For Zoom link, email xinyue.han.th@dartmouth.edu.

"Intraoperative quantification of bone perfusion in lower extremity injury surgery"

Abstract

Orthopaedic surgery is one of the most common surgical categories. In particular, lower extremity injuries sustained from trauma can be complex and life-threatening injuries that are addressed through orthopaedic trauma surgery. Timely evaluation and surgical debridement following lower extremity injury is essential, because devitalized bones and tissues will result in high surgical site infection (SSI) rates. However, the current clinical judgment of what constitutes “devitalized tissue” io subjective and dependent on surgeon experience, so it is necessary to develop imaging techniques for guiding surgical debridement, in order to control infection rates and to improve patient outcome.

In this thesis work, computational models of fluorescence-guided debridement in lower extremity injury surgery will be developed, by quantifying bone perfusion intraoperatively using Dynamic contrast-enhanced fluorescence imaging (DCE-FI) system. Perfusion is an important factor of tissue viability, and therefore quantifying perfusion is essential for fluorescence-guided debridement. In Chapters 3-5 of this thesis, we explore the performance of DCE-FI in quantifying perfusion by pre-clinical and clinical studies: Our first in-human use of dynamic contrast-enhanced texture analysis for orthopaedic trauma classification suggests that spatiotemporal features from DCE-FI can classify bone perfusion intraoperatively with high accuracy and sensitivity; to validate bone perfusion measurements obtained by DCE-FI, we describe a modified fluorescent microsphere quantification technique using cryomacrotome (mQUIC). This technique can measure bone perfusion in periosteal and endosteal separately; First-pass kinetic parameters and arterial input functions have also been investigated for characterization of perfusion changes during lower limb amputation surgery. In conclusion, pharmacokinetic and spatiotemporal patterns of dynamic contrast-enhanced imaging show great potential for quantifying bone perfusion.

In future studies, we will expand the proposed DCE-FI models into predicting bone infection in lower extremity surgery: Firstly, we will develop pre-clinical rodent contaminated fracture model to correlate DCE-FI with infection risk, validate the result by mQUIC, and compare with multi-modality scanning; Secondly, we will establish clinical machine learning predictive model on open fracture surgery, where pixel-scaled prediction on infection risk will be accomplished. The studies will decrease surgical site infection risk and improve successful rates of lower extremity injury surgery.

Thesis Committee

• Jonathan Elliott (chair)
• Kimberley Samkoe
• Vikrant Vaze
• I. Leah Gitajn

Contact

For more information, contact Theresa Fuller at theresa.d.fuller@dartmouth.edu.