Value of diffusion-weighted magnetic resonance imaging in the characterization of complex adnexal masses

Transcription

1 Tu mo ri, 99: , 2013 Value of diffusion-weighted magnetic resonance imaging in the characterization of complex adnexal masses Salvatore Cappabianca, Francesco Iaselli, Alfonso Reginelli, Alfredo D Andrea, Fabrizio Urraro, Roberto Grassi, and Antonio Rotondo Dipartimento di Internistica Clinica e Sperimentale F. Magrassi, A. Lanzara, Unità di Radiologia, Radioterapia e Medicina Nucleare, Seconda Università di Napoli, Naples, Italy ABSTRACT Aims and background. The aim of the study was to define the role of diffusionweighted imaging in the characterization of adnexal complex masses, with particular regard to the distinction between benign and malignant lesions. Conflicting results on this topic have emerged from studies in the last decade, with a consequent substantial limitation to the use of this relatively novel technique in clinical practice. Methods and study design. Magnetic resonance imaging examinations performed on 91 patients with ovarian masses (56 benign, 35 malignant) were retrospectively evaluated by two observers unaware of the final histopathological diagnosis. Diffusionweighted sequences with b values of 0, 500 and 1000 were performed in all cases, and apparent diffusion coefficient maps were automatically generated. The signals of both the cystic and solid components of the ovarian masses were evaluated on T2- weighted and diffusion-weighted images acquired with a b value of Apparent diffusion coefficient values were measured in all cases. Results. With regard to the solid components, hypointensity on both the T2-weighted and diffusion-weighted images has proved to be a reliable indicator of benignancy. In contrast, hyperintensity on both sequences was suggestive of malignancy. Signal intensity of the cystic components and apparent diffusion coefficient values of both components have not proven useful in characterization of the masses. Conclusions. Only the definition of the signal intensities on diffusion-weighted images obtained with the use of high b values on the solid component of a complex adnexal mass is useful to characterize an ovarian mass as benign or malignant. Introduction The incidental identification of an indeterminate complex ovarian mass during a radiological examination is one of the most common clinical problems today. In fact, the integration of clinical data, laboratory tests and findings from conventional imaging is not always sufficient to define the nature of the mass, which cannot be easily obtained even through a biopsy because of its deep location 1,2. Therefore, after the detection of an indeterminate mass, a surgical approach is usually chosen, in many cases proved unnecessary by the subsequent histopathological examination. Ultrasonography, usually with a transvaginal approach, is the first-line diagnostic tool. However, because of its major limitations in terms of sectoriality, spatial resolution and, more importantly, contrast resolution, it is unable to resolve in most of the cases the differential diagnosis between benign and malignant masses 3. Magnetic resonance imaging (MRI) is characterized by a higher information content compared to transvaginal ultrasonography, allowing to achieve high-quality panoramic images in any plane of the space 4,5. Mainly because of the high degree of histological variability typical of ovarian masses and of the consequent overlap of their features in Keywords: diffusion-weighted magnetic resonance imaging, complex adnexal masses. Financial disclosure: The authors declare they have no financial disclosure. Conflict of interest: None. Correspondence to: Francesco Iaselli, Corso Trieste 273, 81100, Caserta, Italy. Tel ; francescoiaselli@hotmail.it Received July 9, 2012; accepted October 15, 2012.

2 DWI IN THE CHARACTERIZATION OF COMPLEX ADNEXAL MASSES 211 conventional sequences, however, findings at MRI are often poorly specific, particularly in the case of masses with a mixed, solid-cystic pattern 6. In the last decade, in the characterization of undeterminate complex ovarian masses, conventional techniques were complemented, with still conflicting results, by diffusion-weighted imaging (DWI) 7,8. The technique, based on the measurement of the Brownian motions of water molecules within a given anatomical district, provides unique information about the biophysical properties of tissues, such as cellular organization and cellular density, microstructure and microcirculation and, more generally, diffusivity of the water. Areas characterized by restricted diffusion or by low values of the apparent diffusion coefficient (ADC) generally correspond with foci of hypercellularity, therefore compatible with uncontrolled malignant proliferation 9,10. In our study, we have tried to define the potentiality of DWI in the characterization of complex ovarian masses and, more in particular, in the often difficult differentiation between benign and malignant masses. We retrospectively analyzed the signal intensity displayed in the T2-weighted images and in the diffusion-weighted images obtained with a b value of 1000 sec/mm 2 and of the ADC value on either the solid or cystic component of 91 masses of undetermined type subsequently defined by histopathological examination. Materials and methods MRI examinations of 91 female patients with 97 initially undetermined complex ovarian masses performed at the Department of Radiology of our University Hospital between September 2008 and March 2012 were retrospectively and independently evaluated by two radiologists experienced in the execution and in the report of MRI examinations including diffusion-weighted sequences (S.C. and F.I.). Both observers were not aware of the definitive histopathological diagnosis. In cases of bilateral masses (6 patients), we considered only the one with the larger diameter. Of the total of 91 complex ovarian masses, 56 were benign (62%, Table 1) and 35 malignant (38%, Table 2). The group of benign lesions included 20 fibromas/fibrotecomas, 13 mucinous adenomas, 9 serous adenomas, 4 Brenner tumors, 4 cases of hydrosalpinx, 1 decidualized endometrioma, 2 cases of polypoid endometriosis, 2 endometrioid cystadenomas and 1 struma ovarii. The group of malignant lesions included 14 serous adenocarcinomas, 3 clear-cell adenocarcinomas, 3 metastatic tumors, 2 endometrioid adenocarcinomas, 3 granulosa cell neoplasms, 2 Sertoli-Leydig cell tumors, 2 mucinous adenocarcinomas, 1 carcinosarcoma, 1 disgerminoma, 1 leiomyosarcoma, 1 borderline mucinous cystadenoma, 1 borderline serous cystadenoma and 1 mixed serous-endometrioid adenocarcinoma. Several authors who carried out studies similar to ours excluded from their series mature cystic teratomas and endometriomas. In fact, for their distinctive histopathological characteristics, such entities displayed at DWI features in some cases totally overlapping those of malignant lesions, resulting in significant reduction of the statistical significance. Moreover, both mature cystic teratomas and endometriomas are in most cases easily identified through the use of conventional sequences (influence of the hemosiderin within endometriomas and of the keratinoid substance within mature cystic teratomas on T1 and T2 relaxation times, possibility to saturate the signal of the fat component of teratomas through the use of specific sequences) 1,11,12-14, without the need for integration with DWI 11,12, On the basis of these findings, we also excluded from our series, although they constitute a large part of ovar- Table 1 - Benign adnexal complex masses: frequency and histopathological features Histotype Frequency Pattern T2 DWI b 1000 ADC S C S C S C Fibroma / fibrothecoma 20 S=15, SC=5 hyper=6 hyper=5 hyper=3, hypo= ± ±0.52 hypo=14 hypo= Mucinous adenoma 13 C=11, SC=2 hyper=2 hyper=11 hypo=2 hypo=11 Struma ovarii 1 SC hyper=1 hyper=1 hypo=1 hypo=1 Serous adenoma 9 C=8, SC=1 hyper=1 hyper=9 hypo=1 hypo=9 Decidualized endometrioma 1 SC=1 hyper=1 hypo=1 hyper=1 hyper=1 Polypoid endometriosis 2 SC=2 hyper=2 hyper=2 hypo=2 hypo=2 Brenner s tumor 4 SC=3, C=1 hypo=3 hyper=4 hyper=1, hypo=4 hypo=2 Endometrioid cystadenoma 2 C=1, SC=1 hypo=1 hyper=2 hyper=1 hypo=2 Hydrosalpinx 4 C=4 X hyper=5 X hypo=5 P at Fisher s exact test $ at Mann-Whitney s test ns ns ns S, solid; SC, solid-cystic; C, cystic; hyper, hyperintensity; hypo, hypointensity; ns, non-significant.

3 212 S CAPPABIANCA, F IASELLI, A REGINELLI ET AL Table 2 - Malignant complex adnexal masses: frequency, histopathological and radiological features Histotype Frequency Pattern T2 DWI b 1000 ADC S C S C S C Serous adenocarcinoma 14 SC=13, hyper=14 hyper=13 hyper=14 hypo= ± ±0.49 S= Clear-cell adenocarcinoma 3 SC=3 hyper=3 hyper=3 hyper=3 hypo=3 Endometrioid carcinoma 2 SC=2 hyper=2 hyper=2 hyper=2 hypo=2 Carcinosarcoma 1 S=1 hyper=1 X hyper=1 X Disgerminoma 1 SC=1 hypo=1 hyper=1 hyper=1 hypo=1 Leiomyosarcoma 1 S=1 hyper=1 X hyper=1 X Metastasis 3 S=3 hypo=2, X hyper=3 X hyper=1 Borderline mucinous cystadenoma 1 SC=1 hypo=1 hyper=1 hyper=1 hypo=1 Borderline serous cystadenoma 1 SC=1 hyper=1 hyper=1 hyper=1 hypo=1 Mucinous adenocarcinoma 2 SC=2 hyper=2 hyper=2 hyper=2 hypo=2 Mixed serous-endometrioid 1 SC=1 hyper=1 hypo=1 hyper=1 hyper=1 adenocarcinoma Neoplasm of Granulosa s cells 3 SC=1, S=2 hyper=3 hyper=1 hyper=3 hypo=1 Neoplasm of Sertoli-Leydig s cells 2 S=2 hyper=2 X hyper=2 X P at Fisher s exact test & Mann-Whitney test ns S, solid; SC, solid-cystic; C, cystic; hyper, hyperintensity; hypo, hypointensity; ns, non-significant. ian pathology, these two entities, except for 1 case of decidualized endometrioma and 2 cases of polypoid endometriosis characterized by abnormal signal intensity at conventional imaging. According to their features at conventional imaging, complex ovarian masses were divided into predominantly cystic (solid component less than 25% of the total), mixed and predominantly solid (cystic component less than 25% of the total). In the case of predominantly cystic or solid masses, signal intensities in T2-weighted images and DWI and ADC values were respectively registered only on the cystic or solid component. In the case of mixed masses, these parameters were instead registered on both the components. For malignant lesions, the most frequent pattern was the mixed one (25 cases, 71%), followed by the predominantly solid one (10 cases, 29%). No lesion with a predominantly cystic pattern was observed in this group. However, the cystic pattern was the most represented type in the group of benign lesions (26 cases, 45%), followed by mixed (16 cases, 28%) and predominantly solid patterns (15 cases, 27%). In all cases, the examinations were performed with the same 1.5 T magnet (Magnetom Symphony, Siemens Medical Solutions, Erlangen, Germany), which provides a maximum gradient strength of 30 mt m -1 with a peak slew rate of 100 mt m -1 ms -1. Both conventional (before and after the administration of a gadolinium-based contrast medium) and diffusion-weighted sequences were performed. In particular, the study protocol (Table 3) included in all cases: axial scans with the use of fast spin-echo T2-weighted sequences (TR/TE 6680/92, msec) from the renal hilum to the symphysis pubis (echo train length, 16; slice thickness, 5 mm; interval, 1 mm; FOV, mm; number of excitations, 2; matrix, 512 x 276); axial scans with the use of GE T1-weighted sequences acquired with the breath-hold technique (TR/TE 168/4, msec; flip angle, 70 ; number of excitations, 1; slice thickness, 5 mm; interval, 1 mm; FOV, mm; matrix, ), and axial scans with the use of GE T1-weighted sequences acquired with the fat-suppression technique (TR/TE 217/5, msec and parameters similar to those described for previous sequences). To these conventional sequences were added in all cases axial scans acquired with the use of a diffusion-weighted sequence (rapid single-shot echo-planar sequence set with the following parameters: TR/TE, 5490/90, msec; matrix, ; FOV, mm; number of excitations, 3; parallel imaging acceleration SENSE factor, 2; chemical shift selective suppression technique, CHESS) from the renal hilum through the pelvis with a slice thickness of 5 mm and a reconstruction interval of 1 mm. Three values of b (0, 500, 1000 sec/mm 2 ) were used, also applying three orthogonal gradients in the space in order to obtain isotropic images. Acquisition time was near 85 seconds. ADC maps were automatically generated. T1-weighted scans in the axial and sagittal planes with the gradient echo technique after intravenous administration of a paramagnetic contrast medium (Dotarem, Guerbet, Aulnay-sous-Bois, Francia alla dose di 0.2 ml/kg) were Subsequently performed with the following acquisition protocol: TR/TE 165/4, msec; flip angle, 70 ; number of excitations, 1; slice thickness, 5 mm; reconstruction interval, 1 mm; FOV, mm; matrix, ). The slice positioning

4 DWI IN THE CHARACTERIZATION OF COMPLEX ADNEXAL MASSES 213 Table 3 - MRI study protocol. Sequence Scan TR, msec TE, msec ETL Slice thickness, Gap, FOV, NEX Matrix mm mm mm TSE T2-weighted Axial x276 GE T1-weighted fat-sat Axial Echo-planar single-shot Axial (diffusion-weighted) fat-sat GE T1-weighted after gadolinium Axial GE T1-weighted after gadolinium Sagittal GE, gradient echo; TSE, turbo spin-echo; TR, repetition time; TE, echo time; ETL, echo train length; FOV, field of view; NEX, no. of excitations. was identical to that set for the images acquired without contrast medium in order to facilitate a comparison. Evaluation of the images obtained with the use of conventional sequences helped us to define the morphostructural features and the signal intensity of the masses, differentiating predominantly cystic, mixed and predominantly solid masses. Considering the aim of our study, the two observers were mainly required to register the signal intensity within the considered masses in the images acquired with the use of T2-weighted sequences and of diffusion-weighted images with a b value of 1000 sec/mm 2 and to measure the ADC value on the maps. T1-weighted images obtained with and without fat suppression, before and after the administration of the gadolinium-based contrast media, were considered only in cases where the findings at T2- weighted imaging and diffusion-weighted imaging were ambiguous so as to prevent the formulation of a suspicion of benignancy or malignancy. On the basis of previous works available in the literature, signal intensity of the solid and cystic components of the ovarian masses was defined on T2-weighted and diffusion-weighted images in relation to that of the myometrium (of the outer layer of the latter in the case of T2-weighted images). With regard to DWI, the evaluation of signal intensity of the masses was exclusively performed on the images obtained with a b value of 1000 sec/mm 2 (greater weighing of the images in diffusion with less influence of the T2 shine-through effect, T2 dark-through effect and perfusion effect, frequent causes of unexpected signal intensity on the images obtained with lower b values, and better signal-to-noise ratio, with consequent increase in the accuracy of the identification of foci of hyperintensity within suspicious areas) Diffusionweighted images obtained with b values of 0 and 500 sec/mm 2 were used only for the measurement of the ADC 20. Regions of interest were manually traced and placed on both the solid and cystic components so as to be as large as possible, cover a region characterized by homogeneous signal, excluding, as far as possible, areas of necrosis and hemorrhage. Values were registered in 10-3 mm 2 /sec. Histotype was defined within a period not exceeding two weeks from the examination. Statistical analysis was based on the comparison of the categorical variables with Fisher s exact test and of the parametric variables with the Mann-Whitney test. P <0.05 was considered statistically significant. Results Results are summarized in Tables 1 and 2. A cystic component was identified in 71% of malignant lesions and in 73% of benign lesions. The mean region of interest surface area for cystic components was mm 2 for malignant lesions and mm 2 for benign lesions. The cystic component of most ovarian masses (both benign and malignant, 65 cases out of 67, 97%) was hyperintense on T2-weighted images, with no signal on diffusion-weighted images. A different appearance, (hypointensity on T2-weighted images and hyperintensity on diffusion-weighted images), instead, had only 1 decidualized endometrioma and 1 mixed serous-endometrioid adenocarcinoma, inside of which even conventional sequences showed a signal compatible with the presence of blood components, then confirmed at histopathology. No significant difference between the signal intensity on T2-weighted images and diffusionweighted images of the cystic components of benign and malignant masses emerged at Fisher s exact test. Mean ADC values of the cystic component of benign and malignant lesions were respectively 2.32 ± mm 2 /sec and 2.41 ± mm 2 /sec, with no significant difference at the Mann-Whitney test. A solid component was instead identified in all malignant masses and in less than half (42%) of benign masses. The mean region of interest surface area for solid components was mm 2 for malignant lesions and mm 2 for benign lesion. Considering the results obtained by previous authors, in the evaluation of the solid component of the ovarian masses we considered and tested as features suggestive for malignancy the hyperintensity on T2-weighted images and on diffusionweighted images 21. In our series, hyperintensity on T2- weighted images was associated with a positive predictive value of 71%, a negative predictive value of 82%, a sensitivity of 88% and a specificity of 61%. Hyperinten-

5 214 S CAPPABIANCA, F IASELLI, A REGINELLI ET AL sity on diffusion-weighted images was associated with a positive predictive value of 85%, a negative predictive value and a sensitivity of 100% and a specificity of 80%. Regarding the value of the hyperintensity on T2-weighted images, the statistical value of the association with malignancy, verified in 31 of 35 masses subsequently proven malignant (91%, P = at Fisher s exact test), was limited by the hypointensity of 4 malignant lesions (1 borderline mucinous cystadenoma, 1 disgerminoma and 2 out of 3 metastasis, 1 Krukemberg tumor and 1 metastasis from colon cancer) and, primarily, by the hyperintensity of 12 benign lesions. Instead, regarding the value of the hyperintensity on diffusion-weighted images, the statistical value of the association with malignancy, verified in 100% of the cases subsequently proven malignant (P = at Fisher s exact test), was limited by the unexpected hyperintensity of 6 benign lesions (3 fibromas/fibrothecomas, 2 of which were characterized by multiple foci of hypercellularity and fibrous component and 1 complicated by diffuse necro-hemorrhage, 1 endometrioid cystadenoma with necrotic component, 1 decidualized endometrioma with large necrohemorrhagic component, and 1 Brenner tumor with large hyaline component). Similarly as described for the cystic component, we found no significant difference at the Mann-Whitney test between the ADC values of the solid component of malignant (1.03 ± mm 2 /sec) and benign masses (1.13 ± mm 2 /sec), although the latter exhibited lower values. Discussion The role of DWI in the characterization of undetermined lesions was initially proposed for disease of the central nervous system 22 and of the head and neck 23. In these districts, DWI has now gained an established role, being part of the study protocols most commonly used and providing an essential diagnostic contribution in a high percentage of cases. Regarding abdominal imaging, the introduction of DWI instead occurred more slowly and incompletely, mainly due to technical difficulties related to motion (breath, pulsation of the great abdominal vessels, intestinal peristalsis) and chemical shift artifacts 24. Nonetheless, advances in the field of parallel imaging in the last 10 years (in particular the decrease in the effective echo time and the use of fewer echo train lengths, with elimination of the more susceptible components) have contributed to reduce image distortion, increase the signal-to-noise ratio and, more importantly, to shorten the duration of the sequences, making DWI more accurate and more easily applicable to common study protocols for the abdomen 12,25. Several authors have attempted to define, with conflicting results, the role of DWI in the characterization of both the cystic and the solid components of ovarian masses 24,26. Our study, as far as we know, is the first in which a combined evaluation of DWI findings was made on both the cystic and the solid components of ovarian masses. In our opinion, this is extremely useful considering that most of the malignancies of the ovary show a cystic-solid mixed pattern 5,27-29 and that the coexistence of these patterns can significantly complicate the possibility of characterization at conventional MRI, which is often sufficient in cases of purely cystic or solid lesions 20. In our study, retrospective evaluation of the features of cystic components on T2- and diffusion-weighted images was not a useful tool for diagnosis formulation. In fact, in most of the cases, both the benign and the malignant masses shared the hyperintensity typically associated with fluid on T2-weighted images and the absence of signal on diffusion-weighted images. Only two lesions with a cystic component, 1 mixed serousendometrioid adenocarcinoma and 1 decidualized endometrioma, showed opposite behavior compared to the remaining cases, with hyperintensity on T2-weighted images, hyperintensity on DWI and low ADC values. Such features are related to the histopathological substrate of the lesions. In fact, in both cases, within the cysts there was a blood component that, as is known, causes a reduction in both the T1 and T2 relaxation times (at additional evaluation with T1-weighted images, both masses were hyperintense) and a restriction of the diffusivity of water molecules. Therefore, irrespective of the histotype of the mass, actually determinant for the signal in diffusion-weighted images and ADC value of a cystic component is its content and, more generally, its degree of viscosity (in the two considered lesions, for example, the paramagnetic effect of methemoglobin was combined with a high protein concentration) 12,13. More complicated, instead, is the discourse about the solid components of the expanded ovary. Most evidence in the literature indicates hyperintensity on T2- and diffusion-weighted images as features suggestive for malignancy of a solid component (Figure 1). In effect, in our series only 4 of the 35 malignant lesions (1 disgerminoma, 1 Krukemberg s tumor, 1 metastasis from adenocarcinoma of the colon and 1 borderline mucinous cystadenoma) showed an abnormal behavior, resulting hypointense to the outer layer of the myometrium on T2-weighted images. In all 4 cases, histopathology revealed an abundant desmoplastic component, responsible for the shortening of the T2 relaxation time and, at the same time, for the restriction of the Brownian motions of water molecules 1. The latter feature is responsible for the signal on diffusionweighted images and for the measured ADC value, which still induced the two observers to formulate an opinion of malignancy, also on the basis of the marked enhancement shown on images acquired after administration of the contrast medium 21,30. Similarly, knowl-

6 DWI IN THE CHARACTERIZATION OF COMPLEX ADNEXAL MASSES A 215 B D C Figure 1 - Right ovarian carcinosarcoma. The mass is T2 hyperintense, DWI hyperintense with ADC = mm2/sec. Histopathology revealed a hypercellular lesion. A) axial T2-image. B) axial diffusionweighted image (b = 1000). C) ADC map. D) 200 original magnification of the carcinosarcoma. The photomicrograph shows the malignant epithelial component of the tumor with densely packed cellular elements. Next to these cells are the fluid and stromal components, quantitatively lower than the epithelial component. edge of the histopathological features of the solid components of 6 of the 31 benign lesions (presence of multiple foci of hypercellularity, necro-hemorrhage and hyaline and fibrous component) allowed us to understand how a normal hypointensity on T2-weighted images could be associated with an unexpected hyperintensity on DWI31. Considering the two situations described above, in particular the discrepancy between the appearance of several lesions on T2-weighted images and in DWI, the importance of an integrated assessment of the findings provided by these two techniques is clear and, where necessary, the evaluation of images from conventional sequences before and after paramagnetic contrast medium injection. As already stated, evaluation of our series did not reveal a significant difference between the ADC values of the cystic component between benign and malignant masses. This result is in line with those obtained by most other authors11,25,32-34, whereas from the evaluation of only a few series emerged a significant difference, with ADC values of the cystic components of the malignant lesions lower than those of the benign lesions6, Regarding the solid component, then, although the values associated with malignancy in gener-

7 216 S CAPPABIANCA, F IASELLI, A REGINELLI ET AL al have proved to be lower as already widely described by other authors 21,34, we noted a considerable overlap with the coefficients exhibited by benign masses associated with restriction to the Brownian motions (lesions composed of spindle cells, hyaline tissue, desmoplastic tissue, necro-hemorrhage), with consequently decreased statistical significance. To make more complex the further integration of the parameters obtained from the use of T2-weighted sequences and DWI with ADC ratios within the solid components was the influence on the DWI signal of the socalled T2 dark-through effect. Similarly to what happens for the best known T2 shine-through effect, the hypointensity of a lesion in T2 also affects its appearance on DWI (low signal), even remaining an ADC value suggestive for reduced water diffusivity. This finding, a potential pitfall to the correct characterization of a solid mass through DWI already reported by Bakir et al. 21 and Fujii et al. 34 in previous studies, occurred in 2 considered fibromas/fibrothecomas, the former characterized by multiple foci of hypercellularity, the latter by intralesional necrosis 38. Conclusions Still today, and probably for many years to come, histopathology is the only completely reliable tool able to provide a definitive diagnosis of the nature of a complex ovarian mass. From our experience, as from the works of other authors, however, emerge encouraging results regarding the possibility to characterize with a high degree of accuracy masses as benign or malignant through the integration of conventional MRI and DWI, thus limiting the number of unnecessary surgeries. A significant correlation between the combined hyperintensity in T2-weighted and diffusion-weighted images and malignancy exists. Instead, in dubious cases (low signal in one of the two sequences), the integration with conventional imaging findings, also obtained after the administration of a paramagnetic contrast agent, may resolve the diagnostic dilemma. Much less useful, instead, is the evaluation through T2-weighted and diffusion-weighted sequences on cystic components of the masses, whose characteristics are influenced rather than by their histotype, by their content. Not very reliable, moreover, is the diagnostic value of the ADC, with no significant difference between the cystic components of benign and malignant lesions and a significant overlap in the values of the solid components of benign and malignant lesions. References 1. Takeuchi M, Matsuzaki K, Nishitani H: Diffusion-weighted magnetic resonance imaging of ovarian tumors: differentiation of benign and malignant solid components of ovarian masses. J Comput Assist Tomogr, 34: , Gatta G, Parlato V, Di Grezia G, Porto A, Cappabianca S, Grassi R, Rotondo A: Ultrasound-guided aspiration and ethanol sclerotherapy for treating endometrial cysts. Radiol Med, 115: , Jung SE, Lee JM, Rha SE, Byun JY, Jung JI, Hahn ST: CT and MR imaging of ovarian tumors with emphasis on differential diagnosis. Radiographics, 22: , Griffin N, Grant LA, Sala E: Adnexal masses: characterization and imaging strategies. Semin Ultrasound CT MRI, 31: , Imaoka I, Wada A, Kaji Y, Hayashi T, Hayashi M, Matsuo M, Sugimura M: Developing an MR imaging strategy for diagnosis of ovarian masses. Radiographics, 26: , Sarty GE, Kendall EJ, Loewy J, Dhir A, Olatunbosun OA, Pierson RA: Magnetic resonance diffusion imaging of ovarian masses: a first experience with 12 cases. MAGMA, 16: , Thomassin-Naggara I, Fournier LS, Roussel A, Marsault C, Bazot M: IRM de diffusion et pelvis feminin. J Radiol, 91: , Thomassin-Naggara I, Touissant I, Perrot N, Rouzier R, Cuenod CA, Bazot M, Darai E: Characterization of complex adnexal masses: value of adding perfusion- and diffusionweighted MR imaging to conventional MR imaging. Radiology, 258: , Malayeri A, El Khouli RH, Zaheer A, Jacobs MA, Corona- Villalobos CP, Kamel IR, Macura KJ: Principles and applications of diffusion-weighted imaging in cancer detection, staging and treatment follow-up. Radiographics, 31: , Yoshikawa T, Kawamitsu H, Mitchell DG, Ohno Y, Ku Y, Seo Y, Fujii M, Sugimura K: ADC measurement of abdominal organs and lesions using parallel imaging technique. Am J Roentgenol, 187: , Nakayama T, Yoshimitsu K, Irie H, Aibe H, Tajima T, Nishie A, Asayama Y, Matake K, Kakihara D, Matsuura S, Nakano H, Honda H: Diffusion-weighted echo-planar MR imaging and ADC mapping in the differential diagnosis of ovarian cystic masses: usefulness of detecting keratinoid substances in mature cystic teratomas. J Magn Reson Imaging, 22: , Namimoto T, Awai K, Nakaura T, Yanaga Y, Hirai T, Yamashita Y: Role of diffusion-weighted imaging in the diagnosis of gynecological disease. Eur Radiol, 19: , Kinoshita T, Ishii K, Naganuma H, Higashiiwai H: MR findings of ovarian tumors with cystic components. Br J Radiol, 73: , Kier R, Smith RC, McCarthy SM: Value of lipid- and watersuppression MR images in distinguishing between blood and lipid within ovarian masses. Am J Roentgenol, 158: , Morita S, Kojima S, Hirata M, Suzuki K, Ueno E: Perfusion fraction of diffusion-weighted MRI for predicting the presence of blood supply in ovarian masses. J Magn Reson Imag, 34: , Stevens SK, Hricak H, Campos Z: Teratomas versus cystic hemorrhagic adnexal lesions: differentiation with protonselective fat-saturation MR imaging. Radiology, 186: , Outwater EK, Siegelman ES, Hunt JL: Ovarian teratomas: tumor types and imaging characteristics. Radiographics, 21: , Chilla B, Hauser N, Singer G, Trippel M, Froehlich JM, Kubik-Huch RA: Indeterminate adnexal masses at ultrasound: effect of MRI imaging findings on diagnostic think-

8 DWI IN THE CHARACTERIZATION OF COMPLEX ADNEXAL MASSES 217 ing and therapeutic decisions. Eur Radiol, 21: , Feuerlein S, Pauls S, Juchems MS, Stuber T, Hoffmann MHK, Brambs H, Ernst AS: Pitfalls in abdominal diffusionweighted imaging: how predictive is restricted water diffusion for malignancy. Am J Roentgenol, 193: , Li W, Chu C, Cui Y, Zhang P, Zhu M: Diffusion-weighted MRI: a useful technique to discriminate benign versus malignant ovarian surface epithelial tumors with solid and cystic components. Abdom Imag doi: /s x, Bakir B, Bakan S, Tunaci M, Bakir VL, Iyibozkurt AC, Berkman S, Bengisu E, Salmaslioglu A: Diffusion-weighted imaging of solid or predominantly solid gynaecological adnexial masses: is it useful in the differential diagnosis? Br J Radiol, 84: , Tsuruda JS, Chew WM, Moseley ME, Norman D: Diffusionweighted MR imaging of the brain: value of differentiating benign extraxial cysts and epidermoid tumors. Am J Roentgenol, 155: , Wang J, Takashima S, Takayama F, Kawakami S, Saito A, Matsushita T, Momose M, Ishiyama T: Head and neck lesions: characterization with diffusion-weighted echo-planar MR imaging. Radiology, 220: , Punwani S: Diffusion weighted imaging of female pelvic cancers: concepts and clinical applications. Eur J Radiol, 78: 21-29, Levy A, Medjhoul A, Caramella C, Zareski E, Berges O, Chargari C, Boulet B, Bidault F, Dromain C, Balleyguier C: Interest of diffusion-weighted echo-planar MR imaging and apparent diffusion coefficient mapping in gynecological malignancies: a review. J Magn Reson Imag, 33: , Whittaker CS, Coady A, Culver L, Rustin G, Padwick M, Padhani AR: Diffusion-weighted MR imaging of female pelvic tumors: a pictorial review. Radiographics, 29: , Sohaib SA, Sahdev A, Van Trappern P, Jacobs IJ, Reznek RH: Characterization of adnexal mass lesions on MR imaging. Am J Roentgenol, 180: , Hricak H, Chen M, Coakley FV, Kinkel K, Yu KK, Sica G, Bacchetti B, Powell CB: Complex adnexal masses: detection and characterization with MR imaging-multivariate analysis. Radiology, 214: 39-46, Thomassin-Naggara I, Darai E, Cuenod CA, Fournier L, Toussaint I, Marsault C, Bazot M.: Contribution of diffusion-weighted MR imaging for predicting benignity of complex adnexal masses. Eur Radiol, 19: , Sala E, Rockall A, Rangarajan D, Kubik-Huch RA: The role of dynamic contrast-enhanced and diffusion weighted magnetic resonance imaging in the female pelvis. Eur J Radiol, 76: , Troiano RN, Lazzarini KM, Scoutt LM, Lange RC, Flynn SD, McCarthy S: Fibroma and fibrotechoma of the ovary: MR imaging findings. Radiology 204: , Kaji Y, Matsuo M, Matsuki M, Yoshida M, Hayashi M, Nanno H, Koshiyama M, Fujii H, Maruyama K, Takizawa O, Sugimura K: Cystic ovarian lesions in SSFP diffusion imaging. Magn Reson Med Sci, 1: , Katayama M, Masui T, Kobayashi S, Ito T, Sakahara H, Nozaki A, Kabasawa H: Diffusion-weighted echo planar imaging of ovarian tumors: is it useful to measure apparent diffusion coefficients? J Comput Assist Tomogr, 26: , Fujii S, Kakite S, Nishihara K, Kanasaki Y: Diagnostic accuracy of diffusion-weighted imaging in differentiating benign from malignant ovarian lesions. J Magn Reson Imag, 28: , Moteki T, Ishizaka H: Evaluation of cystic ovarian lesions using apparent diffusion coefficient calculated from reordered turboflash MR images. Magn Reson Imaging, 17: , Moteki T, Ishizaka H: Diffusion-weighted EPI of cystic ovarian lesions: evaluation of cystic contents using apparent diffusion coefficients. J Magn Reson Imaging, 12: , Moteki T, Horikoshi H, Endo K: Relationship between apparent diffusion coefficient and signal intensity in endometrial and other pelvic cysts. Magn Reson Imaging, 20: , Maldjian JA, Listerud J, Moonis G, Sidiqi F: Computing diffusion rates in T2-dark hematoma and areas of low T2 signal. Am J Neuroradiol, 22: , 2001.