Improving Suturing Skills for Surgical Residents and Advancing Prosthesis Control for Amputees.

Abstract

Proper suturing technique is one of the most important skills a surgical resident should acquire. However, current methods for teaching it rely on subjective performance evaluations. An instrumented training apparatus for abdominal closure could be used to define objective assessments that directly relate to closure quality. I identify a synthetic material that models abdominal fascia using porcine and cadaveric data and design a means to mount the material so that it mimics abdominal closure. Digital images are used to quantify material deformations and provide real-time objective measures regarding the effect of suture placement and tension in the abdominal tissue. In parallel, I develop a finite element model of abdominal fascia and its closure with suture to deduce stresses in the material and forces in the sutures. I find that despite uniform suture spacing, the forces in suture are unevenly distributed along the closure. These findings motivate the development of a surgical learning tool that objectively relays information about suture placement and tension.

In a second body of work, I address the development of a novel interface between an amputee’s peripheral nervous system and a motorized prosthetic device. Conventional myoelectric control cannot produce a sufficient number of independent signals for actuation of modern computerized upper limb prostheses. A compact construct involving grafted muscle surgically prepared at the end of a transected peripheral nerve is envisioned for transducing a nervous signal with fine specificity and sensitivity. Up to 20 such constructs can be prepared in a human arm, and epimysial electrodes on each construct can be used to relay signals encoding 20 independent channels of motor intent. I develop a means of evaluating this construct in awake rats, and demonstrate that the transduced signals suffer minimal crosstalk and are correlated with gait. A decoder is able to reconstruct data produced by motion tracking, and I show that adjacent constructs placed proximal to one another provide the same signals as anatomically intact muscle-nerve antagonist-pair analogs. The correlation between the signals transduced, the walking kinematics, and analogous out of phase activation obtained from adjacent constructs indicates that this technology holds promise for human translation.