A parameterized model for mean urinary inflow rate and its preliminary application in radiotherapy for cervical cancer.

Forty-nine patients with stage II cervical cancer were included to investigate the changes in bladder volume in response to different approaches to maintaining consistent bladder filling. The impacts of age (P ), water consumption (P ), and body mass index (BMI, P ) on the mean urinary inflow rate (v ) were analysed. The bladder volume (BV) increased linearly over time. A large variation in v among individuals was observed, ranging from 0.19 to 5.13 ml/min. The v was correlated with P (R = -0.53, p = 0.01) and P (R = 0.84, p = 0.00), and no correlation between v and P was found (p > 0.05). Therefore, v could be parameterized using two methods: multivariable linear regression and iterative fitting. There was no statistically significant difference between the two methods. The model accuracy was successfully assessed with several validation tests for patients with good compliance (79.2% of all patients), and the proportion of radiotherapy (RT) fractions with zero wait time (one ultrasound (US) scan) increased from 6.5% to 41.2%. The optimal US scanning number and RT time could be provided using this model. This adaptive RT approach could reduce patient discomfort caused by holding onto urine and reduce technician labour as well as cost.