Objective
To estimate the diagnostic accuracy of magnetic resonance in ovarian tumors.
Study Design
A quantitative systematic review was performed. Studies that compared magnetic resonance and paraffin sections within subjects for diagnosis of ovarian tumors were included.
Results
Fifteen primary studies were analyzed, which included 1267 ovarian masses. For borderline or malignant ovarian cancer vs benign ovarian lesions, the pooled likelihood ratio for the occurrence of a positive magnetic resonance result was 6.6 (95% confidence interval, 4.7–9.2) and the posttest probability for borderline or malignant diagnosis was 77% (95% confidence interval, 70–82). Because specificity and likelihood ratio positive were heterogeneous, a random effect model was used and a summary receiver operating characteristic curve was generated. For borderline or malignant ovarian cancer vs benign ovarian lesions, the area under curve was 0.9526.
Conclusion
Magnetic resonance seems to be a useful preoperative test for predicting the diagnosis of pelvic masses.
Malignant ovarian neoplasms are responsible for 4% of all cancers affecting women, are the second most common cause of death from gynecologic cancer and the fourth most common cause of death from all types of cancer affecting women. Recently, to improve patient’s quality of life, laparoscopic surgery has been increasingly used in the treatment of ovarian tumors that are thought to be benign. For malignant ovarian tumors, there is no evidence that laparoscopy for the management of early stage ovarian cancer may improve the quality of life of patients.
Under such circumstance, preoperative characterization of the tumor as benign or malignant has become more important than ever when planning surgery for an ovarian tumor. Ultrasonography with color Doppler is a useful preoperative test for predicting the diagnosis of ovarian lesions. Magnetic resonance imaging (MRI) may be of great help in identifying malignancy before surgery when a sonogram is suboptimal or indeterminate. A new MRI technique with dynamic diffusion-weighted imaging (DWI) is a useful tool for characterizing epithelial ovarian tumors and can help distinguish among benign, borderline, and invasive tumors. This technique assesses the mean and lowest apparent diffusion coefficient (ADC) values of the solid portion for each ovarian lesion. Morphologic criteria such as septa, papillary projections, solid portion, and T2-weighted MRI signal intensity of solid tissue are useful for discriminating invasive from noninvasive ovarian tumors, but none of these morphologic criteria reliably distinguished benign from borderline ovarian tumors.
Our objective was to perform a systematic review and metaanalysis of the literature to ascertain the accuracy of MRI in the diagnosis of ovarian cancer.
Materials and Methods
Data sources
Search strategy. A comprehensive search of the Medline, Cancerlit, Lilacs, and Embase databases, between the dates January 1990 to February 2010, was conducted. The medical subject headings (MeSH) and text words “ovarian neoplasm,” “ovarian lesions,” “ovarian masses,” and “magnetic resonance” were combined with the MeSH term “diagnosis” (sensitivity, specificity, false positive, false negative, predictive value, reference value, receiver operating characteristic curve [ROC], likelihood [LH] ratio, accuracy). The search was limited to human studies but had no language restrictions. In addition, the Cochrane Library was searched. Reference lists of all available primary studies were reviewed to identify additional relevant citations. The authors of the published studies were not contacted.
Criteria for consideration of studies for this review
This review focused on observational studies that evaluated clinically suspected adnexal masses with MRI. All MRI examinations should have been performed on a 1.5 T using the pelvic phased array coil signal reception and T1-weighting for the transverse plane and T2-weighting for the transverse and saggital planes. The increase of signal intensity of the solid components was evaluated on dynamic contrast-enhanced (DCE) MR images when available. In the absence of DCE, the uptake of contrast medium from solid tissue was evaluated comparing pre- and postcontrast MR images. The results of the diagnostic test of interest were compared with the results of a reference standard. The diagnostic test was MRI and the diagnostic reference was the result of the histologic analysis of standard paraffin-embedded sections after surgery. MRI diagnosis was considered correct if it did not differ from that of the paraffin section. For inclusion in this systematic review the final histologic diagnosis of the tumor was designated as benign, borderline, or invasive ovarian tumors. We excluded studies that analyzed only benign, only borderline, or only invasive ovarian tumors as well as reviews and studies that lacked the data needed to construct 2 × 2 contingency tables.
The ovarian lesions were grouped as borderline or invasive ovarian tumors and were compared with benign ovarian lesions. The cases that were deferred due to uncertain MRI were excluded. The primary outcome analyzed was the accuracy of ovarian tumor diagnosis (borderline or invasive ovarian tumors vs benign lesions) by MRI. A secondary outcome was the distribution of histologic types of ovarian tumors according to paraffin-embedded sections (benign, borderline, and invasive). The reviewed studies were analyzed independently by 8 investigators (L.R.M., D.D.R., M.I.R., F.R.S., L.B.F., L.S.S., M.I.E., L.T.B). Final inclusion or exclusion was made with reference to a selection criteria checklist. Disagreements about study inclusion or exclusion were initially solved by consensus, and when this was not possible they were arbitrarily resolved by 1 reviewer (L.R.M.). The agreement statistics among reviewers were computed.
Quality assessment
All articles meeting the eligibility criteria were assessed for methodologic quality. This assessment involved scrutinizing the study designs such as the methods of data collection, the relevant features of the patient population/selection, a description of the MRI test and the histologic reference standard, and the presence of verification biases. The results of the quality assessment of included studies are summarized in the “Scoring of Study Quality” column of Table 1 . Studies were further assessed for methodologic quality with reference to the Oxford Centre for Evidence-Based Medicine Levels of Evidence Classification rubric . Only studies with Oxford Evidence levels 1-3 were considered to be of high quality, whereas those with levels 4 and 5 were excluded.
| Study, y | n | Mean age | Period of study | Adnexal masses and histologic analysis a | Scoring of study quality | Oxford level b |
|---|---|---|---|---|---|---|
| Balan, 2006 | 136 | Not reported | September 1998-February 2001 | 100 | Small population, verification complete, nonblinded, retrospective, test details sufficient, reference test details sufficient, population details insufficient | 3B c 198 |
| Bazot et al,2006 | 136 | 51 (range, 15–82) | January 1999-December 2003 | 168 | Small population, verification complete, nonblinded, consecutive and prospective, test details sufficient, reference test details sufficient, population details sufficient | 2B b |
| Chen et al,2006 | 74 | 53 (range, 18–83) | Not reported | 70 | Small population, verification complete, nonblinded, consecutive and prospective, test details sufficient, reference test details sufficient, population details sufficient | 2B |
| Fenchel et al,2002 | 99 | 46 (SD ± 15) | May 1997-February 1999 | 103 | Small population, verification complete, nonblinded, consecutive and prospective, test details sufficient, reference test details sufficient, population details sufficient | 2B |
| Fujii et al, 2008 | 119 | 52 (range, 16–83) | January 2005-October 2007 | 123 | Small population, verification complete, nonblinded, retrospective, test details sufficient, reference test details sufficient, population details sufficient | 3B |
| Grab et al, 2000 | 101 | 45 (range, 18–82) | Not reported | 88 | Small population, verification complete, nonblinded, consecutive and prospective, test details sufficient, reference test details sufficient, population details sufficient | 2B |
| Huber et al, 2002 | 93 | Not reported | May 1995-January 2001 | 85 | Small population, verification complete, nonblinded, not reported if consecutive and prospective, test details sufficient, reference test details sufficient, population details insufficient | 3B |
| Kawahara et al, 2004 | 49 | Not reported | September 2001-August 2003 | 38 | Small population, verification complete, nonblinded, consecutive and prospective, test details sufficient, reference test details sufficient, population details insufficient | 2B |
| Mascaretti et al,1994 | 53 | Range 21-72 | Not reported | 53 | Small population, verification complete, nonblinded, not reported if consecutive and prospective, test details sufficient, reference test details sufficient, population details sufficient | 3B |
| Reuter et al,1998 | 67 | 48,8 (range 18-84) | January 1994-August 1995 | 42 | Small population, verification complete, nonblinded, not reported if consecutive and prospective, test details sufficient, reference test details sufficient, population details sufficient | 3B |
| Rieber et al, 2001 | 103 | 46 (range 17-82) | Not reported | 103 | Small population, verification complete, nonblinded, consecutive and prospective, test details sufficient, reference test details sufficient, population details sufficient | 2B |
| Sohaib et al, 2005 | 89 | 53 (range 19-86) | Not reported | 72 | Small population, verification complete, nonblinded, consecutive and prospective, test details sufficient, reference test details sufficient, population details sufficient | 2B |
| Stevens et al,1991 | 41 | Not reported | May 1989-May 1990 | 60 | Small population, verification complete, nonblinded, not reported if consecutive and prospective, test details sufficient, reference test details sufficient, population details insufficient | 3B |
| Thomassin-Naggara et al, 2008 | 37 | Range 16-79 | April 2003-December 2004 | 37 | Small population, verification complete, blinded, not reported if consecutive and prospective, test details sufficient, reference test details sufficient, population details insufficient | 3B |
| Thomassin-Naggara et al, 2009 | 77 | 53 (range, 27–87) | March 2006-November 2007 | 77 | Small population, verification complete, nonblinded, consecutive and prospective, test details sufficient, reference test details sufficient, population details sufficient | 2B |
| Thurnher et al,1990 | 60 | 48 (range, 19–82) | Not reported | 54 | Small population, verification complete, nonblinded, consecutive and prospective, test details sufficient, reference test details sufficient, population details sufficient | 2B |
| Tsili et al, 2008 | 75 | 51 (range, 22–80) | February 2004-April 2006 | 89 | Small population, verification complete, nonblinded, consecutive and prospective, test details sufficient, reference test details sufficient, population details sufficient | 2B |
| Yamashita et al, 1995 | 72 | 43 (range, 13–74) | Not reported | 80 | Small population, verification complete, blinded, consecutive and prospective, test details sufficient, reference test details sufficient, population details sufficient | 2B |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
