Abdom Imaging 20:2-8 (1995)
Abdominal Imaging 9 Springer-Verlag New York Inc. 1995
Review CT and MRI of ovarian cancer R. Forstner,* H. Hricak, S. White Department of Radiology,Universityof California, San Francisco, 505 Parnassus Avenue, San Francisco, CA 94143-0628,USA
Ovarian carcinoma is the second most common gynecologic malignancy in the US and is the leading cause of mortality in cancers of the reproductive tract [1]. Its incidence is one in 70 women, 50% occur at under 65 years of age [2]. Since the majority of patients will have advanced disease at time of presentation, early detection and improved characterization of ovarian masses are of great importance. The use of cross-sectional imaging in pretreatment evaluation of ovarian cancer is controversial. Exploratory laparotomy is the mainstay for managing ovarian cancer, as it offers both surgicopathologic staging and tumor debulking [2-4]. However, as many as 30% to 40% of patients are understaged at initial laparotomy [4]. Cross-sectional imaging can provide staging information which can assist in surgical planning and in selection of treatment options. The choice of imaging modality depends on the information required, equipment available, local expertise, and preference [2]. Ultrasound plays a pivotal role in ovarian tumor detection and characterization but its use in staging ovarian cancer is limited. CT is the single best and most practical approach in staging ovarian cancer. The staging accuracy of CT of 70% to 90% is hard to surpass [2]. MR imaging is emerging as a problem-solving modality useful for lesion characterization and staging, including evaluation of locoregional extent, lymph node metastasis, and liver surface implants.
CT Imaging Techniques For evaluation of the pelvis, contiguous 5 - 1 0 mm transaxial images should extend from the inguinal region (below inferior pubic ramus) to the iliac crest. For the evaluation of the abdomen, 10 mm axial sections are * P e r m a n e n t address: C o r r e s p o n d e n c e to:
Zentralr6ntgeninstitut, LKA, Salzburg, Austria
H. Hricak
obtained from the iliac crest to the dome of the diaphragms [3]. Oral and rectal contrast media are necessary to differentiate adnexal and soft-tissue structures from bowel loops [5]. Intravenous contrast media improves tumor delineation and aids in the characterization of tumor architecture [6-8]. Visualization of peritoneal implants is improved with administration of intravenous contrast and use of intraperitoneal contrast has been recommended as well
[9, 10]. Recently, helical or spiral CT has been in clinical use. Its role in imaging the female pelvis has yet to be evaluated. Unlike standard CT, it allows continuous volumetric data acquisition, often within a single breathhold, thus nearly eliminating the problems of volume averaging and respiratory artifacts [11]. Its advantages over standard CT appear to be improved lesion detection and densidometry and optimization of contrast media enhancement [11]. Conventional transaxial images as well as multiplanar or three-dimensional images can be obtained without additional radiation exposure [11]. This may improve lesion detection and assist in defining tumor origin and extent.
MR Imaging Techniques In order to properly detect and characterize the ovaries and ovarian lesions, MR images should be obtained in at least two orthogonal planes. Axial and saggital planes should be routine, but coronal images may prove useful in large ovarian masses which extend into the abdomen. Both T1- and T2-weighted images are necessary. Tl-weighted images are necessary for lesion characterization (e.g., to diagnose fat or hemorrhage within an ovarian mass) and detection of lymph nodes. On T2weighted images the high signal intensity ovarian follicles help in depicting the ovaries, in differentiating ovary from bowel loop, and in helping define the adnexal or uterine origin of a pelvic mass. T2-weighted
R. Forstner et al.: CT and MRI of ovarian cancer sequences are also needed for lesion characterization and staging. Compared with conventional spin echo (CSE), fast spin echo (FSE) technique significantly shortens the acquisition time. Advantages of this method include better resolution due to increased signal-noise per unit time, decreased motion artifact, and increased patient throughput [12-14]. Gadolinium-chelates enhanced Tl-weighted images significantly improve lesion characterization [12, 15, 16]. Contrast-enhanced images also improve the detection of peritoneal and omental implants in patients with ovarian carcinoma. A bolus injection with immediate scanning is recommended. Rarely delayed contrast media enhanced images can be helpful in detecting fistulae. MR imaging is versatile and offers a variety of techniques to improve image quality and answer specific questions. Fat saturation technique can differentiate fat from hemorrhage, and gradient-recalled echo techniques (GRE) can distinguish adenopathy from flowing blood. Intramuscular glucagon (unless medically contraindicated) is recommended to suppress motion artifact from bowel peristalsis. Respiratory compensation should be performed routinely to diminish artifacts due to motion. However, software allowing respiratory compensation with FSE technique is not yet available. Therefore, a tight wrap around the abdomen should always be employed [13]. The use of phased array coils further improves the signal-to-noise levels allowing superb image quality with a reduced field of view of down to 20 cm [12, 17]. However, not all patients are suitable for imaging with phased array coils. In obese patients or patients with abdominal protuberances, proper coil positioning may be difficult, and imaging may be degraded by limited depth of penetration. Furthermore, the field of view is limited; therefore, this technique is less suitable for staging examinations. Finally, additional techniques as fat saturation bands or intensity correction have to be applied to limit artifacts from increased signal in the near field [14].
Ovarian Tumors The most commonly used classification of ovarian tumors is based on the tissue of origin, as set out by the World Health Organization (WHO). Ovarian tumors arise from one of the three ovarian components: the surface epithelium, germ cells, or the stroma of the ovary. Among these three tumor types, germ cell tumors have the best prognosis. They have a higher incidence in children and young females [2, 4]. Secondary tumors of the ovaries are often bilateral and arise most commonly from the gastrointestinal tract, breast, and thyroid primaries [6].
3 Approximately 80% of ovarian tumors are benign, the majority found in women of reproductive age [18]. Mature teratomas (dermoid cysts) constitute one quarter to one half of benign neoplasm presenting between the ages 20-45 years; in patients older than age 45, only 10% of benign tumors are dermoids [19]. Most (90%) of the ovarian tumors are of epithelial type. Serous and mucinous epithelial tumors account for approximately 50% of benign neoplasm prior to menopause and 80% afterward [20]. They present as thinwalled tumors which are bilateral in 10% and 5% of cases, respectively. Serous cystadenoma can be unilocular or multilocular, while mucinous cystadenoma are often multilocular. Fibroma and fibrothecoma, benign tumors of stromal origin, are seen in a small percentage of patients both before and after menopause. Ovarian cancer affects women in the peri- and postmenopausal years; over age 45 years, the chance that an ovarian tumor is malignant rises to 1 in 3 [18]. The most common malignant ovarian tumors are serous or mucinous adenocarcinoma, adenocarcinoma (undifferentiated), and endometroid carcinoma. Malignant ovarian tumors often cause only minor symptoms, such as abdominal distention or pressure. As a result, at the time of diagnosis, approximately 75% of these patients present with cancer spread beyond the ovary-stage greater than II, and 60% have spread beyond the pelvis-stage greater than III [4]. Overall, 5-year survival is variable, from approximately 5% in patients who present late with high stage tumors and have residual disease after initial surgery, to approximately 80% in patients with low stage tumors and no evidence of residual disease [4, 6]. Borderline tumors (ovarian carcinoma of low malignant potential) stage I have been described to have a 5- and 10-year tumor-free survival rate of 95% and 91%, respectively [21].
Ovarian Tumor Characterization Currently, endovaginal ultrasound is the most effective imaging modality in ovarian cancer screening with a high sensitivity (100%) but moderate specificity (83%) [22, 23]. If ultrasound is equivocal or fails, CT or MRI should be performed. In both modalities tumor size, wall thickness, presence of vegetations, and solid structures are criteria for lesion characterization.
CT Appearance of Benign Tumors Benign serous and mucinous cystadenomas are characterized by CT as thin-walled cystic lesions without evidence of soft tissue components, irregular walls, or papillary projections [24] (Fig. 1). Serous cystadenoma are most commonly uni- or bilocular; they may or may
4
R. Forstuer et al.: CT and MRI of ovarian cancer
Fig. 1. Serous cystadenoma. CT demonstrates bilateral multicystic thin-walled ovarian lesions (arrows). Fig. 2. Dermoid. CT shows a fat density ovarian lesion with a mural nodule (arrow) and a punctate calcification (white arrow). Uterus (U); urinary bladder (B).
not show plaquelike calcifications. Mucinous cystadenoma are typically multilocular [24]. Due to the serous nature of the fluid, CT density approaches that of water. High density fluid visualized on CT without contrast enhancement indicates recent hemorrhage (especially in serous cystadenoma) or mucin (in mucinous cystadenoma) [25]. The CT appearance of a fibroma, a benign stromal tumor, is that of a nonspecific solid adnexal lesion. Therefore, by CT it cannot be differentiated from exophytic leiomyomas or other solid tumors. Dermoids can be reliably diagnosed by CT by their fatty density and often a mural nodule, which contains calcifications hair and solid structures, is identified [25] (Fig. 2).
MR Appearance of Benign Tumors The cystic pattern of both serous and mucinous cystadenomas is also well visualized by M R I (Fig. 3 A - C ) .
Fig. 3. Mucinous cystadenoma. A Tl-weighted image. B Contrast enhanced Tl-weighted image. C T2-weighted image. Lesion characterization including its cystic nature and thin wall is best appreciated on contrast-enhanced Tl-weighted image.
Serous fluid with loculations demonstrates low signal intensity on Tl-weighted M R images, high signal intensity on T2-weighted images, and lack of enhancement in contrast enhanced Tl-weighted images. Hemorrhage and mucin within the cysts can be depicted as medium to high signal intensity on both T1- and T2-weighted images. Characteristics specific for benign epithelial tumors such
R. Forstner et al.: CT and MRI of ovarian cancer
5
Fig. 4. Fibrothecoma. A Transaxial Tl-weighted image and B T2-weighted image demonstrate a solid ovarian lesion with low signal intensity on T2-weighted image--a finding consistent with its fibrous nature. Uterus (U); sigmoid (S); uterine fibroid (*). Fig. 5. Dermoid. A Axial T1weighted image. B Axial T1weighted image with fat saturation technique. The combination of Tl-weighted image and fat saturation technique demonstrates the fatty contents of the lesion (*). Uterus (U).
Fig. 6. Endometrioma. A Axial Tl-weighted image. B Axial Tl-weighted image with fat saturation technique. High signal intensity with and without fat saturation technique indicates the hemorrhagic contents within the endometrioma (arrows). Uterine fibroids (*).
as thin walls, m i n i m a l septation, and absence o f papillary projections, are depicted with either T2- or better with contrast enhanced T l - w e i g h t e d images [15, 16, 25]. B e n i g n ovarian fibroma are w e l l - c i r c u m s c r i b e d solid tumors w h i c h demonstrate low signal intensity on both T1- and T 2 - w e i g h t e d images. The low signal on T2 is specific for this lesion and M R I unlike C T allows one to differentiate an ovarian fibroma f r o m an exophytic l e i o m y o m a or other t u m o r (Fig. 4 A and B). F a t saturation techniques are useful in differentiation b e t w e e n d e r m o i d or e n d o m e t r i o m a (Figs. 5A, B and 6A, B). Other findings such as high signal inten-
sity on T l - w e i g h t e d images, shading (seen in endom e t r i o m a ) or c h e m i c a l shift artifact (seen in dermoid) are less reliable in characterization of these lesions [26-28].
CT and MRI Imaging of Malignant Epithelial Tumors C y s t a d e n o c a r c i n o m a s are p r e d o m i n a n t l y cystic tumors usually larger than 4 c m in diameter. Findings suggestive o f m a l i g n a n c y are vegetations (papillary endocystic
6
Fig. 7. Poorly differentiatedadenocarcinoma.CT demonstratesbilateral cystic and solid ovarianmasses. Urinarybladder(B); sigmoid(*).
Fig. 8. Papillary serous adenocarcinoma.A Axial and B sagittal T2weighted images demonstratemulticysticovarian lesions with papillary projections (p). Arrows indicate invasion of the sigmoid colon (s). Urinarybladder (B); uterus (U).
R. Forstner et al.: CT and MRI of ovarian cancer
Although MRI can provide further information on the internal architecture of ovarian malignancies, no significant difference in sensitivity or specificity of staging performance has been noted between MR and CT [15, 29, 35].
Staging of Ovarian Carcinoma projections), solid lesions, large solid parts or necrosis [15, 16] (Figs. 7 and 8A, B). The finding of a thick wall and septa of more than 3 mm thickness are less reliable signs of malignancy. These tumors demonstrate variable signal intensity on T1- and T2-weighted MR images depending on cyst contents. Furthermore, the ancillary findings of pelvic organ and pelvic sidewall invasion, peritoneal and omental involvement, ascites and lymphadenopathy increase the suspicion of malignancy [16]. The presence of tumor vessels on dynamic CT is an additional sign of a malignant mass [25]. Although suggestive of malignancy, the finding of pelvic ascites can also be seen in ovarian torsion, pelvic inflammatory disease, and benign ovarian fibroma and is therefore not specific [25]. Abdominal ascites alone [25] or the finding of ascites anterior to the uterus are highly suggestive for malignant ascites [29]. With both CT and MRI contrast enhancement helps to characterize intratumoral architecture (cystic vs. solid components) and better delineate tumor borders [15, 16, 331. MRI, with its multiplanar capability and superior soft tissue contrast, can facilitate the characterization of ovarian tumors in several situations. MRI enables superior differentiation between adnexal and uterine masses when compared with either ultrasound or CT [29-34]. MR is helpful in patients with a contraindication to iodined contrast material or impaired renal function. MRI is also helpful in pregnant patients with an indeterminate mass on US [30, 31].
In order of decreasing frequency, ovarian carcinoma spreads via locoregional extension, via peritoneal and lymphatic pathways, and hematogenously. Tumor may extend locoregionally to involve the uterus, urinary bladder, sigmoid colon, or pelvic sidewall. Common sites for peritoneal seeding include the peritoneal pouches (pouch of Douglas, Morrisson's pouch, right paracolic gutter) as well as the subphrenic space, liver surface, porta hepatis, intrahepatic fissure, and the omentum and bowel mesenteries. Lymphatic spread may occur to retroperitoneal (periaortic, iliac) or lateral pelvic and inguinal nodes. Hematogenous spread is a late occurrence, and parenchymal liver or splenic metastases are extremely rare upon initial presentation of ovarian cancer. The length and type of surgery depends upon: (1) the extent of locoregional spread, (2) the presence, volume, and location of peritoneal disease, and (3) the presence of lymphadenopathy. Thus, the goal of imaging is to provide an accurate preoperative assessment of the extent of disease, which is of paramount importance for optimal surgical and management planning. For example, in addition to correct staging, the radiologist can provide information about areas difficult to assess surgically (such as the dome of the diaphragm or retroperitoneum), and can alert the surgeon to findings crucial for planning and treatment (such as sigmoid colon invasion, which will necessitate colon resection and/or colostomy, therefore prolonging time for surgery). Imag-
R. Forstner et al.: CT and MRI of ovarian cancer
7
Fig. 9. A C T and B T2-weighted image. Plaque-like diaphragmatic (arrows), interhepatic fissure (*), and splenic surface implants (curved arrow) are well depicted by both methods.
Fig. 10. Omental caking in a patient in which A C T as well as B T1weighted image show similar findings (arrows).
ing is also useful in patient follow-up and for guidance during percutaneous aspiration biopsy. CT is currently recommended as the imaging modality of choice for staging ovarian cancer [2, 3, 29]. This is due to the facts that CT is more widely available than MRI, has shorter imaging and interpretation times, and in general radiologists are more familiar with CT. Also, CT has the advantage of reliably demonstrating calcification, which is of value in characterizing primary and metastatic or serous cystadenocarcinomas. The reported staging accuracy of CT is between 70% and 90% [2, 36, 37]. The main limitation of CT is the inability to detect small peritoneal and mesenteric implants [38]. The detection of peritoneal implants depends on location, surrounding aScites, and implant size. Current CT scanners can detect 50% of peritoneal implants as small as 5 mm [9]. The use of intraperitoneal contrast has been suggested to increase detection of peritoneal tumor implants smaller than 5 mm and has been found to increase detection up to 2 - 4 times that of standard exams [10]. MRI is also capable of depicting spread of ovarian carcinoma throughout the pelvis and the abdomen with a reported accuracy of 75% [16] (Fig. 9A and B). Due
to its multiplanar capability and superior contrast resolution, involvement of the uterus, sigmoid colon, urinary bladder, dome of diaphragm, and liver surface may be assessed better than by CT. Involvement of the pelvic sidewall and retroperitoneum, omental deposits, and peritoneal implants > 1 cm are depicted with similar sensitivity by both MR and CT, while disease in the mesentery or implants on the large and small bowel are better detected by CT [29] (Fig. 10A and B). MR and CT have similar accuracy in depicting lymph nodes, both sharing high sensitivity and lacking specificity in diagnosing malignant involvement of nodes [29, 39].
Summary Among the gynecologic malignancies, ovarian cancer is second most common in incidence. However, unlike the other gynecologic cancers, its mortality has decreased only minimally during the last two decades [1]. Only recently, preliminary studies suggest promising results for ovarian cancer screening using transvaginal ultrasound in combination with serum Ca 125 levels [22, 23].
8
Exploratory laparotomy has been the mainstay in the management of ovarian cancer, as it offers histopathological evaluation as well as cytoreduction. However, it is limited by its inaccuracy with understaging in 3 0 40% at initial presentation. Cross-sectional imaging contributes valuable information toward preoperative surgical and management planning. The proper surgical approach can be selected, the need for preoperative chemotherapeutic debulking can be assessed, and the surgeon will be forewarned of the need for assistance from a gynecologic oncologic surgeon or gastrointestinal oncologic surgeon if a complicated surgical procedure or bowel resection is indicated. CT is established as the primary imaging modality for characterization of ovarian tumors and ovarian cancer staging, while MR is emerging as a problem-solving modality. MR seems to be superior to CT in lesion characterization, in evaluation of local extent of tumor, and in tumor implants involving the hemidiaphragm and liver surface. The role of spiral CT has yet to be explored.
References 1. Boring CC, Squires TS, Tong T, Montgomery S. Cancer statistics, 1994. CA 1994;44:7-26 2. Johnson RJ. Review radiology in the management of ovarian cancer. Clin Radiol 1993;48:75-82 3. Walsh JW. Computed tomography of gynecologic neoplasms. Radiol Clin North Am 1992;30:817-830 4. Young RC, Fuks Z, Hoskins W. Cancer of the ovary. In: De Vita VT, Hellman S, Rosenberg SA, eds. Cancer:principles and practice of oncology, vol I. Philadelphia: Lippincrott, 1989:1162-1191 5. Amman AM, Walsh JW. Normal anatomy and techniques of examination. In: Walsh JW, ed. Computed tomography of the pelvis. New York: Churchill Livingstone, 1985:1 6. Carrington BM. The adnexae. In: Hricak H, Carrington BM, eds. MRI of the pelvis: a text atlas. London: Martin Dunitz, 1991:185 7. Kormano M, Groenroos M. Computer-tomograpNc evaluation of gynecologic rumors. Acta Obstet Gynecol Scand 1984;63:509-516 8. Mamtora H, Ishelwood I. Computed tomography in ovarian carcinoma: patterns of disease and limitations. Radiology 1982;33:165-171 9. Buy JN, Moss AA, Ghossain MA, et al. Peritoneal implants from ovarian tumors: CT findings. Radiology 1988;169:691 10. Halvorsen RA, Panushka C, Oakley G, et al. Intraperitoneal contrast material improves the CT detection of peritoneal metastases. A JR 1991;157:37 11. Heiken JP, Brink JC, Vannier MW. Spiral (helical) CT. Radiology 1993;189:647-656 12. Hricak H. Current trends in MR imaging of the female pelvis. Radiographics 1993;13:913-919 13. Smith RC, Reinhold C, Lange RC, et al. Fast spin echo MR imaging of the female pelvis. Part I. Use of a whole-volume coil. Radiology 1992;184:665 14. Nghiem HV, Herfkens RJ, Francis IR, et al. The pelvis: T2weighted fast spin-echo MR imaging. Radiology 1992; 185:213 15. Sica GT, Stevens SK, Hricak H, et al. Comparison of unenhanced and contrast-enhanced MR images in the evaluation of ovarian lesions (presented at the 78th Scientific Assembly and Annual Meeting of the Radiological Society of North America, 1992) 16. Stevens SK, Hricak H, Stern JL. Ovarian lesions: detection and
R. Forstner et al.: CT and MRI of ovarian cancer
17.
18.
19.
20.
21. 22. 23.
24.
25. 26.
27.
28. 29.
30. 31. 32.
33. 34.
35. 36. 37.
38.
39.
characterization with gadolinium-enhanced MR imaging at 1.5 T. Radiology 1991;181:481 Smith RC, Reinhold C, McCauley TR, et al, Multicoil high-resolution fast spin-echo MR imaging of the female pelvis. Radiology 1992;184:671 Scully RE. General aspects of ovarian tumors. In: Tumors of the ovary and maldeveloped gonads: atlas of tumor pathology, 2nd series, fascicle 16. Washington, D.C.: Armed Forces Institute of Pathology, 1979:30 Scully RE. Teratomas. In: Tumors of the ovary and maldeveloped gonads: atlas of tumor pathology, 2nd series, fascicle 16. Washington, D.C.: Armed Forces Institute of Pathology, 1979:246 Morrow CP, Townsend DE, eds. Tumors of the ovary: neoplasms derived from coelomic epithelium. In: Synopsis of gynecologic oncology, 3rd ed, New York: Churchill Livingstone, 1987:257 Gershenson DM, Silva EG. Serous ovarian tumors of low malignant potential with peritoneal implants. Cancer 1990;65:578-585 Van Nagell Jr, DePriest PD, Gallion HH, Pavlik EJ. Ovarian cancer screening. Cancer 1993;71:1523-1528 Sassone AM, Timor-Tritsch IE, Artner A, et al. Transvaginal sonographic characterization of ovarian disease: evaluation of a new scoring system to predict ovarian malignancy. Obstet Gynecol 1991;78:70-76 Scoutt LM, McCarthy SM, Moss A. Computed tomography and magnetic resonance imaging of the pelvis. In: Moss AA, Gamsu G, Genant HK, eds. Computed tomography of the body with magnetic resonance imaging, vol. III, 2nd ed. Philadelphia: WB Saunders, 1983:1215 Buy JL, Ghossain MA, Sciot C, et al. Epithelial tumors of the ovary: CT findings and correlation with US. Radiology 1991;178:811 Kier R, Smith RC, McCarthy SM. Value of lipid and watersuppression MR Images in distinguishing between blood and lipid within ovarian masses. A JR 1992;158:321-325 Stevens SK, Hricak H, Campos Z. Teratoma versus cystic hemorrhagic adnexal lesions: differentiation with proton-selective fatsaturation MRI. Radiology 1993;186:481-488 Togashi K, Kazumasa N, Kimura I, et al. Endometrial cysts: diagnosis with MR imaging. Radiology 1991;180:73 Forstner R, Hricak H, Occhipinti K, White S, Stern J. Value of CT and MRI in the staging of ovarian cancer. Accepted for presentation at the 1994 Annual Meeting of the Society of Magnetic Resonance. Weinreb JC, Brown CE, Lowe TW, et al. Pelvic masses in pregnant patients MR and US imaging. Radiology 1986; 159:717 Kier R, McCarthy SM, Scoutt LM, Viscarello RR, et al. Pelvic masses in pregnancy: MR Imaging. Radiology 1990;176:709-713 Semelka RC, Lawrence PH, Shoenut JP, Heywood BScM, et al. Primary ovarian cancer; prospective comparison of contrast-enhanced CT and pre- and postcontrast, fat suppressed MRI, with histologic correlation. JMRI 1993;3:99-106 Turnher SA. MRI of pelvic masses in women: contrast enhanced vs unenhanced images. A JR 1992; 159:1243-1250 Mitchell DG, Mintz MC, Spritzer CE, et al. Adnexal masses: MR imaging observations at 1.5 T, with US and CT correlation. Radiology 1987;162:319 Ghossain MA, Buy JN, Ligneres C, et al. Epithelial tumors of the ovary: comparison of MR and CT findings. Radiology 1991;181:863 Amendola MA. The role of CT in the evaluation of ovarian malignancy. Crit Rev Diag Imag 1985;24:329-368 Johnson RJ, Blackledge G, Eddlestone B, et al. Abdomino-pelvic computed tomography in the management of ovarian carcinoma. Radiology 1983;146:447-452 Jaquet P, Jelinek JS, Steves MA, Sugarbaker PH. Evaluation of computed tomography in patients with peritoneal carcinomatosis. Cancer 1993;72:1631 - 1636 Dooms GC, Hricak H, Crooks LE, et al. Magnetic resonance imaging of the lymphnodes: comparison with CT. Radiology 1984;153:719