Diagnosis of Idiopathic Premature Ovarian Failure by Color Doppler Ultrasound under the Intelligent Segmentation Algorithm

Abstract

The aim of this study was to explore the application value of transvaginal color Doppler ultrasound based on the improved mean shift algorithm in the diagnosis of idiopathic premature ovarian failure (POF). In this study, 80 patients with idiopathic POF were selected and included in the experimental group, and 40 volunteers who underwent health examinations during the same period were selected and included in the control group, who underwent transvaginal Doppler ultrasound examination. At the same time, an improved mean shift algorithm was proposed based on artificial intelligence technology and applied to ultrasound image processing. In addition, the ovarian artery parameters of patients were compared in two groups, including peak systolic flow rate (PSV), diastolic flow rate (EDV), resistance index (RI), and pulsatile index (PI). The results showed that the relative difference degree (RDD) of the segmentation results of the algorithm in this study was significantly lower than that of Snake, Live_wire, and the traditional mean shift algorithm, while the relative overlap degree (ROD) and Dice coefficient were opposite, and the differences were significant (P<0.05). The mediolateral diameter of control group was 2.87±0.31cm, and the anteroposterior diameter was 1.86±0.28 cm; while those were 2.11±0.36 cm and 1.13±0.34 cm, respectively, in the experimental group, showing significant differences between the groups (P<0.05). Of the 80 patients in the experimental group, 132 cases with ovarian arteries were found; among 40 patients in the experimental group, 76 cases were found with ovarian arteries, and the hemodynamic detection rate of the experimental group was significantly lower than that of the control group (P<0.05). The ovarian artery parameters PI, RI, and S/D of the experimental group were significantly higher than those of the control group, and the differences were statistically significant (P<0.05). The results showed that the segmentation results of the improved algorithm in this study were more superior to the segmentation results of other algorithms. The regional information loss of the segmentation results was not serious, and the resolution was higher and the definition was higher. The transvaginal color Doppler ultrasound based on the artificial intelligence segmentation algorithm can clearly show the functional status and hemodynamics of the patient’s ovaries. The ovarian artery parameters PI and RI can be used as specific indicators for evaluating the POF.

Three-dimensional transvaginal ultrasonography in the evaluation of diminished ovarian reserve and premature ovarian failure

“…9 Through the use of post-processing software, list marking processing can be implemented for different follicle diameters and volumes, which is convenient for clinicians to observe the subtle follicular structural changes and improves analysis of sinus follicle development. Yu L et al 12 16 This study shows that transvaginal threedimensional ultrasound has good diagnostic efficacy for both DOR and POF, which is consistent with the results of Singhal P et al 19 The use of 3D-TVS technology in the diagnosis and evaluation of DOR and POF does have application value. 15,18 With the progress and update of the examination equipment, ultrasound technology is simple, fast, non-invasive and accurate, and has been gradually cited in clinical work.…”

“…11 These difference may be due to the fact that the inflow site of the right ovarian vein is the inferior vena cava, while the inflow site of the left ovary is the left renal vein. 11,12 Currently, there are few clinical comparative analyses of the left and right ovaries using three-dimensional ultrasound, and further discussion is warranted. 13,14 Our results reported lower AFC, OV, VI, VFI and FI within both the DORgroup and the POF-group versus the Normal ovarian function group (p<0.05).…”

Multi-omics and machine learning for the prevention and management of female reproductive health

“…In order to assess the customized risk of POI following chemotherapy, Chung et al constructed a machine learning-based model that had an accuracy of 88% (area under the ROC 0.87, 95% CI: 0.77-0.96; p < 0.001) (96). An enhanced mean shift algorithm based on artificial intelligence (AI) technology was used to process ultrasound images in women with idiopathic POI, where the functional condition and hemodynamics of patients' ovaries were clearly visible on the transvaginal color Doppler ultrasonography (97). Another study demonstrated that building diagnostic methods for POI prediction may be accomplished using artificial neural networks, where the generalization ability of the train set, validation set, and test set was validated, with a prediction accuracy of over 90% in the test, train, and validation sets (98).…”

A comparison of Doppler measures of ovarian blood flow between women with and without ovarian dysfunction and correlations of Doppler indices with ovarian dysfunction markers: a meta-analysis

“…Ultrasonography has an important role in reproductive medicine and assisted reproductive technologies because of its utility for several purposes including ovarian response monitoring, assessment of endometrial receptivity, transvaginal aspiration of oocytes, transcervical transfer of the embryo to the uterus, pregnancy monitoring, and fetal health assessment. It is also used to measure ovarian volume, antral follicular count (AFC), ovarian stromal blood flow, and to evaluate ovarian function after vascular embolization (3)(4)(5)(6).…”