Evaluation of cystic ovarian lesions using apparent diffusion coefficient calculated from reordered TurboFLASH MR images

Magnetic Resonance Imaging, Vol. 17, No. 7, pp. 955–963, 1999 © 1999 Elsevier Science Inc. All rights reserved. Printed in the USA. 0730-725X/99 $–see front matter

PII S0730-725X(99)00036-3

● Original Contribution

EVALUATION OF CYSTIC OVARIAN LESIONS USING APPARENT DIFFUSION COEFFICIENT CALCULATED FROM REORDERED TURBOFLASH MR IMAGES TAKAO MOTEKI

AND

HIROSHI ISHIZAKA

Department of Diagnostic Radiology, Gunma University Hospital, 3-39-15, Showa-machi, Maebashi, 371 Gunma, Japan Reordered snapshot fast low-angle shot images with, and without, diffusion-perfusion gradients were used for the evaluation of contents of cystic ovarian lesions. Sonographically detected 51 cystic ovarian lesions (13 endometrial cysts, 17 ovarian cysts, 7 serous cystadenomas, 6 mucinous cystadenomas, 8 malignant cystic ovarian tumors) were studied. T1- and T2-weighted images, reordered snapshot fast low-angle shot images with and without diffusion-perfusion gradients (b ⴝ 106 and 0 s/mm2, respectively) were obtained. Using these images, apparent diffusion coefficients (ADCs) were calculated in the cystic contents of these lesions. Endometrial cysts and malignant cystic ovarian tumors showed lower ADC values than ovarian cysts, serous cystadenomas and mucinous cystadenomas (p < 0.02). There was no distinct ADC difference among ovarian cysts, serous cystadenomas, mucinous cystadenomas (p > 0.2). In conclusion, diffusion-weighted magnetic resonance imaging is possible to be useful to evaluate cystic contents of ovarian lesions. © 1999 Elsevier Science Inc. Keywords: Rapid MRI; Pelvis; Neoplasm; Pelvis; Diffusion-perfusion.

Diffusion-weighted magnetic resonance (MR) imaging is sensitive to molecular diffusion, which is due to random, microscopic translational motion of molecules (known as Brownian motion), since random motion in the field gradients produces incoherent phase shifts that result in signal attenuation.1 In the central nervous system, diffusion-weighted and perfusion MRI hold significant promise in detection of brain ischemia at a very early stage, and the noninvasive evaluation of normal brain function and functional disorders.2 Although this technique is difficult to apply to abdominal organs for the reason that it is highly sensitive to motion artifacts, diffusion-weighted echo-planar imaging for hepatic tumor evaluation was recently reported, and concluded that the diffusion-weighted imaging might be useful to evaluate hepatic tumors with marked hyperintensity on T2-weighted images.3 Snapshot fast low-angle shot (FLASH) MR imaging is a method of obtaining image in sub-second acquisition time. This technique is called turboFLASH imaging, and it involves both a preparation period, and a FLASH

sequence with a very short TR (less than 10 ms). It can be performed with standard MR imaging systems. The use of turboFLASH with a preparation pulse sensitive to diffusion (diffusion-weighted turboFLASH) allows diffusion-weighted images to be obtained during a single breath-hold interval, and can greatly reduce the risk of motion artifacts.4,5 It is primarily the ability of MR imaging to accurately identify fat and hemorrhage, and MR imaging can make more precise diagnosis than ultrasound (US) or computed tomography (CT) in the identification of dermoid or endometrial cysts.6,7 However, MR imaging is probably no more accurate than US or CT in differentiation among another cystic ovarian neoplasms, for the reason that no morphologic or tissue characteristics have been described that are exclusive for conventional MR imaging.8,9 We expected that reordered turboFLASH with and without diffusion-perfusion (DP) gradients might be effective for the evaluation of cystic ovarian lesions, in the view that various contents of cystic ovarian tumors may show different apparent diffusion pattern. In the present study, we attempted to evaluate contents of cystic ovarian lesions, using apparent diffusion coefficient (ADC) values calcu-

ACCEPTED 10/10/98; ACCEPTED 3/20/99. Address correspondence to Dr. T. Moteki, Department of

Diagnostic Radiology, Gunma University Hospital, 3-39-15, Showa-machi, Maebashi, Gunma 371, Japan.

INTRODUCTION

955

Magnetic Resonance Imaging ● Volume 17, Number 7, 1999

956

Table 1. Results of ADC values in cystic contents of ovarian lesions

Endometrial cyst Ovarian cyst * The lowest ADC loculus † The highest ADC loculus Serous cystadenoma * The lowest ADC loculus † The highest ADC loculus Mucinous cystadenoma * The lowest ADC loculus † The highest ADC loculus Malignant cystic ovarian neoplasms * The lowest ADC loculus † The highest ADC loculus

No.

Mean diameter (cm)

ADC (10⫺3 mm2/s)

13 17

5.5 ⫾ 3.6 4.1 ⫾ 2.9

4.59 ⫾ 3.15

7

7.7 ⫾ 1.4

6

13.2 ⫾ 7.4

8

13.0 ⫾ 5.67

14.0 ⫾ 7.34 14.3 ⫾ 7.34 13.6 ⫾ 4.74 14.7 ⫾ 4.27 10.5 ⫾ 3.25 16.2 ⫾ 3.35 6.06 ⫾ 1.86 8.92 ⫾ 3.34

Means ⫾ standard deviation. * If a tumor has multilocular cystic compartments, the lowest ADC content among loculi was selected as a representative. † If a tumor has multilocular cystic compartments, the highest ADC content among loculi was selected as a representative.

lated from reordered turboFLASH images with and without DP gradients. MATERIALS AND METHODS We performed turboFLASH imaging with, and without, DP gradients in 48 patients (20 – 82 [mean, 46] years old) 51 cystic ovarian lesions (13 endometrial cysts, 17 ovarian cysts, 7 serous cystadenomas, 6 mucinous cystadenomas, and 8 malignant cystic ovarian tumors) were studied. These lesions were sonographically detected and referred for MR imaging. Twenty-nine ovarian cystic lesions were surgically or endoscopically removed and diagnosed by pathologic examination, as follows; 7 endometrial cysts, 1 ovarian cyst (serous inclusion cyst), 7 serous cystadenomas, 6 mucinous cystadenomas, and 8 malignant ovarian tumors (2 serous cystadenocarcinomas, 2 clear cell carcinomas, 1 endometrioid carcinoma, and 3 mucinous cystadenocarcinomas). Sixteen ovarian cystic lesions were classified as ovarian cysts which are all satisfied all following conditions: 1) no irregularly thickened wall nor enhancing solid component; 2) decreased or stability in size on serial US scans for a minimum of 6 months (mean, 11 months); and 3) no obvious fatty nor hemorrhagic components on T1- and T2-weighted MR images (lower intensity on T1-weighted images and higher intensity on T2-weighted images compared with the adjacent uterine myometrium). Six ovarian cystic lesions were classified as endometrial cysts without operation which were satisfied following conditions: 1) high-intensity contents on T1-weighted images and dissimilar intensity to subcutaneous (s.c.) fat on either T1- or T2-weighted images; 2) having hypointense capsule and septations on T1- and T2-weighted images; and 3) if inhomogeneity area in the lesion,

enhancing solid components were not associated except capsule and septations. We had five cases with dermoid cysts. However, all but one case had abundant sebaceous tissue, and did not show measurable cystic components in them. We stopped evaluation for the group, and excluded them from this study. MR imaging was performed with a 1.0-T magnet and shielded gradients system (Impact: Siemens, Erlangen, Germany). Conventional T1-weighted (600/15 [TR/TE]) spin-echo (SE), and T2-weighted (3,500/90) fast spinecho (FSE) images (350-mm FOV, 256 ⫻ 140 matrix, 5-mm slice thickness, 2 acquisitions) were also obtained in all patients. In 37 patients T1-weighted MR images were also obtained after the i.v. administration of 0.1 mmol/kg gadopentetate dimeglumine (Gd-DTPA) (Mag-

Fig. 1. Scatter plots of ADC values of contents of cystic ovarian lesions. Plots of the highest and lowest ADC loculi at each multiloculated lesion are linked with a line. Data of each group are arranged in size order. Cutoff values used for ADC differentiation between the benign and malignant lesions are presented as dotted lines.

ADC evaluation of cystic ovarian lesions ● T. MOTEKI

ET AL.

957

Fig. 2. Endometrial cyst and serous cystadenoma. A, T1-weighted image shows anteriorly situated serous cystadenoma as mild higher intensity and posteriorly situated endometrial cyst as marked higher intensity than the myometrium. B, T2-weighted image shows serous cystadenoma as marked high intensity and endometrial cyst as mildly lower intensity than the myometrium. Adenomyosis detected as hypointensity is visible in the posterior myometrium. By comparison between the reordered turboFLASH images without DP gradients (C) and with DP gradients (D), there is only mild signal attenuation in the endometrial cyst on the latter image despite marked signal loss in the serous cystadenoma.

nevist, Schering AG, Berlin, Germany). The enhanced study was performed after the SE and FSE, and turboFLASH images scans. Diffusion-weighted turboFLASH imaging was ob-

tained with a preparation pulse; (90°–180°–90°), modified by adding large DP gradients on both sides of the 180° pulse along the readout axis. The preparation pulse was followed by a FLASH sequence with a very short

958

Magnetic Resonance Imaging ● Volume 17, Number 7, 1999

Fig. 3. Ovarian cyst that was unchanged in size and morphologic features during 2 years follow up, sonographically. A, Post-contrast T1-weighted image and B, T2-weighted image show large cystic lesion with mural cyst and smoothly thickened septation. Reordered turboFLASH image without DP gradient. (C) reveals these cysts and bladder as homogeneous high intensity. On reordered turboFLASH image with DP gradients; (D) significant signal attenuation is noted in them all.

TR. Centric reordered sampling of k-spaces was used in the sequence with a TR ms/TE ms of 8.8/4, and 64 lines of data acquired. Single scan time of the turboFLASH is in 0.73 s. The other parameters were a flip angle of 12°, a FOV of 40 cm, a 64 ⫻ 128 matrix, a 15-mm section thickness, and 13 times scans with 1.5 s intervals. Av-

eraged image was made from these last 12 images (excluded the first image to unify the each scan interval for the reason that the first image has a infinite scan interval). Each case involved a series of sequences obtained with and without DP gradients. Total scan time of these 13 turboFLASH images is 19 s, and we instructed patients

ADC evaluation of cystic ovarian lesions ● T. MOTEKI

ET AL.

959

Fig. 4. Mucinous cystadenoma (A). Post-contrast T1-weighted image shows a bilocular cystic lesion with focally thickened capsule and septation (arrows). The reordered turboFLASH image without DP gradient (B) and turboFLASH image with DP gradients (C) demonstrate significant signal attenuation on the latter image.

to hold their breath during the scan. These turboFLASH images were scanned at the same location of the slice showing the maximum diameter of an ovarian lesion on sagittal T2-weighted images. The serial turboFLASH images were visually evaluated at the same window and level setting.

After these all sequences were finished, the signal intensities were measured in the cystic contents of the lesions on the set of turboFLASH images, T1- and T2weighted images. A circular region-of-interest (ROI) was placed in the cystic content of a unilocular lesion or each loculi of multilocular lesion which was more than 2.0 cm

Fig. 5. Mucinous cystadenocarcinoma. A, Post-contrast T1-weighted image shows the ovarian tumor containing non-enhanced central cystic component and marginally situated enhancing solid components. A reordered turboFLASH image without DP gradient (B) and a turboFLASH image with DP gradients (C) demonstrate only mild signal change at the cystic content between these two images, although the bladder shows significant signal attenuation on the latter image.

Magnetic Resonance Imaging ● Volume 17, Number 7, 1999

960

Table 2. Results of morphologic features, signal behavior and ADC classification

Histologic result Serous cystadenoma Serous cystadenoma Serous cystadenoma Serous cystadenoma Serous cystadenoma Serous cystadenoma Serous cystadenoma Mucinous cystadenoma Mucinous cystadenoma Mucinous cystadenoma Mucinous cystadenoma Mucinous cystadenoma Mucinous cystadenoma Mucinous cystadenocarcinoma Endometroid carcinoma Mucinous cystadenocarcinoma Clear cell carcinoma Mucinous cystadenocarcinoma Serous cystadenocarcinoma Serous cystadenocarcinoma Clear cell carcinoma

Wall or Enhanced septation Signal Signal The highest The lowest solid thickness intensity intensity Morphologic ADC loculus ADC loculus Size component (⬎ 3 mm) on T1 WI* on T2 WI* estimation subgroup subgroup 5.7 6.8 7.1 7.8 7.9 9 9.7 4.8 10.0 10.2 10.8 17.5 26.0 5.5 8.2 8.2 9.9 15.8 17.8 18.3 20.4

Non Non Non Non Non Non Non Non Non Non Non Non Non Present Present Present Present Present Present Present Present

Non Non Present Non Non Non Non Non Present Present Non Non Non Present Present Present Present Present Non Non Present

Low High Low-high Low Low Low Low Low Low Low Low Low Low-high High High Low-high Low Low High High Low

High High High High High High High High High High High High High High High High High High High High High

Benign Benign Malignant Benign Benign Benign Benign Benign Malignant Malignant Benign Benign Benign Malignant Malignant Malignant Malignant Malignant Malignant Malignant Malignant

Benign Benign Benign Malignant Benign Benign Malignant Benign Benign Malignant Benign Benign Benign Malignant Malignant Malignant Malignant Malignant Malignant Benign Malignant

Benign Benign Benign Benign Benign Benign Benign Benign Benign Benign Benign Benign Benign Malignant Malignant Malignant Malignant Benign Malignant Malignant Malignant

Note: WI ⫽ weighted images. * Comparison of signal intensity with the uterine myometrium, low-high is representing that the lesion have both high and low intensity loculi.

in a diameter at their cystic components, and outer margin of the ROI was kept at least 5 mm away from the capsule, septation and solid components (enhanced components were regarded as solid components and excluded from ROI) using set of T1-weighted images with, and without, administration of Gd-DTPA, as possible. As in the diffusion-weighted SE technique,2 the ADC for the serial turboFLASH sequences is calculated according to ADC ⫽ (1/b) ln(S/S0); where S0 and S are signal intensities in the ROI, obtained with, and without, the DP gradients, respectively. b ⫽ ␥2G2␦2(⌬ ⫺ ␦/3); where ␥ is the gyromagnetic ratio (42.576 MHz/T), G is the amplitude of the DP gradient pulses (0 and 4.8 mT/m), ␦ is the duration of each gradient pulse (41 ms), ⌬ is the interval between the onset of separating DP gradient pulses, (52 ms) and b is the diffusion sensitivity factor (0 and 106.3 s/mm2). To evaluate ADC difference between the groups of ovarian lesions, we made two subgroups if the multiloculated lesions were present in a group, as follows; 1) the highest ADC loculus subgroup that consists of the highest ADC loculi of multiloculated lesions and unilocular lesions and 2) the lowest ADC loculus subgroup which consists of the lowest ADC loculi of multiloculated lesions and unilocular lesions (overlapping to the highest ADC loculus subgroup). ADC comparisons were per-

formed both between the highest ADC loculus subgroups, and between the lowest ADC loculus subgroups, when either group had multilocutaed lesions. To determine the statistical significance of difference of calculated or measured values between the two groups of cystic ovarian lesions, we used student t test when normal distribution of the data (p ⬎ 0.05 on F analysis). When not, we used Welch t test. Differences were statistically significant for p values less than 0.05 on these t tests. In serous cystadenomas, mucinous cystadenomas and malignant cystic ovarian tumors which were all pathologically confirmed, differentiation between the benign and malignant lesions using conventional MR imaging and ADC values were compared, retrospectively. Benign lesions were composed of serous cystadenomas and mucinous cystadenomas, and malignant lesions were malignant cystic ovarian tumors. A morphologic diagnosis of malignancy was indicated at conventional MR imaging if the following two of three criteria were present10: 1) diameter of more than 4 cm; 2) thickness of the wall or septa of more than 3 mm in a cystic lesion; and 3) presence of nodularity, vegetations, or a large solid component. To describe signal intensity characteristics of these lesions in these three groups, uterine myometrium was

ADC evaluation of cystic ovarian lesions ● T. MOTEKI

used as reference tissue. The signal intensity of the cystic content or each loculi at the identical ROI to the turboFLASH images was compared with adjacent myometrium on T1- and T2-weighted images. Discriminant analysis was applied to obtain the best ADC differentiation between the benign and malignant lesions, and to perform comparative study with morphologic differentiation. RESULTS Mean and standard deviation of the calculated ADC values are presented in Table 1. Scatter plots of ADC values of cystic ovarian lesions are demonstrated on Fig. 1. Ovarian cysts and endometrial cysts are smaller than the other groups (p ⬍ 0.01), except between endometrial cysts and serous cystadenomas (p ⬎ 0.07). There are no significant size differences among serous cystadenomas, mucinous cystadenomas and malignant cystic neoplasms (p ⬎ 0.06), except between serous cystadenomas and malignant cystic ovarian tumors (p ⬍ 0.05). Endometrial Cysts The calculated ADC values of the endometrial cysts (Fig. 2) had a range of 0.46 –12.3 ⫻ 10⫺3 mm2/s. Endometrial cysts showed lower ADC values than any other groups (p ⬍ 0.01), except for the malignant cystic ovarian tumors when the lowest loculi subgroup was selected as a representative (p ⬎ 0.25). Ovarian Cysts The calculated ADC values of the ovarian cysts (Fig. 3) had a range of 6.1–28.3 ⫻ 10⫺3 mm2/s. There were no significant ADC difference among ovarian cysts, serous cystadenomas and mucinous cystadenomas (p ⬎ 0.20). Ovarian cysts showed higher ADC values than malignant cystic ovarian neoplasms, if the highest or lowest ADC loculi subgroups were representative of these lesions (p ⬍ 0.02). Serous Cystadenomas The calculated ADC values of the serous cystadenomas (Fig. 2) had a range of 8.7–20.7 ⫻ 10⫺3 mm2/s. This group showed higher ADC values than malignant cystic ovarian neoplasms, if the highest or lowest ADC loculi subgroups were representative of them (p ⬍ 0.02). Mucinous Cystadenomas The calculated ADC values of the mucinous cystadenomas (Fig. 4) had a range of 8.5–20.6 ⫻ 10⫺3 mm2/s. This group showed lower ADC values than malignant cystic ovarian neoplasms, when the lowest ADC loculi or the highest ADC loculi subgroups were representative of them (p ⬍ 0.01).

ET AL.

961

Malignant Cystic Ovarian Tumors The calculated ADC values of the malignant cystic ovarian tumors (Fig. 5) had a range of 3.8 –16.1 ⫻ 10⫺3 mm2/s. Standard Imaging Features and ADC Patterns in Serous Cystadenomas, Mucinous Cystadenomas and Malignant Cystic Ovarian Tumors When the lowest loculi subgroup was selected as a representative, the cutoff value of ADC between the benign and malignant lesions obtained from discriminant analysis was 0.0083 mm2/s. When the highest loculi subgroup was selected, the cutoff value of ADC between the benign and malignant lesions was 0.0117 mm2/s. Standard imaging features (signal behavior and morphologic features), and ADC patterns in operatively proved serous cystadenomas, mucinous cystadenomas and malignant cystic ovarian tumors were presented in Table 2. A diagnosis with MR images and ADC values of benign tumors versus malignant tumors were presented as follows. With conventional MR imaging, sensitivity was 100% (8 of 8 tumors), specificity was 77% (10 of 13), positive predictive value was 73% (8 of 11), and negative predictive value was 100% (8 of 8). With ADC values of the highest loculi subgroups, sensitivity was 88% (7 of 8 tumors), specificity was 77% (10 of 13), positive predictive value was 70% (7 of 10), and negative predictive value was 91% (10 of 11). With ADC values of the lowest loculi subgroups, sensitivity was 87.5% (7 of 8 tumors), specificity was 100% (13 of 13), positive predictive value was 100% (7 of 7), and negative predictive value was 93% (13 of 14). For reference, if cystic lesions with higher intensity content or a loculus than uterine myometrium on T1weighted images were classified as malignancy, sensitivity was 62.5% (5 of 8 tumors), specificity was 77% (10 of 13), positive predictive value was 63% (5 of 8), and negative predictive value was 77% (10 of 13). Two of six mucinous cystadenomas (Fig. 4) and one of seven serous cystadenoma which were morphologically classified malignancy were correctly distinguish as benign on ADC classification when the lowest ADC loculi subgroups were selected. Such case in which ADC was more advantageous than morphologic features was also present in one ovarian cyst (Fig. 3) showed higher ADC than the cut-off values, although the lesion was not pathologically proved. By contrast, 1 of 8 malignant cystic ovarian tumors that was morphologically classified malignancy was misdiagnosed as benign on ADC classification when the lowest or the highest ADC loculi subgroups were selected.

962

Magnetic Resonance Imaging ● Volume 17, Number 7, 1999

DISCUSSION MR imaging has been shown a high degree of diagnostic specificity for certain types of ovarian masses, such as dermoid cyst and endometrial cyst. However, there is no specific part for conventional MR imaging to characterize another cystic ovarian tumors. Namely, though the presence of solid tissue, septations, indistinct margins, mural nodules, ascites, or a combination of these makes an ovarian mass more likely to be malignant,10 –12 these features can be obtained on another modality such as US and CT8,9. On the contrary, apparent diffusion in cystic ovarian lesions cannot be evaluated on any other modalities. There were no significant size difference between ovarian cysts and endometrial cysts, and between mucinous cystadenomas and malignant cystic ovarian tumors. However, endometrial cysts and malignant ovarian tumors tended to show larger ADC values than ovarian cysts, serous cystadenomas and mucinous cystadenomas, when the highest or the lowest ADC loculi subgroups were representative of them. These results are representing that ADC difference could occur even between size equivalent groups such as endometrial cysts and ovarian cysts, and that the ADC difference could be caused from the cystic contents difference. The ADC differences in the cystic contents may be mainly caused by the difference of viscosity,13 and/or the difference of concentration of protein (such as lysozyme, albumin and fibrinogen), sugar and nucleic acid14 –16 in the contents. The possibility of usefulness of ADC value to evaluate cystic ovarian lesion is that among larger cystic ovarian lesions, malignant cystic ovarian tumors are apt to have lower ADC contents than serous or mucinous cystadenomas, although there are some ADC overlapping among loculi of them. The problem inherent in the turboFLASH technique using sequential phase encoding should be considered that T1 relaxation could occur during the time of sequential acquisition of the phase-encoding steps.4 The differential diffusion effect acquired at the end of the preparation period may be diminished for tissues with short T1 values in a moment, due to the faster longitudinal magnetization recovery (T1 contamination); ADC value at higher intense area on T1-weighted image is possible to be underestimated. In this regard, it is known that contents of endometrial cysts, mucinous cystic neoplasms and malignant cystic neoplasms occasionally have higher intensity than serous cystadenomas or ovarian cysts on T1-weighted images. Therefore, though there were ADC differences between them, ADC values may be modified in some degree by the T1 contamination, and this difference may, if present, be spuriously larger. To minimized T1 contamination from turboFLASH images, we introduce centric-reordered acquisition of k spaces to turbo-

FLASH, instead of sequential k space sampling. This method enables us to first obtain central k spaces which are mainly related to image contrast in MR images and diminished T1 contamination. Some images of turboFLASH images (Figs. 2D, 5C, and 5D) were compromised by the appearance of artifactual bands parallel to the frequency encode direction. These band structures are supposed to be caused by resonant offset angle differences among successive repetition intervals in the turboFLASH sequence.4,17 This imaging technique that uses body coil on a 1.0 tesla system, however, reveals lower soft tissue resolution, and averaging of images should be required. In the event of morphologic evaluation of cystic ovarian tumors, whether benign or malignancy, sensitivity for malignancy is adequately high. However, some benign cystic ovarian tumors having thickened capsules or septations which were morphologically diagnosed as malignancy were correctly diagnosed on ADC values when the lowest ADC loculus subgroups were used (Table 2, Figs. 3 and 4). The ADC differentiation showed higher sensitivity and specificity than signal intensity classification. These results show the possibility that diffusionweighted MR imaging might be one more diagnostic tool in addition to conventional morphologic features of cystic ovarian lesions, although our study was favorably designed for ADC values in the point that the cutoff value for ADC was retrospectively defined, and the number of studied cystic ovarian lesions is still small. Further clinical trials are needed to establish the limits of this technique, and its precise role in differential diagnosis of cystic ovarian pathologies. In conclusion, calculated ADC values would be useful to evaluate cystic contents of ovarian lesions, for the results show that endometrial cysts and malignant cystic ovarian tumors tended to show lower ADC values than another benign cystic ovarian lesions. Acknowledgments—We thank Yukio Totsuka and Hisatsugu Hirota for their help in technical assistance.

REFERENCES 1. Le Bihan, D.; Breton, E.; Lallemand, D.; Aubin, M.L.; Vignaud, J.; Laval-Jeantet, M. Separation of diffusion and perfusion in intravoxel incoherent motion MR Imaging. Radiology 168:497–505; 1998. 2. Le Bihan, D.; Turner, R.; Douek, P.; Patronas, N. Diffusion MR imaging: Clinical applications. AJR 159:591–599; 1992. 3. Namimoto, T.; Yamashita, Y.; Sumi, S.; Tan, Y.; Takahashi, M. Focal liver masses: Characterization with diffusion-weighted echo-planar MR imaging. Radiology 204: 739 –744; 1997. 4. Coremans, J.; Spanoghe, M.; Budinsky, L.; Sterckx, J.; Luypaert, R.; Eisendrath, H.; Osteaux, M. A comparison

ADC evaluation of cystic ovarian lesions ● T. MOTEKI

5.

6.

7.

8.

9.

10.

between different imaging strategies for diffusion measurements with the centric-encoded turboFLASH sequences. J. Magn. Reson. 124:323–342; 1997. Moteki, T.; Ishizaka, H.; Horikoshi, H.; Matsumoto, M. Differentiation between hemangiomas and hepatocellular carcinomas with the apparent diffusion coefficient calculated from turboFLASH MR images. J. Magn. Reson. 5:187–191; 1995. Mitchell, D.G.; Mintz, M.C.; Spritzer, C.E.; Gussman, D.; Arger, P.H.; Coleman, B.G.; Axel, L.; Kressel, H.Y. Adnexal masses: MR imaging observations at 1.5 T, with US and CT correlation. Radiology 162:319 –324; 1987. Nyberg, D.A.; Porter, B.A.; Olds, M.O.; Olson, D.O.; Andersen, R.; Wesby, G.E. MR imaging of hemorrhagic adnexal masses. J. Comput. Assist. Tomogr. 11:664 – 669; 1987. Smith, F.W.; Cherryman, G.R; Bayliss, A.P.; Fullerton, W.T.; Law, A.N.; Robertson, E.M.; Weir, J; Donald, R.T.; Mallard, J.R. A comparative study of accuracy of ultrasound, X ray CT and low field MRI in the demonstration of cervical and uterine malignancy. Magn. Reson. Imaging 7:677– 679; 1989. Scoutt, L.M.; MacCarthy, S.M. Imaging of ovarian masses: Magnetic resonance imaging. Clin. Obstet. Gynecol. 34:443– 451; 1991. Yamashita, Y.; Torashima, M.; Hatanaka, Y.; Harada, M.; Higashida, Y.; Takahashi, M.; Mizutani, H.; Tashiro, H.; Iwamasa, J.; Miyazaki, K.; Okamura, H. Adnexal masses:

11.

12.

13.

14.

15.

16.

17.

ET AL.

963

Accuracy of characterization with transvaginal US and pre-contrast and post-contrast MR imaging. Radiology 194:557–565; 1995. Stevens, S.K.; Hricak, H. Ovarian lesions: Detection and characterization with gadolinium-enhanced MR imaging at 1.5T. Radiology 181:481– 488; 1991. Ghossain, M.A.; Buy, J.N; Ligneres, C.; Bazot, M.; Hassen, K.; Malbec, L., Hugol, D.; Truc, J.B.; Decroix, Y.; Poitout, P; et al. Epithelial tumors of the ovary: Comparison of MR and CT findings. Radiology 181:863– 870; 1991. Yamashita, Y.; Tang, Y.; Takahashi, M. Ultrafast MR imaging of the abdomen: Echo planar imaging and diffusion-weighted imaging. J. Magn. Reson. 8:367–374; 1998. Baranowska, H.N.; Olszewski, K.J. The hydration of proteins in solutions by self-diffusion coefficients. NMR study. Biochim. Biophys. Acta. 1289:312–314; 1996. Brosio, E.; D’Ubaldo, A.; Verzegnassi, B. Pulsed field gradient spin-echo NMR measurement of water diffusion coefficient in thickening and gelling agents: guar galactomannan solutions and pectin gels. Cell. Mol. Biol. Noisy. le grand. 40:569 –573; 1994. Lapham, J.; Rife, J.P.; Moore, P.B.; Crothers, D.M. Measurement of diffusion constants for nucleic acids by NMR. J. Biomol. NMR. 10:255–262; 1997. Elster, A.D. Gradient echo imaging: Techniques and acronyms. Radiology 186:1– 8; 1993.