[Artificial intelligence: to a better predictive strategy for testicular sperm extraction outcome in azoospermia]

Azoospermia, defined as the absence of sperm in the semen, is found in 10-15 % of infertile patients. Two-thirds of these cases are caused by impaired spermatogenesis, known as non-obstructive azoospermia (NOA). In this context, surgical sperm extraction using testicular sperm extraction (TESE) is the best option and can be offered to patients as part of fertility preservation, or to benefit from in vitro fertilization. The aim of the preoperative assessment is to identify the cause of NOA and evaluate the status of spermatogenesis. Its capacity to predict TESE success remains limited. As a result, no objective and reliable criteria are currently available to guide professionals on the chances of success and enable them to correctly assess the benefit-risk balance of this procedure. Artificial intelligence (AI), a field of research that has been rapidly expanding in recent years, has the potential to revolutionize medicine by making it more predictive and personalized. The aim of this review is to introduce AI and its key concepts, and then to examine the current state of research into predicting the success of TESE.

A Machine Learning Approach for the Prediction of Testicular Sperm Extraction in Nonobstructive Azoospermia: Algorithm Development and Validation Study

BACKGROUND

Testicular sperm extraction (TESE) is an essential therapeutic tool for the management of male infertility. However, it is an invasive procedure with a success rate up to 50%. To date, no model based on clinical and laboratory parameters is sufficiently powerful to accurately predict the success of sperm retrieval in TESE.

OBJECTIVE

The aim of this study is to compare a wide range of predictive models under similar conditions for TESE outcomes in patients with nonobstructive azoospermia (NOA) to identify the correct mathematical approach to apply, most appropriate study size, and relevance of the input biomarkers.

METHODS

We analyzed 201 patients who underwent TESE at Tenon Hospital (Assistance Publique-Hôpitaux de Paris, Sorbonne University, Paris), distributed in a retrospective training cohort of 175 patients (January 2012 to April 2021) and a prospective testing cohort (May 2021 to December 2021) of 26 patients. Preoperative data (according to the French standard exploration of male infertility, 16 variables) including urogenital history, hormonal data, genetic data, and TESE outcomes (representing the target variable) were collected. A TESE was considered positive if we obtained sufficient spermatozoa for intracytoplasmic sperm injection. After preprocessing the raw data, 8 machine learning (ML) models were trained and optimized on the retrospective training cohort data set: The hyperparameter tuning was performed by random search. Finally, the prospective testing cohort data set was used for the model evaluation. The metrics used to evaluate and compare the models were the following: sensitivity, specificity, area under the receiver operating characteristic curve (AUC-ROC), and accuracy. The importance of each variable in the model was assessed using the permutation feature importance technique, and the optimal number of patients to include in the study was assessed using the learning curve.

RESULTS

The ensemble models, based on decision trees, showed the best performance, especially the random forest model, which yielded the following results: AUC=0.90, sensitivity=100%, and specificity=69.2%. Furthermore, a study size of 120 patients seemed sufficient to properly exploit the preoperative data in the modeling process, since increasing the number of patients beyond 120 during model training did not bring any performance improvement. Furthermore, inhibin B and a history of varicoceles exhibited the highest predictive capacity.

CONCLUSIONS

An ML algorithm based on an appropriate approach can predict successful sperm retrieval in men with NOA undergoing TESE, with promising performance. However, although this study is consistent with the first step of this process, a subsequent formal prospective multicentric validation study should be undertaken before any clinical applications. As future work, we consider the use of recent and clinically relevant data sets (including seminal plasma biomarkers, especially noncoding RNAs, as markers of residual spermatogenesis in NOA patients) to improve our results even more.

P-057 Machine learning-based prediction of testicular sperm extraction: comparison of different preprocessing and models, required sample size and relevance of input biomarkers

Study question

Can advanced machine learning applied to the preoperative assessment predict the testicular sperm extraction outcome in azoospermic context and how many patients are required?

Summary answer

Despite encouraging results (AUC = 92.0%, sensitivity = 83.9% and specificity = 84.2%), integrating new biomarkers would probably be more relevant than enrolling additional patients.

What is known already

Testicular sperm extraction (TESE) is an essential therapeutic tool for the male infertility management and is often the “last hope” before gamete donation for these patients. However, it is an invasive procedure and is successful in up to 50%. Until now, no model is sufficiently powerful to accurately predict the success of sperm retrieval in TESE. Among the few models already developed, the findings are highly disparate despite having common input data (preoperative assessment). Moreover, only few types of machine learning models and procedures have been investigated. Performances were mostly capped despite the inclusion sometimes of more than 1000 patients.

Study design, size, duration

Data of 175 patients who underwent TESE between 2012 and 2021 were retrospectively analyzed. The performances of a wide range of preprocessing methods and machine learning models (state-of-the-art methods in machine learning) we explored, evaluated, and compared. The objective was to predict the presence or absence of spermatozoa, using 17 parameters (clinical, hormonal, genetic, history) from the preoperative assessment. The study protocol was approved by a local ethics committee (IRB CER-2021-041).

Participants/materials, setting, methods

After data preprocessing (standardization…), Machine Learning models (Bayesian Naive Classification, logistic regression, k-nearest neighbor classifier, support vector machine, random forests, GradientBoosting and XGBoost) and Deep Learning models were tested. The validation procedure consisted of splitting the dataset into a training set and test set. Beyond the standard metrics (sensitivity, specificity, AUC-ROC), the identification of the most relevant variables and the learning curve to determine the optimal patient number to be included were performed.

Main results and the role of chance

At least one live spermatozoon was found in the testicular tissue of 104 (59.4%) patients (positive TESE) out of 175. The best performing model (Random Forest with appropriate preprocessing) obtained the following results on the test set: AUC = 92.0%, sensitivity = 83.9% and specificity = 84.2%, leading to an efficient tool, which gives additional and more relevant information than the different variables taken separately. Inhibin B, FSH and history of cryptorchidism were the variables with the most discriminating power. However, a plateau in the model performance was observed (beyond 110 patients), whatever the approach or the preprocessing used. A trend curve shows that beyond 110 patients, no improvement can be observed and cast doubt about the power of the traditional preoperative parameters assessed before TESE. The classic preoperative assessment can probably not fully predict the TESE outcomes. Further work is needed to be enhance with new hypothesis and the use of new biomarkers to be integrated into the models.

Limitations, reasons for caution

The main limitation was the monocentric design and the use of retrospective data.

Wider implications of the findings

Machine learning models can provide the basis for an enhanced decision support system tool in the context of azoospermia. Indefinitely increasing the number of participants is not likely to be the solution: further hypotheses and biomarkers integration into the models will probably be necessary to improve performance.

Trial registration number

not applicable

Long-term outcomes after penile prosthesis placement for the Management of Erectile Dysfunction: a single-Centre experience

Background

Penile prothesis (PP) is the gold-standard treatment of drug-refractory erectile dysfunction (ED). While postoperative outcomes have been widely described in the literature, there are few data about patient satisfaction and intraoperative events. We aimed to assess long-term patient satisfaction and perioperative outcomes after PP implantation in a single-centre cohort of unselected patients using validated scales.

Results

A total of 130 patients received a PP (median age: 62.5 years [IQR: 58–69]; median International Index of Erectile Function (IEEF-5) score: 6 [IQR: 5–7]). Median follow-up was 6.3 years [IQR: 4–9.4]. Thirty-two (24.6%) patients underwent surgical revision, of which 20 were PP removals (15.4%). Global PP survival rate was 84.6% and previous PP placement was a risk factor for PP removal (p = 0.02). There were six (4.6%) non-life-threatening intraoperative events including two which resulted in non-placement of a PP (1.5%). EAUiaic grade was 0 for 124 procedures (95.4%), 1 for four procedures (3.1%) and 2 for two procedures (1.5%). Of patients who still had their PP at the end of the study, 91 (80.5%) expressed satisfaction.

Conclusions

PP implantation is a last-resort treatment for ED with a satisfactory outcome. PPs are well accepted by patients.

RIDME distance measurements using Gd(iii) tags with a narrow central transition

Methods based on pulse electron paramagnetic resonance allow measurement of the electron–electron dipolar coupling between two high-spin labels.

Solid and liquid phase 59Co NMR studies of cobalamins and their derivatives

We describe the application of 59Co NMR to the study of naturally occurring cobalamins. Targets of these investigations included vitamin B12, the B12 coenzyme, methylcobalamin, and dicyanocobyrinic acid heptamethylester. These measurements were carried out on solutions and powders of different origins, and repeated at a variety of magnetic field strengths. Particularly informative were the solid-state central transition NMR spectra, which when combined with numerical line shape analyses provided a clear description of the cobalt coupling parameters. These parameters showed a high sensitivity to the type of ligands attached to the metal and to the crystallization history of the sample. 59Co NMR determinations also were carried out on synthetic cobaloximes possessing alkyl, cyanide, aquo, and nitrogenated axial groups, substituents that paralleled the coordination of the natural compounds. These analogs displayed coupling anisotropies comparable to those of the cobalamins, as well as systematic up-field shifts that can be rationalized in terms of their stronger binding affinity to the cobalt atom. Cobaloximes also displayed a higher regularity in the relative orientations of their quadrupole and shielding coupling tensors, reflecting a higher symmetry in their in-plane coordination. For the cobalamines, poor correlations were observed between the values measured for the quadrupole couplings in the solid and the line widths observed in the corresponding solution 59Co NMR resonances.

Interactions between natural polyamines and tRNA: an 15N NMR analysis

15N NMR spectroscopy was used to explore the interactions between natural polyamines and Escherichia coli tRNA. It was found that when tRNA is added to solutions of 15N-labeled spermine or spermidine, there is a considerable decrease in the relative heights of the -NH(2+)--resonances with respect to the signals arising from the -NH3+ groups. The presence of tRNA was also found to reduce the longitudinal relaxation times T1 of the nitrogens, mainly those of the -NH(2+)- groups. The longitudinal relaxation times of the nitrogens were used to characterize the temperature dependence of the binding, and they allowed us to calculate the activation energies that determine the correlation times of amino groups in the presence of tRNA. Both the thermodynamic and the relaxation results indicate that (i) spermine binds more strongly to tRNA than spermidine does and (ii) within each of these molecules the -NH(2+)- groups bind more strongly to tRNA than the more electropositive -NH3+ moieties. This specificity suggests that the interaction between polyamines and tRNA cannot be described exclusively in terms of electrostatic forces and that other interactions (most likely, hydrogen bonding) are very important for establishing the polyamine-tRNA link. Some of the factors that may conspire against the binding of -NH3+ groups to tRNA are briefly discussed.