Automated quantification of penile curvature using artificial intelligence

Male Infertility: New Developments, Current Challenges, and Future Directions

There have been many significant scientific advances in the diagnostics and treatment modalities in the field of male infertility in recent decades. Examples of these include assisted reproductive technologies, sperm selection techniques for intracytoplasmic sperm injection, surgical procedures for sperm retrieval, and novel tests of sperm function. However, there is certainly a need for new developments in this field. In this review, we discuss advances in the management of male infertility, such as seminal oxidative stress testing, sperm DNA fragmentation testing, genetic and epigenetic tests, genetic manipulations, artificial intelligence, personalized medicine, and telemedicine. The role of the reproductive urologist will continue to expand in future years to address different topzics related to diverse questions and controversies of pathophysiology, diagnosis, and therapy of male infertility, training researchers and physicians in medical and scientific research in reproductive urology/andrology, and further development of andrology as an independent specialty.

Unaided visual assessment of ventral curvature during hypospadias repair is inferior to objective measurement using app goniometry

OBJECTIVE

Failure to perform artificial erection or objectively assess ventral curvature (VC) during primary hypospadias repair is an important reason for residual/ recurrent chordee. The present study compares the accuracy of unaided visual inspection (UVI) with objective VC assessment using smartphone application (app) goniometry.

METHODS

All patients who underwent primary hypospadias repair between January 2021 and September 2022 were included. Assistant surgeons were asked to grade the degree of VC on UVI (after degloving and an artificial erection test) into: none, mild (<30 degree), severe(>30 degree). Lateral profile photograph was taken and angle measurement was performed on an android mobile application (Angulus). Correlation was performed with both methods of assessment.

RESULTS

During this period a total of 210 patients were analyzed; VC was noted in 40/138 (29%) cases of distal and in 62/72 (86%) cases of proximal hypospadias. Erroneous visual inspection was noted in 41/210 (20%; 95% CI 14-25%) on UVI (15 erroneously marked none while 26 marked mild). Among those found to have chordee, UVI assessed 39/82 (47%) as severe while app goniometry assessed 65/97 (67%) as severe. There was significant relative risk of labelling severe chordee as a mild one by UVI: 1.4 (95%CI 1-1.8; p=0.01).

CONCLUSIONS

UVI was erroneous in 20% of cases. UVI was less accurate in differentiating severe chordee from mild one. In 60% patients UVI alone could have led to erroneous VC assessment and thus wrong selection of technique. Further studies are required to validate our findings and standardize VC measurement using an app goniometry.

Automated measurement of penile curvature using deep learning-based novel quantification method

Objective

Develop a reliable, automated deep learning-based method for accurate measurement of penile curvature (PC) using 2-dimensional images.

Materials and methods

A set of nine 3D-printed models was used to generate a batch of 913 images of penile curvature (PC) with varying configurations (curvature range 18° to 86°). The penile region was initially localized and cropped using a YOLOv5 model, after which the shaft area was extracted using a UNet-based segmentation model. The penile shaft was then divided into three distinct predefined regions: the distal zone, curvature zone, and proximal zone. To measure PC, we identified four distinct locations on the shaft that reflected the mid-axes of proximal and distal segments, then trained an HRNet model to predict these landmarks and calculate curvature angle in both the 3D-printed models and masked segmented images derived from these. Finally, the optimized HRNet model was applied to quantify PC in medical images of real human patients and the accuracy of this novel method was determined.

Results

We obtained a mean absolute error (MAE) of angle measurement <5° for both penile model images and their derivative masks. For real patient images, AI prediction varied between 1.7° (for cases of ∼30° PC) and approximately 6° (for cases of 70° PC) compared with assessment by a clinical expert.

Discussion

This study demonstrates a novel approach to the automated, accurate measurement of PC that could significantly improve patient assessment by surgeons and hypospadiology researchers. This method may overcome current limitations encountered when applying conventional methods of measuring arc-type PC.