Presented by: Hadar Tamir MD
Rabin Medical Center

Introduction

The treatment of choice for proximal ureteral stones is ureteroscopy or shock wave lithotripsy. When discussing primary ureteroscopy (pURS) with a patient it is important to recognize the scenario of a tight ureter requiring ureteral stent placement. To our knowledge, a comprehensive model that utilizes potential predictors to determine the presence of a tight ureter has not yet been developed.

Materials

We retrospectively reviewed 168 patients undergoing pURS for proximal urolithiasis from 2014 to 2023 in RMC. The primary outcome was the completion rate of the URS. We collected patients’ demographic, medical history, and surgical factors from medical records. We trained 3 classical learning multi-feature machine learning classifiers to predict pURS success relying on patient's characteristics.

Results

,

To develop predictive models, we employed three machine learning classifiers: naive Bayes, decision trees, and logistic regression. These models achieved accuracy rates of 73.5%, 88.2%, and 79.4% respectively. Interestingly, certain features that were initially deemed insignificant in the single-feature analysis played a significant role in predicting the outcome of pURS. Factors that were included in the most accurate predictive model were Cr, age, BMI, largest ureteral stone diameter, Hb, and previous ipsilateral procedure.

In a multivariable analysis, the highest rate of success was found in male gender with previous procedure and stone expulsion (p=0.024).

Conclusion

pURS has shown a high success rate for treating proximal ureteral stones. Utilizing machine learning techniques, clinicians can accurately assess the likelihood of completing the treatment in a single procedure based on patient features. This has the potential to significantly impact pre-procedure discussions with patients, providing valuable insights into the chances of requiring a two-procedure treatment approach.

Funding

None.

Co-Authors

Sagi A. Shpitzer, MD
Rabin Medical Center

Shayel Bercovich, MD
Rabin Medical Center

Abd E. Darawsha, MD
Rabin Medical Center

Ron Gilad, MD
Rabin Medical Center

Yaron Ehrlich, MD
Rabin Medical Center

David Lifshitz, PHD
Rabin Medical Center