Bone age assessment is a task performed daily by paediatricians in hospitals worldwide Bone age assessment is a task performed daily by paediatricians in hospitals worldwide

The main reasons for bone age assessment to be performed are to: 

  • Monitor skeletal development and therefore diagnose growth disorders. 
  • Predict final adult height.
  • To verify age claims made by asylum seekers who have invalid age documents.

Manually predicting bone age from radiographs is a difficult and time consuming task. The aim of this project is to produce an automated bone age assessment system. 

figure 1 bone agingThis procedure is undertaken by obtaining a radiograph of the patient's non-dominant hand. An examination of the skeletal development of the hand is then undertaken using one of two methods: Greulich and Pyle or Tanner and Whitehouse. These are found to be subjective and time-consuming, respectively, hence the need for an automated solution.

We propose the Automated Skeletal Maturity Assessment algorithm. This consists of 6 distinct stages:

a. Hand Outline Segmentation

b. Hand Outline Verification

c. Bone Segmentation

d. Feature Extraction

e. Bone Segmentation Verification

f. Age Prediction

Figure 1 summarises these stages. Using this modular approach allows us to investigate and make improvements to each stage individually.

 

 

Code and Data

The paper Automated Skeletal Maturity Assessment (ASMA): Modelling Bone Age from Hand Radiographs is currently under review.

ASMA zip file Publication [4]

Please send any enquiries to Luke Davis

References

Research Team

Mr. Luke Davis, Dr. Anthony Bagnall, Dr. Barry Theobald

Dr. Andoni Toms (Consultant Radiologist - Norwich Radiology Academy, Norfolk and Norwich Hospital)