Proportional vascularization along the fallopian tubes and ovarian fimbria: assessment by confocal microtomography

Objective To evaluate and reconstruct three-dimensional images of vascularization along the fallopian tube (FT), as well as to determine its relationship with the ovary and ovarian fimbria, and to quantify the blood vessels along the FT according to its anatomical segments, using confocal microtomography (micro-CT). Materials and Methods Nine specimens (six FTs and three FTs with ovaries) were fixed in a solution of 10% formalin for > 24 h at room temperature. Iodine staining was performed by soaking the specimens in 10% Lugol’s solution for 24 h. All specimens were evaluated using micro-CT. A morphometric analysis was performed on the reconstructed images to quantify the vascular distribution along the FT. Results In the FTs evaluated, the density of blood vessels was significantly greater in the fimbrial segments than in the isthmic segments (p < 0.05). The ovarian fimbria was clearly identified, demonstrating the important relationship between these vessels and the FT fimbriae. Conclusion We believe that the vascularization in the fimbriae is greater than and disproportional that in the other segments of FT, and that the ovarian fimbria plays an important role in the development of that difference.


INTRODUCTION
The human oviduct has been present in the medical literature for almost 3,000 years and was first described, by the Greek physician Soranus of Ephesus (circa AD 100), as the conduit linked to fertilization. In his masterpiece carcinomas (the most aggressive and deadly gynecologi cal malignancies) originate from the distal FT rather than from the ovarian surface epithelium (2)(3)(4)(5) . Consequently, the International Federation of Gynecology and Obstet rics staging classification for ovarian, tubal, and peritoneal cancers was revised in 2014 (6) . That rekindled interest in FT physiology and pathology. There have been increasing numbers of studies involving the cellular and molecular aspects of the FT, mainly the mucosa in the fimbriae, the most common location for tubal malignant transforma tion (7,8) . Distinctive stem cell markers were recently iden tified in the fimbriae, suggesting that the fimbriae have a unique embryonic development, separate from other tubular segments (9,10) . However, information about vascu larization along the FT is scarce. One report, by Stange (11) , described the vascular distribution along the FT as "dis proportional", increasing among the fimbriae.
In this study, we aimed to observe the distribution of blood vessels along the FT, focusing on the infundibulum and its relationship with the ovary. To that end, we em ployed confocal microtomography (microCT), a nonde structive Xray technology that provides highresolution threedimensional images of ex vivo specimens.

MATERIALS AND METHODS Patients
We included nine patients who were scheduled to undergo surgical procedures for the treatment of benign gynecological conditions, such as uterine adenomyosis and uterine fibroid, or surgical sterilization. The study was approved by the Research Ethics Committee of Vassouras University (Reference no. 56031916.0.0000.5290), in the municipality of Vassouras, Brazil, and all participating pa tients gave written informed consent. Patients who had a macroscopically abnormal FT or ovary were excluded, as were those suspected of having malignant conditions.

Specimen preparation
Nine specimens (six FTs and three FTs with ovaries) were fixed in 10% formalin for > 24 h at room tempera ture (12) . After washing twice with distilled water, iodine staining was performed by soaking the specimens in 10% Lugol's solution for 24 h, as previously described for bio logical samples (13) . The specimens were removed from the staining solution and rinsed with phosphatebuffered for malin to remove excess stain and prevent surface satura tion. To hold the specimens in place, ensure the mechani cal stability, and avoid movement artifacts during micro CT, the specimens were fixed onto Styrofoam. After the micro-CT scan, the specimens were returned to 10% for malin to destain and to prepare for histological analysis.

Micro-CT and images analysis
For the acquisition of FT images, we employed a mi croCT system (SkyScan 1173 v.1.6.9.4; Bruker microCT, Kontich, Belgium). The parameters (Xray energy, current, and exposure time) were individually adapted to optimize the images for each specimen. The following parameters were used for image acquisition: energy, 40 kV; current, 200 µA; exposure time, 250-100 ms; scan duration, 24-77 min; and voxel size, 11.04-21.01 µm. Postprocessing analysis and image reconstructions were performed using the software CTan, version 1.16 (Bruker microCT).

Morphometric analysis-quantifying vessel density
To quantify the vascular distribution along the FT, a morphometric analysis was performed on images recon structed using the software CTan. Images were classified into regions of interest: fimbriae, ampulla, and isthmus. Following classification, the FT lumen was excluded and the total volume of each segment was quantified. The quantitative measurement of vascular volume was calcu lated using the morphometric parameter of connectivity (14) . To remove the maximum number of artifacts and isolated radiopaque signals, continuous, connected radiopaque images of the vessels were selected. The total volume of the vessels within each segment was reported in voxels.
To quantify the vessel density within each region of the specimen, the ratio between the total FT volume and the FT blood volume was analyzed. Statistical analysis was performed using the Predictive Analytics Software pack age, version 18.0 for Windows (SPSS Inc., Chicago, IL, USA) and GraphPad Prism (GraphPad Software Inc., San Diego, CA, USA). We calculated the mean, standard deviation, median, 25th and 75th percentiles, as well as minimum and maximum values. Since the sample size was too small to properly evaluate the distribution, nonpara metric tests were used. Friedman and Dunn's tests were used in order to compare the different regions of the uter ine tubes. Statistical significance was defined as p < 0.05.

RESULTS
The vascular distribution of diseasefree FTs from women of reproductive age was evaluated after patients underwent a total abdominal hysterectomy or salpingec tomy. The uterus was separated from the FT because of the lack of morphological references and methods to cal culate its vascularization. The easily recognizable FT mu cosa was vital to selecting the FT segments.
The vascular distribution of the FT was visible in all the specimens. The specimens presented outstanding resolution and demonstrated a clear threedimensional vessel network. Although the Lugol's solution was read ily absorbed into the larger vessels, very small vessels and capillaries could not be detected. The voxel size and in adequacies in connectivity limited our ability to study the smaller vessels. The vessel endothelium was also visible, with greater intensity than the muscle segments of the FT. The vessel endothelium was visible even in the vessels without blood (Figure 1). Table 1 shows the descriptive statistics for total vol ume, vascular volume, and vascular proportion in the fim brial, ampullary, and isthmic segments of the FT. Differ ences between the fimbrial and isthmic segments of the FT were significant (p < 0.05). The values of the fimbriae and ampulla were not statistically different, although they did demonstrate increasing vascularization toward the dis tal part of the FT (Table 2). The presence of vessels in the ovarian fimbria was effortlessly visualized and reconstructed in threedimen sional imaging (Figure 2), demonstrating the important vascular relationship between the FT fimbriae and the ovarian fimbria. The characteristics of the ovarian fim bria, such as the number and thickness of vessels, were not evaluated.

DISCUSSION
The present study aimed to observe the distribution of blood vessels along the FT. In addition to its function as a conduit between the uterus and ovary for gamete fertiliza tion, the FT has been associated with many gynecological disorders. It is known to be the pathway for retrograde menstruation and development of ovarian endometrioid or clear cell carcinoma (15) , it is associated with the spread of tubal and ovarian highgrade serous carcinoma into the endometrium (16) , and it is frequently involved with as cending infections in pelvic inflammatory disease. Recent conceptualizations for the pathogenesis of serous ovarian tumors have focused on their possible origins within the FT (5) . After seminal sources suggested that the FT is a source of ovarian and peritoneal tumors, many researchers discovered a stem cell-like enrichment of the fimbriae and suggested that the fimbriae have a different developmental origin from the rest of the tube (9) . This was also previously described, in cases of isolated ovarian agenesis without the accompanying fimbriae but with an otherwise normal  FT (17,18) . In the present study, we were able to demonstrate increased vascularization in the distal segments of FT. The absence of a statistically different fimbriae-ampulla ratio notwithstanding, the method employed was able to quan tify variability in the vasculature and to emphasize the as sociation between the ovarian fimbria and FT fimbriae. The ovarian fimbria establishes the ovarian surface epi thelium-ligament-fimbriae epithelial continuum, which plays a recognized role in FT vascularization (19) . In our study, the ovarian fimbria demonstrated an intense, local ized vascular connection with the FT fimbriae and is likely the driving force behind the increased vascularization in this segment. To our knowledge, Stange (11) was the first to describe this disproportional, increased vascularization and considered the observation to be a result of the ab sence of a bursa ovarica and of fimbriae. Other studies have described various temperature gradients along the FT and ovary, which they have attributed to relative blood flow distribution, because sperm has the ability to be di rectionally guided through the oviduct by temperature changes (20,21) .
A number of recent studies conducted in Brazil have highlighted the importance of imaging methods in the evaluation of the female reproductive system (22)(23)(24)(25)(26)(27) . The use of microCT was pivotal to observing the vasculariza tion in detail. Over the past few years, since the applica tion of contrast agents to increase radiographic contrast of small organic samples became common, impressive progress has been made in the study of small softtissue samples with microCT. Contrast varies depending on the tissue specificity of the contrast agent, as well as its cost, toxicity, and overall effect on the histologically prepared specimen. Iodine, which is the most frequently used con trast agent in microCT, has several advantages, such as low toxicity, low cost, rapid staining, and perceived nondestructiveness, allowing biological specimens to be analyzed and subsequently used in conventional micros copy (13,28) . In human tissues, the ex vivo use of micro CT includes unenhanced studies-for intraoperative breast cancer (29,30) , lungs (31) , bones (32) , and dental pa thologies (33,34) -and contrastenhanced studies-for the brain (35) , fetal heart (36) , postmortem lower digestive tract obstruction imaging (37) , and human autopsy (38) . Some  Surface reconstruction of the adnexa, demonstrating part of the myometrium on the right. The ovary was normal (asterisk), and the arrow shows the ovarian fimbria. B: CT image of the left adnexa. Note the intense vascularization of the fimbriae compared with the other segments of the tube and the vascular connection from the ovary (asterisk) to the fimbriae (arrow, ovarian fimbria).
A B methods used for phenotyping FTs have good histopatho logical correlation (39,40) ; however, those methods are not able to reconstruct the FT vasculature microscopically.
Micro-CT allows the quantification of soft-tissue morphology, including linear and volumetric data, and has been used for many vascular reconstructions in ani mals (41) . In the present study, an "angiogramlike" image was produced due to the high affinity between iodine and blood. However, smaller vessels and vessels not contain ing blood may have been missed, the observed affinity be tween iodine and endothelium being the exception. The absence of small vessels in the continuum image pre cluded the use of an automated system to reconstruct the vessels. Further studies are needed in order to optimize fixation, the staining process, and image acquisition for gynecologic specimens. The present study was also limited by the small number of specimens and therefore could not measure the influence of patient age, hormone therapy, and the various surgical techniques performed during the specimen collection process.
In summary, we believe that the FT fimbrial segment has a disproportional and increased vascularization com pared with the other segments and that the ovarian fim bria plays an important role in the development of those differences. Larger, multicenter studies are needed in or der to confirm these findings, to determine the influences of the ovarian surface epithelium-ligament-fimbriae epi thelial continuum, and to identify the roles that both play in tubal and ovarian physiology and pathologies.