Novel application to evaluate endometrial blood flow using transvaginal superb microvascular imaging: A preliminary study describing physiological changes from ovulation to mid‐luteal phase

HM, data acquisition and analysis, manuscript writing. TY, conception and design, manuscript writing. TF and, MF, data acquisition. AN, and TH data analysis and interpretation. KU, manuscript reviewing and project administration. Final manuscript approved by all authors.

This work was supported in part by the Supporting Fund of Obstetrics and Gynecology, Kurume University.

Among 45 images, the evaluations of the EBF grade were concordant between the two raters in 37 images (grade 1: 19 images, grade 2: 7 images, and grade 3: 11 images) and discordant in the remaining eight images (grade 1 vs. 2: 3 images, grade 2 vs. 3: 4 images, and grade 1 vs. 3: 1 image). Cohen's kappa coefficient was 0.724, showing substantial agreement between the two raters. Cohen's kappa coefficient was 0.694, showing substantial agreement between the different menstrual cycles. Among 17 cycles in which EBF was measured on both ovulation and D5–7, EBF on ovulation was grade ≥2 (grade 2: 7 cycles and grade 3: 10 cycles) (Table  S1 ). In contrast, EBF on D5–7 had the following grades (grade 1: 11 cycles, grade 2: 5 cycles, and grade 3: 1 cycle). The EBF from ovulation to D5–7 showed a downgrade of depth in 14 cycles (82.3%) (grade 3 to 2: 3 cycles, grade 3 to 1: 6 cycles, and grade 2 to 1: 5 cycles) and no changes in the remaining three cycles (17.6%) (grade 3 to 3: 1 cycle and grade 2 to 2: 2 cycles) ( p  = 0.001) (Figure  3 ). Changes in the grade of endometrial microvascular flow from ovulation (OV) to 5–7 days after ovulation (D5–7) in the same menstrual cycle. The endometrial thickness differed between the EBF grades on ovulation (grades 1 vs. 2, 1 vs. 3, and 2 vs. 3: p  = 0.557, 0.014, and 0.017, respectively). However, no differences were found between the EBF grades on D5–7 (Figure  4 ). Relation between the grade of endometrial blood flow and the endometrial thickness on ovulation (OV) and 5–7 days after ovulation (D5–7). Open circle, thin group; closed circle, non‐thin group. The dots represent individual images obtained from the patients. Among the cycles in which EBF was observed on ovulation, 7 cycles showed an endometrial thickness of <8.0 mm (thin group) and the remaining 19 cycles showed an endometrial thickness of ≥8.0 mm (non‐thin group). The median endometrial thickness in the thin and non‐thin groups was 6.6 mm (range, 5.4–7.7 mm) and 9.9 mm (range, 8.4–16.5 mm), respectively. The EBF grades were higher in the non‐thin group than in the thin group ( p  < 0.001).

This study is the first to show the physiological changes of endometrial microvascular flow from ovulation to the mid‐luteal phase using SMI. Many reports to date have described the changes in uterine blood flow during the menstrual cycle by ultrasonographic Doppler methods targeting the uterine artery and its peripheral branches. Although the results vary among the reports, almost all reports concluded that the uterine blood flow was maximized around the time of implantation. 1 , 2 , 3 , 4 , 5 , 6 By contrast, our study demonstrated that EBF decreased from ovulation to the mid‐luteal phase. One of the reasons for this discrepancy is that the targeted vessels differed between previous reports and our study. Blood flow parameters such as the pulsatility index of the uterine and arcuate arteries are thought to reflect the vascular resistance of peripheral blood flow (i.e., spiral arteries); however, the blood flow dynamics of the uterine and arcuate arteries may not represent those of the spiral arteries. Two reports focused on endometrial and subendometrial blood flow profiles during the normal menstrual cycle using three‐dimensional (3‐D) power Doppler methods in which a quantitative evaluation was performed by analyzing the histogram of the total Doppler signals of the endometrium and tissues underneath the endometrium. 7 , 8 The endometrial vascularity indices were found to increase throughout the follicular phase, decrease after ovulation, and then increase again during the luteal phase. There are some limitations in describing the EBF by power Doppler methods. The EBF mainly originates from spiral arteries, which have low velocities, and the Doppler power signals are often beyond resolution, yielding underestimation of the EBF profile. Even if detected, power Doppler signals reportedly blot over the “actual” blood vessels, thereby resulting in overestimation, especially in quantitative analyses of vascularity. Three‐D imaging is the one of the most promising methods to describe the entire endometrial vessels. However, to date, 3‐D SMI imaging has not been available on a commercial basis. The alternative approach is to reconstruct 3‐D imaging off‐line by piling up 2‐D SMI images of the vessels. For these approaches, the resolutions for describing tiny vessels, such as radial arteries, would be reduced spatiotemporally. The main objective of this study was not to describe the whole picture of endometrial vessels, including spiral arteries, quantitatively, but rather to delineate the endometrial arterial flow profiles in the different phases of menstrual cycle. This study demonstrated flow profiles for the first time using the currently most‐advanced ultrasound Doppler technique. In this study, we described EBF qualitatively by evaluating the depth of Doppler flow signals penetrating the endometrium. However, in future studies, quantitative evaluation of EBF can be made by measuring vascular density using Doppler flow signals distributed in the endometrium. On B‐mode transvaginal ultrasonography of the uterus in women of reproductive age, endometrial tissues can be delineated by a few layers with hyperechogenicity and/or hypoechogenicity. These layers are clearly separated from the myometrium by a thin hypoechoic layer called the subendometrial halo. 15 SMI shows that tiny vessels connected to the radial arteries run into the tissues beneath the subendometrial halo (i.e., the endometrium). These vessels are considered mostly spiral arteries running through the functional layer of the endometrium 16 (Figure  2 ). In this study, the endometrial microvascular flow profile was qualitatively assessed using a grading system according to the depth of Doppler signals in the endometrium. First, although the grading system criteria are defined, evaluation tends to be subjective. Therefore, we assessed the reliability of the EBF grading system; the kappa coefficient of interobserver differences proved substantial, showing that the EBF grading system we proposed was appropriate for subsequent analyses. Second, we evaluated the difference in EBF between the different menstrual cycles in the same patient, and the substantial kappa coefficient showed high reproducibility between the different cycles. Third, the Doppler signals obtained from the endometrium are so weak that various conditions in the measurements may influence signal detections. To minimize such differences in signal detection derived from the measurements, we compared the EBF within the same patient. We found that EBF decreased from ovulation to the mid‐luteal phase, which is consistent with the report by Raine‐Fenning et al. 7 Possible reasons for the reduction in EBF in the mid‐luteal phase include structural and/or functional changes of the spiral arteries, reduction of the blood supply from the proximal arteries, and/or compression of adjacent tissues such as that due to stromal edema. During the luteal phase, spiral arteries become more tortuous, yielding slow flow velocities and emission of lower Doppler power signals. In rat experiments, spiral arteries constricted after ovulation, resulting in diminished blood flow to the surface endometrium. 17 Raine‐Fenning et al. 7 showed that the EBF changed in accordance with the subendometrial blood flow during the menstrual cycle, suggesting that endometrial flow is influenced by subendometrial blood flow. They also demonstrated that the vascular indices in the endometrial and subendometrial regions, reflecting microvascular spatial density, decreased in the mid‐luteal phase, suggesting that this reduction may be due to an increase in the distance between individual vessels. 7 Further studies of the subendometrial blood flow together with the vascular density of EBF using SMI may explain the mechanisms underlying the decremental change in EBF from ovulation to the mid‐luteal phase. Embryos do not develop well at high oxygen tension because of the production of reactive oxygen species on the surface endometrium, preventing cell growth and differentiation of the embryo. 18 , 19 Therefore, relative hypoperfusion of the endometrium on ovulation is pertinent to provide better circumstances for the fertilized egg and embryo. Our study demonstrated the correlation between the endometrial thickness and the EBF on ovulation but not in the mid‐luteal phase. Angiogenesis is reportedly important in endometrial proliferation, 20 , 21 and a decrease in EBF has been cited as a cause of a thin endometrium. Takasaki et al. 22 reported a good correlation between the endometrial thickness and persistently high blood flow impedance of the uterine radial arteries throughout the menstrual cycle in patients with a thin endometrium. However, because our study showed that the EBF is reduced in the mid‐luteal phase, the differences in EBF, if any, may be beyond the capacity of SMI detection. The relation between endometrial thickness and the pregnancy rate has been reported in assisted reproduction technology, 23 , 24 with significantly lower pregnancy rates when the endometrium is less than 7–8 mm thick. 25 , 26 , 27 , 28 Therefore, we divided the endometrium into thin and non‐thin groups with a cutoff level of 8 mm. Interestingly, in the non‐thin group, no patients had grade 1 EBF, and among patients with endometrial thickness of ≥10 mm, most had grade 3 EBF. These findings suggest that the endometrial perfusion at ovulation is important for implantation and subsequent pregnancy. Successful implantation of fertilized eggs and subsequent pregnancy, conceived either naturally or after in vitro fertilization, are dependent on endometrial receptivity as well as embryonal factors. Endometrial receptivity can be associated with various factors, including the thickness and shape of the endometrium at implantation, morphological characteristics of endometrium and myometrium, such as uterine anomalies, endometrial polyp and submucosal myoma, and functional aberrations, such as peristalsis disorders of subendometrium and localized inflammation (i.e., chronic endometritis). 29 , 30 , 31 In addition, EBF is considered to be closely related to endometrial receptivity. 29 , 30 Therefore, we believe our findings concerning the physiological changes of EBF from ovulation to implantation provide the basis for further studies of the elucidation of implantation disorders due to endometrial malperfusion. The limitations of this study include its retrospective nature and very small sampling size. Because this was a preliminary study to evaluate physiological changes of EBF from ovulation to the mid‐luteal phase, future studies will focus on the changes during the whole menstrual cycle, and endometrial function at ovulation.

SMI showed a decrease in EBF from ovulation to the mid‐luteal phase and established a relation between endometrial thickness in the ovulatory phase and endometrial perfusion in the normal menstrual cycle. These findings may pave the way for the elucidation of the pathophysiology of implantation disorders and the clinical applications for evaluating the efficacy of various therapeutic agents for infertility.

The uterine blood perfusion in women of reproductive age reportedly varies during the menstrual cycle. 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 Initially, the blood flow in the uterine artery was assessed by pulsed‐Doppler waveform analysis with the assumption that the proximal arterial blood flow profile reflected the peripheral vascular resistance of the targeted tissue (i.e., endometrium). 1 , 2 , 3 , 4 , 5 , 6 However, evidence demonstrated that the blood flow profile of the uterine artery and its downstream branches was not representative of the entire endometrial blood flow (EBF). 7 , 8 This is partly because of the unique blood flow system and anatomical structure of the uterine artery. Unlike the placental vascular system, in which the whole chorionic vascular bed is fed by two umbilical arteries running almost straight in the umbilical cord, endometrial perfusion mainly arises from the bilateral uterine and ovarian arteries. For instance, the flow profile of the uterine artery differs between the side containing the ovary that bears the developing ovarian follicle and the contralateral side, suggesting that the uterine arterial flow is influenced by the ovarian arterial flow. 6 The uterine artery runs tortuously along the uterus; therefore, the blood flow profile differs depending on the sampling sites. Three‐dimensional power Doppler methods have recently been applied to describe the vascularity of the whole endometrium and subendometrial regions. 7 , 8 Power Doppler is superior to conventional pulsed Doppler in terms of detecting vessels with low flow rates; power Doppler is independent of the velocity and direction of targeted blood flows, allowing the detection of lower velocities than color Doppler. However, it seems difficult to obtain an accurate and reliable measurement of tiny vascular flows such as those of spiral arteries in the endometrium. By observing the arterioles running through the subendometrial regions and endometrium, EBF profiles were reported to be different during the menstrual cycle, and EBFs were reduced at the time of implantation during the menstrual cycle. Therefore, we hypothesized that the EBF would decrease from ovulation to implantation periods. Superb microvascular imaging (SMI) is a new ultrasound technology to detect low Doppler power signals with high frame rates while neglecting motion artifacts. 9 SMI makes it possible to map the blood flow of various organs and tissues, including the uterine endometrium, with high resolution. 10 , 11 , 12 To our knowledge, no reports have described assessment of EBF by SMI during the normal menstrual cycle. This study aimed to describe the physiological changes of the EBF from ovulation to the mid‐luteal phase using transvaginal SMI.

The authors declare that they have no conflict of interest.

The participants of this study comprised 17 women with regular menstrual cycles of 26–35 days who were managed in our institute from February to August 2022 and retrospectively analyzed. Patients with smoking habits and those with uterine fibroids, adenomyosis, endometrial polyp, tubal edema, diabetes, and cardiovascular disorders were excluded because of the potential impacts of these conditions on uterine blood flow. None of the patients had clinical findings suggestive of intrauterine inflammatory diseases. Their median age was 32.5 years (first to third interquartile range, 29.8–40.0 years). During screening of the patients, SMI was performed to delineate the EBF; the follicular size before ovulation and the endometrial thickness were also assessed. The day of ovulation was estimated by ultrasonographic follicular sizes, urinary luteinizing hormone (LH) assay, and/or basal body temperature. When patients' follicular sizes measured using transvaginal ultrasonography were ≥16 mm in maximum dimension, urinary LH concentrations were titrated semi‐quantitatively with an immunochromatography assay kit (Gold Sign, Morinaga) at 12‐h intervals. This assay kit shows positive staining when urinary LH concentrations exceed 20 mIU/mL, indicating the LH surge. The day of ovulation was determined to be the day following the observation of maximum staining with the LH kit. The event of ovulation was further confirmed by ultrasonographic observations indicating that the follicular size was diminished by 90% in dimension and/or a rise in the basal body temperature. For each patient, the observations were performed by one examiner (H.M.) within 1 day of the estimated ovulation and/or 5–7 days after the estimated ovulation (D5–7). During the analysis of menstrual cycles, no infertility treatments were provided, and mechanical contraception and/or abstaining from sexual intercourse were proposed. The uterus was delineated at the sagittal section using transvaginal ultrasonography (Aplio a450; Canon Medical Systems, Otawara, Japan) incorporated with SMI (Figure  1 and Video  S1 ). The region of interest covered the whole endometrium with an SMI frequency of 4.8 MHz and velocity range of 1.0–1.6 cm/s. The images were recorded as a video clip, and still images during the systolic phase of blood flow were retrieved. Dual views of B‐mode and power Doppler imaging (top) and superb microvascular imaging (SMI) of the uterine sagittal section (middle) and transverse section (bottom) of the same patient. Note that SMI demonstrates that the arteries connected to the radial artery run beyond the outer hypoechoic area surrounding the endometrium (subendometrial halo), forming a dense vascular layer in the endometrium that cannot be detected by power Doppler imaging. In 11 patients, images on ovulation and/or D5–7 were obtained during two menstrual cycles (Table  S1 ). Because some patients could not visit the clinic during the estimated periods of ovulation or mid‐luteal phase as mentioned above, 28 menstrual cycles in 17 patients were measured (both ovulation and D5–7, 17 cycles obtained from cases #5, 6, 7, 8, 9, 10, 11, 13, 14, 15, 16 and 17; ovulation only, nine cycles from cases #1, 2, 3, 4, 6, 9, 12 and 15; and D5–7 only, two cycles from cases #3 and 17). Thus, 26 and 19 images were obtained on ovulation and D5–7, respectively. Endometrial thickness was measured as the end‐to‐end distance of the hypervascular or vascular hypoechoic layer along the endometrial cavity. EBF was evaluated by the depth signals. We modified the original classification according to a report by Applebaum 13 as follows (Figure  2 and Video  S2 ): signals penetrating the outer hypoechoic area surrounding the endometrium (subendometrial halo) but not entering the hyperechoic outer margin of the endometrium (basal layer of the endometrium) (grade 1), signals reaching up to half of the endometrium (grade 2), and signals covering the whole endometrium (grade 3). Classification of endometrial microvascular flow by depth. Grade 1 (top), signals penetrate the outer hypoechoic area surrounding the endometrium (subendometrial halo) but do not enter the hyperechoic outer margin of the endometrium (basal layer of the endometrium); grade 2 (middle), signals reach up to half the endometrium; and grade 3 (bottom), signals cover the whole endometrium. To evaluate the reliability of the EBF grading system, correlations of the evaluations performed by two different examiners (T.F. and M.F.) on a total of 45 images were analyzed with Cohen's kappa coefficient. The raters were blinded to each other's results until the evaluation was completed. The following indices by Landis and Koch 14 were used for interpretation of the agreement: 0.81–1.00, almost perfect; 0.61–0.80, substantial; 0.41–0.60, moderate; 0.21–0.40, fair; and 0.00–0.20, slight. EBF was measured on ovulation and D5–7 in nine and six patients, respectively, during two menstrual cycles (Table  S1 ). To evaluate the reproducibility of EBF between the different menstrual cycles, the correlation of the EBF grade in the nine and six patients measured on ovulation and D5–7, respectively, combined, was analyzed with Cohen's kappa coefficient. To evaluate the changes in EBF from ovulation to D5–7, the difference in the EBF grade was assessed between ovulation and D5–7 in 17 menstrual cycles measured during the same cycle. The change was evaluated using Wilcoxon's signed rank test. The relationship between the EBF grade and endometrial thickness was analyzed in 26 images on ovulation and 19 images on D5–7. Comparisons of the endometrial thickness between the different EBF grades were made using Dunn's non‐parametric comparison for post hoc testing followed by the Kruskal‐Wallis test. In addition, the data were divided into a thin group (endometrial thickness of <8.0 mm) and non‐thin group (endometrial thickness of ≥8.0 mm), and the EBF grade was compared between the two groups using the Wilcoxon rank‐sum test. The significance level was set at p  < 0.05. The statistical analyses were performed using IBM SPSS Statistics Version 20 (IBM Corp.). All patients gave informed consent to participate in this study, which was approved by the institutional review board of our hospital (#21291, February 21, 2022).

Video S1. Click here for additional data file. Video S2. Click here for additional data file. Table S1. Click here for additional data file.