Segmentation of mature human oocytes provides interpretable and improved blastocyst outcome predictions by a machine learning model.

Within the medical field of human assisted reproductive technology, a method for interpretable, non-invasive, and objective oocyte evaluation is lacking. To address this clinical gap, a workflow utilizing machine learning techniques has been developed involving automatic multi-class segmentation of two-dimensional images, morphometric analysis, and prediction of developmental outcomes of mature denuded oocytes based on feature extraction and clinical variables. Two separate models have been developed for this purpose-a model to perform multiclass segmentation, and a classifier model to classify oocytes as likely or unlikely to develop into a blastocyst (Day 5-7 embryo). The segmentation model is highly accurate at segmenting the oocyte, ensuring high-quality segmented images (masks) are utilized as inputs for the classifier model (mask model). The mask model displayed an area under the curve (AUC) of 0.63, a sensitivity of 0.51, and a specificity of 0.66 on the test set. The AUC underwent a reduction to 0.57 when features extracted from the ooplasm were removed, suggesting the ooplasm holds the information most pertinent to oocyte developmental competence. The mask model was further compared to a deep learning model, which also utilized the segmented images as inputs. The performance of both models combined in an ensemble model was evaluated, showing an improvement (AUC 0.67) compared to either model alone. The results of this study indicate that direct assessments of the oocyte are warranted, providing the first objective insights into key features for developmental competence, a step above the current standard of care-solely utilizing oocyte age as a proxy for quality.

An artificial intelligence tool predicts blastocyst development from static images of fresh mature oocytes.

RESEARCH QUESTION: Can a deep learning image analysis model be developed to assess oocyte quality by predicting blastocyst development from images of denuded mature oocytes? DESIGN: A deep learning model was developed utilizing 37,133 static oocyte images with associated laboratory outcomes from eight fertility clinics (six countries). A subset of data (n = 7807) was allocated to test model performance. External model validation was conducted to assess generalizability and robustness on new data (n = 12,357) from two fertility clinics (two countries). Performance was assessed by calculating area under the curve (AUC), balanced accuracy, specificity and sensitivity. Subgroup analyses were performed on the test dataset for age group, male factor and geographical location of the clinic. Model probabilities of the external dataset were converted to a 0-10 scoring scale to facilitate analysis of correlation with blastocyst development and quality. RESULTS: The deep learning model demonstrated AUC of 0.64, balanced accuracy of 0.60, specificity of 0.55 and sensitivity of 0.65 on the test dataset. Subgroup analyses displayed the highest performance for age group 38-39 years (AUC 0.68), a negligible impact of male factor, and good model generalizability across geographical locations. Model performance was confirmed on external data: AUC of 0.63, balanced accuracy of 0.58, specificity of 0.57 and sensitivity of 0.59. Analysis of the scoring scale revealed that higher scoring oocytes correlated with higher likelihood of blastocyst development and good-quality blastocyst formation. CONCLUSION: The deep learning model showed a favourable performance for the evaluation of oocytes in terms of competence to develop into a blastocyst, and when the predictions were converted into scores, they correlated with blastocyst quality. This represents a significant first step in oocyte evaluation for scientific and clinical applications.

Can endometrial thickness at baseline or maximum stimulated levels during IVF predict the presence of endometrial polyps in women with two failed embryo transfers?

PURPOSE: To evaluate whether different measurements of endometrial thickness pre-IVF cycle and during the IVF cycles as measured by transvaginal ultrasound are associated with the absence or presence of endometrial polyps. DESIGN: A retrospective cohort study was conducted in a university-affiliated fertility center. Patients were women who underwent two embryo transfer cycles and failed to conceive. INTERVENTIONS: hysteroscopic evaluation and resection of any masses. RESULTS: There was no difference on comparing the groups with and without polyps in the mean endometrial thicknesses at baseline scans pre-treatment or during IVF cycle. For women who failed two embryo transfer cycles, at any given endometrial thickness the probability of the presence of a polyp was 30-40%. ROC curves failed to detect an actionable relationship with different endometrial thicknesses and the relationship with an endometrial polyp, with most areas under the curve being just above 0.5. However, once the maximum stimulated endometrial thickness was ≥ 13 mm, there was a 70% chance of a polyp being noted at hysteroscopy. This was a statistical difference in the probability of a polyp being present as compared to the lesser thicknesses (p = 0.05). CONCLUSION: Baseline or maximum stimulated endometrial thickness at IVF fails to predict with accuracy the presence of a polyp. However, if the maximum stimulated thickness was at least 13 mm, there was a higher probability of a uterine polyp being present. Such a cutoff would nevertheless miss most polyps. At any baseline thickness on CD 2-5, a polyp has a 30-40% probability of being present in women who failed two embryo transfers. ROC curves suggest that at baseline, or maximum stimulated endometrial thickness, the ability to predict a polyp is no better than flipping a coin. As such, endometrial cavity evaluation for polyps is legitimate in women with two embryo transfers irrelevant of the baseline or stimulated thickness.

Endometrial compaction before frozen euploid embryo transfer improves ongoing pregnancy rates.

OBJECTIVE: To assess whether the calculated difference in endometrial thickness from the end of the estrogen phase to the day of ET (after 6 days of P in hormonally prepared cycles) is associated with ongoing pregnancy rates in euploid frozen ETs (FETs). DESIGN: An observational cohort study. SETTING: Single tertiary care medical center. PATIENT(S): Ultrasound images from 234 hormonally prepared FET cycles were assessed. All the transfers were elective single ETs of a euploid embryo, post-preimplantation genetic testing for aneuploidy (PGT-A). INTERVENTION(S): Ultrasound measurements of peak endometrial thickness at the end of the estrogen phase and again after 6 days of P at the time of ET. MAIN OUTCOME MEASURE(S): Ongoing pregnancy rate in relation to the delta between endometrial thickness at the end of estrogen phase and at the time of ET. RESULT(S): We calculated the ongoing pregnancy rate in cycles where the endometrial lining decreased (compacted) after addition of P by 5%, 10%, 15%, and 20% and demonstrated a significantly higher pregnancy rate after all rates of compaction of the endometrial lining in comparison with cycles where the endometrial lining did not compact. The ongoing pregnancy rate in this cohort, after compaction of 15% or more, was 51.5%, compared with 30.2% in cycles where the endometrial lining did not compact. CONCLUSION(S): There is a significant correlation between endometrial lining compaction and ongoing pregnancy rate in FET cycles of euploid embryos. These findings help to explain why some euploid embryos may fail to implant.

Endometrial compaction (decreased thickness) in response to progesterone results in optimal pregnancy outcome in frozen-thawed embryo transfers.

OBJECTIVE: To evaluate whether the change in endometrial thickness between the end of the estrogen phase and the day of embryo transfer has an impact on the pregnancy rate in frozen-thawed embryo transfer (FET) cycles. DESIGN: Retrospective observational cohort study. SETTING: Single tertiary care medical center. PATIENT(S): Ultrasound images in 274 FET cycles were reviewed. All patients underwent endometrial preparation with the use of hormonal therapy. INTERVENTIONS(S): Ultrasound measurements of endometrial thickness at the end of the estrogen phase and the day of embryo transfer. MAIN OUTCOME MEASURE(S): The change in endometrial thickness and ongoing pregnancy rate. RESULT(S): We calculated the ongoing pregnancy rate in patients whose endometrial thickness decreased (compacted) after starting progesterone by 5%, 10%, 15%, or 20% compared with patients with no change or increased endometrial thickness. The ongoing pregnancy rate was significantly increased at all levels of compaction compared with no compaction. The ongoing pregnancy rate showed a significant increase with each decreasing quartile of change in thickness (increased percentage of compaction) in the progesterone phase compared with the estrogen phase. CONCLUSION(S): There is a highly significant inverse correlation between the ongoing pregnancy rate and the change of endometrial thickness between the end of estrogen administration and the day of embryo transfer.

Oral contraceptive pills as an option for non-surgical management of retained products of conception - a preliminary study.

Many Patients with persistent retained products of conception prefers to avoid surgical interventions, such as a dilatation and curettage (D&C) that might pose an additional future risk to their already compromised fertility or obstetric performance. The aim of this study was to the possibility of induced withdrawal bleeding following oral contraceptive administration as a non-surgical treatment for patients with persistent retained products of conception (RPOC). A retrospective study of patients presenting with retained products of conception (RPOC) after failed expectant management or after treatment with PGE1 was performed. Twelve women presenting with RPOC at ≤8 weeks gestation with minimal to mild vaginal bleeding and no signs of infection were treated with oral contraceptive pill (OCP) containing 0.03 mg ethinylestradiol and 0.15 mg of desogestrel for 3 weeks. Out of the 12 patients treated, nine women (75%) successfully expelled the RPOC after completing the three-week course of OCPs. The three cases (25%) that did not resolve following OCP treatment had pregnancy products with positive blood flow on Doppler examination. We conclude that OCPs may be a useful medical treatment option for persisting RPOC in selected patients with absence blood flow on Doppler examination wishing to avoid surgical intervention.

Intramuscular versus Vaginal Progesterone Administration in Medicated Frozen Embryo Transfer Cycles: A Randomized Clinical Trial Assessing Sub-Endometrial Contractions.

OBJECTIVE: The study aimed to assess whether sub-endometrial contractility is reduced by the use of intramuscular (IM) progesterone. DESIGN: This is a randomized clinical trial. Patients assigned to a medicated day 5 frozen embryo transfer (FET) were randomly allocated to "vaginal progesterone" or "IM progesterone": patients randomized to the vaginal arm were treated with 200 mg micronized progesterone 3 times daily while patients randomized into the IM progesterone arm were treated with a single daily injection of 50 mg progesterone in oil. The main outcome measure was the number of sub-endometrial contractions (waves) per minute 1 day before a blastocyst embryo transfer. RESULTS: Thirty-four patients were enrolled. The progesterone serum concentration was significantly higher in patients using the IM progesterone (85.2 ± 50.1 vs. 30.3 ± 11.2 nmol/L, respectively) but this did not translate into a lower sub-endometrial contractility (2.4 ± 4.8 vs. 1.4 ± 1.1 contraction/min, respectively). Clinical pregnancy rates were comparable between groups. The number of sub-endometrial waves was significantly lower among pregnant patients (p = 0.02). CONCLUSIONS: The use of IM progesterone in medicated FET cycles does not reduce the sub-endometrial activity compared to vaginal progesterone administration. Our data support a poor clinical pregnancy outcome with high wave activity, regardless of the progesterone mode.

Current tools for the optimization of embryo transfer technique for recurrent implantation failure.

Recurrent implantation failure (RIF) is the name of a clinical condition coined following the widespread use of in-vitro fertilization (IVF), which has allowed compartmentalization of several different fertility treatments. Its definition is dynamic and depends on the population of patients studied, as well as the type and quality of clinical practice. In this review we survey the tools which are currently used in order to improve treatment outcome in patients with recurrent implantation failure. Some of these practices are more commonly or firmly established than others, however the beneficial contribution of most of these tools to improve reproductive outcomes among patients with recurrent implantation failure still lacks proper scientific validation.

Obstetric and perinatal outcome from single cleavage transfer and single blastocyst transfer: a matched case-control study.

AIM: To evaluate a possible relationship between extended embryo culture and outcome of pregnancies resulting from single embryo transfers (SETs). DESIGN: A retrospective matched case-control study Setting: University fertility center Patients: About 106 live births from single cleavage embryo transfers were matched 1:2 with 212 live births from single blastocyst transfers. INTERVENTIONS: A cohort of 3522 fresh SETs using non-donor oocytes in women ≤40 years old from August 2010 to December 2013. Live births were matched by maternal age, body mass index, smoking and parity. Adjustments were made for gender of the baby and embryo quality. Obstetric and perinatal outcomes including birth weight, low birth weight, small for gestational age, preterm delivery, preeclampsia, placental abruption and neonatal complications were compared. RESULTS: Matched live birth outcomes showed no increased risk of obstetric or perinatal complications in pregnancies resulting from single blastocyst transfers compared to single cleavage transfers. CONCLUSIONS: Extended culture was not associated with increased adverse obstetric and perinatal outcome in pregnancies resulting from fresh SETs in this study.

Do measurements of uterine septum using three-dimensional ultrasound and magnetic resonance imaging agree?

OBJECTIVE: To evaluate the agreement between measurements of uterine septum by 3-D ultrasound (3-D US) and magnetic resonance imaging. METHODS: Fourteen women with a uterine septum underwent both a 3-D US and an MRI followed by hysteroscopic metroplasty in a tertiary care infertility clinic within a university teaching hospital. The agreement between 3-D US and MRI measurements of septum length (SL), septum width (SW) and the distance between fundal serosa and septal tip (SSD) was compared using the Bland Altman limit of agreement analysis. RESULTS: 3-D US correctly identified all women who had a uterine septum identified with MRI. The mean difference in SL was 2.7 mm (upper limit of agreement and lower limit of agreement were 10.3 mm and -4.8 mm, respectively). The mean difference in SW was 2.3 mm (upper limit of agreement and lower limit of agreement of 10.4 mm and -9.2 mm, respectively). The SSD mean difference was 2.8 mm (upper limit of agreement and lower limit of agreement of 10.6 mm and -5 mm, respectively). CONCLUSIONS: Categorical agreement in diagnosis was 100% and agreement between 3-D US and MRI measurements was good with narrow differences. Variations were unlikely to be greater than intra- and inter-observer variations anticipated in MRI readings. 3-D US can replace MRI for diagnosis and measurement of uterine septa.

Objectif : Évaluer la concordance des mesures de la cloison utérine par échographie 3-D (EG 3-D) et par imagerie par résonance magnétique. Méthodes : Quatorze femmes présentant une cloison utérine ont fait l'objet d'une EG 3-D et d'une IRM, le tout ayant été suivi par la tenue d'une métroplastie hystéroscopique au sein d'une clinique de fertilité (soins tertiaires) située dans un hôpital universitaire. La concordance des mesures par EG 3-D et par IRM de la longueur de la cloison, de la largeur de la cloison et de la distance entre la séreuse fundique et l'extrémité septale a été évaluée au moyen de l'analyse de limite de concordance Bland Altman. Résultats : L'EG 3-D a permis d'identifier correctement toutes les femmes qui présentaient une cloison utérine ayant été identifiée par IRM. La différence moyenne en ce qui concerne la longueur de la cloison était de 2,7 mm (les limites de concordance supérieure et inférieure étaient de 10,3 mm et de -4,8 mm, respectivement). La différence moyenne en ce qui concerne la largeur de la cloison était de 2,3 mm (les limites de concordance supérieure et inférieure étaient de 10,4 mm et de -9,2 mm, respectivement). La différence moyenne en ce qui concerne la distance entre la séreuse fundique et l'extrémité septale était de 2,8 mm (les limites de concordance supérieure et inférieure étaient de 10,6 mm et de -5 mm, respectivement). Conclusions : La concordance de catégorie en ce qui concerne le diagnostic était de 100 % et la concordance des mesures par EG 3-D et par IRM était bonne (le tout s'accompagnant de faibles différences). Les variations étaient peu susceptibles d'être supérieures aux variations intraobservateurs et interobservateurs anticipées pour ce qui est des valeurs issues de l'IRM. L'EG 3-D peut remplacer l'IRM pour l'établissement d'un diagnostic de cloison utérine et pour la mesure de celle-ci.

Live birth following serial vitrification of embryos and PGD for fragile X syndrome in a patient with the premutation and decreased ovarian reserve.

PURPOSE: To present a live birth resulting from serial vitrification of embryos and pre-implantation genetic diagnosis (PGD). METHODS: A 31-year-old with primary infertility, fragile-X premutation, and decreased ovarian reserve (DOR) (baseline FSH level 33 IU/L), presented after failing to stimulate to follicle diameters >10 mm with three cycles of invitro fertilization (IVF). After counseling, the couple opted for serial in-vitro maturation (IVM), embryo vitrification, and genetic testing using array comparative genomic hybridization (aCGH) and PGD. Embryos were vitrified 2 days after intra-cytoplasmic sperm injection (ICSI). Thawed embryos were biopsied on day-three and transferred on day-five. RESULTS: The couple underwent 20 cycles of assisted reproductive technology. A total of 23 in-vivo mature and five immature oocytes were retrieved, of which one matured in-vitro. Of 24 embryos, 17/24 (71 %) developed to day two and 11/24 (46 %) survived to blastocyst stage with a biopsy result available. Four blastocysts had normal PGD and aCGH results. Both single embryo transfers resulted in a successful implantation, one a blighted ovum and the other in a live birth. CONCLUSIONS: Young patients with DOR have potential for live birth as long as oocytes can be obtained and embryos created. Serial vitrification may be the mechanism of choice in these patients when PGD is needed.

Endometrial thickness in 1,500 asymptomatic postmenopausal women not on hormone replacement therapy.

OBJECTIVE: To determine the normal endometrial thickness (ET) on transvaginal ultrasound (TVUS) of asymptomatic postmenopausal women not on hormone replacement therapy. A subgroup that was determined to be suspicious for having an endometrial polyp was compared with the remainder. METHODS: This prospective study selected 1,500 consecutive asymptomatic postmenopausal women receiving TVUS assessment from January to August 2010. ET was recorded. Results were divided into those with a normal-appearing lining (n = 1,399) and those suspicious for polyp (n = 101). Results for the entire sample were obtained and the groups were compared using independent samples t tests. RESULTS: Of 1,500 women aged 45-95 years, 77.1% had an ET of ≤4 mm and 92% were ≤5 mm. Independent samples t tests were performed to compare the mean age and mean ET based on polyp status (i.e. with or without a possible polyp). There was a significant difference in mean age, 67.71 vs. 62.36 years (p < 0.01) and mean ET 8.02 vs. 3.40 mm (p < 0.01) between groups. CONCLUSIONS: 92% of asymptomatic postmenopausal women not on hormone replacement therapy had an ET of ≤5 mm. The mean ET was 3.71 ± 1.9 mm. However, a significant group, 6.7%, had an endometrial lining suspicious for polyp. These women had a significant increase in mean age and ET.

Nimodipine, a calcium channel blocker, delays the spontaneous LH surge in women with regular menstrual cycles: a prospective pilot study.

BACKGROUND: Currently GnRH analogue injections are used to prevent premature LH surges in women undergoing assisted reproductive technology. This was a pilot study to determine the safety and effectiveness of nimodipine, an oral calcium channel blocker, to delay the mid-cycle spontaneous LH surge in women with regular menstrual cycles. METHODS: Eight women with regular menstrual cycles self-monitored three consecutive cycles for the day of an LH surge by daily urine assay. The first and third cycles were observatory. In the second cycle, subjects took nimodipine 60 mg by mouth three times daily for four days, starting two days prior to the expected LH surge day based on cycle one. RESULTS: The LH surge day in cycle 2 (nimodipine) was significantly delayed in comparison to both observatory cycle 1 (15.5+/-3.4 vs 14.0+/-2.8 days; p=0.033) and cycle 3 (15.1+/-3.5 vs 13.1+/-2.4 days; p=0.044). There was no difference in the LH surge day between the two observatory cycles (13.4+/-2.4 vs 13.1+/-2.4 days; p=0.457). Three patients experienced a mild headache. CONCLUSIONS: There was a statistically significant delay in the spontaneous LH surge day in the treatment cycle in comparison to both observatory cycles. Nimopidine should be further investigated as an oral alternative to delay a spontaneous LH surge.

The effect of acute aromatase inhibition on breast parenchymal enhancement in magnetic resonance imaging: a prospective pilot clinical trial.

OBJECTIVE: The breast is highly hormonally sensitive especially to the sex steroid hormone estrogen. Both physiological and iatrogenic steroid hormone modifications could affect how the breast tissue may appear in breast imaging techniques. We hypothesized that estrogen deprivation therapy could reduce breast nonspecific enhancement on magnetic resonance imaging (MRI). METHODS: This study was a prospective pilot phase II clinical trial. The study was approved by Health Canada and the institutional research ethics board, and participants signed informed consent forms. Sixteen healthy postmenopausal women were enrolled, and 14 completed the study. Baseline breast MRI was done followed 1 month later by administration of a high-dose aromatase inhibitor (letrozole 12.5 mg/day) for 3 successive days before a second breast MRI. Background breast parenchymal enhancement was compared between the pretreatment and posttreatment studies. RESULTS: There was a statistically significant reduction of the average background breast enhancement after treatment with aromatase inhibitors compared with baseline MRI. Of particular interest, specific areas of benign breast enhancement were reduced after aromatase inhibitor treatment. No significant adverse effects were recorded using this relatively high dose of the aromatase inhibitors. CONCLUSIONS: This preliminary study provided evidence that aromatase inhibitors could reduce the parenchymal background enhancement of benign breast tissue during MRI and may improve the specificity of the technique.

Modified oxygen mask to induce target levels of hyperoxia and hypercarbia during radiotherapy: a more effective alternative to carbogen.

PURPOSE: Carbogen has long been under investigation as an adjuvant to radiotherapy of tumors. A major factor confounding its evaluation is its inconsistency in raising blood partial pressure of CO(2) (pCO(2)). We investigated whether a new partial rebreathing method would provide better control of pCO(2) than carbogen. METHODS AND MATERIALS: We compared the efficacy of each method in 10 healthy volunteers. Volunteers breathed 1.5, 3 and 5% carbogen in 5-min stages via the usual non-rebreathing circuit. All the volunteers then breathed 100% O(2) through a commercial sequential gas delivery (SGD) circuit modified by attaching a reservoir to its exhalation port. Hypercarbia was induced by step reductions in oxygen flow to the SGD circuit. We monitored minute ventilation and end-tidal pCO(2) (ETpCO(2)) as a surrogate for its arterial value. RESULTS: Inhalation of 1.5 and 3% carbogen did not increase ETpCO(2) from baseline (40 +/- 1.5 mmHg); 5% carbogen increased ETpCO(2) to 45 +/- 1.6 mmHg (p < 0.001). With the SGD circuit, reducing O(2) flow to 4.3 +/- 0.7 l/min increased ETpCO(2) in all subjects from 41 +/- 2.0 mmHg (baseline) to 46 +/- 2.1 mmHg (p < 0.001). Voluntary hyperventilation reduced ETpCO(2) with 5% carbogen but not with SGD (p = 0.379). CONCLUSIONS: We confirm previous observations that carbogen inhalation does not result in a predictable rise in ETpCO(2) and suggest that a precise and stable target ETpCO(2) can instead be induced by simply controlling O(2) flow into a modified SGD circuit. We hoped that the reliable control of pCO(2) will enable studies that address first, the efficacy of raising ETpCO(2) on specific tumor blood flow, and eventually, its benefit as an adjuvant to radiotherapy.