Orthopaedic Research Laboratory Alumni Council

Alumni Laboratories

Visit the research laboratories of the alumni of Dr. Woo.

Board of Directors

Richard Debski, Ph.D. President
Caroline Wang, M.S. Secretary
Jamie Pfaeffle, M.D., Ph.D. Treasurer
Doug Boardman, M.D.
Thay Lee, Ph.D.
Patrick McMahon, M.D.
Karen Ohland, M.S.
Christos Papgeorgiou, M.D.
Masataka Sakane, M.D.
Sven Scheffler, M.D.
Jennifer Wayne, Ph.D.

2004 Mrs. Ho-Tung Cheong Grant Recipient

Job N. Doornberg
Harvard Medical School

The Influence of Three-Dimensional Computed Tomography Scans on the Characterization and Treatment of Distal Humerus Fractures

Job Doornberg (left) with his advisor Dr. David Ring


I would like to thank the Orthopaedic Research Laboratory Alumni Council for the opportunity to work in the 3D-Lab of the Massachusetts General Hospital in Boston during the summer of 2004. It is a great honor to be the first Dutch Medical Student to receive the Mrs. Ho-Tung Cheong Grant. The grant was awarded to study the influence of three-dimensional computed tomography scans on the characterization and treatment of distal humerus fractures.

Study Abstract

Fractures of the distal humerus that involve the articular surface occur in specific patterns, the identification of which helps guide optimal treatment.   While evaluation of complex fractures with three-dimensional computed tomography reconstructions has proved useful for fractures at other sites, the influence of computed tomography on the characterization and classification of fractures of the distal humerus has not been studied. We hypothesized that three-dimensional (3D) reconstructions of computed tomography scans provide an improved view of the articular surface of the fractured distal humerus and can make identification of fracture fragments more accurate.  

Five independent observers evaluated 30 consecutive intra-articular fractures of the distal humerus for the presence of six fracture characteristics.   Fractures were also classified according to the AO Comprehensive Classification of Fractures and the Mehne/Matta classification.   Two rounds of evaluation were compared: plain radiographs and 2D-CT scans initially; versus radiographs, 2D-CT scans and 3D-CT reconstructions two weeks later.   These rounds were repeated after an interval of two weeks to evaluate intra-observer reliability.

We found that 3D-CT improved both the intra-observer and inter-observer reliability for the AO classification and the Mehne/Matta classification compared to reliability of 2D imaging.   3D-CT reconstructions improved the intra-observer agreement from moderate to substantial for all fracture characteristics compared to 2D-CT (average ? 2D   = 0.55, range 0.15 to 0.89 vs. average ? 3D   = 0.79, range 0.58 to 0.94).   Addition of 3D-CT improved categorical inter-observer agreement for a coronal fracture line, but not for any of the fracture characteristics (average ? 2D = 0.40, range 0.14 to 0.60 vs. average K 3D = 0.44, range 0.13 to 0.57).   Addition of 3D-CT images had limited influence on sensitivity, specificity and accuracy for recognition of fracture characteristics.   3D images also improved intra-observer reliability for choice of treatment (? 2D = 0.62 vs. ? 3D = 0.75), but there was no statistical significance.   Inter-observer agreement for treatment was only fair for both imaging modalities (? 2D = 0.24 versus ? 3D = 0.28)

We concluded that three-dimensional reconstructions of CT scans of distal humerus fractures improve intra-observer agreement, but have a limited influence on inter-observer agreement.   This suggests that the 3D reconstructions do help with characterization of the fracture, but that different observers see different things in the scans--most likely a reflection of the training, knowledge, and experience with these relatively uncommon injuries.

In other words

When surgeons use 3D reconstructions of the distal humerus they are more consistent in their characterization, classification and treatment of fractures (better intra-observer variability).   Agreement between different observers did not improve as much as we expected using 3D.   Fracture classification systems are only useful in decision-making and evaluation of outcome if there is agreement and consistency among observers.   Unfortunately, observers did not agree as much with each other as they agree with themselves.

I would like to thank the Orthopaedic Research Laboratory Alumni Council, as well as my mentors Professor Jesse Jupiter and Associated Professor David Ring for a great opportunity!   The original paper has recently been submitted for publication in the Journal of Bone and Joint Surgery, and if accepted we will gratefully acknowledge ORLAC's support!   Thanks!


© Orthopaedic Research Laboratory Alumni Council 2013. | Template Design by Download Website Templates