Using Artificial Intelligence to Predict Surgical Shunts in Men with Ischemic Priapism

Area of Science:

• Urology
• Artificial Intelligence in Medicine
• Emergency Medicine

Background:

• Ischemic priapism is a critical condition demanding immediate treatment to prevent permanent erectile dysfunction.
• Predicting the need for surgical intervention, such as shunting, remains challenging.

Purpose of the Study:

• To identify factors influencing the escalation to surgical shunting in ischemic priapism cases.
• To develop and validate machine learning models for predicting which patients require shunting.

Main Methods:

• Retrospective analysis of 334 ischemic priapism cases from a single county hospital (2010-2019).
• Data collection included demographics, etiology, duration, and interventions.
• Machine learning models were trained and tested to predict the need for surgical shunting.

Main Results:

• Most cases (78%) resolved with non-surgical methods like phenylephrine injection or aspiration.
• Surgical shunting was necessary in 10% of patients.
• Longer duration of priapism before intervention (median 48 vs 7 hours) and non-sickle cell etiology were associated with increased shunting necessity.

Conclusions:

• Extended duration of ischemic priapism correlates with a higher likelihood of requiring shunting, though non-surgical options remain effective for up to 36 hours.
• Sickle cell disease appears to be a protective factor against the need for shunting.
• Developed AI models achieved 87.2% accuracy in predicting shunting needs, with an online calculator available.