DEVELOPMENT OF A RENAL TUMOR MIMIC MODEL FOR LEARNING MINIMALLY INVASIVE NEPHRON SPARING SURGICAL TECHNIQUES

PURPOSE: New laparoscopic partial nephrectomy and tumor ablation techniques are continuously being developed and evaluated in large animal models. However, to our knowledge no reliable renal tumor model exists to evaluate procedure efficacy. We developed and assessed the reliability of a tumor mimic model to serve as a training tool for minimally invasive kidney surgery. MATERIALS AND METHODS: An agarose based model was created using a mixture of 3% agarose, 3% cellulose, 7% glycerol and 0.05% methylene blue. It is liquid when heated, but solidifies after cooling to physiological temperatures. The agar was injected (0.7 cc) into porcine renal parenchyma to make endophytic or completely intraparenchymal lesions. Three-dimensional ultrasound images of the lesions were obtained during initial development to ensure spherical lesions. A group of 20 lesions was initially placed in an ex vivo setting to assess size consistency and define baseline impedance characteristics. An additional 20 tumor mimics each were established in a laparoscopic model in a laparoscopic box trainer and an in vivo laparoscopic model. They were ablated with a temperature based radio frequency generator to assess impedance characteristics but the efficacy of ablation was not assessed. The in vivo model consisted of placing the agar lesion percutaneously under direct laparoscopic vision. RESULTS: The agarose mixture was easily injected and readily visible on ultrasound as hyperechoic distinct lesions. Lesions had a mean size of 10.8 +/- SD 1.3 mm on ultrasound and 10.9 +/- 1.2 mm grossly. The impedance of normal renal parenchyma and unablated lesions was similar. Mean lesion size in the ex vivo model after radio frequency ablation was 9.8 +/- 2.0 mm on ultrasound, which was similar to the gross mean lesion size of 9.7 +/- 1.0 mm. Similar results were obtained for the in vivo model with a mean size of 10.1 +/- 2.1 and 10.4 +/- 1.5 mm, respectively. The lesions were easily identified grossly as blue solid lesions that replaced renal parenchyma. CONCLUSIONS: The described renal tumor mimic model reproducibly creates ex vivo and in vivo porcine kidney lesions. Lesion size and impedance do not change with the application of radio frequency energy. This model should be a valuable adjunct in the development, assessment and teaching of novel, nephron sparing, minimally invasive surgical techniques.