Machine Learning for High-Fidelity 3D-Printed Organ Models in Surgery

Machine Learning Enabled Design and Optimization for 3D-Printing of High-Fidelity Presurgical Organ Models

Unmet Need: A better 3D printed presurgical organ model

Presurgical organ models enable students and surgeons to practice new or difficult techniques prior to ever setting foot in an operating room. However, the practice experience is most beneficial when the models are as lifelike as possible. Many models currently exist that vary along this spectrum – from practicing injection technique on an orange purchased from the grocery store to elaborate, 3D printed models that are based on the patient’s own imaging data. Recent advances in imaging, materials, and 3D printing techniques have improved our capacity to produce lifelike models, but the optimization process often overlooks variations caused by object geometry, material properties, and printing technique. Thus, producing a truly useful model remains labor-intensive, inefficient, and unable to keep up with current needs.

The Technology 3D-printing optimized by machine learning produces more lifelike models, reduces waste, and saves time.

To overcome the limitations of existing 3D printing methods, WSU inventors have developed a machine learning algorithm that quickly identifies optimal settings. The algorithm is capable of simultaneously addressing layer height, nozzle travel speed, and material dispensing pressure to produce a high-fidelity model. These models require fewer iterations, produce less waste, and save time compared to current techniques.

Applications:

Manufacturing precise and customized organ models for surgical planning and rehearsal
Medical device evaluation
Manufacturing organ models for training, educational, and student use

Advantages: