Wo m e n ’s I m a g i n g • O r i g i n a l R e s e a r c h Ma et al. MR Spectroscopy of Solid Adnexal Tumors

Downloaded from www.ajronline.org by NYU Langone Med Ctr-Sch of Med on 05/23/15 from IP address 128.122.253.212. Copyright ARRS. For personal use only; all rights reserved

Women’s Imaging Original Research

MR Spectroscopy for Differentiating Benign From Malignant Solid Adnexal Tumors Feng Hua Ma1 Jin Wei Qiang1 Song Qi Cai1 Shu Hui Zhao1 Guo Fu Zhang2 Ya Min Rao 2 Ma FH, Qiang JW, Cai SQ, Zhao SH, Zhang GF, Rao YM

Keywords: adnexal tumors, MRI, ovarian tumors, spectroscopy DOI:10.2214/AJR.14.13391 Received July 2, 2014; accepted after revision November 7, 2014. This study was supported by National Natural Science Fuondation (grant no. 81471628), the Shanghai Municipal Commission of Science & Technology (grant no. 124119a3300), and Shanghai Municipal Commission of Health and Family Planning (grant no. 2013SY075, no. ZK2012A16). 1 Department of Radiology, Jinshan Hospital, Shanghai Medical College, Fudan University, 1508 Longhang Rd, Jinshan District, Shanghai 201508, China. Address correspondence to J. W. Qiang ([email protected]). 2 Department of Radiology, Obstetrics and Gynecology Hospital, Shanghai Medical College, Fudan University, 419 Fangxie Rd, Huangpu District, Shanghai 200011, China. Address correspondence to G. F. Zhang ([email protected]).

WEB This is a web exclusive article. AJR 2015; 204:W724–W730 0361–803X/15/2046–W724 © American Roentgen Ray Society

W724

OBJECTIVE. The purpose of this article is to investigate the proton MR spectroscopy (1H-MRS) features of solid adnexal tumors and to evaluate the efficacy of 1H-MRS for differentiating benign from malignant solid adnexal tumors. MATERIALS AND METHODS. Sixty-nine patients with surgically and histologically proven solid adnexal tumors (27 benign and 42 malignant) underwent conventional MRI and 1H-MRS. Single-voxel spectroscopy was performed using the point-resolved spectroscopy localization technique with a voxel size of 2 × 2 × 2 cm3. Resonance peak integrals of choline, N-acetyl aspartate (NAA), creatine, lactate, and lipid were analyzed, and the choline-tocreatine, NAA-to-creatine, lactate-to-creatine, and lipid-to-creatine ratios were recorded and compared between benign and malignant tumors. RESULTS. A choline peak was detected in all 69 cases (100%), NAA peak in 67 cases (97%, 25 benign and 42 malignant), lipid peak in 47 cases (17 benign and 30 malignant), and lactate peak in eight cases (four benign and four malignant). The mean (± SD) choline-tocreatine ratio was 5.13 ± 0.6 in benign tumors versus 8.90 ± 0.5 in malignant solid adnexal tumors, a statistically significant difference (p = 0.000). There were no statistically significant differences between benign and malignant tumors in the NAA-to-creatine and lipid-to-creatine ratios (p = 0.263 and 0.120, respectively). When the choline-to-creatine threshold was 7.46 for differentiating between benign and malignant tumors, the sensitivity, specificity, and accuracy were 94.1%, 97.1%, and 91.2%, respectively. CONCLUSION. Our preliminary study shows that the 1H-MRS patterns of benign and malignant solid adnexal tumors differ. The choline-to-creatine ratio can help clinicians differentiate benign from malignant tumors.

O

varian cancer is the leading cause of death due to gynecologic malignancy. Because there are no obvious symptoms or only nonspecific symptoms in the early stage, which results in delayed tumor detection and diagnosis, the 5-year survival rate is less than 20% [1]. Although MRI plays a significant role in the detection, determination of origin and extension, and characterization of adnexal masses owing to its multiplanar imaging, superior capability of soft-tissue contrast, and functional imaging techniques such as diffusion-weighted imaging [2–5], there are considerable overlaps on conventional MRI and diffusion-weighted imaging appearances between benign and malignant tumors because of the numerous pathologic types and, consequently, complicated morphologic features of adnexal masses [6, 7]. For example, benign ovarian tumors may be

mixed cystic-solid or solid and similar to malignant ovarian tumors, whereas malignant tumors may be mainly cystic and resemble benign tumors [5, 6]. Proton MR spectroscopy (1H-MRS) is a noninvasive functional in vivo imaging technique that can explore the biochemical metabolism by measuring the contents of proton-containing compounds in tissues [8, 9]. It has been shown that 1H-MRS can distinguish between malignant and benign tissues ex vivo [10]. However, because of its technical limitations, the in vivo application of 1H-MRS in ovarian tumors is still in its infancy, and the few preliminary studies have reported inconsistent results [11–13]. Furthermore, to our knowledge, the differentiation between benign and malignant solid adnexal tumors has not yet been investigated by use of 1H-MRS. The purpose of this study was to evaluate 1H-MRS in the differentia-

AJR:204, June 2015

Downloaded from www.ajronline.org by NYU Langone Med Ctr-Sch of Med on 05/23/15 from IP address 128.122.253.212. Copyright ARRS. For personal use only; all rights reserved

MR Spectroscopy of Solid Adnexal Tumors tion of benign and malignant solid adnexal tumors, thereby improving the characterization of solid adnexal tumors and assisting in surgical planning. Materials and Methods Clinical Data The institutional review boards of our hospitals approved this retrospective study and waived the informed consent requirement. From February 2013 to April 2014, a total of 180 consecutive patients with adnexal tumors (≥ 5 cm in maximal diameter) underwent 1H-MRS as well as conventional MRI. We excluded 106 patients with cystic or mixed cystic-solid ovarian or fallopian tumors and five patients with images of poor spectral quality. The remaining 69 patients had adnexal solid masses and were included in this study. They included 27 patients with benign masses and 42 patients with malignant masses; the mean patient age was 54 years (range, 21–80 years). All tumors were proven by surgery (laparoscopy in 41 cases and laparotomy in 28 cases) and histopathology within 2 weeks after the MRI scan and were staged according to the International Federation of Gynecology and Obstetrics staging system [14].

MRI Scanning MRI was performed under active monitoring by a radiologist using a 1.5-T unit (Avanto, Siemens Healthcare) with a phased-array pelvic coil. The patient lay in the supine position and breathed freely during the acquisition. The conventional unenhanced, MRS, and contrastenhanced scans were performed sequentially. The parameters of various sequences are listed in Table 1. The contrast-enhanced T1-weighted FLASH 2D sequence with fat suppression (TR/ TE, 196/2.9) was performed in the axial and sagittal planes after the IV administration of 0.2 mmol/kg gadopentetate dimeglumine (Magnevist, Bayer Schering) at a rate of 2–3 mL/s. The

TABLE 1: Parameters of Various MRI Sequences for Evaluating Adnexal Masses T1-Weighted Imaging

T2-Weighted Imaging

Contrast-Enhanced T1-Weighted Imaging

Axial

Axial

Coronal

Sagittal

Axial

Sagittal

Single-Voxel Spectroscopy

Sequence

2D SE

2D TSE

2D FSE

2D FSE

2D SPGR

2D FSE

PRESS

TR/TE

761/10

8000/83

4000/98

8000/98

196/2.9

761/4.8

1500/135

FOV (mm2)

28–34

28–34

35–36

25–28

28–34

25–28

2×2×2

5

5

5

5

5

5

20

1.5

1.5

1.2

1.5

1.5

1.5

Parameters

Slice thickness (mm) Gap (mm) Matrix

480 × 640 256 × 256 320 × 320 256 × 256 256 × 256 256 × 256

0 256 × 256

Note—SE = spin-echo, TSE = turbo spin-echo, FSE = fast spin-echo, SPGR = spoiled gradient recalled-echo, PRESS= point-resolved spectroscopy sequence.

scanning parameters were as follows: slice thickness, 5 mm; gap, 1.5 mm; matrix, 256 × 256; FOV, 20–25 × 34 cm; and four excitations. The scanning range was from the inferior pubic symphysis to the renal hilum in the coronal and sagittal planes and extended to the whole abdomen in cases of suspected malignant masses. After the identification of solid adnexal masses on axial or sagittal T2-weighted imaging, singlevoxel spectroscopy was performed using a pointresolved spectroscopy sequence. The 2 × 2 × 2-cm3 voxel was placed in the target area without obvious hemorrhage, necrosis, and cystic component by referring to conventional orthogonal MR images, by a technologist under the guidance of a radiologist. The six outer volume suppression bands were applied at the voxel edges. After automatic shimming, spectral acquisition was performed using 192 averages, and the total acquisition time was 4.54 minutes.

Image Analysis MR images were reviewed independently by two radiologists with 11 and 30 years of experience in gynecologic imaging. Discrepancies were resolved in consensus. The MRI features (tumor size, shape, signal intensity, and enhancement pattern) and asso-

ciated findings (ascites, peritoneal implant, lymphadenopathy, and distant metastasis) were assessed. Spectral reconstruction was performed using software provided by the manufacturer (Syngo, Siemens Healthcare) and consisted of water reference, Hanning filter (center, 0 ms; width, 512 ms), zero-filling from 512 to 1024 points, Fourier transformation, automatic frequency shift correction, automatic six-order polynomial baseline correction, automatic voxel-specific phase correction based on choline and creatine peaks, and an automatic voxel-specific curve fitting. The spectral data quality was assessed by measuring the full-width half maximum on the Fourier-transformed non– water-suppressed spectrum collected at each voxel location, whereas spectroscopic signal-to-noise ratio was considered acceptable when the amplitude of metabolite peaks of interest was greater than three times the amplitude of the background noise. Semiquantitative spectroscopic analysis of the tumors was performed by assessing the signals from lactate (1.31 ppm), lipid (1.33 ppm), N-acetyl aspartate (NAA; 2.02 ppm), creatine (3.04 ppm), choline-containing compounds (3.23 ppm), and other prominent metabolite signals. The areas under the specific resonance peaks (integrals) were determined assuming gaussian fits for choline, NAA,

20

10

Choline I:49.7 Creatine I:11.8

NAA I:14.6

Lipid I:51.3

0

3

A

2

1

B

Fig. 1—51-year-old woman with right ovarian thecoma. A, Sagittal T2-weighted image with fat saturation shows slightly hyperintense solid mass with 2 × 2 × 2-cm3 voxel in center (square). B, MR spectroscopy (TR/TE, 1500/135; average, 192) shows prominent choline peak (3.23 ppm, integral = 49.7) and lipid peak (1.33 ppm, integral = 51.3). Cholineto-creatine and lipid-to-creatine ratios are 4.21 and 4.34, respectively. NAA = N-acetyl aspartate.

AJR:204, June 2015 W725

Downloaded from www.ajronline.org by NYU Langone Med Ctr-Sch of Med on 05/23/15 from IP address 128.122.253.212. Copyright ARRS. For personal use only; all rights reserved

Ma et al. Fig. 2—62-year-old woman with left ovarian serous carcinoma. A, Sagittal T2-weighted image with fat saturation shows slightly hyperintense solid mass with 2 × 2 × 2-cm3 voxel in center (square). B, MR spectroscopy (TR/TE, 1500/135; average, 192) shows extremely prominent choline peak (3.23 ppm, integral = 230) and prominent lipid peak (1.33 ppm, integral = 62.5). Choline-to-creatine and lipid-tocreatine ratios are 7.78 and 2.11, respectively. NAA = N-acetyl aspartate.

Choline I:230

30

20 Creatine I:29.6

10

NAA I:21.0

Lipid I:62.5

0

–10 3

2

1

A

B

lipid, creatine, and a J-coupled fit for lactate. A relatively stable creatine peak was used as an internal reference, and the choline-to-creatine, NAA-tocreatine, lipid-to-creatine, and lactate-to-creatine ratios were calculated for each voxel.

regard to choline-to-creatine, NAA-to-creatine, lipid-to-creatine, and lactate-to-creatine ratios were compared using an independent two-sample t test. A p value less than 0.05 was considered statistically significant. ROC curve analysis was used to evaluate the diagnostic performance of cholineto-creatine ratio for differentiating between benign and malignant tumors, and the corresponding sensitivity, specificity, and accuracy were calculated.

Statistical Analysis Statistical analysis was performed using SPSS (version 17.0 for Windows, SPSS). Differences between malignant and benign tumors in size, shape, signal intensity, and enhancement patterns were compared using the independent two-sample t test, Pearson chi-square test, or Fisher exact test; differences between malignant and benign tumors with

Results MRI morphologic features of 69 cases of solid adnexal masses are shown in Table 2. Statistically significant differences

TABLE 2: MRI Morphologic Characteristics of Benign and Malignant Tumors MRI Features Maximum diameter (cm), mean ± SD

Benign Tumors (n = 27)

Malignant Tumors (n = 42)

p

11.02 ± 0.9

10.71 ± 0.6

0.772a 0.000a

Shape Round or oval

20 (74)

13 (31)

Irregular

7 (26)

29 (69)

Signal intensity on T2-weighted imagesb

0.001a

Low

11 (41)

2 (5)

Slightly high

10 (37)

28 (67)

Mixed

6 (22)

12 (29)

Enhancementb Mild

0.975a 1 (4)

2 (5)

Moderate

5 (19)

8 (19)

Obvious

21 (78)

32 (76)

Ascites

0 (0)

9 (21)

0.010c

Peritoneal implant

0 (0)

8 (19)

0.016c

Lymphadenopathy

0 (0)

6 (14)

0.040c

Note—Except where noted otherwise, data are number (%) of tumors. aIndependent two-sample t test. bSignal intensity and enhancement refer to that of the myometrium. cChi-square or Fisher exact test.

W726

were found between the benign and malignant tumor groups with regard to shape (p = 0.000), signal intensity (p = 0.001), ascites (p = 0.010), peritoneal implant (p = 0.016), and lymphadenopathy (p = 0.040). According to the International Federation of Gynecology and Obstetrics staging system, 31 patients had malignant epithelial ovarian tumors at the following stages: 10 were stage I (32%), six were stage II (19%), 13 were stage III (42%), and two were stage IV (7%). Good quality spectral data was obtained for all 69 patients, with a good signal-to-noise ratio. The full-width half maximum was collected at each voxel location at a mean of 8.15 Hz (range, 3–18 Hz). A choline peak was detected in all 69 solid tumors, NAA peak in 67 tumors (25 benign and 42 malignant), lipid peak in 47 tumors (17 benign and 30 malignant), and lactate peak in only eight tumors (four benign and four malignant). The histologic type and the ratios of the integrals of choline-to-creatine, NAA-to-creatine, and lipid-to-creatine in benign and malignant tumors are shown in Table 3. The mean (± SD) choline-to-creatine ratio was 5.13 ± 0.6 in benign tumors (Figs. 1 and 2) versus 8.90 ± 0.5 in malignant tumors (p = 0.000; Figs. 3–5). There were no statistically significant differences in the NAA-to-creatine and lipid-to-creatine ratios between the two groups (p = 0.263 and 0.120, respectively). The ROC curve analysis of choline-to-creatine ratio yielded an AUC of 0.981 and a threshold of 7.46 for differentiating malignant from benign tumors, with a sensitivity, specificity, and accuracy of 94.1%, 97.1%, and 91.2%, respectively. Discussion Because a solid mass or a large solid component of a cystic-solid mass is one of the most important MRI criteria for malignant ovarian tu-

AJR:204, June 2015

MR Spectroscopy of Solid Adnexal Tumors

Downloaded from www.ajronline.org by NYU Langone Med Ctr-Sch of Med on 05/23/15 from IP address 128.122.253.212. Copyright ARRS. For personal use only; all rights reserved

TABLE 3: Histologic Type and Ratios of the Metabolite Integrals in Benign and Malignant Adnexal Tumors Mean Ratio of Metabolites Histologic Type

No. of Tumors

Choline to Creatinea

NAA to Creatineb

Lipid to Creatineb

Benign tumors

27

5.13

1.70

6.74

Ovarian thecomas

12

5.75

1.16

6.61

Fibromas

3

2.17

1.29

3.68

Sclerosing stromal tumor

1

5.71

1.15

9.44

Leiomyomas of broad ligament

11

5.67

2.16

5.63

Malignant tumors

42

8.90

2.45

8.99 8.89

Serous carcinomas

19

8.95

1.89

Clear cell carcinomas

8

10.10

2.22

7.14

Mucinous carcinomas

3

9.60

7.17

5.37

Endometrial ovarian cancer

1

8.69

1.38

6.77

Metastatic ovarian tumors

6

5.53

2.24

12.99

Fallopian tube carcinomas

3

8.66

1.02

0

Malignant germ cell tumor

1

7.15

1.55

0

Intestinal stromal tumor

1

10.58

1.39

5.72

aThere were statistically significant differences between benign and malignant tumors in the choline-to-­

creatine ratio (p = 0.000). NAA = N-acetyl aspartate.

bNot statistically significant.

mors, some benign ovarian tumors, such as sex cord–stromal tumors, adenofibromas, Brenner tumors, and broad ligament leiomyomas, often appear as solid or cystic-solid masses, and leiomyomas, fibrothecomas, and sclerosing stromal tumors often show mixed low-to-high signal intensity on T2-weighted images and marked or delayed enhancement. Therefore, they may have considerable morphologic overlap with malignant counterparts, as shown in the present study, which is a diagnostic pitfall [5–7, 15]. Proton MR spectroscopy (1H-MRS) MRS is a noninvasive diagnostic tool for the investigation of tumor metabolism and is a meaningful supplement to conventional MRI. It has been shown that MRS is useful for the detection, differentiation, staging, and therapeutic monitoring of brain, breast, and prostate tumors [16–18]. In the application of ovarian tumors, 1H-MRS is still in the preliminary stage, with some inconsistent results, probably due to the diversity of samples and the complicated tumor histopathologic and morphologic features [11–13, 15]. For instance, a study conducted by Stanwell et al. [12] showed a higher choline-tocreatine ratio in malignant ovarian cancer than

20 15

Choline I:71.3

10 5

Creatine I:14.6

NAA I:17.2

0 –5 3

A

C

2

1

B

D

Fig. 3—41-year-old woman with right leiomyomas of broad ligament. A, Sagittal T2-weighted image with fat saturation shows isointense solid mass with 2 × 2 × 2-cm3 voxel in center (square). B, MR spectroscopy shows high choline peak (3.23 ppm, integral = 71.3). Choline-to-creatine ratio is 4.88. NAA = N-acetyl aspartate. C and D, Axial T2-weighted image with fat saturation (C) and contrast-enhanced axial T1-weighted image with fat saturation (D) show round isointense solid mass (arrow) with marked enhancement.

AJR:204, June 2015 W727

Downloaded from www.ajronline.org by NYU Langone Med Ctr-Sch of Med on 05/23/15 from IP address 128.122.253.212. Copyright ARRS. For personal use only; all rights reserved

Ma et al. in benign cystic ovarian tumors, and another study [11] found no difference in the presence of choline between benign and malignant gynecologic disease. In our study, we ruled out cystic and mixed cystic-solid adnexal tumors to avoid the comparison between cystic and solid tumors. To our knowledge, no study has so far investigated 1H-MRS in discriminating benign from malignant solid adnexal tumors. Studies have shown that changes in the phospholipid metabolism are a common feature of cancer and that choline levels are significantly higher in malignant tumors than in benign tumors [19–22]. Because the spectral resolution of in vivo 1H-MRS is lower than that of in vitro 1H-MRS, studies use the ratio of peak integral to describe, quantitatively analyze, and compare metabolic changes among different tumors. Stanwell et al. [12] reported that an integral choline-to-creatine ratio greater than 3.09 indicated a malignant gynecologic tumor, whereas the absence of a choline peak or an integral choline-to-creatine ratio less than 1.15 indicated a benign tumor. Li et al. [13] found that a threshold of 2 in the choline peak-tonoise ratio could accurately discriminate between benign and malignant adnexal tumors. In this study, the integral choline-to-creatine ratio was 5.13 ± 0.6 in benign adnexal tumors versus 8.90 ± 0.5 in malignant counterparts, and the difference was statistically significant. The choline-to-creatine threshold was 7.46 for differentiating between benign and malignant adnexal tumors, with a sensitivity, specificity, and accuracy of 94.1%, 97.1%, and 91.2%, respectively. The ratios were higher than those of previous studies [12, 13], probably because of the solid tumors in all patients. NAA is a recognized marker of neurons that is observed at 2.02 ppm. The NAA peak decreases significantly in brain gliomas because of the absence of or damage to neurons. However, an unassigned and prominent resonance of 2.0–2.1 ppm has frequently been found on in vivo MRS of breast, cervical, prostate, and ovarian cancers and is assigned to the −CH3 moiety of sialic acid or N-acetyl groups of glycoproteins [17–20]. In the study of Kolwijck et al. [23], in vivo and in vitro MRS were used to analyze this unassigned and prominent resonance in ovarian cyst fluids, and the NAA and N-acetyl groups from glycoproteins or glycolipids were conformed to contribute to the 2.0to 2.1-ppm resonance complex. Stanwell et al. [12] reported that a resonance at 2.07 ppm was present in all of the teratomas and serous cystadenomas, as well as a portion of the serous carcinomas. Its amplitude was greater in ma-

W728

150

A

B

C

D

Choline I:886

100

50 Creatine I:77.0

NAA I:73.1

Lipid I:302

0 3

2

1

Fig. 4—59-year-old woman with left ovarian clear cell carcinoma. A and B, Axial (A) and sagittal (B) T2-weighted images with fat saturation show irregular mixed hyperintense solid mass, with 2 × 2 × 2-cm3 voxel in homogeneous area (square, B). C, MR spectroscopy (TR/TE, 1500/135; average, 192) shows extremely prominent choline peak (3.23 ppm, integral = 886) and prominent lipid peak (1.33 ppm, integral = 302). Choline-to-creatine and lipid-to-creatine ratios are 11.51 and 5.11, respectively. NAA = N-acetyl aspartate. D, Contrast-enhanced sagittal T1-weighted image with fat saturation shows irregular markedly enhanced mass (arrow).

lignant than in benign tumors. No statistically significant difference in the NAA-to-creatine ratio was found between benign and malignant solid tumors in the current study. Further investigations are warranted to determine whether the results correlate with the different pathologic types of benign and malignant tumors. To some extent, a lactate peak can also reflect tumor biologic behavior. The increased glucose consumption in malignant tumors produces more lactate than in benign tumors. Okada et al. [24] found that lactate was detected in all malignant pelvic tumors and in some benign tumors, with a significantly higher lactate content in the former. However, pelvic abscesses should be excluded. Hascalik et al. [8] suggested that a high lactate level in a pelvic abscess was the result of anaerobic glycolysis as well as increased glucose consumption. In the

study of Massuger et al. [25], the lactate levels of malignant ovarian cysts were six times higher than those of benign ovarian cysts, which is considered to be the result of a reduced blood flow. In this study, a lactate peak was detected in only four malignant and four benign tumors, and no statistical analysis was performed for lactate-to-creatine ratio. The discrepant results between our study and other studies were probably related to the solid tumors in all our cases. Lipid may be observed in various malignant tumors. Cho et al. [19] showed an intense lipid peak at 1.3 ppm in malignant ovarian tumors and benign teratomas, whereas benign epithelial ovarian tumors showed no detectable lipid peak. Elevated lipid levels were detected by in vivo and ex vivo MRS in cervical cancer with a sensitivity and specificity of 77.4% and 93.8%, respectively [26]. Takeuchi et al. [27] reported

AJR:204, June 2015

MR Spectroscopy of Solid Adnexal Tumors

Downloaded from www.ajronline.org by NYU Langone Med Ctr-Sch of Med on 05/23/15 from IP address 128.122.253.212. Copyright ARRS. For personal use only; all rights reserved

References Choline I:171 20

10

Creatine I:20.4

NAA I:27.7

Lipid I:52.7

0

3

2

1

A

B

C

D

Fig. 5—60-year-old woman with left ovarian endometrial carcinoma. A, Sagittal T2-weighted image with fat saturation shows slightly hyperintense solid mass with 2 × 2 × 2-cm3 voxel in center (square). B, MR spectroscopy (TR/TE, 1500/135; average, 192) shows prominent choline peak (3.23 ppm, integral = 171) and medium lipid peak (1.33 ppm, integral = 52.7). Choline-to-creatine and lipid-to-creatine ratios are 8.38 and 2.64, respectively. NAA = N-acetyl aspartate. C, Axial T1-weighted image shows irregular solid mass. D, On contrast-enhanced axial T1-weighted image with fat saturation, solid component shows mild enhancement.

that the presence of a lipid peak for the diagnosis of thecomas or fibrothecomas had a sensitivity of 100% and specificity of 92%. The high lipid peak reflects abundant intracellular lipid contents and is considered a specific metabolite for thecomas or fibrothecomas, which may contribute to distinguishing thecomas or fibrothecomas from other benign ovarian fibrous tumors or subserosal uterine leiomyomas. In this study, a lipid peak was detected in 11 of 12 thecomas or fibrothecomas, two of three fibromas, and six of 11 leiomyomas versus 30 of 42 malignant tumors. The lipidto-creatine ratio was 8.99 ± 2.6 in malignant tumors, greater than the 6.74 ± 0.9 found in benign tumors, but the difference was not statistically significant. A higher proportion of thecomas in our benign group may compromise the

difference. More cases are needed to further clarify the value of the lipid peak in discriminating benign from malignant tumors. Our study had some limitations. First, a selection bias was inevitable because of the retrospective nature. Second, lactate, which was detected in only eight cases in this study, may be obscured by overlapping with more intense lipid resonances. Third, only the ratio of metabolites with creatine was compared, and no quantitative analysis was performed. In conclusion, in vivo 1H-MRS can quantitatively evaluate the metabolite concentration in solid adnexal tumors, and the choline-to-creatine ratio is helpful for differentiating benign from malignant solid adnexal tumors, thereby provide assistance in surgical planning.

1. Siegel R, Naishadham D, Jemal A. Cancer statistics, 2012. CA Cancer J Clin 2012; 62:10–29 2. Mohaghegh P, Rockall AG. Imaging strategy for early ovarian cancer: characterization of adnexal masses with conventional and advanced imaging techniques. RadioGraphics 2012; 32:1751–1773 3. Bazot M, Daraï E, Nassar-Slaba J, Lafont C, Thomassin-Naggara I. Value of magnetic resonance imaging for the diagnosis of ovarian tumors: a review. J Comput Assist Tomogr 2008; 32:712–723 4. Sohaib SA, Sahdev A, Van Trappen P, Jacobs IJ, Reznek RH. Characterization of adnexal mass lesions on MR imaging. AJR 2003; 180:1297–1304 5. Zhao SH, Qiang JW, Zhang GF, et al. Diffusionweighted MR imaging for differentiating borderline from malignant epithelial tumors of the ovary: pathological correlation. Eur Radiol 2014; 24:2292–2299 6. Fujii S, Kakite S, Nishihara K, et al. Diagnostic accuracy of diffusion-weighted imaging in differentiating benign from malignant ovarian lesions. J Magn Reson Imaging 2008; 28:1149–1156 7. Thomassin-Naggara I, Daraï E, Cuenod CA, et al. Contribution of diffusion-weighted MR imaging for predicting benignity of complex adnexal masses. Eur Radiol 2009; 19:1544–1552 8. Hascalik S, Celik O, Sarac K, Meydanli MM, Alkan A, Mizrak B. Metabolic changes in pelvic lesions: findings at proton MR spectroscopic imaging. Gynecol Obstet Invest 2005; 60:121–127 9. Canese R, Pisanu ME, Mezzanzanica D, et al. Characterisation of in vivo ovarian cancer models by quantitative 1H magnetic resonance spectroscopy and diffusion-weighted imaging. NMR Biomed 2012; 25:632–642 10. Boss EA, Moolenaar SH, Massuger LF, et al. High-resolution proton nuclear magnetic resonance spectroscopy of ovarian cyst fluid. NMR Biomed 2000; 13:297–305 11. Booth SJ, Pickles MD, Turnbull LW. In vivo magnetic resonance spectroscopy of gynaecological tumours at 3.0 T. BJOG 2009; 116:300–303 12. Stanwell P, Russell P, Carter J, Pather S, Heintze S, Mountford C. Evaluation of ovarian tumors by proton magnetic resonance spectroscopy at three Tesla. Invest Radiol 2008; 43:745–751 13. Li WH, Chu CT, Zhang P, et al. MRI and MRS analysis of ovarian endometrioid carcinomas. J Clin Radiol China 2008; 27:470–472 14. Prat J. Staging classification for cancer of the ovary, fallopian tube, and peritoneum. Int J Gynecol Obstet 2014; 124:1–5 15. Tanaka YO, Okada S, Satoh T, et al. Solid non-invasive ovarian masses on MR: histopathology and a diagnostic approach. Eur J Radiol 2011; 80:e91–e97 16. Bulik M, Jancalek R, Vanicek J, Skoch A, Mechl M. Potential of MR spectroscopy for assessment of glioma grading. Clin Neurol Neurosurg 2013;

AJR:204, June 2015 W729

Downloaded from www.ajronline.org by NYU Langone Med Ctr-Sch of Med on 05/23/15 from IP address 128.122.253.212. Copyright ARRS. For personal use only; all rights reserved

Ma et al. 115:146–153 17. Bolan PJ. Magnetic resonance spectroscopy of the breast: current status. Magn Reson Imaging Clin N Am 2013; 21:625–639 18. Selnaes KM, Gribbestad IS, Bertilsson H, et al. Spatially matched in vivo and ex vivo MR metabolic profiles of prostate cancer-investigation of a correlation with Gleason score. NMR Biomed 2013; 26:600–606 19. Cho SW, Cho SG, Lee JH, et al. In-vivo proton magnetic resonance spectroscopy in adnexal lesions. Korean J Radiol 2002; 3:105–112 20. Podo F, Sardanelli F, Iorio E, et al. Abnormal choline phospholipid metabolism in breast and ovary cancer: molecular bases for noninvasive imaging

W730

approaches. Curr Med Imag Rev 2007; 3:123–137 21. Iorio E, Mezzanzanica D, Alberti P, et al. Alterations of choline phospholipid metabolism in ovarian tumor progression. Cancer Res 2005; 65:9369–9376 22. Iorio E, Mezzanzanica D, Alberti P, et al. Activation of phosphatidylcholine cycle enzymes in human ovarian cancer cells. Cancer Res 2010; 70:2126–2135 23. Kolwijck E, Engelke UF, van der Graaf M, et al. N-acetyl resonances in in vivo and in vitro NMR spectroscopy of cystic ovarian tumors. NMR Biomed 2009; 22:1093–1099 24. Okada T, Harada M, Matsuzaki K, Nishitani H, Aono T. Evaluation of female intrapelvic tumors by clinical proton MR spectroscopy. J Magn Re-

son Imaging 2001; 13:912–917 25. Massuger LF, van Vierzen PB, Engelke U, Heerschap A, Wevers R. 1H-magnetic resonance spectroscopy: a new technique to discriminate benign from malignant ovarian tumors. Cancer 1998; 82:1726–1730 26. Mahon MM, Cox IJ, Dina R, et al. 1H Magnetic resonance spectroscopy of preinvasive and invasive cervical cancer: in vivo-ex vivo profiles and effect of tumor load. J Magn Reson Imaging 2004; 19:356–364 27. Takeuchi M, Matsuzaki K, Harada M. Preliminary observations and diagnostic value of lipid peak in ovarian thecomas/fibrothecomas using in vivo proton MR spectroscopy at 3T. J Magn Reson Imaging 2012; 36:907–911

AJR:204, June 2015