Fat-containing adnexal masses on MRI: solid tissue volume and fat distribution as a guide for O-RADS Score assignment

Abstract

Purpose:

To explore ways to improve O-RADS MRI scoring for fat-containing adnexal masses, by investigating methods for quantifying solid tissue volume and fat distribution and evaluating their associations with malignancy.

Methods:

This retrospective, single-center study included patients with fat-containing adnexal masses on MRI during 2008–2021. Two radiologists independently reviewed overall size (Size_overall), size of any solid tissue (Size_anysolid), size of solid tissue that was not Rokitansky nodule (Size_{non-Rokitansky}), and fat distribution. Wilcoxon test, Fisher-exact test, and ROC curve analysis were performed. Reference standard was pathology or follow-up >24 months.

Results:

188 women (median age 35 years) with 163 benign and 25 malignant lesions were included. Size_overall (R1, 9.9 cm vs 5.9 cm; R2, 12.4 cm vs 6.0 cm), Size_anysolid (R1, 5.1 cm vs 1.2 cm; R2, 3.2 cm vs 0.0 cm), Size_{non-Rokitansky} (R1, 5.1 cm vs 0.0 cm; R2, 3.1 cm vs 0.0 cm), and fat distribution differed significantly between malignant and benign lesions (p<0.01). Area under ROC curve was greatest using Size_{non-Rokitansky} (R1, 0.83; R2, 0.86) vs Size_overall (R1, 0.78; R2, 0.81) or Size_anysolid (R1, 0.79; R2, 0.81), though differences were non-significant (p=0.480.93). Cutoffs for Size_{non-Rokitansky} (R1, ≥1.2 cm; R2, ≥1.0 cm) yielded sensitivity and specificity of 0.72 and 0.93 (R1) and 0.76 and 0.95 (R2). Among immature teratomas, 90% displayed scattered fat.

Conclusion:

Overall size, size of (any or non-Rokitansky-nodule) solid tissue, and fat distribution differed between benign and malignant fat-containing adnexal masses. Incorporating these would constitute simple and practical approaches to refining O-RADS MRI scoring.

Introduction

Adnexal masses are often detected in women, either incidentally or when investigating vague pelvic symptoms. Although it is standard of care to use ultrasound (US) as the first-line imaging modality to characterize such lesions, approximately 10–30% remain indeterminate after the initial US evaluation (1). With recent technological advances, multiparametric magnetic resonance imaging (MRI) incorporating conventional anatomical sequences (e.g., T1- and T2-weighted imaging) and functional sequences (e.g., diffusion-weighted imaging and dynamic contrast-enhanced MRI) has shown promise for further characterizing sonographically indeterminate adnexal masses, providing further clues for pin-pointing the origin (ovarian vs. extra-ovarian), differentiating malignant and benign lesions, and predicting histological subtypes (2–4). Building upon these findings, a five-point risk stratification system for assessing the likelihood of ovarian malignancy on MRI in sonographically indeterminate lesions – the Ovarian-Adnexal Reporting and Data System (O-RADS) MRI score – was recently introduced by a group of international experts (5).

The O-RADS MRI score has been prospectively validated in a multi-national multi-center cohort of 1340 women; it demonstrated excellent diagnostic performance for determining ovarian cancer, with a sensitivity of 0.93 (95% confidence interval [CI], 0.89–0.96) and a specificity of 0.91 (95% CI, 0.89–0.93) (6). Despite these initial promising results, there are some areas within the O-RADS MRI scoring that require clarification and refinement to further improve its diagnostic performance and interobserver agreement. One of these areas is the assessment of fat-containing adnexal masses. The current O-RADS MRI system assigns a score of 2 (“almost certainly benign”) to a fat-containing lesion with no solid enhancing solid tissue, while a score of 4 (“intermediate risk”) is given when there is a large volume of enhancing solid tissue. At first glance, this is intuitive, considering that malignant fat-containing adnexal lesions (e.g., mature cystic teratomas [MCT] with malignant transformation, immature teratomas, or other types of malignant germ cell tumors) commonly have large and irregular solid tissue in contrast with the more common and benign mature cystic teratomas (MCT) (7). Nevertheless, the current O-RADS MRI scoring system raises a few unanswered questions: (1) How much is a “large volume” of solid tissue? (2) Should cases with a “small volume” of solid tissue (i.e., Rokitansky nodule [RN]) be assigned a score of 2? (3) Should we only consider the volume of solid tissue or can we incorporate other features such as the distribution of fat? Although many different patterns of fat (e.g., fat-fluid level, fat balls, or scattered fat) have been described for various types of fat-containing adnexal masses, it has not been established whether these are associated with the probably of malignancy, and evaluating such features is not part of the current O-RADS MRI scoring system.

The purpose of this study was to assess possible methods for clarifying and refining the O-RADS MRI scoring system for assessing fat-containing adnexal masses, by investigating what is the optimal method (and threshold) for quantifying volume of solid tissue and whether evaluation of the distribution of fat can additionally be used to determine malignancy.

Methods

Patients

This retrospective study was performed after obtaining approval from our institutional review board and was compliant with the Health Insurance Portability and Accountability Act. Our institutional radiology and pathology databases were searched to identify patients with fat-containing adnexal masses on MRI between April 2008 and October 2021 based on MRI reports that contained the key words fat, lipid, adnexa and/or ovary. Exclusion criteria were (1) no fat-containing adnexal masses, (2) inadequate MRIs for imaging assessment (e.g., due to artifacts, lack of intravenous contrast), (3) prior treatment for adnexal lesion, and (4) lack of adequate reference standard (i.e., histopathology or >24 months of imaging follow-up (6)). Figure 1 illustrates the patient selection process.

Fig. 1.

Flowchart of patient selection process

MRI Analysis

MRI examinations were performed on 1.5- and 3.0-Tesla scanners. Due to the long inclusion period, MRI protocols varied, but MRI scans were generally acquired according to the parameters provided in Table 1. MRIs were independently reviewed by two radiologists specialized in gynecologic oncologic imaging (-- and – [initials blinded for review], with 7 and 5 years of post-residency experience, respectively); the radiologists were aware that the patients had fat-containing adnexal masses on MRI, but were otherwise blinded to clinical and pathological data. MRIs were specifically reviewed for (1) overall lesion size (Size_overall), (2) size of any solid tissue (Size_anysolid), (3) whether this solid tissue was considered a typical RN or not, and (4) distribution of fat. Solid tissue was considered a RN if it appeared as a rounded or relatively smooth-shaped structure protruding into the lesion lumen and forming acute angles with the lesion wall (8). Solid tissue was considered not a RN (i.e., non-RN soft tissue) and thus suspicious if it demonstrated one or more of the following features: obtuse angle with lesion wall, irregular shape, papillary appearance, transmural growth (9). The radiologists were instructed to assess all available sequences and planes to comprehensively decide whether the solid tissue was RN or not. The size of non-RN soft tissue was labeled “Size_{non-Rokitansky}”. Fat distribution was categorized as pure fat, fat-fluid level, focal fat, scattered fat, or fat balls as shown in Figure 2 (10).

Table 1.

Pelvic MRI parameters

Sequence	Axial T1WI	Sagittal T2WI	Axial T2WI	Coronal T2WI	Axial DWI	Axial FS Pre-contrast T1WI	Axial FS DCE-MRI*	Axial FS post-contrast T1†
TE (ms)	Min	102	102	102	Min	Min	Min	Min
TR (ms)	400–650	3500	3500	3500	6000	Default		Default
Slice thickness (mm)	5	4	4	4	4	4	3.4–4.0	4
Gap (mm)	1	0	0	0	0	-2	0	-2
Field-of-view (cm)	32	22–26	22–26	22–26	26–28	32	28–30	32
Matrix	448 x 224	288 x 224 or 256	288 x 224 or 256	288 x 224 or 256	128x128	320 x 192	260x212	320 x 192
No. of excitations	1	3–4	3–4	3–4	4 (b = 0) 14 (b = 800)	1	1	1

B = b-value; DCE = dynamic contrast-enhanced; FS = fat-saturated; Min = minimum; MRI = magnetic resonance imaging; TE = echo time; TR = repetition time; T1WI = T1-weighted imaging; T2WI = T2-weighted imaging.

^*15 phases during total 3 minutes.

^†After DCE-MRI

Fig. 2.

Categorization of fat pattern in adnexal masses. Axial T1-weighted imaging showing fat (arrows) in the adnexal masses in the patterns of (a) pure fat, (b) fat-fluid level (with incidental Rokitansky nodule [*]), (c) focal fat, (d) scattered fat, and (e) fat balls.

Clinicopathological Data

The following information was obtained from the electronic medical records: age, gender, race, details of pathologic specimens, and follow-up. Histopathologic analysis or imaging follow-up of greater than 6 months served as the reference standard for determining whether the adnexal mass was benign or malignant (11).

Statistical Analysis

Inter-reader agreement on MRI assessment was evaluated using intra-class correlation coefficients (ICC) and Cohen k statistics for continuous and categorical variables. The level of agreement was defined as follows for ICC and k values: 0.00–0.39, poor; 0.40–0.59, fair; 0.60–0.74, good; and 0.75–1.00, excellent (12, 13). The following MRI findings of Size_overall, Size_anysolid, and Size_{non-Rokitansky} were compared between malignant and benign adnexal masses by using the Wilcoxon rank sum test for categorical variables and Fisher’s exact test for continuous variables. Receiver operating characteristic (ROC) curve analysis was done to determine the diagnostic performance of each of these findings. The Youden index was used to determine optimal cutoffs and corresponding sensitivity and specificity with the 95% confidence intervals calculated (14). The statistical software R was used for the analysis, and p-values <0.05 were considered significant.

Results

Patients and Clinicopathological Characteristics

Table 1 describes the patient and clinicopathological characteristics. There were 188 women with a median age of 35 years (interquartile range [IQR] 26–45), most of whom were white (n = 129 [68.6%]); 173 (92.0%) patients underwent surgical procedures and the diagnosis was based on biopsy in 2 (1.1%) and imaging follow-up in 13 (6.9%) patients. Median interval between MRI and the reference standard was 31 days (IQR 13.5–69.5) for pathological analysis and 55 months (IQR 40–59) for imaging follow-up. Diagnoses were most commonly mature cystic teratomas (n = 155 [82.4%]) followed by immature teratomas with or without yolk sac differentiation (n = 10 [5.3%] and 4 [2.1%], respectively).

Inter-reader Agreement of MRI Findings

The inter-reader agreement was excellent for all size measurements: Size_overall (ICC = 0.965 [95% CI 0.954–0.974]); Size_anysolid (ICC = 0.811 [95% CI 0.715–0.870]); Size_{non-Rokitansky} (ICC = 0.877 [95% CI 0.836–0.908]). Agreement was good for the assessment of Rokitansky nodule (k = 0.67 [95% CI 0.54–0.80]). The agreement for fat distribution considering all patterns was fair (k = 0.432 [95% CI 0.347–0.517]); however, when dichotomizing between scattered fat and other patterns, it was excellent (k = 0.895 [95% CI 0.752–1.00]).

MRI Findings Between Malignant and Benign Entities

MRI features of malignant and benign ovarian masses are summarized in Table 3, and representative images are shown in Figures 3–5. For both radiologists, malignant masses had greater Size_overall, Size_anysolid, and Size_{non-Rokitansky} than did benign masses. Median measurements for malignant vs. benign masses were as follows: Siz_eoverall, 9.9 cm (IQR 8.4–14.8) vs 5.9 cm (IQR 4.2–8.3), respectively, for R1 and 12.4 cm (IQR 8.5–15.2) vs 6.0 cm (IQR 4.2–8.3), respectively for R2 (p <0.001 for both); Size_anysolid, 5.1 cm (IQR 2.0–10.0) vs 1.2 cm (IQR 0.0–2.2) for R1, respectively, and 3.2 cm (IQR 1.5–6.7) vs 0.0 cm (IQR 0.0–1.5), respectively, for R2 (p <0.001for both); and Size_{non-Rokitansky}, 5.1 cm (IQR 0.0–10.0) vs 0.0 cm (IQR 0.0–0.0), respectively, for R1 and 3.1 cm (IQR 1.0–6.7) vs 0.0 cm (IQR 0.0–0.0), respectively, for R2 (p <0.001 for both). The distribution of fat was significantly different between malignant and benign ovarian masses for both radiologists (p <0.01 for both). Benign masses more commonly showed fat-fluid levels than did malignant masses (46.0% [75/163] vs 20.0% [5/25] for R1; 21.5% [35/163] vs 16.0% [4/25] for R2) or pure fat pattern (41.7% [68/163] vs 20.0% [5/25] for R1; 59.5% [97/163] vs 12.0% [3/25] for R2) whereas a pattern of scattered fat was more common in malignant masses than in benign masses (52.0% [13/25] vs 2.5% [4/163] for R1; 52.0% [13/25] vs 0.6% [1/163] for R2). Notably, scattered fat was found in most (9/10) immature teratomas by both radiologists.

Table 3.

MRI findings in malignant and benign ovarian masses.

MRI Findings		RI			R2
		Malignant	Benign	P-value	Malignant	Benign	P-value
Overall size (cm)		9.9(8.4–14.8)	5.9 (4.2–8.3)	<0.01	12.4 (8.5–15.2)	6.0 (4.2–8.3)	<0.01
Size of any solid tissue (cm)		5.1 (2.0–10.0)	1.2 (0.0–2.2)	<0.01	3.2 (1.5–6.7)	0.0 (0.0–1.5)	<0.01
Size of non-RN solid tissue (cm)		5.1 (0.0–10.0)	0.0 (0.0–0.0)	<0.01	3.1 (1.0–6.7)	0.0 (0.0–0.0)	<0.01
Fat distribution*	Fat-fluid level	5 (20.0)	68 (41.7)	<0.01	4(16.0)	35 (16.0)	<0.01
	Focal	2 (8.0)	14(8.6)		5 (20.0)	27 (16.6)
	Scattered	13 (52.0)	4(2.5)		13 (52.0)	1 (0.6)
	Fat balls	0 (0.0)	2(1.2)		0 (0.0)	3(1.8)
	Pure fat	5 (20.0)	75 (46.0)		3 (12.0)	97 (59.5)

^*Presented as number of patients and percentage; other findings are presented as median and interquartile range

RI = Radiologist 1; R2 = Radiologist 2; RN = Rokitansky nodule

Fig. 3.

Forty-two-year-old woman with fat-containing benign adnexal mass. (a) Axial T1-weighted image shows right adnexal mass, measuring 2.5 cm and 2.4 cm according to radiologists 1 and 2, respectively, and demonstrating pure fat (*) except for a central area (arrow). (b) Sagittal T2-weighted image shows that this central area (arrow) is well defined and forms acute angles (broken white lines) with wall of the mass. (c) Contrast-enhanced fat-suppressed axial T1-weighted image shows central area (arrow) that enhances and was considered solid tissue, measuring 1.3 cm and 1.1 cm according to radiologists 1 and 2, respectively. This was considered a Rokitansky nodule given features in (B). Follow-up MRI over approximately 2 years demonstrated no change, and the mass was considered benign and probably a mature cystic teratoma.

Fig. 5.

Twenty-seven-year-old woman with fat-containing immature teratoma. (a) Axial T1-weighted image shows left adnexal mass, measuring 12.9 cm and 14.0 cm according to radiologists 1 and 2, respectively, demonstrating pattern of scattered fat (white arrows). (b) Axial T2-weighted image shows heterogeneous central soft tissue with mixed components of soft tissue, fluid, and fat (broken line). (c) Contrast-enhanced fat-suppressed axial T1-weighted image shows that mass (broken line) demonstrates strong and heterogeneous enhancement and measures 10.4 cm and 8.0 cm according to radiologists 1 and 2, respectively, and was not considered Rokitansky nodule. Ovarian cystectomy revealed immature teratoma.

Diagnostic Performance of MRI Findings

Figure 6 shows the ROC curves using different MRI findings for differentiating malignant and benign ovarian masses according to both readers. Although there were no statistically significant differences (p = 0.481–0.928), for both readers, the area under the ROC curves were greater when using Size_{non-Rokitansky} (0.83 [95% CI, 0.74–0.93] for R1; 0.86 [95% CI, 0.77–0.95] for R2) compared to Size_overall (0.78 [95% CI, 0.67–0.90] for R1; 0.81 [95% CI, 0.71–0.91] for R2) and Size_anysolid (0.79 [95% CI, 0.67–0.91] for R1; 0.81 [95% CI, 0.70–0.92] for R2). Optimal cutoffs using Size_{non-Rokitansky} were ≥1.2 cm (R1) and ≥1.0 cm (R2), yielding sensitivity and specificity of 0.72 (95% CI, 0.51–0.88) and 0.93 (95% CI, 0.88–0.97), respectively for R1, and 0.76 (95% CI, 0.55–0.91) and 0.95 (95% CI, 0.90–0.97), respectively, for R2. For Size_anysolid, the optimal cutoffs were ≥3.9 cm (R1) and ≥2.1 cm (R2), with sensitivity and specificity of 0.60 (95% CI, 0.39–0.79) and 0.90 (95% CI, 0.84–0.94), respectively, for R1], and 0.72 (95% CI, 0.51–0.88) and 0.85 (95% CI, 0.78–0.90), respectively, for R2. For Size_overall, the optimal cutoffs were ≥8.4 cm (R1) and ≥8.5 cm (R2), yielding sensitivity and specificity of 0.76 (95% CI, 0.55–0.91) and 0.77 (95% CI, 0.69–0.83), respectively, for R1 and 0.76 (95% CI, 0.55–0.91) and 0.77 (95% CI, 0.69–0.83), respectively, for R2.

Fig. 6.

Receiver operating characteristic (ROC) curves differentiating malignant and benign ovarian masses using 3 different MRI assessments by radiologist 1 (a) and radiologist 2 (b). Black solid line = Overall size of mass; Blue line = size of any solid tissue; Red line = size of solid tissue not representing a Rokitansky nodule. Black broken line = line of equality.

Discussion

In this study, we explored strategies for reducing the ambiguity of the current O-RADS MRI scoring system for fat-containing adnexal masses, by evaluating associations of quantitative lesion/solid tissue volume measurements as well as fat distribution patterns with malignant histology. We found that the general premise of the current O-RADS MRI scoring system is valid, in that a large volume of soft tissue should raise concerns for malignancy when encountering fat-containing ovarian masses on MRI (5). The results of our study may provide guidance on how to optimally measure soft tissue; rather than measuring any solid tissue of adnexal masses, excluding typical RN from measurements could be a better solution. Additionally, incorporating evaluation of the distribution of fat could provide incremental value for determining malignancy, without overcomplicating the scoring system. If the findings of our study are validated, they will provide the basis for further updates and iterations of the O-RADS MRI scoring system, which could potentially decrease ambiguities, improve diagnostic performance, and facilitate more widespread use of this tool.

One of the main goals of the O-RADS MRI scoring system is to enhance reproducibility. The document provides standardized terms (i.e., ‘lexicons’) with specific definitions for assessing adnexal lesions, in order to ensure consistency among radiologists and to improve communication between different stakeholders (e.g., radiologists and gynecologists) (5, 15). One of the most common sources of O-RADS-MRI score misclassification (and, in turn, inter-reader variability) was found to stem from interpretive errors caused by the difficulty of recognizing the enhancing solid component(i.e., ‘soft tissue) (16). This is again reflected in our study, where the agreement between the two radiologists was excellent for assessing measurements of Size_anysolid and Size_{non-Rokitansky}. Similarly, when considering reproducibility for assessing fat distribution within the adnexal mass, it may be advantageous to use dichotomous interpretation—i.e., scattered fat vs. not scattered fat. This approach not only proved to be more relevant for correlation with malignancy, but also yielded a higher level of inter-reader agreement when compared to assigning all categories of fat distribution (k = 0.895 vs 0.432).

The ambiguities related to assessment of solid tissue in fat-containing adnexal masses in the current O-RADS MRI scoring system are perhaps related to the variable results in the literature. Rokitansky nodules are characteristically associated with mature cystic teratomas; however, this solid tissue may be comprised of different types of tissues including fibrous stroma, vessels, and glial or thyroid tissue (17). It has been noted that even when mature cystic teratomas do not harbor malignant degeneration, the amounts of these solid tissues can be quite large, with reported ranges of up to 4.2 cm (17). Similarly, in our study, the upper quartile measurements of any solid tissue in benign adnexal masses (which predominantly consisted of mature cystic teratomas) were 2.1 cm (R1) and 1.5 cm (R2). Nevertheless, it has consistently been shown that solid tissue components associated with malignancy are generally larger and show suspicious features (e.g., obtuse angle with the lesion wall or transmural involvement) (18, 19). In line with this, our results showed that the areas under the ROC curves, sensitivities and specificities for determining malignancy were higher (though not to a statistically significant degree) when size measurements were based on ‘suspicious’ non-RN solid tissue as compared with any solid tissue for both readers, resulting in moderate sensitivities (0.72 and 0.76) and high specificities (0.93 and 0.95) using cutoff values of ≥1.2 cm (R1) and ≥1.0 cm (R2(, respectively. Based on these results, radiologists may be encouraged to incorporate the gestalt impression regarding the presence or absence of a Rokitansky nodule when measuring solid components, but it would be completely justifiable to measure any solid tissue when in doubt for determining the malignant potential of a fat-containing adnexal mass.

Our study suggests that assessing the distribution of fat may serve as a simple and effective approach to obtaining additional information for determining whether a fat-containing adnexal mass is benign or malignant. While fat-fluid levels or pure fat were more commonly seen in benign entities, a scattered distribution of fat was predominantly noted in the immature teratomas. Scrutinizing the distribution of fat may be helpful especially when suspecting immature teratomas, as we noted that there was difficulty ascertaining the exact volume of soft tissue in such cases, because the solid tissue in immature teratomas is often intermixed with scattered soft tissue, calcifications, fat, and fluid. All cross-sectional imaging modalities, including MRI and even CT, are highly sensitive for detecting and assessing the fat pattern, and these features of fat distribution are in agreement with previous literature describing various types of fat-containing adnexal masses (20). Prior studies have noted that imaging findings suggestive of immature teratomas tend to show scattered small foci of fat amongst irregular solid tissue interspersed with coarse calcifications (21). As such, integrating the assessment of fat distribution as part of the O-RADS MRI scoring system in fat-containing adnexal masses might be a practical measure that could improve the current scoring scheme without overcomplicating it.

Our study had several limitations. First, it was a retrospective, single-institution study, with the inherent limitations of such a study design. In particular, the number of malignant adnexal masses was small, and therefore, the thresholds derived from the current cohort may not be universally applicable in cohorts with a larger predominance of malignancies. Nevertheless, the cutoff values suggested here (i.e., >1.0 or >1.2 cm for Size_{non-Rokitansky}) were simple to obtain, are reproducible and could be used as an example or starting point for further investigations. Second, we were unable to assess the pattern of enhancement of the solid tissue, as we included a cohort that spanned a long period of time with varying MRI protocols. In the current O-RADS MRI scoring system, assessment of the pattern of enhancement is not part of the assessment for fat-containing ovarian masses. However, recent studies have debated whether using dynamic contrast-enhanced MRI or qualitatively assessing the enhancement curve types could be beneficial for determining malignancy in indeterminate ovarian masses in general (22, 23). In addition, quantitative analysis of other MRI parameters such as diffusion-weighted imaging could not be done as part of our analysis for the same reason (i.e., long period of inclusion). Recent studies have shown potential incremental value of apparent diffusion coefficient values measured in cystic and solid components of the adnexal masses (24, 25). Further studies will be needed to ascertain if this works specifically in fat-containing adnexal masses.

In conclusion, overall size, size of (any or non-Rokitansky-nodule) solid tissue, and fat distribution differed between benign and malignant fat-containing adnexal masses. The incorporation of these imaging features in fat-containing ovarian masses would constitute a simple and practical approach to refining the O-RADS MRI scoring system and guiding the determination of ovarian malignancy.

Fig. 4.

Seventy-five-year-old woman with fat-containing malignant adnexal mass. (a) Axial T1-weighted image shows right adnexal mass measuring 14.8 cm and 15.2 cm according to radiologists 1 and 2, respectively, demonstrating fat-fluid level (white arrow) and another smaller focal area of fat along the wall (broken white arrow). (b) Axial T2-weighted image shows intermediate signal mass (broken double lines) forming obtuse angle (broken black lines) with wall of the mass. Irregular contour (black arrow heads) raises suspicion for transmural extension beyond the wall. (c) Contrast-enhanced fat-suppressed axial T1-weighted image shows that mass (broken double lines) demonstrates strong and heterogeneous enhancement and measures 11.3 and 10.9 cm according to radiologists 1 and 2, respectively. Hysterectomy with bilateral salpingo-oophorectomy was done, revealing squamous cell carcinoma arising from teratoma.

Table 2.

Clinicopathological characteristics of patients with fat-containing adnexal masses

Characteristics	Category		No. of patients (%)*
Age (years)			35 (26, 45)*
Gender	Female		188 (100.00)
Ethnicity	White		129 (68.6)
	Black		81 (9.6)
	Asian		22 (11.7)
	Other		19 (10.1)
Surgery type	Cystectomy		89 (47.3)
	Oophorectomy		3 (1.6)
	Hysterectomy and bilateral salpingo-oophorectomy		33 (17.6)
	Salpingo-oophorectomy		48 (25.5)
	None	Biopsy	2 (1.1)
		Imaging follow-up	13 (6.9)
Histological subtype	Mature cystic teratoma		155 (82.4)
	Immature teratoma	With yolk sac differentiation	10 (5.3)
		Without yolk sac differentiation	4 (2.1)
	Mature teratomas with malignant Transformation		5 (2.7)
	Other malignancies†		6 (3.2)
	Other benign entities††		8 (4.3)

^*Presented as median and interquartile range; other characteristics are presented as number of patients and percentage

^†Carcinoid tumerlet (n = 1), metastatic serous carcinoma (n = 1), intestinal mucinous neoplasm of uncertain malignant potential (n = 1), bilateral mature teratomas and collision with low-grade appendiceal-type mucinous tumor (n = 1), collision tumor between serous borderline tumor and mature cystic teratoma (n = 1), poorly differentiated carcinoma with neuroendocrine differentiation (n = 1)

^††Fat necrosis (n = 1); lipoleiomyoma (n = 5); struma ovarii (n = 2)