-like excrescences (45%) and predominally solid with exophytic papillary projections (16%) (10). Comparing with serous BOT, Bazot et al recently reported that MBOT were more frequently multilocular, with loculi of different signal intensity, and with a high number of septa (>10), which were commonly grouped, irregular, and thickened (11). In our MBOT series, three MRI morphological categories were identified: unilocular cyst with papillary projections (17%); multilocular cyst (76%); solid tumor (7%). In accordance with previous studies, multilocular cystic mass was the commonest finding (3,10,11). MBOT have been separated into two distinct subtypes: I- and E-MBOT. They differ in their
Table 1 MRI Features of Mucinous Borderline Ovarian Tumors MRI features Tumor diameter Multilocularity Signal discrepancy Thickened septa or wall Honeycomb loculi Projection Solid portion
Total (n ¼ 30)
I-MBOT (n ¼ 20)
E-MBOT (n ¼ 10)
P Valuey
16.5 6 6.5 23 (76%) 18 (60%) 16 (53%) 15 (50%) 10 (33%) 17% (5)
18.3 6 5.6 18 (90%) 15 (75%) 11 (55%) 13 (65%) 5 (25%) 2 (10%)
12.8 6 7.0 5 (50%) 3 (30%) 5 (50%) 2 (20%) 5 (50%) 3 (30%)
0.028 0.026 0.024 0.550 0.025 0.231 0.300
*Except for tumour size (Student’s), all P values were calculated using Fisher’s exact test.
6
clinicopathologic features, immunophenotype, clinical behavior, and surgery approach (14,17,18). Up to date, however, we have found no study concerning the imaging features of the two distinct subtypes. I-MBOT account for approximately 85% and the remaining 15% are E-MBOT. In our series, less I-MBOT (67%) were found, happen in older patients and have a poorer prognosis than E-MBOT (1,14,15). Multilocularity was more frequent in I-MBOT (90%) than in EMBOT (50%). Honeycomb loculi, which were composed of grouped innumerable little loculi, were seen in 50% of MBOT and more frequently in I-MBOT (65%) than in E-MBOT (20%). Signal discrepancy was another distinguished aspect of MBOT which was more common in I-MBOT (75%) than in E-MBOT (30%). It is also known as “stain-glass appearance”, which is defined as cystic masses in which the loculi show variable signal intensity on both T1WI and T2WI, and results from the viscosity of contents, blood product, or debris (19). However, our study suggests that the different patterns of mucus may be the underlying cause for these signal changes. I-MBOT are histopathologically heterogeneous, which often simultaneously contain areas of cystadenoma and intraepithelial carcinoma (1,14,15). The tumor is typically composed of numerous dilated glands and cysts of various sizes with internal epithelial tufts and papillations lined by a typical intestinaltype mucinous epithelium. These glands and cysts often produce different types of mucus, and consequently forming a complex cystic pattern (1,2,14). In our series, the most characteristic signal discrepancy was hyperintense on T1WI and hypointense on T2WI, which was correlate with pathological colloid and often seen in honeycomb loculi. E-MBOT, also known as seromucinous or M€ ullerian BOT, contain only cells resembling endocervical glands (1,14,15). They are clinically and pathologically closely related to serous BOT, and have little in common with I-MBOT (1,14,15). Because of the 40% rate of synchronous bilaterality and the 17% risk of leaving behind a borderline tumor in a grossly normal ovary, complete surgical staging is suggested (14). The characteristic feature of E-MBOT is unilocular or paucilocular cyst with multiple papillary projections, an appearance resembling that of serous BOT (1,3,14). In our series, all two cases of bilateral tumors are E-MBOT and 40% of E-MBOT are unilocular cyst. A solid mass with a branching papillary architecture and internal hypointense branching on T2WI is highly suggestive of serous BOT (10,11). In the present study, however, one case of E-MBOT had this similar appearance. MBOT are thought to represent an intermediate stage in the stepwise progression from benign to malignant tumors (1,14). Compared with MBOT, the mucinous cystadenomas tend to have less loculi, more uniform wall and septa, less frequent signal discrepancies and absent papillae (20). The primary ovarian mucinous carcinomas are uncommon, representing approximately 5% of ovarian surface epithelial carcinomas (2,9,14). The invasive carcinomas are usually mixed cystic-solid or multiloculated cystic
Ma et al.
with significantly greater and vigorously enhanced solid elements than MBOT (9,12). In addition, the peritoneal implants, metastatic retroperitoneal lymph nodes and large amount of ascities are more often seen in invasive tumors. There were some limitations in our study. First, the relative smaller sample size was in sufficient to subdivide the MRI features of MBOT. Further study increased sample size is warranted. Second, a selective bias was inevitably present owing to the retrospective nature of the study. Some small ovarian tumors characterized by US did not enroll in MRI study. Third, the potential value of functional MRI will be studied further. In conclusion, the present study demonstrated that MBOT typically presented as a multilocular cystic mass with honeycomb loculi, signal discrepancy, and irregular thickened septa or wall. Other MBOT appeared as uni- or paucilocular or cyst with multiple papillary projections and solid mass. REFERENCES 1. Hart WR. Mucinous tumours of the ovary: a review. Int J Gynecol Pathol 2005;24:4–25. 2. Hart WR, Norris HJ. Borderline and malignant mucinous tumours of the ovary: histologic criteria and clinical behavior. Cancer 1973;31:1031–1045. 3. Lalwani N, Shanbhogue AKP, Vikram R, Nagar A, Jagirdar J, Prasad SR. Current update on borderline ovarian neoplasms. AJR Am J Roentgenol 2010;194:330–336. 4. Houck K, Nikrui N, Duska L, et al. Borderline tumours of the ovary: correlation of frozen and permanent histopathologic diagnosis. Obstet Gynecol 2000;95:839–843. 5. Song T, Choi CH, Kim HJ, et al. Accuracy of frozen section diagnosis of borderline ovarian tumours. Gynecol Oncol 2011;122: 127–131. 6. Bazot M, Nassar-Slaba J, Thomassin-Naggara I, Cortez A, Uzan S, Darai E. MR imaging compared with intraoperative frozensection examination for the diagnosis of adnexal tumours: correlation with final histology. Eur Radiol 2006;16:2687–2699. 7. Park JY, Kim DY, Kim JH, Kim YM, Kim YT, Nam JH. Surgical management of borderline ovarian tumours: the role of fertilitysparing surgery. Gynecol Oncol 2009;113:75–82. 8. Vargas HA, Barrett T, Sala E. MRI of ovarian masses. J Magn Reson Imaging 2013;37:265–281. 9. deSouza NM, O’Neill R, Mcclndoe GA, Dina R, Soutter WP. Borderline tumours of the ovary: CT amd MRI features and tumour markers in differentiation from stage I disease. AJR Am J Roentgenol 2005;184:999–1003. 10. Bent CI, Sahdev A, Rockall AG, Singh N, Sohaib SA, Reznek RH. MR appearances of borderline ovarian tumours. Clin Radiol 2009;64:430–438. 11. Bazot M, Haouy D, Dara€ı E, Cortez A, Dechoux-Vodovar S, Thomassin-Naggara I. Is MRI a useful tool to distinguish between serous and mucinous borderline ovarian tumours? Clin Radiol 2013;68:e1–e8. 12. Thomassin-Naggara I, Toussaint I, Perror N, et al. Characterization of complex adnexal masses: value of adding perfusion- and diffusion-weighted MR imaging to conventional MR imaging. Radiology 2011,258:793–803. 13. Carter JS, Koopmeiners JS, Kuehn-Hajder JE, et al. Quantitative multiparametric MRI of ovarian cancer. J Magn Reson Imaging 2013 doi:0.1002/jmri.24119. [Epub ahead of print]. 14. Acs G. Serous and mucinous borederline (low malignant potential tumours of the ovary. Am J Clin Pathol 2005;123(Suppl 1):s13– s57. 15. Hart WR. Borderline epithelial tumours of the ovary. Mod Pathol 2005;18:s33–s50. 16. Anfinan N, Sait K, Ghatage P, Nation J, Chu P. Ten years experience in the management of borderline ovarian tumours at
MRI Appearances and Pathology of Ovarian Tumors Tom Baker Cancer Centre. Arch Gynecol Obstet 2011;284:731– 735. 17. Rutgers JL, Scully RE. Ovarian mullerian mucionous papillary cystadenomas of borederline malignancy: aclinicopathologic analysis. Cancer 1988;61:340–348. 18. Rodriguez IM, Irving JA, Prat J. Endocervical-like mucinous borderline tumours of the ovary: a clinicpathologic analysis of 31 cases. Am J Surg Pathol 2004;28:1311–1318.
7 19. Tanaka YO, Nishida M, Kurosaki Y, Itai Y, Tsunoda H, Kubo T. Differential diagnosis of gynecological “stained glass” tumours on MRI. Br J Radiol 1999;72:414–420. 20. Okamoto Y, Tanaka YO, Tsunoda H, Yoshikawa H, Minami M. Malignant or borderline mucinous cystic neoplasms have a larger number of loculi than mucinous cystadenoma: a retrospective study with MR. J Magn Reson Imaging 2007;26:94– 99.
Technical Note
MRI Appearances of Mucinous Borderline Ovarian Tumors: Pathological Correlation Feng Hua Ma, MD,1 Shu Hui Zhao, MD,1 Jin Wei Qiang, MD, PhD,1* Guo Fu Zhang, MD, PhD,2y Xue Zhen Wang, MD,2 and Li Wang, MD3 MUCINOUS TUMOR IS a common epithelial tumor of the ovary, which can be classified as benign, borderline, and malignant tumors according to their pathological features. Mucinous borderline ovarian tumors (MBOT) account for 17% of mucinous ovarian tumors and 32% of all borderline ovarian tumors respectively (1). The first comprehensive clinicopathologic report on MBOT appeared in 1973 (2). MBOT are characterized as tumors with low malignant potential, mitotic activity and nuclear atypia, and without obvious invasion of the stroma (2). However, MBOT may present with penitoneal implants, lymph node metastases, and a high recurrence rate after the resection (1,3). As they tend to occur in young women of childbearing age and have a good prognosis, the current management guidelines for MBOT are different from that of benign or malignant tumors, with a trend toward fertility-sparing surgery. Gynecologists often determine the surgical approach based on the results of intraoperative frozen section diagnosis; nevertheless, because of their often large size and internal heterogeneity, the accuracy of frozen section diagnosis is only 60–75% and MBOT are more likely to be misdiagnosed (4–7). Therefore, the accurate characterization of MBOT with the preoperative imaging would be extremely helpful in surgical planning. Studies have suggested that MRI can detect and characterize ovarian morphological features and is a useful tool for distinguishing benign from malignant ovarian tumors (8–11). Recently, the novel techniques such as perfusion- and diffusion-weighted imaging (PWI and DWI) have achieved the promising results in the characterization of ovarian masses (12,13). However, little data are available on the imaging feature of MBOT (6,9–11). This study describes the MRI morphological features of MBOT and correlated them with pathological findings to improve the accuracy of preoperative diagnosis.
Purpose: The purpose of this study is to evaluate the MRI features of mucinous borderline ovarian tumors (MBOT). Materials and Methods: MRI morphology of 30 MBOT proven MBOT by surgery and pathology was retrospectively studied and correlated with the histopathological findings. On MRI, tumors were classified into three morphological categories: (i) unilocular cyst in five (17%) tumors. (ii) multilocular cyst in 23 (76%) tumors. (iii) solid mass in 2 (7%) tumors. MRI features of tumors were identified including the multilocularity (23/30, 77%), honeycomb loculi (15/30, 50%), signal discrepancy (different signal intensity on T1WI and T2WI) (19/30, 63%), thickened wall or septa (>3 mm) (16/30, 53%). Results: Intestinal type and endocervical type of MBOT, two distinctly histologic subtypes, were found in 20 (67%) and 10 (33%) tumors respectively. There were a higher prevalence of multilocularity (P ¼ 0.026), honeycomb loculi (P ¼ 0.025), and signal discrepancy (P ¼ 0.024) in intestinal type than endocervical type of MBOT. Conclusion: Typical MRI features of MBOT are large multilocular tumors with honeycomb loculi, heterogeneous signal intensity of the loculi, and thickened wall or septa. Key Words: ovary; mucinous tumors; borderline; magnetic resonance imaging J. Magn. Reson. Imaging 2013;00:000–000. C 2013 Wiley Periodicals, Inc. V
1 Department of Radiology, Jinshan Hospital, Shanghai Medical College, Fudan University, Shanghai, China. 2 Department of Radiology, Obsterics and Gynecology Hospital, Shanghai Medical College, Fudan University, Shanghai, China. 3 Department of Pathology, Jinshan Hospital, Shanghai Medical College, Fudan University, Shanghai, China. Contract grant sponsor: Shanghai Municipal Science and Technology Commission; Contract grant number: 124119a3300; Contract grant sponsor: Shanghai Municipal Health Bureau; Contract grant number: 2008-196. *Address reprint requests to: J.W.Q., Department of Radiology, Jinshan Hospital, Shanghai Medical College, Fudan University, 1508 Longhang Road, Jinshan District, Shanghai 201508, China. E-mail: [email protected] y Co-corresponding atuhor: G.F.Z., Department of Radiology, Obsterics and Gynecology Hospital, Shanghai Medical College, Fudan University, 419 Fangxie Road, Huangpu District, Shanghai 200011, China. E-mail: [email protected] Received June 14, 2013; Accepted August 21, 2013. DOI 10.1002/jmri.24408 View this article online at wileyonlinelibrary.com. C 2013 Wiley Periodicals, Inc. V
MATERIALS AND METHODS Subjects The institutional review boards of our hospitals approved this retrospective study and informed consent was obtained from each patient for academic use 1
2
of their clinical data. Patients with the suspected ovarian tumors (by gynecologic examination, biomarkers, ultrasound, or computed tomography) were enrolled in a prospective MRI study project of ovarian tumor from February 2008 to November 2012 at Jinshan Hospital and Obstetrics and Gynaecology Hospital of Fudan University. Among 530 cases proven by surgery and histology, we found a total of 28 patients with MBOT (8 cases from J.S.H and 20 cases from O.G.H) and retrospectively studied their MRI features. The age of the patients ranged from 16 to 75 years (mean age, 39 years). Twenty-one (75%) of them were premenopausal and 7 (25%) postmenopausal. All patients underwent surgery (laparoscopy in 11 cases, laparotomy in 17 cases) within the 1 week after completing the MRI scan. MRI Scanning MRI was performed with a 1.5 Tesla (T) MR unit (Symphony or Avanto, Siemens, Erlangen, Germany) with a phased-array pelvic coil. The patient lay in the supine position and breathed freely. The following sequences were obtained: axial T1weighted spin-echo imaging (T1WI) (repetition time/ echo time [TR/TE] ¼ 340/10 ms); T1WI flash twodimensional (2D) with fat saturation (TR/TE ¼196/ 2.9 ms); T2-weighted turbo spin-echo imaging (T2WI) with and without fat saturation (TR/TE ¼ 8000/83 ms and TR/TE ¼ 4000/98 ms), and sagittal and coronal T2WI turbo spin-echo (TR/TE ¼ 8000/98 ms). The contrast-enhanced T1WI flash 2D with fat saturation (TR/TE ¼ 196/2.9 ms) was performed in the axial, sagittal and coronal planes after the intravenous administration of 0.1 mmol.kg1 Gadopentetate dimeglumine (Gd-DTPA, Magnevist; Bayer Schering, Guangzhou, China) injected at a rate of 2–3 mL.s1. The scanning parameters were as follows: slice thickness ¼ 5 mm; gap ¼ 1.5 mm; matrix ¼ 256 256; field of view ¼ 20–25 cm 34 cm; excitations ¼ 4. The scanning range was from the inferior public symphysis to the renal hilum and was extended beyond the dome of tumor in the cases with huge masses. Image Analysis The MR images were reviewed independently by two radiologists (F.H.M. and J.W.Q.), with 10 and 25 years of experience in gynecological imaging. Discrepancies were resolved in consensus. The following potential MR features (i) the laterality, shape, and size; (ii) the presence and number of loculus; (iii) the thickness of the septa and wall; (iv) the presence and types of solid component (mural nodule, papillary projection, solid area, or mass); (v) the signal intensity of cystic and solid components; (vi) the enhancement pattern; and (vii) the presence of pelvic ascites and associated ovarian lesions, were assessed. Tumors were classified as purely cystic, predominantly cystic, mixed cystic-solid, or solid. Cystic tumors were classified as unilocular and multilocular (2 loculi). Signal discrepancy was defined as the different signal intensity between the adjacent loculus on both T1WI and T2WI.
Ma et al.
The thickness of wall or septa greater than 3 mm was defined as thickening. The size of tumor, wall, septa and solid component were measured on the contrastenhanced T1WI. Histopathological Examination The histopathological evaluations were performed by a pathologist who specialized in gynecological pathology. A radiologist who specialized in gynecological imaging worked together with the pathologist to evaluate the specimen grossly. After determining the shape and size of the tumor, the fresh specimen was dissected into 1- to 5-cm sections. Grossly, the tumor was assessed for cystic, solid or mixed cystic-solid; the presence of loculus; the nature of the intracystic content, including the color and the viscosity; the thickness of the wall and septa, and the size, shape, architecture, and distribution of the solid component. Microscopically, the tumor was classified into one of two distinct subtypes according its histologic characteristic: intestinal type and endocervical type of MBOT (I-MBOT and E-MBOT). The staging of tumor was performed according to 2002 International Federation of Gynecology and Obstetrics (FIGO). Statistical Analysis Statistical analysis was performed with SPSS 17.0 for Windows (SPSS Inc., Chicago, IL). Continuous parametric variables were compared using Student’s t-test between I-MBOT and E-MBOT, and categorical variables were compared using the Fisher’s exact test. All tests were two-sided. A P-value less than 0.05 were considered statistically significant. RESULTS Clinical and Pathological Findings A total of 30 tumors in 28 patients were found. The tumor size ranged from 4.2 cm to 28.3 cm (mean, 18.3 6 5.6 cm). There were bilateral tumors in two patients and unilateral tumors in 26 patients. Histologically, 20 (67%) tumors were I-MBOT and 10 (33%) tumors were E-MBOT. There were five associated lesions, including three endometriomas, one serous cystadenoma, and one mature teratoma. All cases were at stage Ia (FIGO). MRI Features According to the presence and number of septa or solid component, three MRI morphological patterns were identified. (i) Unilocular cyst: 5 (17%) tumors were unilocular with distinct border and oval shape. The cystic component was pathologically confirmed clear mucus, and its signal intensity was similar to that of urine on both T1WI and T2WI. The cystic wall was thin and uniform with a thickness less than 3 mm. There was one or more papillary projection from the cyst wall, varied from 0.6 cm to 2.5 cm and demonstrated enhancement after the administration of
MRI Appearances and Pathology of Ovarian Tumors
3
Figure 1. A 39-year-old woman with a MBOT. The tumor shows a unilocular cyst with a large intracystic papillary projection (arrow). The signal intensity of the cystic component is homogenously urine-like. The papillary projection is moderate signal intensity on axial T2-weighted fat-saturated image (a) and avidly enhanced on contrast-enhanced T1-weighted fat-saturated image (b).
Gd-DTPA (Fig. 1). (ii) Multilocular cyst: 23 (76%) tumors appeared this pattern consisting of multilocular (over 7 loculi) cystic mass. Honeycomb loculi, which composed of grouped innumerable little loculi (3 mm), which enhanced moderately or vigorously, was seen in 16 (53%) tumors (Fig. 4). A solid component was found in 5 (17%) tumors with the largest diameter ranged from 1.5 cm to 4.5 cm and moderate enhancement. Gross pathological sections showed that the innumerable tiny loculi were found within the solid component. (iii) Solid tumor: 2 of 30 tumors (7%) appeared as a solid
Figure 2. A 33-year-old woman with a MBOT. The tumor shows a multilocular cyst with honeycomb loculi (arrows) which demonstrates a mix of low, moderate and high signal intensities on T2-weighted fat-saturated image (a). The low signal on T2-weighted fat-saturated image appears as a solid area (arrowheads) on contrast-enhanced image (b) and a fine honeycomb structure in the dissected specimen. Histology reveals numerous loculi (yellow arrowheads) in this intestinal type of MBOT (c) (H&E 100).
4
Ma et al.
Figure 3. A 26-year-old woman with a MBOT. Axial T1-weighted image (a), T2-weighted fat-saturated image (b), and coronal T2-weighted image (c) demonstrate a multilocular cyst with various signal intensities between adjacent loculi. The loculi show no obvious enhancement on contrast-enhanced T1-weighted fat-saturated image (d).
lobulated mass with hypointense on T1WI and a slightly hyperintense on T2WI. The tumors were enhanced moderately and heterogeneously, and pathologically compact papillary architecture (Fig. 5). The MRI features of the subtypes of MBOT are summarized in Table 1. There was a significant difference
between I-MBOT and E-MBOT in the maximum diameter (P ¼ 0.028), with the I-MBOT tending to be larger. Multilocularity was present in 23 of 30 (76%) MBOT, a higher incidence being found in the I-MBOT (18/20, 90%) than in E-MBOT (5/10, 50%) (P ¼ 0.026). Honeycomb loculi were present in 15 of 30 (50%), a higher
Figure 4. A 43-year-old woman with a MBOT. Sagittal T2-weighted fat-saturated image (a) and contrast-enhanced T1weighted fat-saturated image (b) demonstrate a unilocular cyst with an incomplete septa, thickened wall (arrow) and avidly enhanced intracystic multiple papillary projections.
MRI Appearances and Pathology of Ovarian Tumors
5
Figure 5. A 60-year-old woman with a MBOT. The tumor appears as a solid lobulated mass which is heterogeneously hyperintense on T2-weighted image (a) and moderately enhanced on contrast-enhanced T1-weighted fat-saturated image (b).
incidence in the I-MBOT (13/20, 63%) than in the E-MBOT (2/10, 20%) (P ¼ 0.025). Signal discrepancy was present in 18 of 30 (60%) MBOT, a higher incidence in I-MBOT (15/20, 75%) than in E-MBOT (3/ 10, 30%) (P ¼ 0.024). Correlation Between Pathology and MRI Findings Gross pathology confirmed MRI findings and four patterns of mucus were noted: (i) clear mucus, which was similar to signal intensity of urine, was found in 17 tumors; (ii) yellowish and viscous mucus in 8 tumors, seen frequently in the larger louculi, appeared a moderately intense on T1WI and a hyperintense on T2WI; (iii) clear jelly-like mucus in 8 tumors, seen commonly in the smaller loculi, was hypointense on T1WI and moderately intense on T2WI; (iv) white colloid in 9 tumors, seen in the honeycomb loculi, was hyperintense on T1WI and hypointense on T2WI. DISCUSSION MBOT account for 30–50% of BOT and happen in patients ranging in age from 9 to 88 years, but almost half of women were young ranging from 35 to 47 years (1,14,15). MBOT present at earlier stages than invasive tumors, with approximately 90% at stage I. In our series, the mean age of patients was 39 years, and all cases were at stage Ia. Owing to a favorable prognosis of MBOT, the conservative surgery of unilateral salpingo-oophorectomy is often performed to
preserve patient’s fertility (4,7,16). However, a consequently high recurrence rate calls for the need of complete surgical staging, which consists of total abdominal hysterectomy and bilateral salpingooophorectomy, cytological analyses of the ascites or the peritoneal lavage fluid, a multiple peritoneal biopsy, and an appendectomy for I-MBOT (4,7,16). Although studies have suggested that MRI is useful in characterizing complex adnexal masses (8–11) and quantitative multiparametric MRI (PWI, DWI,T2 map, etc.) have been confirmed having value in distinguishing malignant from benign ovarian tumors (12,13), the comprehensive MRI features specific to MBOT have not been studied.. Bent et al reviewed the MRI appearances of BOT and classified them into four morphological categories: unilocular cyst (19%); minimally septate cyst with papillary projections (19%); markedly septate lesion with plaque-like excrescences (45%) and predominally solid with exophytic papillary projections (16%) (10). Comparing with serous BOT, Bazot et al recently reported that MBOT were more frequently multilocular, with loculi of different signal intensity, and with a high number of septa (>10), which were commonly grouped, irregular, and thickened (11). In our MBOT series, three MRI morphological categories were identified: unilocular cyst with papillary projections (17%); multilocular cyst (76%); solid tumor (7%). In accordance with previous studies, multilocular cystic mass was the commonest finding (3,10,11). MBOT have been separated into two distinct subtypes: I- and E-MBOT. They differ in their
Table 1 MRI Features of Mucinous Borderline Ovarian Tumors MRI features Tumor diameter Multilocularity Signal discrepancy Thickened septa or wall Honeycomb loculi Projection Solid portion
Total (n ¼ 30)
I-MBOT (n ¼ 20)
E-MBOT (n ¼ 10)
P Valuey
16.5 6 6.5 23 (76%) 18 (60%) 16 (53%) 15 (50%) 10 (33%) 17% (5)
18.3 6 5.6 18 (90%) 15 (75%) 11 (55%) 13 (65%) 5 (25%) 2 (10%)
12.8 6 7.0 5 (50%) 3 (30%) 5 (50%) 2 (20%) 5 (50%) 3 (30%)
0.028 0.026 0.024 0.550 0.025 0.231 0.300
*Except for tumour size (Student’s), all P values were calculated using Fisher’s exact test.
6
clinicopathologic features, immunophenotype, clinical behavior, and surgery approach (14,17,18). Up to date, however, we have found no study concerning the imaging features of the two distinct subtypes. I-MBOT account for approximately 85% and the remaining 15% are E-MBOT. In our series, less I-MBOT (67%) were found, happen in older patients and have a poorer prognosis than E-MBOT (1,14,15). Multilocularity was more frequent in I-MBOT (90%) than in EMBOT (50%). Honeycomb loculi, which were composed of grouped innumerable little loculi, were seen in 50% of MBOT and more frequently in I-MBOT (65%) than in E-MBOT (20%). Signal discrepancy was another distinguished aspect of MBOT which was more common in I-MBOT (75%) than in E-MBOT (30%). It is also known as “stain-glass appearance”, which is defined as cystic masses in which the loculi show variable signal intensity on both T1WI and T2WI, and results from the viscosity of contents, blood product, or debris (19). However, our study suggests that the different patterns of mucus may be the underlying cause for these signal changes. I-MBOT are histopathologically heterogeneous, which often simultaneously contain areas of cystadenoma and intraepithelial carcinoma (1,14,15). The tumor is typically composed of numerous dilated glands and cysts of various sizes with internal epithelial tufts and papillations lined by a typical intestinaltype mucinous epithelium. These glands and cysts often produce different types of mucus, and consequently forming a complex cystic pattern (1,2,14). In our series, the most characteristic signal discrepancy was hyperintense on T1WI and hypointense on T2WI, which was correlate with pathological colloid and often seen in honeycomb loculi. E-MBOT, also known as seromucinous or M€ ullerian BOT, contain only cells resembling endocervical glands (1,14,15). They are clinically and pathologically closely related to serous BOT, and have little in common with I-MBOT (1,14,15). Because of the 40% rate of synchronous bilaterality and the 17% risk of leaving behind a borderline tumor in a grossly normal ovary, complete surgical staging is suggested (14). The characteristic feature of E-MBOT is unilocular or paucilocular cyst with multiple papillary projections, an appearance resembling that of serous BOT (1,3,14). In our series, all two cases of bilateral tumors are E-MBOT and 40% of E-MBOT are unilocular cyst. A solid mass with a branching papillary architecture and internal hypointense branching on T2WI is highly suggestive of serous BOT (10,11). In the present study, however, one case of E-MBOT had this similar appearance. MBOT are thought to represent an intermediate stage in the stepwise progression from benign to malignant tumors (1,14). Compared with MBOT, the mucinous cystadenomas tend to have less loculi, more uniform wall and septa, less frequent signal discrepancies and absent papillae (20). The primary ovarian mucinous carcinomas are uncommon, representing approximately 5% of ovarian surface epithelial carcinomas (2,9,14). The invasive carcinomas are usually mixed cystic-solid or multiloculated cystic
Ma et al.
with significantly greater and vigorously enhanced solid elements than MBOT (9,12). In addition, the peritoneal implants, metastatic retroperitoneal lymph nodes and large amount of ascities are more often seen in invasive tumors. There were some limitations in our study. First, the relative smaller sample size was in sufficient to subdivide the MRI features of MBOT. Further study increased sample size is warranted. Second, a selective bias was inevitably present owing to the retrospective nature of the study. Some small ovarian tumors characterized by US did not enroll in MRI study. Third, the potential value of functional MRI will be studied further. In conclusion, the present study demonstrated that MBOT typically presented as a multilocular cystic mass with honeycomb loculi, signal discrepancy, and irregular thickened septa or wall. Other MBOT appeared as uni- or paucilocular or cyst with multiple papillary projections and solid mass. REFERENCES 1. Hart WR. Mucinous tumours of the ovary: a review. Int J Gynecol Pathol 2005;24:4–25. 2. Hart WR, Norris HJ. Borderline and malignant mucinous tumours of the ovary: histologic criteria and clinical behavior. Cancer 1973;31:1031–1045. 3. Lalwani N, Shanbhogue AKP, Vikram R, Nagar A, Jagirdar J, Prasad SR. Current update on borderline ovarian neoplasms. AJR Am J Roentgenol 2010;194:330–336. 4. Houck K, Nikrui N, Duska L, et al. Borderline tumours of the ovary: correlation of frozen and permanent histopathologic diagnosis. Obstet Gynecol 2000;95:839–843. 5. Song T, Choi CH, Kim HJ, et al. Accuracy of frozen section diagnosis of borderline ovarian tumours. Gynecol Oncol 2011;122: 127–131. 6. Bazot M, Nassar-Slaba J, Thomassin-Naggara I, Cortez A, Uzan S, Darai E. MR imaging compared with intraoperative frozensection examination for the diagnosis of adnexal tumours: correlation with final histology. Eur Radiol 2006;16:2687–2699. 7. Park JY, Kim DY, Kim JH, Kim YM, Kim YT, Nam JH. Surgical management of borderline ovarian tumours: the role of fertilitysparing surgery. Gynecol Oncol 2009;113:75–82. 8. Vargas HA, Barrett T, Sala E. MRI of ovarian masses. J Magn Reson Imaging 2013;37:265–281. 9. deSouza NM, O’Neill R, Mcclndoe GA, Dina R, Soutter WP. Borderline tumours of the ovary: CT amd MRI features and tumour markers in differentiation from stage I disease. AJR Am J Roentgenol 2005;184:999–1003. 10. Bent CI, Sahdev A, Rockall AG, Singh N, Sohaib SA, Reznek RH. MR appearances of borderline ovarian tumours. Clin Radiol 2009;64:430–438. 11. Bazot M, Haouy D, Dara€ı E, Cortez A, Dechoux-Vodovar S, Thomassin-Naggara I. Is MRI a useful tool to distinguish between serous and mucinous borderline ovarian tumours? Clin Radiol 2013;68:e1–e8. 12. Thomassin-Naggara I, Toussaint I, Perror N, et al. Characterization of complex adnexal masses: value of adding perfusion- and diffusion-weighted MR imaging to conventional MR imaging. Radiology 2011,258:793–803. 13. Carter JS, Koopmeiners JS, Kuehn-Hajder JE, et al. Quantitative multiparametric MRI of ovarian cancer. J Magn Reson Imaging 2013 doi:0.1002/jmri.24119. [Epub ahead of print]. 14. Acs G. Serous and mucinous borederline (low malignant potential tumours of the ovary. Am J Clin Pathol 2005;123(Suppl 1):s13– s57. 15. Hart WR. Borderline epithelial tumours of the ovary. Mod Pathol 2005;18:s33–s50. 16. Anfinan N, Sait K, Ghatage P, Nation J, Chu P. Ten years experience in the management of borderline ovarian tumours at
MRI Appearances and Pathology of Ovarian Tumors Tom Baker Cancer Centre. Arch Gynecol Obstet 2011;284:731– 735. 17. Rutgers JL, Scully RE. Ovarian mullerian mucionous papillary cystadenomas of borederline malignancy: aclinicopathologic analysis. Cancer 1988;61:340–348. 18. Rodriguez IM, Irving JA, Prat J. Endocervical-like mucinous borderline tumours of the ovary: a clinicpathologic analysis of 31 cases. Am J Surg Pathol 2004;28:1311–1318.
7 19. Tanaka YO, Nishida M, Kurosaki Y, Itai Y, Tsunoda H, Kubo T. Differential diagnosis of gynecological “stained glass” tumours on MRI. Br J Radiol 1999;72:414–420. 20. Okamoto Y, Tanaka YO, Tsunoda H, Yoshikawa H, Minami M. Malignant or borderline mucinous cystic neoplasms have a larger number of loculi than mucinous cystadenoma: a retrospective study with MR. J Magn Reson Imaging 2007;26:94– 99.
