The physiological sonographic features of the ovary in healthy subjects: a joint systematic review and meta-analysis by the Italian Society of Gynecology and Obstetrics (SIGO) and the Italian Society of Endocrinology (SIE)

Purpose There is a lack of uniformity in the definition of normal ovary ultrasound parameters. Our aim was to summarize and meta-analyze the evidence on the topic. Full-text English articles published through December 31, 2020 were retrieved via MEDLINE and Embase. Data available for meta-analysis included: ovarian follicular count, ovarian volume, and ovarian Pulsatility Index (PI) assessed by Doppler ultrasound. Methods Cohort, cross-sectional, prospective studies with a single or double arm were considered eligible. Interventional studies were included when providing baseline data. Both studies on pre- and post-menopausal women were screened; however, data on menopausal women were not sufficient to perform a meta-analysis. Studies on pre-pubertal girls were considered separately. Eighty-one papers were included in the meta-analysis. Results The mean ovarian volume was 6.11 [5.81–6.42] ml in healthy women in reproductive age (5.81–6.42) and 1.67 ml [1.02–2.32] in pre-pubertal girls. In reproductive age, the mean follicular count was 8.04 [7.26–8.82] when calculated in the whole ovary and 5.88 [5.20–6.56] in an ovarian section, and the mean ovarian PI was 1.86 [1.35–2.37]. Age and the frequency of the transducers partly modulated these values. In particular, the 25–30-year group showed the higher mean follicular count (9.27 [7.71–10.82]), followed by a progressive age-related reduction (5.67 [2.23–9.12] in fertile women > 35 years). A significant difference in follicular count was also found according to the transducer’s upper MHz limit. Conclusion Our findings provide a significant input to improve the interpretation and diagnostic accuracy of ovarian ultrasound parameters in different physiological and pathological settings.


Introduction
Ultrasound examination is the standard imaging method to analyze ovarian morphology, while providing also some important functional information or to identify patients with polycystic ovary morphology (PCOM) [1]. The existing Rotterdam guidelines define the ultrasound characteristics of PCOM by the presence of ovarian volume > 10 ml or the presence of 12 or more follicles measuring 2-9 mm in diameter in each ovary [1]. Other typical ultrasound features of PCOM have been widely studied including central stromal echogenicity [2], increased ovarian blood flow [reduced F. Moro and I. Scavello have contributed equally.
Members of the Women's Endocrinology Group of the Italian Society of Endocrinology are listed in the Acknowledgements section.

3
pulsatility index (PI) and reduced resistance index] [3], stromal index and stromal to ovarian area ratio [4,5]. In addition, the Androgen Excess Society guidelines criteria have increased the threshold count of small ovarian follicles to 25 [6]. However, the accurate determination of numerous follicles can be obtained only with the new-generation US machines, not available in many centers.
The structure of the ovary is basically made up of an outer cortical and an inner medullary region. The cortex consists primarily of follicles in different stages of maturation, the medulla of stromal cells, lymphatics, blood vessels, and nerves. The sonographic features of the ovaries are highly variable, depending on the cyclic influence of the hypothalamic-pituitary hormonal axis, which determines ovarian hormone production, follicular maturation, and degeneration [7].
Ultrasound examination is also an excellent diagnostic tool to discriminate between benign and malignant ovarian masses in the hands of experienced examiners using subjective assessment [8]. A consensus opinion on terms, definitions, and measurements to describe the sonographic features of adnexal tumors was established by the International Ovarian Tumor Analysis (IOTA) Group [9]. The IOTA group created ultrasound-based models with similar accuracy to that of expert ultrasound examiners to characterize the ultrasound appearance of benign ovarian tumors (i.e., endometriomas, dermoid cysts, cystadenofibromas) [10][11][12], and to differentiate from the borderline [13,14], and the malignant ones [15][16][17].
In contrast, the sonographic appearance of the ovary in physiological conditions has been poorly investigated. Only some prospective studies described ovarian characteristics in healthy pre-and post-menopausal women in terms of volume and vascularization patterns [18][19][20]. However, there is a lack of uniformity in the definition of normal ovary ultrasound parameters and no consensus statement has been established.
Two Italian societies-the Italian Society of Gynecology and Obstetrics (SIGO) and the Italian Society of Endocrinology (SIE)-agreed on the urgent need to produce a consensus statement to define normal ovary ultrasound parameters. To reach this goal, a joint commission of the 2 societies promoted a systematic review and meta-analysis of the existing evidence on ultrasound parameters of the normal ovary. Therefore, the aim of this review is to define the sonographic parameters of the normal ovary, including ovarian follicular count, ovarian volume, and vascular indices. This represents a fundamental and critical step for orientating clinicians not only in interpreting sonographic data, but also to substantiate future research in the field of pathological conditions such as Polycystic Ovary Syndrome (PCOS).
Even though majority of retrieved records enrolled pre-menopausal women, we did not exclude the few studies involving post-menopausal ones, and presented data accordingly.

Research question
This study aimed to answer this question: which are the normal ovarian follicular count, ovarian volume, ovarian stroma, and vascular indices assessed by ultrasound and Doppler ultrasound?

Study outcomes
The study outcomes were the mean values of the following parameters: mean ovarian follicular count (whole ovary; number), mean ovarian follicular count (ovarian section; number), mean ovarian volume (ml), and mean ovarian artery PI (number). We were not able to provide a metaanalysis of ovarian stroma volume or other vascular indices [i.e., PSV (peak systolic velocity) and RI (Resistance Index)] due to the lack of an adequate number of eligible studies reporting these measures.

Type of study design included
Studies that are considered eligible were cohort studies, cross-sectional studies, and prospective studies with a single arm or including two groups (i.e., healthy controls). We included interventional studies when they provided baseline data (obtained before any treatment, i.e., ovarian stimulation for Assisted Reproductive Technology, ART).

Study population
Both studies on pre-menopausal and post-menopausal women were considered eligible for the qualitative analysis; however, data on menopausal women were not sufficient to perform a meta-analysis. Studies on pre-pubertal girls were considered separately. Records were selected when presenting data obtained from healthy volunteers (including control arms) or from the general population (i.e., screening studies). Regarding data on women from infertile couples, they were included only when a male factor or a tubal factor was specified as the only identified infertility factor.

Search strategy
We performed a systematic review of the literature using methodological approaches previously published [21], and following a protocol written prior to starting the review (PROSPERO registration protocol: CRD42022300584). An extensive search was performed in the following databases: MEDLINE and Embase. Only articles in English and with full-text were included. The search was performed using the words "ovarian", "Doppler" and "ultrasound" [All Fields], accruing all records on human beings published between January 2000 and December 31, 2020.

Study selection
All the team members independently screened records for inclusion, blinded to each other's' decisions. Two of the team members (IS and EM) checked decisions and resolved eventual conflicts. Selections were recorded in a dedicated Excel spreadsheet.

Data extraction
The following data were recorded: number of subjects, ovarian volume (ml; mean and standard deviation-SD), ovarian follicular fount (in the whole ovary or in the ovarian section, mean and SD), ovarian artery Pulsatility Index (PI) (mean and SD), reproductive stage of enrolled subjects, (pre-pubertal/reproductive age/menopausal), age, 2D (2 dimensional) or 3D ultrasound technique, use of transabdominal (TA) or transvaginal (TV) approach, menstrual phase in reproductive age women, description of the size of the follicles included in the follicular count, upper limit of the frequency of the transducer used in MHz, and localization of the artery sampled for Doppler analysis. All the team members independently extracted data and two of the team members (IS and EM) checked them. Disagreements between individual judgements was resolved by collective discussion during dedicated online meetings. Study investigators were contacted for unreported data or additional details. Data were recorded in a dedicated Excel spreadsheet.

Quality assessment
The quality of studies was assessed according to the Oxford Centre for Evidence-Based Medicine: Levels of Evidence [22]. As shown in Table 1, the vast majority of studies were judged as Level 3 (non-randomized or cohort studies).

Statistical analysis
Four separate meta-analyses were conducted, one for every outcome (ovarian volume, ovarian follicular count-whole ovary, ovarian follicular count-ovarian section, ovarian artery PI). Heterogeneity was assessed using I 2 statistics and a random-effects model was applied for all analyses. The effect measures were expressed as mean [lower limit; upper limit]. When the mean was not provided, the closest approximation of mean and SD (standard deviation) from median and IQR (interquartile range) was calculated [23].

Analysis of available studies
The initial literature search produced 1032 potentially relevant citations. After screening and detailed assessment (see flow chart- Fig. 1 for reasons for exclusion), 50 studies were included in the meta-analysis of follicular count in the whole ovary, 11 in the meta-analysis of follicular count per ovarian section, 73 in the meta-analysis of ovarian volume, and 16 in the meta-analysis of ovarian PI. Some papers provided data for more than one of the 4 meta-analyses. The flow chart of study selection is reported in Fig. 1 and the details of the retrieved studies are reported in Table 1.
To determine the age-related differences in ovarian parameters, studies enrolling pre-pubertal girls or fertile women were considered separately. Data on menopausal women were not sufficient to perform a meta-analysis. Moreover, in the analysis of the studies including reproductive-aged women, when information on age was available, the results were stratified into 5 age bands: < 20 years old (1st group); ≥ 20 and < 25 years old (2nd group); ≥ 25 and < 30 years old (3rd group); ≥ 30 and < 35 years old (4th group); and ≥ 35 years old (5th group).

Ovarian follicular count: whole ovary
Studies reporting information on follicular count calculated on the whole ovary were included in a separate meta-analysis from those reporting the same data obtained within an ovarian section. In general, we included studies reporting information on all visible follicles measuring ≥ 2.0 mm, with an upper limit of 8.0-10.0 mm (see Table 1).
Fifty studies were identified including information on follicular count calculated on the whole ovary. After excluding the 2 studies enrolling pre-pubertal girls [35a and b], the overall mean follicular number was 8.04 [7.26-8.82] (n = 5013 subjects, mean age 29.66 years; Fig. 2).

Ovarian follicular count: a section
Eleven studies including information on ovarian follicular count calculated within an ovarian section (maximum length in sagittal section) were identified. All of them considered women in their reproductive years. The pooled ovarian follicular number was 5.88 [5.20-6.56] (Fig. 3).

Ovarian volume
Seventy-three studies were identified, which reported information on ovarian volume calculated using the formula for a prolate ellipsoid. Among them, 65 studies enrolled women in their reproductive years, whereas 6 [31, 35a and b, 48a, 57b, 59] enrolled girls of pre-pubertal age, one newborn [75] and one post-menopausal woman [30]. When considering the 65 studies including women in their reproductive years (n = 4107 subjects, mean age 27.42 years), the overall mean ovarian volume was 6.11 [5.81-6.42] ml (Fig. 4). After excluding one study enrolling newborns [75], in the remaining six studies considering subjects of pre-pubertal age (n = 358 subjects, mean age 7.42 years), the mean ovarian volume was 1.67 [1.02-2.32] ml (Fig. 4), which was significantly lower than in women in reproductive age (Q = 147.05, p < 0.0001).

Discussion
The present systematic and meta-analytic approach demonstrates for the first time that the overall mean ovarian volume was 6.11 ml in women in reproductive age and 1.67 ml in pre-pubertal girls, with a range of 5.81-6.42 and of 1.02-2.32, respectively. In reproductive age, the overall mean follicular count was 8.04 when calculated in the whole ovary and 5.88 when calculated in an ovarian section, with a range of 7.26-8.82 and of 5.20-6.56, respectively. However, age and the frequency of the transducers were found to significantly modulate these values. In contrast, the authors agreed that all the other sonographic parameters (i.e., ovarian stroma) could not be considered eligible for OVERALL meta-analysis, mainly due to the paucity and heterogeneity of data. The standardization of ultrasound parameters is of paramount clinical relevance since it contributes to the diagnostic workflow of several endocrine conditions, including Premature Ovarian Insufficiency, PCOS and poor ovarian responders (POR) in ART procedures (Bologna criteria: antral follicle count ≤ 5-7 follicles [104]. Noteworthy, we hereby demonstrated that the population of women with a normal ovarian function showed a follicular count ranging between 7.26 and 8.82, thus corroborating the proposed criteria for either PCOM (> 12 according to Rotterdam criteria) or POR (≤ 5-7 according to Bologna criteria).
Another relevant finding of the present study is the apparent age-dependent modulation of the total ovarian follicular count (as calculated in the whole ovary). Despite being not able to highlight a significant difference among the 5 age sub-groups, we noted that the age range 25-30 years was the one with the higher mean follicular count (9.27), followed by a progressive age-related reduction (5.67 in fertile women older than 35 years). In addition, pubertal girls younger than 20 showed a higher mean follicular count (7.87) than those in the 20-25-year group (7.26). In line with this trend, multifollicular ovaries are seen commonly in girls with a gynecological age of ˂ 8 years, and should be considered as a physiologic condition during early adolescence [105]. This is relevant to avoid a misdiagnosis of PCOS in this population. A similar age-modulated trend, although without significant differences between groups, was observed when the follicles were counted in an ovarian section. Our study, meta-analyzing data on follicular count obtained in different ultrasound settings, could be of relevance since the analysis of follicular count per ovarian section is commonly performed in clinical practice. Another important parameter that we were able to metaanalyze was the ovarian volume. As expected, a significant difference was found when comparing pre-pubertal girls and women in their reproductive years (mean values 1.67 vs. 6.11, respectively), whereas our data did not reveal a clear modulation by age in fertile women. As observed for the follicular count, even when we considered the mean upper limit for normal ovarian volume (6.42 ml), this value fell well below the proposed PCOM threshold of 10 ml, thus supporting the appropriateness of this cut-off in defining PCOM.
Even though the ovarian PI is a Doppler ultrasound parameter with a more limited clinical application, we were able to meta-analyze available data while finding again an age-dependent modulation, which revealed two peaks: one pertaining the 20-25 years and one pertaining the > 35 years group. Intriguingly, PI of arterial blood vessels within the genital district has been previously reported to increase as a function of metabolic risk factors, including body mass index, waist circumference, and insulinresistance biomarkers [106]. However, since data on cardiometabolic risk factors or relative treatments were not systemically available in the included studies, the potential contributions of these mediators could only be mentioned without any inference.
Regarding the role of ultrasound methodology, the TV approach has been proved as more sensitive and specific than the TA one, not only in the diagnosis of pelvic disease of gynecologic origin, but also in cases of ovarian follicle monitoring and evaluation for PCOM [107]. In the 2014 "task force report from the Androgen Excess and Polycystic Ovary Syndrome Society", which proposed a threshold of ≥ 25 follicle number per ovary to define PCOM when using transducers with a frequency ≥ 8 MHz, the TA route was described as "not suitable for recording a precise follicle count" [6]. It should be noted that, in situations when the TA approach is the only possible, it can provide a reliable assessment of ovarian volume. As for the transrectal route, in adolescent patients, a 3D version combined with the TA technique has been showed to improve the precision of PCOM definition [108]. In the reviewed studies, all those performed in pre-pubertal girls used a TA-only approach, whereas none reported the use of the transrectal one. Regarding the studies on women in reproductive age, only 4 employed only TA ultrasound [48,57,73,98], while a very few others [25,29,47,55,74,87,97] reported the use of both methods (TA and TV) (see Table 1). None of the 4 studies conducted with TA ultrasound provided data for the meta-analysis of Ovarian Follicular Count in the whole ovary, and only one [73] provided data for the meta-analysis of Ovarian Follicular Count per section. Therefore, no subgroup analysis was performed.
The main strength of our research is the standardized, meta-analytic approach, which, to our knowledge, is the first ever conducted on this topic. In addition, we decided to perform several sub-analyses according to potential modulating factors, including follicular count obtained per ovarian section, which is commonly performed in clinical practice. In addition, we refrained from interpreting our findings in a pathological setting. The key guidance of the present document was to provide informative and user-friendly data, that could be used and critically considered in future research on pathological ovarian conditions. It is advisable that the much-needed standardization of ovarian ultrasonography will allow the development of more evidence-based, universally accepted criteria for the diagnosis of ovarian disorders, especially PCOM.  Among the limitations, we would like to mention the lack of an indication pertaining the cycle phase for ovarian ultrasound; however, ad highlighted in the summarizing Table, the vast majority of records reported to have conducted the ultrasound study during the follicular phase.

Conclusions
In conclusion, our systematic review and meta-analysis provides a relevant clinical information for a more accurate assessment of physiological ultrasound ovarian parameters in pre-pubertal girls and women in reproductive age. Each center should standardize ovarian US according to the available machines, at least when analyzing those parameters which resulted to be significantly modulated by the quality of the transducer. However, we strongly believe that such guidance should improve the interpretation and diagnostic accuracy of ovarian ultrasound parameters in different physiological and pathological settings.

Informed consent Informed consent is not required.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.