For a study, researchers created a new quantitative method for determining vesicoureteral reflux (VUR) based solely on voiding cystourethrograms (VCUG). Individual renal units were rated as low-grade (I-III) or high-grade (IV–V) using an online dataset of VCUGs. Using 3 simple user annotations, they created an image analysis and machine learning approach to automatically calculate and normalize the ureteropelvic junction (UPJ) width, ureterovesical junction (UVJ) width, and maximum ureter width and ureter tortuosity. A random forest classifier was trained to distinguish between low- and high-grade VUR. The institutional imaging repository was used to create an external validation cohort. Receiver-operating-characteristic and precision-recall curve analyses were used to assess discriminative capacity. To analyze the model’s predictions, investigators employed Shapley Additive exPlanations. Around 44 renal units were acquired from the institutional imaging repository, and 41 renal units were abstracted from an internet dataset. Between low- and high-grade VUR, significant changes in UVJ width, UPJ width, maximum ureter width, and tortuosity were identified. A random-forest classifier performed well on leave-one-out cross-validation, with an accuracy of 0.83, AUROC of 0.90, AUPRC of 0.89, and accuracy of 0.84, AUROC of 0.88, and AUPRC of 0.89 on external validation. The most critical parameter was tortuosity, followed by maximum ureter width, UVJ width, and UPJ width. The study group created qVUR (quantitative VUR), a web application that allowed users to input any VCUG for automated grading using the model described here. Even in non-standardized VCUG datasets, quantitative indicators could be used to grade VUR. In addition, they could use machine learning technologies to rate VUR.