2017 ISAKOS Biennial Congress ePoster #412


Osteochondritis Dissecans of the Capitellum: Lesion Size and Pattern Analysis Using Quantitative 3-Dimensional Computed Tomography and Mapping Technique

Rens Bexkens, MD, Amsterdam, Noord-Holland NETHERLANDS
Jacobien Oosterhoff, BSc, Amsterdam NETHERLANDS
Tsung-Yuan Tsai, PhD, Boston, MA UNITED STATES
Job Doornberg, MD, Amsterdam NETHERLANDS
Michel Van Den Bekerom, MD, Amsterdam NETHERLANDS
Denise Eygendaal, Prof., MD, PhD, Breda NETHERLANDS
Luke S. Oh, MD, Orlando, FL UNITED STATES

Massachusetts General Hospital, Boston, Massachusetts, UNITED STATES

FDA Status Not Applicable


State-of-the-art quantitative three-dimensional computed tomography and mapping are reliable techniques that provide greater understanding of lesion morphology and location in patients with capitellar osteochondritis dissecans



Surgical decision-making for capitellar osteochondritis dissecans (OCD) is influenced by lesion size, location, and stability. The goals of this study were to evaluate the reliability of a quantitative 3-dimensional computed tomography (Q3DCT) technique for measurement of capitellar osteochondritis dissecans (OCD) surface area, analyze OCD distribution using mapping technique, and investigate associations between Q3DCT lesion quantification, demographics and clinical examination.


We identified patients with capitellar OCD who presented to our orthopaedic sports medicine practice between 2001 and 2016 who had a preoperative CT scan performed (slice thickness =1.25 mm). This resulted in 17 patients with a median age of 15 years (range=12-23). Polygon 3D models were reconstructed after marking osseous structures in three planes. Surface areas of the OCD lesion as well as the capitellum were measured. Observer agreement was assessed with the use of Intraclass Correlation Coefficient (ICC). Heat maps were created to visualize OCD distribution.


Measurements of OCD surface area demonstrated almost perfect intra-observer (ICC=0.99; CI=0.98-0.99) and inter-observer agreement (ICC=0.93; CI=0.86-0.97). Measurements of capitellar surface area also showed almost perfect intra-observer (ICC=0.97; CI=0.91-0.99) and inter-observer agreement (ICC=0.86; CI=0.46-0.96). The median OCD surface area was 101 mm2 (range=49-217). Based on OCD heat mapping, the posterolateral zone of the capitellum was most frequently affected, located 56° anteriorly relative to the humeral shaft. The lateral capitellar wall was involved in four OCDs (24%). OCDs in which the lateral wall was involved were associated with larger lesion size (p=.041), longer duration of symptoms (p=.030), and worse elbow extension (p=.013).


The ability to quantify capitellar OCD surface area and lesion location in a reliable manner using Q3DCT and mapping technique should be considered when detailed knowledge of lesion size and location is desired.