Paul Michael Bunch MD, Presenter: Nothing to Disclose

Cree Michael Gaskin MD, Abstract Co-Author: Research Consultant, Johnson & Johnson Author with royalties, Oxford University Press Author with royalties, Thieme Medical Publishers, Inc

Talissa A. Altes MD, Abstract Co-Author: Research Grant, Vertex Pharmaceuticals Incorporated

James Patrie MS, Abstract Co-Author: Nothing to Disclose

Joan McIlhenny MD, Abstract Co-Author: Nothing to Disclose

Greulich and Pyle's Radiographic Atlas of Skeletal Development of the Hand and Wrist (G&P) has long been a reference standard for determination of skeletal maturity. Skeletal Development of the Hand and Wrist: Digital Bone Age Companion (DBAC) is commercially available bone age software which is based upon G&P. We sought to determine performance of the digital atlas (DBAC) relative to the standard method (G&P).

For this IRB approved study, we identified 96 bone age studies (52% male, 48% female) which were interpreted for clinical purposes by two experienced faculty pediatric radiologists using the G&P method. The study group comprised 65.9% normal, 17.7% advanced, and 16.7% delayed skeletal maturity. The same radiologists were blinded and asked to re-read their own studies randomized to G&P method (n=48) versus DBAC method (n=48). They recorded interpretation time, bone age standard selection, and impression (normal vs. delayed vs. advanced). When a new read matched the old read, this was considered a gold standard result. When a new read disagreed with the old read, the radiologists performed an unblinded third read using G&P to determine a gold standard. The data were analyzed by use of binomial exact tests and McNemar tests.

The mean interpretation time for DBAC (26.1 seconds) was significantly less than for G&P (31.9 seconds) (P=0.012). There was 100% agreement between DBAC reads and the gold standard for both bone age determination and final impression. There was 95.8% agreement between G&P re-reads and the gold standard for both bone age determination and final impression. There was no significant difference between agreement with gold standard for DBAC compared with G&P for both bone age (P=0.495) and final impression (P=0.495).

There is no difference in bone age determination between DBAC and G&P methods. Radiologists experienced with G&P interpret studies faster with DBAC than with G&P.

An available digital atlas yields similar results to the existing standard for bone age determination. Radiologists may read bone age studies faster with the digital atlas.

Bunch, P, Gaskin, C, Altes, T, Patrie, J, McIlhenny, J, Can Commercially Available Bone Age Software Outperform the Greulich & Pyle Book in the Determination of Skeletal Age by Experienced Readers?.  Radiological Society of North America 2012 Scientific Assembly and Annual Meeting, November 25 - November 30, 2012 ,Chicago IL. http://archive.rsna.org/2012/12022579.html