Abstract
Dynamic contrast-enhanced (DCE) MRI is not available in all imaging centres to investigate adnexal masses. We proposed modified magnetic resonance (MR) scoring system based on an assessment of the enhancement of the solid tissue on early phase postcontrast series and diffusion-weighted imaging (DWI) with apparent diffusion coefficient (ADC) map and investigated the validity of this protocols in the current study.
In this cross-sectional retrospective study, pelvic MRI of a total of 245 patients with 340 adnexal masses were studied based on the proposed modified scoring system and ADNEX MR scoring system.
Modified scoring system with the sensitivity of 87.3% and specificity of 94.6% has an accuracy of 92.1%. Sensitivity, specificity, and accuracy of ADNEX MR scoring system is 96.6%, 91%, and 92.9%, respectively. The area under the receiver operating characteristic curve for the modified scoring system and ADNEX MR scoring system is 0.909 (with 0.870-0.938 95% confidence interval [CI]) and 0.938 (with 0.907-0.961 95% CI), respectively. Pairwise comparison of these area under the curves showed no significant difference (P = .053).
Modified scoring system is less sensitive than the ADNEX MR scoring system and more specific but the accuracy is not significantly different.
According to our study, MR scoring system based on subjective assessment of the enhancement of the solid tissue on early phase postcontrast series and DWI with ADC map could be applicable in imaging centres that DCE is not available.
Introduction
Ovarian cancer is one of the most frequent cancers in women (3.6% of all malignancies), with a 4.3% fatality rate.1 Late diagnosis is one of the main factor leading to high mortality because the illness is frequently asymptomatic in the early stages or has vague symptoms. A multidisciplinary approach based on clinical findings, laboratory findings, and imaging modalities is required for the optimal assessment of ovarian masses. In assessing adnexal lesions, ultrasound is the first-line step modality.2 Most ovarian masses are benign, and most are correctly diagnosed as benign or malignant after an ultrasound scan. However, ultrasonography studies indicate that 18%-31% of ovarian masses are indeterminate.3,4 Despite the low malignancy rate among resected ovarian masses (8%-20% in resected lesions),5,6 a considerable percentage of women with benign adnexal masses is reported indeterminate on ultrasonography, endure unnecessary surgical procedures, and are at risk of infertility and morbidity.7,8
In contrast, there is a danger of deteriorating the prognosis with restricted, primary, and noncancerous surgery for malignant tumours reported indeterminate on ultrasonography.9 As a result, the characterization of masses and preoperative risk evaluation of indeterminate masses on ultrasound is a clinical requirement that necessitates a reliable scoring system to standardize imaging reports and evaluate the risk of malignancy.10,11 Utilizing the scoring system; we could detect lesions requiring surgery and the extent of the operation. MRI is beneficial in assessing lesions that are reported indeterminate on ultrasonography. The accuracy of MRI is significantly more (88.9%) than in transvaginal ultrasound (63.9%).12 With the inclusion of perfusion and diffusion-weighted sequences, compared to traditional MRI, total accuracy has increased by more than 90%.13–16
Several MRI protocols have been developed to investigate adnexal masses.16-15 In the diagnosis of malignant adnexal masses, Thomassin-Naggara’s MRI scoring system for adnexal lesions (ADNEX MR scoring system) based on dynamic contrast-enhanced images showed a sensitivity of 93.5%-94.9% and a specificity of 92.9%-96.6%.17–20
Dynamic contrast-enhanced (DCE) is unavailable in all imaging centres due to time limitations or lack of perfusion curve analysis software. In such centres, visual assessment of the enhancement of the solid tissue on early phase postcontrast series can be performed as a substitute. Some imaging centres do not perform time-intensity curve analysis because the potential for improved diagnostic performance is questionable, and the technique requires an investment of resources and time constraints. Based on the Guide for Assessing Adnexal Lesions from the ACR O-RADS Committee in 2022 and a related letter published recently in October 2022, visual analysis of the enhancement of the solid tissue suggested on the 30-40-s postcontrast series.21,22
In this study, we focused on adnexal masses utilizing 2 different protocols: the ADNEX MR scoring system introduced in 2013 by Thomassin-Naggara17 and the modified magnetic resonance (MR) scoring system based on the enhancement of the solid tissue on the 30-s postcontrast series and diffusion-weighted imaging (DWI) with apparent diffusion coefficient (ADC) map. We investigated the validity of these 2 protocols in the current study.
Materials and methods
Patients
In this multicentric cross-sectional retrospective study, pelvic MRI was performed on all women over 18 years old due to indeterminate ultrasound findings from March 2016 to August 2021. Patients whose MRI was evaluated had been referred to a tertiary imaging centre with a 3-T MR imaging unit (GE 3Tdiscovery750Gem, phase array pelvic coil surface). We included all patients whose histopathological result after adnexal mass surgery was accessible or who underwent follow-up for at least 1 year and whose follow-up information was accessible. Patients whose subsequent surgical or follow-up data were unavailable, and cases with inappropriate MRI images were excluded (Figure 1).
Flowchart of study population.
A total of 245 individuals with a mean age of 42.77 ± 13.19 years and the range of 18 (except for one case of 7 years old girl with fat-containing adnexal mass indicative of mature teratoma) to 84 years with 340 adnexal masses were studied, of which 87 patients (35.5%) had more than one adnexal mass. Also, 198 masses underwent surgery, and their histopathological findings were available.
MRI protocol and evaluation
All of the patients had been referred to an imaging centre with a 3-T MR imaging unit (GE 3Tdiscovery750Gem, phase array pelvic coil surface). Before the MRI, patients had to stay fasting for at least 3 h.
T2-weighted fast spin echo was conducted in the axial, sagittal, and coronal planes, as well as axial T2-weighted sequences with fat suppression and T1-sequences with and without fat suppression. Diffusion-weighted images in the axial plane with b-values of 50, 500, and 1000 s/mm were also obtained, as is common in clinical practice. Contrast material gadolinium chelate was injected with a dose of 0.2 ml/kg body weight at a rate of 3.5 ml/s using an injector followed by a 10 ml infusion of normal saline. After injecting gadolinium, dynamic contrast-enhanced images were acquired, including 5 images 30 s apart.
Image analysis
Two radiologists with at least 10 years of expertise in pelvic MRI reporting independently and without knowing clinical and laboratory findings, histopathology, or follow-up data examined the MRI pictures and interpreted the results using the ADNEX MR scoring system and modified scoring system.
If there was a solid part, the time-intensity curve (TIC) was studied in comparison with the external myometer in dynamic contrast-enhanced sequences in the work station.
In the proposed modified scoring system, just a subjective assessment of solid part enhancement in the early phase (at 30 s) compared to external myometrium was considered for characterization.
The signal intensity in DWI (containing the different b-values and ADC map) is also assessed for solid-enhancing tissues. Blood-degradation products and fat typically show low ADC due to the shortened T1 values, which makes mature teratomas and endometriomas common pitfalls on DWI. To avoid such misinterpretation specifically for fatty or haemorrhagic lesions, T2, T1, and T1 with fat-suppression weighted sequences were considered in DWI analysis. Also, adnexal lesions with typical imaging features of dermoid cyst or endometrioma were classified as benign(score = 2), and further evaluation was not performed. Furthermore, the signal intensity on DWI (containing the different b-values and ADC map) is assessed only for solid-enhancing tissues. In heterogeneous masses, the most restricted part was chosen for assessment. According to previous studies, solid parts include solid masses, papillary projections, and thickened irregular septa.17,23 In these enhancing tissues, a combination of high b1000 and low ADC based on subjective assessment,24 was considered malignant according to the proposed modified scoring system as mentioned in Table 1 (Figure 2).
Images from 43-year-old woman with left-sided solid-cystic adnexal mass indeterminate in transvaginal ultrasound. (A) Axial T1-weighted and (B) axial T2-weighted magnetic resonance (MR) images show left-sided multiloculated adnexal mass, with largest diameter of 85 mm, with intermediate signal-intensity locules and intermediate-high signal-intensity solid tissues. (C and D) Axial diffusion-weighted image showing, high signal intensity at b = 1000 s/mm2 and low apparent diffusion coefficient signal intensity of solid enhancing tissues. (E and F) dynamic contrast-enhanced MRI yielded high risk time-intensity curve in comparison with adjacent external myometrium. Histology confirmed malignant granulosa cell tumour.
ADNEX MR scoring system and modified scoring system.
| ADNEX MR score | Classical scoring system criteria | Modified scoring system criteria |
|---|---|---|
| 1. No mass | Not included | Not included |
| 2. Benign mass | Purely cystic mass/purely endometriotic mass/purely fatty mass/absence of wall enhancement/low b1000 and low T2 within solid tissue/complex cystic-fatty lesions/lesions without enhancing solid tissue | Similar criteria to classical scoring system |
| 3. Possibly benign mass | Absence of solid tissue TIC curve type 1 within solid tissue | Enhancing solid tissue ≤myometrium at 30 s on non-DCE MRI And Low b1000 and high ADC |
| 4. Indeterminate mass | TIC curve type 2 within solid tissue | Enhancing solid tissue ≤myometrium at 30 s on non-DCE MRI And High b1000 and low ADC |
| 5. Possibly malignant mass | Peritoneal implants TIC curve type 3 within solid tissue | Enhancing solid tissue >myometrium at 30 s on non-DCE MRI |
| ADNEX MR score | Classical scoring system criteria | Modified scoring system criteria |
|---|---|---|
| 1. No mass | Not included | Not included |
| 2. Benign mass | Purely cystic mass/purely endometriotic mass/purely fatty mass/absence of wall enhancement/low b1000 and low T2 within solid tissue/complex cystic-fatty lesions/lesions without enhancing solid tissue | Similar criteria to classical scoring system |
| 3. Possibly benign mass | Absence of solid tissue TIC curve type 1 within solid tissue | Enhancing solid tissue ≤myometrium at 30 s on non-DCE MRI And Low b1000 and high ADC |
| 4. Indeterminate mass | TIC curve type 2 within solid tissue | Enhancing solid tissue ≤myometrium at 30 s on non-DCE MRI And High b1000 and low ADC |
| 5. Possibly malignant mass | Peritoneal implants TIC curve type 3 within solid tissue | Enhancing solid tissue >myometrium at 30 s on non-DCE MRI |
ADNEX MR scoring system and modified scoring system.
| ADNEX MR score | Classical scoring system criteria | Modified scoring system criteria |
|---|---|---|
| 1. No mass | Not included | Not included |
| 2. Benign mass | Purely cystic mass/purely endometriotic mass/purely fatty mass/absence of wall enhancement/low b1000 and low T2 within solid tissue/complex cystic-fatty lesions/lesions without enhancing solid tissue | Similar criteria to classical scoring system |
| 3. Possibly benign mass | Absence of solid tissue TIC curve type 1 within solid tissue | Enhancing solid tissue ≤myometrium at 30 s on non-DCE MRI And Low b1000 and high ADC |
| 4. Indeterminate mass | TIC curve type 2 within solid tissue | Enhancing solid tissue ≤myometrium at 30 s on non-DCE MRI And High b1000 and low ADC |
| 5. Possibly malignant mass | Peritoneal implants TIC curve type 3 within solid tissue | Enhancing solid tissue >myometrium at 30 s on non-DCE MRI |
| ADNEX MR score | Classical scoring system criteria | Modified scoring system criteria |
|---|---|---|
| 1. No mass | Not included | Not included |
| 2. Benign mass | Purely cystic mass/purely endometriotic mass/purely fatty mass/absence of wall enhancement/low b1000 and low T2 within solid tissue/complex cystic-fatty lesions/lesions without enhancing solid tissue | Similar criteria to classical scoring system |
| 3. Possibly benign mass | Absence of solid tissue TIC curve type 1 within solid tissue | Enhancing solid tissue ≤myometrium at 30 s on non-DCE MRI And Low b1000 and high ADC |
| 4. Indeterminate mass | TIC curve type 2 within solid tissue | Enhancing solid tissue ≤myometrium at 30 s on non-DCE MRI And High b1000 and low ADC |
| 5. Possibly malignant mass | Peritoneal implants TIC curve type 3 within solid tissue | Enhancing solid tissue >myometrium at 30 s on non-DCE MRI |
In the proposed modified scoring system, the score is measured as in Table 1. The calculation of scores in the classic ADNEX MR scoring system is also presented in Table 1.
Reference standard
The standard indicator in this study is histopathology.25 In statistical analysis, masses with borderline histopathology reports were classified as malignant.24 Cases that were not candidates for surgery and whose histopathological results were unavailable were evaluated for at least 1 year based on clinical monitoring and imaging.
Statistical analysis
SPSS Statistics, version 22, and MedCalc software, version 20.1.4, were used to analyse the data. We reported quantitative variables with mean and standard deviation and qualitative variables with number and age. Using the chi-square test, we investigated the relationship between patients’ baseline information and histology type and mass type. We assessed the connection between the scoring systems findings and the type of adnexal masses using chi-squared and Fisher’s exact tests. We calculated diagnostic parameters, including sensitivity, specificity, positive and negative predictive values, accuracy, and area under the receiver operating characteristic (ROC) for the ADNEX MR scoring system and proposed modified MR scoring system. P < .05 was considered significant in all tests. We considered a score ≥4 as a cut-off for malignancy.
Interobserver agreement was assessed based on unweighted and Fleiss kappa indices.
Results
Patients with malignant masses were substantially older than those with benign masses (40.89 years, whereas for 46.30 years, P < .001). The average mass size was 73.12 ± 47.88 mm, ranging from 7 to 360 mm. The size of malignant masses was significantly larger than benign masses (88.36, whereas for 64.97 mm, P < .001). Also, 58.2% (198) underwent surgery, and histopathology results are available. In addition, 41.8% (142) were followed for at least 1 year, and 81.8% of adnexal masses originated from the ovary (Table 2).
The association between patients’ baseline information and adnexal mass types in patients.
| Variable | Mean ± SD/N (%) | |
|---|---|---|
| Age (year) | – | 42.77 ± 13.19 |
| Largest diameter (mm) | – | 73.12 ± 47.88 |
| Pathology | – | Number (%) |
| Benign | 222 (65.3) | |
| Malignant | 94 (27.6) | |
| Borderline | 24 (7.1) | |
| – | Follow up | 142 (41.8) |
| Surgery | 198 (58.2) | |
| Location | Ovarian | 278 (81.8) |
| Non-ovarian | 62 (18.2) | |
| Early enhancement | Negative | 47.1 (160) |
| Positive | 52.9 (180) |
| Variable | Mean ± SD/N (%) | |
|---|---|---|
| Age (year) | – | 42.77 ± 13.19 |
| Largest diameter (mm) | – | 73.12 ± 47.88 |
| Pathology | – | Number (%) |
| Benign | 222 (65.3) | |
| Malignant | 94 (27.6) | |
| Borderline | 24 (7.1) | |
| – | Follow up | 142 (41.8) |
| Surgery | 198 (58.2) | |
| Location | Ovarian | 278 (81.8) |
| Non-ovarian | 62 (18.2) | |
| Early enhancement | Negative | 47.1 (160) |
| Positive | 52.9 (180) |
The association between patients’ baseline information and adnexal mass types in patients.
| Variable | Mean ± SD/N (%) | |
|---|---|---|
| Age (year) | – | 42.77 ± 13.19 |
| Largest diameter (mm) | – | 73.12 ± 47.88 |
| Pathology | – | Number (%) |
| Benign | 222 (65.3) | |
| Malignant | 94 (27.6) | |
| Borderline | 24 (7.1) | |
| – | Follow up | 142 (41.8) |
| Surgery | 198 (58.2) | |
| Location | Ovarian | 278 (81.8) |
| Non-ovarian | 62 (18.2) | |
| Early enhancement | Negative | 47.1 (160) |
| Positive | 52.9 (180) |
| Variable | Mean ± SD/N (%) | |
|---|---|---|
| Age (year) | – | 42.77 ± 13.19 |
| Largest diameter (mm) | – | 73.12 ± 47.88 |
| Pathology | – | Number (%) |
| Benign | 222 (65.3) | |
| Malignant | 94 (27.6) | |
| Borderline | 24 (7.1) | |
| – | Follow up | 142 (41.8) |
| Surgery | 198 (58.2) | |
| Location | Ovarian | 278 (81.8) |
| Non-ovarian | 62 (18.2) | |
| Early enhancement | Negative | 47.1 (160) |
| Positive | 52.9 (180) |
Notably, 65.3% of the adnexal masses were benign, and 34.7% were malignant. The prevalence of borderline masses was 7.1%. Tumours of epithelial origin were the most prevalent (81 masses/68.6%) among 118 masses with malignant pathology. Metastases were the second most prevalent type of malignant tumour (28 masses/23.9%), followed by tumours of SEX CORD origin (7 masses/6%) and GERM CELL (2 masses/1.7%) (Table 3). The enhancing solid part was depicted in 52.6% (179) of the masses.
Adnexal masses subtypes based on MRI finding, histopathology, and follow-up imaging.
|
| Serous cystadenoma | N = 14 |
| Serous adenofibroma | N = 3 | ||
| Mucinous cystadenoma | N = 6 | ||
| Seromucinous cystadenoma | N = 1 | ||
| Endometriotic cyst | N = 56 | ||
| Brenner tumour | N = 2 | ||
| Mature teratoma | N = 33 | ||
| Ovarian fibrothecoma | N = 9 | ||
| Ovarian thecoma | N = 2 | ||
| Ovarian fibroma | N = 2 | ||
| Ovarian leiomyoma | N = 2 | ||
| Luteinized follicular cyst | N = 1 | ||
| Serous inclusion cyst of the ovary | N = 1 | ||
| Tumour-like lesions (functional and corpus luteum cyst, OHSS, and tubo-ovarian abscess) | N = 28 | ||
| Hydrosalpinx, pyosalpinx, hematosalpinx, salpingitis, peritoneal inclusion cyst, broad ligament leiomyomas, fibroma, hematoma and EP | N = 62 | |
| Serous borderline tumour | N = 10 | |
| Mucinous borderline tumour | N = 12 | ||
| Seromucinous borderline tumour | N = 1 | ||
| Endometrioid borderline tumour | N = 1 | ||
| Serous cystadenocarcinoma | N = 37 | |
| Endometrioid carcinoma | N = 6 | ||
| Clear cell carcinoma | N = 2 | ||
| Mucinous cystadenocarcinoma | N = 10 | ||
| Seromucinous Cystadenocarcinoma | N = 2 | ||
| Granulosa cell tumour | N = 7 | ||
| Germ cell tumour | N = 2 | ||
| Metatstatic lesions | N = 28 | ||
|
| Serous cystadenoma | N = 14 |
| Serous adenofibroma | N = 3 | ||
| Mucinous cystadenoma | N = 6 | ||
| Seromucinous cystadenoma | N = 1 | ||
| Endometriotic cyst | N = 56 | ||
| Brenner tumour | N = 2 | ||
| Mature teratoma | N = 33 | ||
| Ovarian fibrothecoma | N = 9 | ||
| Ovarian thecoma | N = 2 | ||
| Ovarian fibroma | N = 2 | ||
| Ovarian leiomyoma | N = 2 | ||
| Luteinized follicular cyst | N = 1 | ||
| Serous inclusion cyst of the ovary | N = 1 | ||
| Tumour-like lesions (functional and corpus luteum cyst, OHSS, and tubo-ovarian abscess) | N = 28 | ||
| Hydrosalpinx, pyosalpinx, hematosalpinx, salpingitis, peritoneal inclusion cyst, broad ligament leiomyomas, fibroma, hematoma and EP | N = 62 | |
| Serous borderline tumour | N = 10 | |
| Mucinous borderline tumour | N = 12 | ||
| Seromucinous borderline tumour | N = 1 | ||
| Endometrioid borderline tumour | N = 1 | ||
| Serous cystadenocarcinoma | N = 37 | |
| Endometrioid carcinoma | N = 6 | ||
| Clear cell carcinoma | N = 2 | ||
| Mucinous cystadenocarcinoma | N = 10 | ||
| Seromucinous Cystadenocarcinoma | N = 2 | ||
| Granulosa cell tumour | N = 7 | ||
| Germ cell tumour | N = 2 | ||
| Metatstatic lesions | N = 28 | ||
OHSS = ovarian hyperstimulation syndrome; EP = ectopic pregnancy.
Adnexal masses subtypes based on MRI finding, histopathology, and follow-up imaging.
|
| Serous cystadenoma | N = 14 |
| Serous adenofibroma | N = 3 | ||
| Mucinous cystadenoma | N = 6 | ||
| Seromucinous cystadenoma | N = 1 | ||
| Endometriotic cyst | N = 56 | ||
| Brenner tumour | N = 2 | ||
| Mature teratoma | N = 33 | ||
| Ovarian fibrothecoma | N = 9 | ||
| Ovarian thecoma | N = 2 | ||
| Ovarian fibroma | N = 2 | ||
| Ovarian leiomyoma | N = 2 | ||
| Luteinized follicular cyst | N = 1 | ||
| Serous inclusion cyst of the ovary | N = 1 | ||
| Tumour-like lesions (functional and corpus luteum cyst, OHSS, and tubo-ovarian abscess) | N = 28 | ||
| Hydrosalpinx, pyosalpinx, hematosalpinx, salpingitis, peritoneal inclusion cyst, broad ligament leiomyomas, fibroma, hematoma and EP | N = 62 | |
| Serous borderline tumour | N = 10 | |
| Mucinous borderline tumour | N = 12 | ||
| Seromucinous borderline tumour | N = 1 | ||
| Endometrioid borderline tumour | N = 1 | ||
| Serous cystadenocarcinoma | N = 37 | |
| Endometrioid carcinoma | N = 6 | ||
| Clear cell carcinoma | N = 2 | ||
| Mucinous cystadenocarcinoma | N = 10 | ||
| Seromucinous Cystadenocarcinoma | N = 2 | ||
| Granulosa cell tumour | N = 7 | ||
| Germ cell tumour | N = 2 | ||
| Metatstatic lesions | N = 28 | ||
|
| Serous cystadenoma | N = 14 |
| Serous adenofibroma | N = 3 | ||
| Mucinous cystadenoma | N = 6 | ||
| Seromucinous cystadenoma | N = 1 | ||
| Endometriotic cyst | N = 56 | ||
| Brenner tumour | N = 2 | ||
| Mature teratoma | N = 33 | ||
| Ovarian fibrothecoma | N = 9 | ||
| Ovarian thecoma | N = 2 | ||
| Ovarian fibroma | N = 2 | ||
| Ovarian leiomyoma | N = 2 | ||
| Luteinized follicular cyst | N = 1 | ||
| Serous inclusion cyst of the ovary | N = 1 | ||
| Tumour-like lesions (functional and corpus luteum cyst, OHSS, and tubo-ovarian abscess) | N = 28 | ||
| Hydrosalpinx, pyosalpinx, hematosalpinx, salpingitis, peritoneal inclusion cyst, broad ligament leiomyomas, fibroma, hematoma and EP | N = 62 | |
| Serous borderline tumour | N = 10 | |
| Mucinous borderline tumour | N = 12 | ||
| Seromucinous borderline tumour | N = 1 | ||
| Endometrioid borderline tumour | N = 1 | ||
| Serous cystadenocarcinoma | N = 37 | |
| Endometrioid carcinoma | N = 6 | ||
| Clear cell carcinoma | N = 2 | ||
| Mucinous cystadenocarcinoma | N = 10 | ||
| Seromucinous Cystadenocarcinoma | N = 2 | ||
| Granulosa cell tumour | N = 7 | ||
| Germ cell tumour | N = 2 | ||
| Metatstatic lesions | N = 28 | ||
OHSS = ovarian hyperstimulation syndrome; EP = ectopic pregnancy.
The ADNEX MR scoring system had an 85% concordance with the gold standard of pathology in discriminating adnexal benign and malignant tumours (kappa coefficient = 0.85). The modified scoring system has an 82% agreement rate with the gold standard of pathology in discriminating between benign and malignant adnexal tumours (kappa coefficient = 0.82). There was a concordance of 88% between the 2 scoring systems (kappa coefficient = 0.88). The area below the ROC curve (AUROC curve) to predict the malignancy of the lesion for the ADNEX MR scoring system index was 0.938 with 0.907-0.961 95% confidence interval (CI), and for the Modified scoring system index, it was 0.909 with 0.870-0.938 95% CI (Figure 3). A pairwise comparison of these area under the curves showed no significant difference (P = 0.053).
ROC curve in the scoring systems.
By considering malignancy for score ≥4, the sensitivity and specificity for the ADNEX MR scoring system were determined at 96.6% and 91%, respectively. The modified scoring system had sensitivity and specificity of 87.3% and 94.6%, respectively (Table 4).
Diagnostic parameters of ADNEX MR scoring system and modified scoring system.
| Parameter | ADNEX MR scoring system | Modified scoring system |
|---|---|---|
| Sensitivity (%) | 96.6 | 87.3 |
| Specificity (%) | 91 | 94.6 |
| PPV (%) | 85.1 | 89.6 |
| NPV (%) | 98.1 | 93.3 |
| Accuracy (%) | 92.9 | 92.1 |
| Parameter | ADNEX MR scoring system | Modified scoring system |
|---|---|---|
| Sensitivity (%) | 96.6 | 87.3 |
| Specificity (%) | 91 | 94.6 |
| PPV (%) | 85.1 | 89.6 |
| NPV (%) | 98.1 | 93.3 |
| Accuracy (%) | 92.9 | 92.1 |
Diagnostic parameters of ADNEX MR scoring system and modified scoring system.
| Parameter | ADNEX MR scoring system | Modified scoring system |
|---|---|---|
| Sensitivity (%) | 96.6 | 87.3 |
| Specificity (%) | 91 | 94.6 |
| PPV (%) | 85.1 | 89.6 |
| NPV (%) | 98.1 | 93.3 |
| Accuracy (%) | 92.9 | 92.1 |
| Parameter | ADNEX MR scoring system | Modified scoring system |
|---|---|---|
| Sensitivity (%) | 96.6 | 87.3 |
| Specificity (%) | 91 | 94.6 |
| PPV (%) | 85.1 | 89.6 |
| NPV (%) | 98.1 | 93.3 |
| Accuracy (%) | 92.9 | 92.1 |
Two pathology-proven benign masses (serous cyst adenoma and serous cyst adenofibroma) had enhancing solid parts with TIC curve type 2 without restriction in DWI upgraded to SCORE = 4 using ADNEX MR scoring system but considered SCORE = 3 using the modified scoring system.
Four pathology-confirmed malignant masses (mucinous cyst adenocarcinoma, well-differentiated serous carcinoma, and granulosa cell tumour) and 5 masses with tissue diagnosis of borderline tumour (Figure 4) had enhancing solid tissue with TIC curve type 2 without restriction in DWI considered SCORE = 4 according to ADNEX MR scoring system but downgraded to SCORE = 3 using the modified scoring system as a false negative.
Images from 66-year-old woman with left-sided cystic adnexal mass indeterminate in transvaginal ultrasound. (A) Axial T1-weighted and (B) axial T2-weighted magnetic resonance (MR) images show left-sided multiloculated adnexal cystic mass, with largest diameter of 45 mm, with intermediate-high signal-intensity solid tissues. (C and D) Axial diffusion-weighted image showing, high signal intensity at b = 1000 s/mm2 and high apparent diffusion coefficient signal intensity of solid enhancing tissues. (E and F) Dynamic contrast-enhanced MRI yielded type 2 time-intensity curve in comparison with adjacent external myometrium. In subjective assessment (G) enhancement of the solid parts is almost equal to external myometrium. This lesion was classified as score = 4 based on ADNEX MR scoring system and as score = 3 based on modified scoring system. Histology confirmed border line serous tumour.
Images from 38-year-old woman with cystic adnexal mass on the left side. (A) Sagittal T2-weighted magnetic resonance (MR) images noted left-sided multiloculated adnexal mass, with largest diameter of 52 mm, with intermediate signal-intensity locules and intermediate-signal-intensity thickened and irregular internal septa. (B and C) Axial diffusion-weighted image showing, intermediate signal intensity at b = 1000 s/mm2 and high apparent diffusion coefficient signal intensity of solid enhancing tissues. (D) In subjective assessment enhancement of the solid tissue is not more than external myometrium. Dynamic contrast-enhanced MRI (E and F) shows intermediate time-intensity curve with type-2 plateau in comparison with adjacent external myometrium. Histology proved benign infected endometrioma. ADNEX MR score was upgraded to 4 due to type-2 curve in modified scoring system the score is 3 as possibly benign.
Eight pathology-proven non-malignant adnexal masses (including tub-ovarian abscess and infected endometrioma, Figure 4) (Figure 5) with enhancing solid tissue with TIC curve type 2 without restriction in DWI were considered SCORE = 4 according to ADNEX MR scoring system as false positive but downgraded to SCORE = 3 using the modified scoring system.
The interobserver agreement on classifying adnexal lesions based on both scoring systems was high (κ = 0.92).
Discussion
As previously stated, DCE MRI is unavailable in all imaging centres due to time restrictions and a lack of specialized software. Thus, we investigated the scoring system by subjective and visual assessment of the enhancement in the early phase and the DWI findings (DWI with ADC map). Although several studies have shown that the examination of enhancement as TIC in DCE MRI has greater accuracy,26,27 the proposed modified scoring system in the current study has an acceptable accuracy of 92.1% with a sensitivity of 87.3% and a specificity of 94.6%. In the present investigation, the sensitivity, specificity, and accuracy of the ADNEX MR scoring system were 96.6%, 91%, and 92.9%, respectively, which is not significantly different. Based on the current study, the proposed modified scoring system has a lower false positive rate, whereas the ADNEX MR scoring system has a lower false negative rate.
In non-DCE MRI, due to lack of TIC, it is impossible to establish differences between MRI scores 3 and 4 lesions.21 In such cases, in the modified scoring system, we suggest using DWI findings (DWI with ADC map) as mentioned in Table 1.
The modified scoring system in our study was less sensitive and more specific than the previous study in 2020. The scoring system based on DWI images in the study by Hottat et al24 had a sensitivity of 95.7% versus 87.3% in our study and a specificity of 93.3% versus 94.6%.
In our investigation, the sensitivity and specificity of the ADNEX MR scoring system were determined to be 96.6% and 91%, respectively. According to Thomassin-Naggara et al,17 conducted a study on 198 adnexal lesions considering score ≥4 as malignancy in 2013, the sensitivity and specificity for the ADNEX MR scoring system were 93.5% and 96.6%, respectively. Ruiz et al28 conducted a similar investigation on 148 adnexal lesions in 2016 and found that the sensitivity and specificity were 91.7% and 92.7%, respectively. In another study conducted by Pereira et al18 in 2018, the ADNEX MR scoring system based on a simplified MRI protocol showed a sensitivity of 94.9% and specificity of 97.5%. Hottat et al24 detrained the sensitivity and specificity of the ADNEX MR scoring system to be 95.5% and 86.6%, respectively. In another similar study published in 2021, ADNEX MR SCORING system showed a sensitivity of 92.9% and specificity of 94.9%.29 Our study reported that the ADNEX MR scoring system was more sensitive than earlier researches. ADNEX MR scoring system had less specificity in the current study compared to above mentioned studies except for the study by Hottat et al in 2020. These findings may be related to variances in the research population and the distribution of malignant lesions.
Out of 24 masses with borderline pathology, 7 cases with solid part with type 3 curve and score = 5 in the ADNEX MR scoring system, were downgraded as score = 3 in the modified scoring system as false negatives according to the low b1000 and high ADC. It is notable that among 24 pathology proven borderline masses, 13 cases had score 5 based on modified scoring system and properly considered possibly malignant. Although the ADNEX MR scoring system appears more helpful in identifying borderline tumours in our investigation, this conclusion cannot be generalized and further studies with larger sample sizes are required.
Among borderline masses which were downgraded using modified scoring system, 3 masses had 1-4 enhancing papillary projections and one mass was multilocular cyst with largest diameter of 10 cm< with enhancing inner wall. Above mentioned masses were categorized as O-RADS US 4 ≤ based on O-RADS US (Ovarian-Adnexal reporting and Data System Ultrasound). So, as a hypothesis, in the absence of DCE MRI for better evaluation of borderline tumours, use of combination of some O-RADS US scoring system items and DWI may be helpful. More investigations with larger number of cases for confirmation of this hypothesis are required.
For the masses with score = 3 including downgraded borderline tumours in basic exam, follow-up imaging is requested as further evaluation. These masses may possibly upgrade to score 4 ≤ with subsequent proper management. Also, expert opinion of gynaecologic oncologists and clinical characteristics of the patient should be considered to have the best approach.
Limitations of the study
One year of follow-up information in adnexal masses without surgery may be short for some slow-growing borderline masses. Also, the study population could be larger for more reliable generalizations. As another limitation, visual assessment of solid part enhancement compared to external myometrium and interpretation of signal intensities in DWI images are subjective approach like colour score evaluation in O-RADS US.
Conclusion
In our proposed modified scoring system, a subjective assessment was performed on the enhancement of solid tissue in early phases (30 s after contrast material injection). Also, to differentiate between MRI score 3 and 4 lesions, DWI with an ADC map was utilized. The modified scoring system is less sensitive and more specific than the ADNEX MR scoring system, but the accuracy is not significantly different. Although further studies are required, the proposed modified scoring system could be suggested in imaging centres where DCE MRI is unavailable.
Funding
Nothing to declare.
Conflicts of interest
Nothing to declare.
Ethical approval
This study was approved by the local ethics committee (IR.TUMS.IKHC.REC.1399.535) and the need for written informed consent was waived by the ethical approval committee.
