Abstract
Objective
To explore the value of magnetic resonance imaging (MRI) and clinical features in identifying ovarian thecoma-fibroma (OTF) with cystic degeneration and ovary adenofibroma (OAF).
Methods
A total of 40 patients with OTF (OTF group) and 28 patients with OAF (OAF group) were included in this retrospective study. Univariable and multivariable analyses were performed on clinical features and MRI between the two groups, and the receiver operating characteristic (ROC) curve was plotted to estimate the optimal threshold and predictive performance.
Results
The OTF group had smaller cyst degeneration degree (P < .001), fewer black sponge sign (20% vs. 53.6%, P = .004), lower minimum apparent diffusion coefficient value (ADCmin) (0.986 (0.152) vs. 1.255 (0.370), P < .001), higher age (57.4 ± 14.2 vs. 44.1 ± 15.9, P = .001) and more postmenopausal women (72.5% vs. 28.6%, P < .001) than OAF. The area under the curve of MRI, clinical features and MRI combined with clinical features was 0.870, 0.841, and 0.954, respectively, and MRI combined with clinical features was significantly higher than the other two (P < .05).
Conclusion
The cyst degeneration degree, black sponge sign, ADCmin, age and menopause were independent factors in identifying OTF with cystic degeneration and OAF. The combination of MRI and clinical features has a good effect on the identification of the two.
Advances in knowledge
This is the first time to distinguish OTF with cystic degeneration from OAF by combining MRI and clinical features. It shows the diagnostic performance of MRI, clinical features, and combination of the two. This will facilitate the discriminability and awareness of these two diseases among radiologists and gynaecologists.
Keywords: ovarian thecoma-fibroma, ovary adenofibroma, magnetic resonance imaging, clinical feature, differential diagnosis
Introduction
Ovarian thecoma-fibroma (OTF) and ovarian adenofibroma (OAF) are two rare ovarian tumours that account for 4% and 1.7% of primary ovarian tumors, respectively.1,2 They originate from completely different ovarian tissues. OTF originates from the sex cord-stromal of ovary3 and consists of theca cells and fibroblasts in varying proportions. OAF is derived from the epithelium stroma of ovary,4 which is histologically characterized by both epithelial components and prominent fibrous stroma. Cystic degeneration is the most common degenerative change in OTF5 and is one of the most important factors leading to misdiagnosis by radiologists. Studies have shown that tumours ≥6 cm in diameter are more likely to show cystic or cystic-solid masses,6 which are atypical in MRI and can easily be misdiagnosed as other cystic-solid and cystic ovarian tumours.
The OTF and OAF have a similar proportion of ovarian stroma in the solid components, and both contain fibrous components. Similar histological components determined the similarity in MRI findings. These two tumours of different origins differ in clinical features, such as the fact that OTF secretes oestrogen and thus may cause oestrogen-related diseases,7 which may help to distinguish them. Because of the rarity, few studies have dialectically compared the MRI and clinical features of OTF with cystic degeneration and OAF so far. Therefore, the systematic and detailed analysis and identification of MRI and clinical features are of great significance in improving the understanding of these two tumours by radiologists and gynaecologists.
The purpose of this study was to analyse the MRI and clinical features in OTF with cystic degeneration and OAF in detail, combined with pathology. To determine the best features to distinguish the two diseases, and to evaluate the diagnostic performance of MRI combined with clinical features. To improve the accuracy of preoperative differential diagnosis of OTF with cystic degeneration versus OAF.
Methods
Patients and clinical data
In this retrospective study, data from 56 patients with OTF and 47 patients with OAF who were pathologically diagnosed from June 2014 to November 2023 were obtained by searching the picture archiving and communication system (PACS) of our hospital. The recruitment pathway is shown in Figure 1. The inclusion criteria were as follows: (1) no biopsy or therapy was performed before MRI examination; (2) pelvic enhanced MRI was performed in our hospital within 2 weeks before surgery; (3) the image quality was adequate for diagnosis and measurement; (4) the clinical data of the patients were complete. The exclusion criteria were as follows: (1) the volume of the solid part is too small (<1 cm in diameter) to measure the ADC value; (2) masses without cystic component; (3) concomitantly suffering from other malignant tumors. Finally, there are 40 patients with OTF and 28 patients with OAF were included in this study. The study was approved by the Ethics Committee of our hospital and exempted from the informed consent of the patients.
Figure 1.
Flowchart of patient recruitment pathway.
The following clinical data were recorded: age, body mass index (BMI), menopause, symptoms or signs, parity, carbohydrate antigen 125 (CA125), human epididymal protein 4 (HE-4), Rome index (ROM, normal reference range, premenopause, 0%-11.4%, postmenopause, 0-29.9%), carcino-embryonic antigen (CEA), alpha fetoprotein (AFP), and carbohydrate antigen 19-9 (CA19-9).
MRI examination
All patients underwent preoperative pelvic examination using a unitary system (GE Signa HDXT 3.0 T MRI scanner, GE Healthcare, United States) equipped with an 8-channel phased-array body coil. All of them fasted for at least 4 h and were given an intramuscular injection of scopolamine hydrochloride half an hour before the examination to reduce gastrointestinal peristaltic artifacts. The main plain MRI protocols for the evaluation of ovarian lesions are axial fast spin echo (FSE) T1-weighted image (T1WI), axial FSE T2-weighted image (T2WI), axial fat suppression T2-weighted image (FS-T2WI), sagittal FSE T2WI and axial diffusion weighted imaging (DWI) (b = 0 or 1000 s/mm2) sequences. Contrast-enhanced scans (arterial, venous, and delayed phase) included axial, sagittal, and coronal delayed late liver acquisition with volume acceleration (LAVA) sequences (scan times were 25, 60, 130, 160, and 180 s after contrast injection). The contrast agent (0.1 mmol/kg, Gadodiamide, General Electric Healthcare, United States) was injected via the anterior elbow vein using a high-pressure syringe (Optistar LE; Mallinckrodt). Detailed scan parameters are shown in Table 1.
Table 1.
MRI scanning protocols.
| Sequence | TR/TE (ms) | FOV (cm) | Matrix | Slice gap (mm) | Thickness (mm) | NEX |
|---|---|---|---|---|---|---|
| Axial T1WI | 525/7.5 | 34 × 24 | 352 × 224 | 2 | 6 | 2 |
| Axial T2WI | 3000/120 | 24 × 24 | 288 × 256 | 1 | 3 | 4 |
| Sagittal T2WI | 3220/105 | 26 × 24 | 320 × 256 | 1 | 6 | 1 |
| Axial FS-T2WI | 4600/105 | 24 × 24 | 320 × 256 | 2 | 6 | 2 |
| Axial DWI | 4000/65.2 | 40 × 28 | 96 × 130 | 2 | 6 | 6 |
| Axial T1WI LAVA-FLEX | 5.4/1.3 | 34 × 30 | 296 × 224 | 0 | 4 | 1 |
| Sagittal T1WI LAVA-FLEX | 6.8/1.3 | 28 × 25 | 268 × 224 | 0 | 4 | 1 |
| Coronal T1WI LAVA-FLEX | 4.5/1.4 | 36 × 36 | 272 × 224 | 0 | 4 | 1 |
Abbreviations: TR = repetition time, TE = echo time, FOV = field of view, NEX = number of excitations.
Imaging analyses
Two radiologists with 12 and 8 years of experience in gynaecologic imaging analysed the MRI features respectively, ignoring the pathological findings. The ADC values of the tumour were measured using Function Tool software, and the area of interest (ROI) was outlined manually in the level containing the largest solid area of the tumour and in the upper and lower levels, avoiding visible cystic lesions, necrosis or haemorrhage. Five ROIs were placed at each level and the ROI of the lowest ADC value is considered as the ADCmin, the ROI of the highest ADC value is considered as the ADCmax, and the ADCmean is the average of all measurements. The following MRI characteristics of the two types of diseases were recorded (When both ovaries presented with similar masses, the most complex mass was evaluated on MRI. Otherwise evaluate the largest mass.): tumour haemorrhage (① present, ② absent); capsule (① present, ② absent); tumour angiogenesis (① present: on contrast enhancement, blood vessels were seen in the solid part of the mass, ② absent); pelvic fluid (① none, ② little: confined to peri-uterine, peri-adnexal, or Douglas fossa, ③ moderate: confined to the pelvis and not exceeding the upper pelvic rim, ④ large: exceeding the upper pelvic rim); border (① well defined, ② ill defined); mean diameter [(transverse diameter + vertical diameter + anteroposterior diameter)/3 cm]; side of the lesion (① left, ② right, ③ bilateral); morphology (① multinodular: multiple nodules or chambers, ② orbicular: approximately round or oval, ③ lobulated: composed of lobules without nodular appearance); cystic degeneration degree (① grade 1: cystic lesion area < 1/3, ② grade 2: 1/3-2/3, ③ grade 3: >2/3); cyst pattern (① mainly large: all cystic components straight through ≥1 cm, ② mainly small: all cystic components straight through < 1 cm, ③ mixed); endometrial hyperplasia (① present: endometrial thickness >5 mm in premenopausal women or > 16 mm in postmenopausal women. ② absent); uterine fibroids or adenomyosis (① present, ② absent); signal intensity of solid areas on T1WI, T2WI, FS-T2WI and DWI compared to gluteus (① hypointensity, ② isointensity, ③ hyperintensity, ④ mixed); the degree of enhancement of the solid areas in the arterial phase (25 s), venous phase (60 s) and delayed phase (160 s) (① mild: smaller than the gluteus ② moderate: similar to the gluteus, ③ severe: more than the gluteus); black sponge sign (① present: a solid component of very low signal intensity containing tiny foci of high T2 signal intensity, ② absent).
Statistical analyses
Statistical analyses were implemented with SPSS (version 24.0; IBM Corporation, Armonk, NY, United States) and Python (version 3.6.0). The agreement between the two radiologists in assessing the MRI features of the OTF group and the OAF group was evaluated by Kappa coefficient (Kappa coefficient >0.61 was indicative of good agreement), and the differences in the assessments were discussed to reach a consensus, and the consensus data were used for data analysis. Intraclass correlation coefficient (ICC) was used to assess intra- and interobserver agreement for continuous variables such as ADC values (ICC >0.75 was indicative of good agreement). Continuous variables with normal distribution, expressed as mean ± standard deviation (SD), were tested with Student t-test, while non-normal distribution was tested with Mann-Whitney U test, and the results were expressed as median (interquartile range). The Chi-square test or Fisher exact test was used for categorical variables, and the statistical results were expressed in terms of number (percentage). Logistic regression analysis was performed on the significant variables in univariate analysis to screen the independent factors of differential diagnosis. The receiver operating characteristic (ROC) curves of MRI, clinical features, and MRI combined with clinical features (MRI + Clinical features) were drawn. The area under the curve (AUC) was calculated to distinguish OTF with cystic degeneration from OAF. DeLong test was used to compare differences in AUCs. A two-tailed P value < 0.05 indicated statistical significance. The Hosmer-Lemeshow test was used to verify the goodness of fit of MRI + Clinical features (P > .05, representing the ideal calibration result), and the calibration curve was drawn to show the results. The average ADC values measured by two radiologists were used for data analysis. Spearman correlation analysis was used to analyse the correlation between CA125 levels and two tumour sizes. Post hoc analysis was performed using the Bonferroni method.
Result
Consistency checks of data
The Kappa coefficients (intra and inter) of the two radiologists ranged from 0.665 to 0.901 (P < .05) for qualitative data, and ICC (intra and inter) for quantitative data ranged from 0.781 to 0.919 (P < .05) for quantitative data, indicating good reproducibility.
Clinical features and differential diagnostic efficacy
A total of 68 female patients (mean age 51 years; age range 15-94 years) were included in this study and were divided into the OTF group (40 cases) and OAF group (28 cases) based on pathology. Table 2 summarized and compared some of the clinical characteristics of the OTF and OAF groups. The age (57.4 ± 14.2 vs. 44.1 ± 15, P = .001), BMI (23.9 ± 3.3 vs. 21.9 ± 2.9, P = .009), CEA (1.8 (0.9) vs. 1.1 (1.6), P = .017) and AFP (3.0 (1.6) vs. 2.0 (2.5), P = .041) in OTF group were significantly higher than those in OAF group. There were significantly more postmenopausal patients (72.5% vs. 28.6%, P < .001), more patients with endometrial hyperplasia (27.5% vs. 7.1%, P = .036), more right-sided lesions and fewer bilateral lesions (60% vs. 35.7%, P < .05; 2.5% vs. 21.4%, P < .05) in OTF group than in OAF group. Lower HE-4 (43.6 (17.4) vs. 63.8 (21.3), P = .012) was found in the OTF group, when compared to OAF group. However, the median values of HE-4, CEA, and AFP in both groups were in the normal range. No significant statistical differences were observed in the variables of symptoms or signs, parity, ROM, CA19-9 and CA125. Logistic regression analysis of clinical features showed that age and menopause were independent factors in distinguishing OTF from OAF. The AUC value was 0.841, with accuracy, sensitivity, and specificity of 0.824, 0.786, and 0.850, respectively (Table 5).
Table 2.
Clinical features of the patients.
| Features | OTF | OAF | χ2/t/U | P | |
|---|---|---|---|---|---|
| Age | 57.4 ± 14.2 | 44.1 ± 15.9 | 3.488 | .001‡ | |
| BMI (kg/m2) | 23.9 ± 3.3 | 21.9 ± 2.9 | 2.703 | .009‡ | |
| Menopause | Postmenopause | 29 (72.5%) | 8 (28.6%) | 12.813 | <.001* |
| Premenopause | 11 (27.5%) | 20 (71.4%) | 13.233 | ||
| Symptoms or signs | Asymptomatic | 21 (52.5%) | 13 (46.4%) | 4.293 | .354† |
| Pain or discomfort | 8 (20.0%) | 11 (39.3%) | |||
| Palpable mass | 5 (12.5%) | 3 (10.7%) | |||
| Frequent urination | 1 (2.5%) | 0 | |||
| Vaginal bleeding | 5 (12.5%) | 1 (3.6%) | |||
| Parity | Multipara | 37 (92.5%) | 21 (75.0%) | 2.747 | 0.097* |
| Nullipara | 3 (7.5%) | 7 (25.0%) | |||
| CA125 (U/mL) | 29.2 (28.5) | 26.5 (49.4) | 535 | 0.755§ | |
| HE-4 (pmol/L) | 43.6 (17.4) | 63.8 (21.3) | 337.500 | 0.012§ | |
| ROM | Normal | 31 (77.5%) | 17 (60.7%) | 2.235 | 0.135* |
| Abnormal | 9 (22.5%) | 11 (39.3%) | |||
| CEA (ng/mL) | 1.8 (0.9) | 1.1 (1.6) | 517 | 0.017§ | |
| AFP (ng/mL) | 3.0 (1.6) | 2.0 (2.5) | 690 | 0.041§ | |
| CA19-9 (U/mL) | 8.5 (9.4) | 11.5 (13.7) | 426 | 0.171§ |
Abbreviations: BMI = body mass index, CA125 = carbohydrate antigen 125, HE-4 = human epididymal protein 4, ROM = Rome index, CEA = carcino-embryonic antigen, AFP = alpha fetoprotein, CA19-9 = carbohydrate antigen 19-9.
Chi-squared test.
Fisher exact test.
Student t test.
Mann-Whitney U test.
Table 5.
Diagnostic efficacy of MRI, clinical features, and MRI combined with clinical features to discriminate between OTF with cystic degeneration and OAF.
| Model | Sensitivity | Specificity | Accuracy | F1-Score | AUC (95% CI) |
|---|---|---|---|---|---|
| MRI | 0.825 | 0.857 | 0.838 | 0.857 | 0.870 (0.785-0.954) |
| Clinical features | 0.786 | 0.850 | 0.824 | 0.786 | 0.841 (0.744-0.937) |
| MRI + clinical features | 0.893 | 0.925 | 0.912 | 0.893 | 0.954 (0.909-0.998) |
Among the 40 OTF, CA125 was elevated in 14 (35.0%) patients (mean, 81.893 U/mL, range, 3.7-895.5 U/mL, normal value, <35 U/mL). In the 14 patients with elevated CA125, 85.7% (12/14) of the tumours were larger than 5 cm in diameter. The results of Spearman correlation analysis showed a statistically significant correlation between tumour diameter and CA125 levels in the OTF group (r = 0.344, P = .030). In contrast, no significant correlation was seen between CA125 level and tumour diameter in the OAF group.
MRI findings and differential diagnostic efficacy
We compared the MRI features of the OTF and OAF groups, and representative cases of these two diseases are shown in Figures 2 and 3. As shown in Tables 3 and 4, the following MRI features were statistically different between the OTF and OAF groups: ① pelvic fluid (χ2 = 11.011, P = .012), ② cystic degeneration degree (χ2 = 16.656, P < .001), ③ cyst pattern (χ2 = 11.180, P = .003), ④ black sponge sign (χ2 = 8.293, P = .004), ⑤ ADCmin (χ2 = 271, P < .001), ⑥ ADCmean (χ2 = 268, P = .001), ⑦ T2WI signal (χ2 = 11.926, P = .008), and ⑧ FS-T2WI signal (χ2 = 14.158, P = .003). In this study, the OTF and OAF groups showed similar performance in terms of tumour haemorrhage, tumour angiogenesis, border, mean diameter, morphology, uterine fibroid or adenomyosis, ADCmax, T1WI signal, DWI signal, and contrast enhancement did not reflect significant differences.
Figure 2.
A 94-year-old female with pathologically proven right ovarian thecoma-fibroma. (A) Sagittal T2WI showed a well-defined multilocular cystic-solid mass with partially thickened, dark-signal-intensity walls (thin arrow) and a solid component (thick arrow) with low signal intensity (similar to skeletal muscle). (B, C) The solid area showed hyperintensity on DWI (b = 1000 s/mm2). The ADCmin at this level was 1.030 × 10−3 mm2/s. (D) Axial T1WI dynamic enhancement showed mild enhancement of the wall (thick arrow) and solid parts (thin arrow). Abbreviations: U = uterus, BL = urinary bladder.
Figure 3.
A 51-year-old female with an adenofibroma of the left ovary. (A) The T2-weighted, sagittal image showed a complex mass with an oval contour and multiple cysts (thick arrow). Due to the presence of collagen fibres, both solid areas and walls (thin arrow) showed very low T2 signal strength. b. Multiple tiny cystic signals were seen within the parenchyma on FS-T2WI (arrow). (C) The DWI (b = 1000 s/mm2) showed that the wall (thin arrow) was a high signal, while the solid part of the signal was uneven. The ADCmin at this level was 1.460 × 10−3 mm2/s. (D) After gadodiamide injection, fat-suppressed T1WI showed slight enhancement of the parenchymal area (thick arrow) and wall (thin arrow). Abbreviations: U = uterus, BL = urinary bladder.
Table 3.
Univariate analysis of MRI features between OTF group and OAF group.
| Features | OTF | OAF | χ 2/t | P | |
|---|---|---|---|---|---|
| Tumor haemorrhage | Present | 8 (20.0%) | 11 (39.3%) | 3.043 | .081* |
| Absent | 32 (80.0%) | 17 (60.7%) | |||
| Pelvic fluid | None | 8 (20.0%) a | 15 (53.6%) b | 11.011 | .012* |
| Little | 19 (47.5%) a | 10 (35.7%) a | |||
| Moderate | 10 (25.0%) a | 3 (10.7%) a | |||
| large | 3 (7.5%) a | 0 a | |||
| Tumour angiogenesis | Present | 8 (20.0%) | 11 (39.3%) | 3.043 | .081* |
| Absent | 32 (80.0%) | 17 (60.7%) | |||
| Border | Well defined | 38 (95.0%) | 28 (100%) | 0.223 | .637* |
| Ill defined | 2 (5.0%) | 0 | |||
| Mean diameter (cm) | 6.2 (6.5) | 5.8 (6.1) | 513.5 | .815§ | |
| Side of the lesion | Left | 15(37.5%) a | 12 (42.9%) a | 8.107 | .017* |
| Right | 24 (60.0%) a | 10 (35.7%) b | |||
| Bilateral | 1 (2.5%) a | 6 (21.4%) b | |||
| Morphology | Multinodular | 12 (30.0%) | 14 (50.0%) | 2.782 | .281* |
| Orbicular | 22 (55.0%) | 11 (39.3%) | |||
| Lobulated | 6 (15.0%) | 3 (10.7%) | |||
| Cystic degeneration degree | Grade 1 | 13 (32.5%) a | 0 b | 16.656 | <.001* |
| Grade 2 | 17 (42.5%) a | 9 (32.1%) a | |||
| Grade 3 | 10 (25.0%) a | 19 (67.9%) b | |||
| Cyst pattern | Mainly small | 17 (42.5%) a | 2 (7.1%) b | 11.180 | .003* |
| Mainly large | 11 (27.5%) a | 16 (57.1%) b | |||
| Mixed | 12 (30.0%) a | 10 (35.8%) a | |||
| Endometrial hyperplasia | Present | 11 (27.5%) | 2 (7.1%) | 4.414 | .036* |
| Absent | 29 (72.5%) | 26 (92.9%) | |||
| Uterine fibroid or adenomyosis | Present | 17 (42.5%) | 9 (32.1.0%) | 0.748 | .387* |
| Absent | 23 (57.5%) | 19 (67.9%) | |||
| Black sponge sign | Present | 8 (20.0%) | 15 (53.6%) | 8.293 | .004* |
| Absent | 32 (80.0%) | 13 (46.4%) | |||
| ADCmin (× 10−3 mm2/s) | 0.986 (0.152) | 1.255 (0.370) | 271 | <0.001§ | |
| ADCmean (× 10−3 mm2/s) | 1.317 (0.460) | 1.577 (0.382) | 268 | .001§ | |
| ADCmax (× 10−3 mm2/s) | 1.789 (0.223) | 1.798 (0.320) | 492.5 | 0.613§ |
Abbreviations: ADCmin = minimum ADC value, ADCmean = average ADC value, ADCmax = maximum ADC value.
The same subscript letters represent no significant difference between groups, while completely different subscript letters represent significant difference between groups.
Chi-squared test.
Fisher exact test.
Student t test.
Mann-Whitney U test.
Table 4.
Comparison of MRI signals between OTF group and OAF group.
| Sequence | OTF | OAF | χ2 | P |
|---|---|---|---|---|
| T1WI (Hypo-/Iso-/Hyper-/mixed) | 24/15/0/1 | 13/13/1/1 | 2.721 | .437* |
| T2WI (Hypo-/Iso-/Hyper-/mixed) | 9a/0a/7a/24a | 13b/3b/4a/8b | 11.926 | .008* |
| FS-T2WI (Hypo-/Iso-/Hyper-/mixed) | 7a/1a/6a/26a | 12b/5b/4a/7b | 14.158 | .003* |
| DWI (Hypo-/Iso-/Hyper-/mixed) | 9/3/15/13 | 10/2/10/6 | 1.773 | .621* |
| CE-A (mild/moderate/severe) | 25/14/1 | 12/16/0 | 4.057 | .132* |
| CE-V (mild/moderate/severe) | 12/26/2 | 12/15/1 | 1.199 | .549* |
| CE-D (mild/moderate/severe) | 12/25/3 | 5/22/1 | 2.079 | .364* |
Abbreviations: Hypo- = hypointensity, Iso = isointensity, Hyper- = hyperintensity, CE-V = contrast enhancement venous phase, CE-A = contrast enhancement arterial phase, CE-D = contrast enhancement delayed phase.
The same subscript letters represent no significant difference between groups, while completely different subscript letters represent significant difference between groups.
Chi-squared test.
All the variables with significant differences between the two groups were included in logistic regression analysis. The results show that cystic degeneration degree, black sponge sign, and ADCmin are the independent features that distinguish OTF group from OAF group. These representative MRI features distinguished the two diseases with an AUC of 0.870, with an accuracy, sensitivity, and specificity of 0.838, 0.825, and 0.857, respectively (Table 5).
Differential diagnostic efficacy of MRI combined with clinical features
We investigated the variability in MRI and clinical features between the OTF and OAF groups using univariate analysis. Logistic regression analysis was performed for features with significant differences in univariate analysis. The age, menopause, cystic degeneration degree, black sponge sign and ADCmin were the independent features to distinguish the two diseases. The AUC of MRI + Clinical features was 0.954, with accuracy, sensitivity, and specificity of 0.912, 0.893, and 0.925, respectively (Table 5). Delong test results showed that the AUC of MRI + Clinical features was significantly superior to the MRI or clinical features (Table 6). The ROC curve analysis yielded the following cut-offs for distinguishing OTF from OAF: ADCmin = 1.150 × 10−3 mm2/s, age = 56.5 years. Hosmer-lemeshow test was used to evaluate the calibration ability of MRI + Clinical features. The results showed that there was no significant difference between the predicted probability and the actual probability of MRI + Clinical features (χ2 = 12.059, P = .149), suggesting that the MRI + Clinical features had a high goodness of fit. The ROC curve and calibration curve are shown in Figure 4.
Table 6.
Pairwise comparison of ROC curves.
| AUCs | Difference between areas | 95% CI | P |
|---|---|---|---|
| MRI vs. clinical features | 0.029 | −0.105 to 0.163 | .670 |
| Clinical features vs. MRI + clinical features | 0.113 | 0.0269-0.199 | .010 |
| MRI vs. MRI + clinical features | 0.0838 | 0.0174-0.150 | .014 |
Figure 4.
(A) ROC curves and AUC for MRI, clinical features, and MRI + Clinical features to identify OTF with cystic degeneration and OAF. The red line represents MRI + Clinical features, the blue line represents clinical features, and the green line represents MRI. (B) Calibration curve of the MRI + Clinical features model. (The x-axis is the predicted probability of the model; the y-axis is the true probability. The dashed line is a perfect calibration curve, which indicates that the predicted probability is consistent with the true probability; The blue line is the calibration curve of the MRI + Clinical features model, and the closer it is to the dashed line, the better the predictive performance of the model.)
Discussion
Ultrasound is usually considered the first-line imaging method for ovarian tumours. However, the ultrasonic features of both OTF with cystic degeneration and OAF show a low-echo ovarian mass with attenuation of the posterior wall.8,9 The OAF and OTF with cystic degeneration often present as multilocular cystic or solid cystic masses with clear boundaries on CT, and the solid components are mildly and moderately enhanced.4,10 Therefore, it is difficult for ultrasound and CT to distinguish OAF and OTF with cystic degeneration. MRI has proven to be the preferred technique for characterizing tumours due to its good soft tissue resolution. MRI can show tumour components well, based on the multi-parameter, multi-sequence features. Both OTF and OAF are relatively rare ovarian tumors.11,12 At present, systematic studies are still lacking, and there is no literature on the comparative MRI findings of these two diseases specially. In this study, we evaluated the ability of MRI, clinical features, and a combination of both to distinguish OTF with cystic degeneration from OAF.
OTF is rarely seen in women before the age of 40, with postmenopausal women accounting for 73.1% of the total number of cases.13,14 This is similar to the results of the present study (29/40, 72.5%). The peak of OAF incidence was 40-49 years old,15 and the premenopausal prevalence of OAF was higher in this group. The results of this study suggest that 56.5 years old is the threshold age to distinguish between the two diseases. According to previous studies, OTF was a functional tumour, and it is associated with elevated estradiol, such as abnormal vaginal bleeding, thickening of the endometrium, and uterine leiomyoma or adenomyosis.7 As a result, endometrial thickening occurred in about 42.5% of OTF which was significantly higher than that in OAF. Nevertheless, there was no significant difference in the incidence of uterine fibroid or adenomyosis between the two groups.
Benign fibroma or fibroma-like tumours of the ovary (such as thecoma and granulosa cell tumour) with pleural and ascites are known as Meigs Syndrome.16 In this study, pelvic fluid was more common in the OTF group than in the OAF group. The 80.0% (32/40) of our OTF cases presented with pelvic fluid, and moderate and large pelvic fluid accounted for 25% (10/40) and 7.5% (3/40) respectively. The pathophysiological mechanisms of pelvic fluid are not well understood and may be related to direct tumour secretion, obstruction of lymphatic vessels due to compression, and direct stimulation of the peritoneum.17,18 Unlike the OTF group, only 46.4% (13/28) of OAF had pelvic fluid, and only 3 (10.7%, 3/28) had moderate pelvic fluid. In the present study, there was no statistically significant difference in CA125 between the OTF and OAF groups. We performed a Spearman correlation analysis of the mean tumour diameter and CA125 levels in the OTF group. The correlation between mean tumour diameter and CA125 level in the OTF group (r = 0.344, P = .030) was found to be statistically significant, which is consistent with the findings of Zhang et al.19 It further validates that tumour irritation of the peritoneal may be responsible for the elevated CA125 level.20
The typical OAF is composed of both epithelium and fibrous stroma. The OTF with cystic degeneration and OAF could locally show a very low signal-intensity appearance in T2WI and FS-T2WI, which is the T2 shortening effect of collagen. The solid component of OAF may exhibit the characteristic black sponge sign, that is, a solid component of very low signal intensity containing tiny foci of high T2 signal intensity, which are actually small cystic glandular structures scattered in the dense fibrous interstitium.21 In our study, this manifestation was detected in both diseases on T2WI and FS-T2WI sequences. There were 15 (15/28, 53.6%) cases of the masses in the OAF group that exhibited the black sponge sign, significantly more than the 8 (8/40, 20.0%) cases in the OTF group (P = .004). The reason may be that the microcystic changes in OTF are arranged in the fibrous stroma, which is different from OAF in pathology. Logistic regression analysis showed that the black sponge sign was an independent indicator to differentiate OTF with cystic degeneration from OAF. When the black sponge sign is present, we believe that OAF should be considered as the first diagnosis.
The OTF are prone to degenerative changes, especially when the tumour is large, such as cystic degeneration, oedema, myxoid degeneration, haemorrhagic necrosis, and so on, of which cystic change is the most common.22 Cystic degeneration is caused by insufficient blood supply to the tumour and ischemia. At this time, the tumour may appear as a cystic-solid mass or cystic mass. According to the proportion of cystic area to tumour area, the cystic degeneration was classified into grade 1,2, and 3. The majority of the cystic degeneration in OTF with cystic degeneration was grade 2 (42.5%, 17/40), but the number of the three grades was not too different. However, there were no grade 1 cystic lesions in all OAF cases in this group. By logistic regression analysis, we further confirmed that cystic degeneration degree was an independent indicator to distinguish OTF with cystic degeneration from OAF. It may be related to the strong secretory activity of the interstitial cells contained in OAF.
DWI assesses tissue and its microenvironment based on Brownian motion of intracellular and extracellular water molecules and examines the degree of diffusion motion of water molecules.23 It is widely used in gynaecologic malignant and benign diseases, plays an important role in diagnosis and auxiliary diagnosis, and provides useful supplementary information for its management.24,25 Fang et al26 reported that the mean ADC of thecomas (1.30 ± 0.25 × 10−3 mm2/s) was significantly greater than that of adult granulosa cell tumours (0.87 ± 0.28 × 10−3 mm2/s) (P < .05). Türkoğlu et al27 retrospectively analysed the ability of DWI and other multiparametric MRI to identify benign and malignant ovarian masses in 43 patients and found that DWI and dynamic enhancement (DCE) MRI were highly accurate in identifying benign and malignant ovarian masses. It can help women with ovarian masses to choose the best treatment. A study by Wei et al28 found that uterine fibroids in the adnexal area had significantly higher ADC values than thecomas/fibrothecomas (P < .05). No signal increase was observed in 42.9% (12/28) of OAF patients on DWI, which may be the result of T2 blocking effect. The reason for the 35.7% (10/28) increase in signal is unclear, it may be that some OAF epithelial cells are poorly differentiated and form borderline and malignant lesions, resulting in a relative increase in cell density,29 or the restricted diffusion of water within the microcapsule lumen due to the viscous mucus contents.21 In OTF Group, 37.5% (15/40) patients showed high signal intensity in DWI, which may be related to the enrichment of membrane cells in these cases. We also investigated the ability of ADC values (including ADCmean, ADCmin, ADCmax) to discriminate between these two groups. The OTF group had lower ADC values than OAF, especially ADCmin. By plotting the ROC curve, we obtained a cut-off value for ADCmin, and when ADCmin is less than 1.150 × 10−3 mm2/s, the diagnosis may be more favourable to OTF.
A combination of MRI and clinical features for discriminating OTF from OAF yielded an AUC of 0.954, and the accuracy, sensitivity, and specificity were improved to a certain extent compared to MRI or clinical features alone. Hence, we believe that MRI combined with clinical features is a promising tool and offers a window of opportunity to distinguish OTF with cystic degeneration from OAF and improve the awareness of radiologists and gynaecologists about these two diseases.
Our study has several limitations. First, the number of patients in our study was limited. The results may not be robust enough. However, considering that such patients are rare, and the results are somewhat interpretable, they are still presented. More cases are needed to further validate the value of these MRI features in discriminating OTF with cystic degeneration from OAF. Second, this study is a single-centre retrospective study, and selection bias is inevitable. Third, this study used a single index model of DWI, which did not exclude the effect of blood perfusion on the diffusion of water molecules. In the future, there is a need to combine intravoxel incoherent motion diffusion weighted imaging based on a double exponential model that can reflect both true diffusion (molecular motion) and pseudo-diffusion (blood perfusion) of the tissue.
Conclusion
In summary, OTF with cystic degeneration has lower cystic degeneration degree and ADCmin, less black sponge, higher age, and more postmenopausal woman compared with OAF. Cystic degeneration degree, black sponge sign, ADCmin, age and menopause were independent features in the differential diagnosis of these two diseases. The MRI combined with clinical features have a good ability to distinguish OTF with cystic degeneration from OAF. The predictive capacity of MRI combined with clinical features is superior to individual MRI and clinical features.
Acknowledgements
The authors are grateful to all involved female patient for their participation in the study.
Contributor Information
Juan Bo, Department of Radiology, Anhui Provincial Hospital Affiliated to Anhui Medical University, Hefei, Anhui 230001, China.
Mingjie Sun, Faculty of Graduate Studies, Wannan Medical College, Wuhu, Anhui 241002, China.
Chao Wei, Department of Radiology, Western District, First Affiliated Hospital of University of Science and Technology of China, No.107 Huanhu East Road, Shushan District, Hefei, Anhui, 230031, China.
Longyu Wei, Faculty of Graduate Studies, Bengbu Medical College, Bengbu, Anhui 233030, China.
Baoyue Fu, Faculty of Graduate Studies, Bengbu Medical College, Bengbu, Anhui 233030, China.
Bin Shi, Department of Radiology, Western District, First Affiliated Hospital of University of Science and Technology of China, No.107 Huanhu East Road, Shushan District, Hefei, Anhui, 230031, China.
Xin Fang, Department of Radiology, Western District, First Affiliated Hospital of University of Science and Technology of China, No.107 Huanhu East Road, Shushan District, Hefei, Anhui, 230031, China.
Jiangning Dong, Department of Radiology, Anhui Provincial Hospital Affiliated to Anhui Medical University, Hefei, Anhui 230001, China; Department of Radiology, Western District, First Affiliated Hospital of University of Science and Technology of China, No.107 Huanhu East Road, Shushan District, Hefei, Anhui, 230031, China.
Author contributions
JD and JB conceptualized and designed the study. JB and MS collected and arranged the data. JB, MS, CW, LW, BF, and BS analysed the data and wrote the manuscript. XF implemented MR examination. All authors contributed to the article and approved the submitted version. JB and MS are co-first authors and contributed equally to this study.
Funding
This study was supported by Key Research and Development Projects of Anhui Province (no. 2022e07020008) and The University Science Technology Research Project of Anhui Educational Committee (no. 2022AH051262).
Conflicts of interest
The authors declare that they have no conflicts of interest regarding the publication of this article.
References
- 1. Troiano RN, Lazzarini KM, Scoutt LM, Lange RC, Flynn SD, McCarthy S.. Fibroma and fibrothecoma of the ovary: MR imaging findings. Radiology. 1997;204(3):795-798. [DOI] [PubMed] [Google Scholar]
- 2. Cho SM, Byun JY, Rha SE, et al. CT and MRI findings of cystadenofibromas of the ovary. Eur Radiol. 2004;14(5):798-804. 10.1007/s00330-003-2060-z [DOI] [PubMed] [Google Scholar]
- 3. Moch H. WHO Classification of Tumours Editorial Board. Female Genital Tumours: WHO Classification of Tumours, 5th edn, volume 4. IARC Press; 2020.105-106. [Google Scholar]
- 4. Taylor EC, Irshaid L, Mathur M.. multimodality imaging approach to ovarian neoplasms with pathologic correlation. Radiographics. 2021;41(1):289-315. 10.1148/rg.2021200086 [DOI] [PubMed] [Google Scholar]
- 5. Lee YR. CT imaging findings of ruptured ovarian endometriotic cysts: emphasis on the differential diagnosis with ruptured ovarian functional cysts. Korean J Radiol. 2011;12(1):59-65. 10.3348/kjr.2011.12.1.59 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Chung BM, Park SB, Lee JB, Park HJ, Kim YS, Oh YJ.. Magnetic resonance imaging features of ovarian fibroma, fibrothecoma, and thecoma. Abdom Imaging. 2015;40(5):1263-1272. 10.1007/s00261-014-0257-z [DOI] [PubMed] [Google Scholar]
- 7. 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(4):907-911. 10.1002/jmri.23711 [DOI] [PubMed] [Google Scholar]
- 8. Virgilio BA, De Blasis I, Sladkevicius P, et al. Imaging in gynecological disease (16): clinical and ultrasound characteristics of serous cystadenofibromas in adnexa. Ultrasound Obstet Gynecol. 2019;54(6):823-830. 10.1002/uog.20277 [DOI] [PubMed] [Google Scholar]
- 9. Chen H, Liu Y, Shen LF, Jiang MJ, Yang ZF, Fang GP.. Ovarian thecoma-fibroma groups: clinical and sonographic features with pathological comparison. J Ovarian Res. 2016;9(1):81. 10.1186/s13048-016-0291-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Zhang Z, Wu Y, Gao J.. CT diagnosis in the thecoma-fibroma group of the ovarian stromal tumors. Cell Biochem Biophys. 2015;71(2):937-943. 10.1007/s12013-014-0288-7 [DOI] [PubMed] [Google Scholar]
- 11. Chechia A, Attia L, Temime RB, Makhlouf T, Koubaa A.. Incidence, clinical analysis, and management of ovarian fibromas and fibrothecomas. Am J Obstet Gynecol. 2008;199(5):473.e1-473.e4734. 10.1016/j.ajog.2008.03.053 [DOI] [PubMed] [Google Scholar]
- 12. Avesani G, Caliolo G, Gui B, et al. Pearls and potential pitfalls for correct diagnosis of ovarian cystadenofibroma in MRI: a pictorial essay. Korean J Radiol. 2021;22(11):1809-1821. 10.3348/kjr.2020.1312 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Wu B, Peng WJ, Gu YJ, Cheng YF, Mao J.. MRI diagnosis of ovarian fibrothecomas: tumour appearances and oestrogenic effect features. Br J Radiol. 2014;87(1038):20130634. 10.1259/bjr.20130634 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Li X, Zhang W, Zhu G, Sun C, Liu Q, Shen Y.. Imaging features and pathologic characteristics of ovarian thecoma. J Comput Assist Tomogr. 2012;36(1):46-53. 10.1097/RCT.0b013e31823f6186 [DOI] [PubMed] [Google Scholar]
- 15. Shukla S, Srivastava D, Acharya S, Dhote S, Vagha S.. Serous adenofibroma of ovary: an eccentric presentation. J Cancer Res Ther. 2015;11(4):1030. 10.4103/0973-1482.150419 [DOI] [PubMed] [Google Scholar]
- 16. Saha S, Robertson M.. Meigs' and pseudo-Meigs' syndrome. Australas J Ultrasound Med. 2012;15(1):29-31. 10.1002/j.2205-0140.2012.tb00140.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Mohammed SA, Kumar A.. Meigs syndrome. In: StatPearls. StatPearls Publishing; 2022:3. [PubMed] [Google Scholar]
- 18. Kortekaas KE, Pelikan HM.. Hydrothorax, ascites and an abdominal mass: not always signs of a malignancy—Three cases of Meigs' syndrome. J Radiol Case Rep. 2018;12(1):17-26. 10.3941/jrcr.v12i1.3209 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Zhang S, Li Y, Wang M, et al. Pelvic effusion in patients with OTF associated with the tumor size and plaste CA-125 level: A retrospective magnetic resonance imaging study. Clin Imaging. 2022;81:62-66. 10.1016/j.clinimag.2021.06.007 [DOI] [PubMed] [Google Scholar]
- 20. Shen Y, Liang Y, Cheng X, Lu W, Xie X, Wan X.. Ovarian fibroma/fibrothecoma with elevated serum CA125 level: A cohort of 66 cases. Medicine (Baltimore). 2018;97(34):e11926. 10.1097/MD.0000000000011926 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Takeuchi M, Matsuzaki K, Harada M.. Ovarian adenofibromas and cystadenofibromas: magnetic resonance imaging findings including diffusion-weighted imaging. Acta Radiol. 2013;54(2):231-236. 10.1258/ar.2012.120516 [DOI] [PubMed] [Google Scholar]
- 22. Kitajima K, Kaji Y, Sugimura K.. Usual and unusual MRI findings of ovarian fibroma: correlation with pathologic findings. Magn Reson Med Sci. 2008;7(1):43-48. [DOI] [PubMed] [Google Scholar]
- 23. Caroli A. Diffusion-Weighted magnetic resonance imaging: clinical potential and applications. J Clin Med. 2022;11(12):3339. 10.3390/jcm11123339 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Bonde A, Andreazza Dal Lago E, Foster B, Javadi S, Palmquist S, Bhosale P.. Utility of the diffusion weighted sequence in gynecological imaging: review article. Cancers (Basel). 2022;14(18):4468. 10.3390/cancers14184468 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Bakir B, Bakan S, Tunaci M, et al. Diffusion-weighted imaging of solid or predominantly solid gynaecological adnexial masses: is it useful in the differential diagnosis? Br J Radiol. 2011;84(1003):600-611. 10.1259/bjr/90706205 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Fang M, Dong J, Zhong Q, et al. Value of diffusion-weighted imaging combined with conventional magnetic resonance imaging in the diagnosis of thecomas and their differential diagnosis with adult granulosa cell tumors. Acta Radiol. 2019;60(11):1532-1542. 10.1177/0284185119830280 [DOI] [PubMed] [Google Scholar]
- 27. Türkoğlu S, Kayan M.. Differentiation between benign and malignant ovarian masses using multiparametric MRI. Diagn Interv Imaging. 2020;101(3):147-155. 10.1016/j.diii.2020.01.006 [DOI] [PubMed] [Google Scholar]
- 28. Wei C, Chen YL, Li XX, et al. Diagnostic performance of MR imaging-based features and texture analysis in the differential diagnosis of ovarian thecomas/fibrothecomas and uterine fibroids in the adnexal area [published correction appears in Acad Radiol. 2021 Sep;28(9):1329]. Acad Radiol. 2020;27(10):1406-1415. 10.1016/j.acra.2019.12.025 [DOI] [PubMed] [Google Scholar]
- 29. Kozawa E, Inoue K, Takahashi M, Kato T, Yasuda M, Kimura F.. Diffusion-weighted MR imaging findings of ovarian adenocarcinofibromas and adenofibromas. Clin Imaging. 2014;38(4):483-489. 10.1016/j.clinimag.2014.01.014 [DOI] [PubMed] [Google Scholar]