IB Lab PANDATM

Bone age assessment for prognosing the height of children

With a radiological evaluation of the bone age, physicians can predict conditions that affect the growth of children. Manual estimations by comparing digital radiographs with references in the Greulich & Pyle atlas is tedious and suffer from a high degree of inter-rater variability.
PANDA
provides a swift automated method to estimate bone age, as well as monitor child growth and development.

IB Lab KOALATM

IB Lab LAMATM

IB Lab PANDATM

Bone age assessment for prognosing the height of children

5 radiological findings, measurements and results including:
Pediatric bone age according to Greulich & Pyle, natural standard deviation, delayed / advanced bone age patient status,  growth potential and height estimation according to Bayley & Pinneau

This product is CE-certified.

Not for clinical use in the USA.

Situation

Radiographic skeletal age (or bone age) assessment is widely used by pediatricians and endocrinologists as part of the clinical assessment of skeletal maturity to diagnose growth disorders. This measurement is the basis of decision-making for therapies delaying or preventing physical growth and development. Measurement of hand radiographs for assessing bone age according to the Greulich & Pyle method is tedious taking up to 9 minutes [1]. Using lookup tables and comparing pictures manually with the Greulich & Pyle reference atlas requires high focus and precision, often resulting in a significant intra- and inter-reader variability [2].

Product

IB Lab’s diagnostic support tool PANDA uses deep learning technology to report bone age based on the Greulich & Pyle scale and presents the results within less than 5 seconds saving valuable time. PANDA’s automated bone age measurement according to Greulich & Pyle is precise to ±4.3 months [3]. The derived adult height estimation according to Bailey and Pineau is precise to ±2.5 cm [4]. This provides accurate data for decision making, also to the non-expert. Standardized measurements and reporting schemes facilitate monitoring of treatment progress.

PANDA highlights relevant clinical findings by applying latest international medical standards to enable timely and accurate decision making. The findings are summarized in a visual output report, attached to the original x-ray image and saved automatically in the PACS system. The AI-results are fed as text into your pre-defined RIS-template for accelerated reporting. The AI facilitates monitoring of disease progression by facilitating comparison of radiographic disease parameters over time. See how it works.

Benefits

  • Saves time: less than 5 seconds for determining pediatric bone age according to Greulich&Pyle
  • Enhances the consistency in reading and reporting of hand radiographs for bone age and height estimation
  • Facilitates monitoring and forecasting of physicians’ therapeutic success

Training & Validation

  • Trained on over 12,000 hand radiographs from two institutions in the US (Lucile Packard Children’s Hospital at Stanford University and Children’s Hospital Colorado) [5]
  • PANDA uses an ensemble of decision models to report bone age based on the Greulich & Pyle atlas
  • Standard derivation for a given chronological age is determined by rounding down to the next age in the Brush table for the appropriate sex [6]
  • Validated on the Digital Hand Atlas dataset consisting of 1384 DICOM left hand radiographs of normal children between 0 and 228 months of age, from different ethnicities (African, Asian, Caucasian, Hispanic). The data was collected from the Children’s Hospital of Los Angeles between the late 90s and until late 2000’s.

Note: product description above is based on the product version V1.04

Example Cases

Expert opinion without AI measurements

AP view of a left hand with visible growth plates of the distal forearm and finger rays I-V. All carpal bones are fully visible.

Automated AI measurements

PANDA’s overlay summarizes the relevant measurements and reference values to provide a starting point for the radiologists assessment

Automated AI report

PANDA assesses the patient’s bone age and suggests it as being within the range of normal skeletal maturity, based on reference standards. Growth potential achieved and projected adult height are also calculated.

Expert opinion without AI measurements

AP view of a left hand with visible growth plates of the distal forearm and finger rays I-V. All carpal bones are fully visible.

Automated AI measurements

PANDA’s toggleable overlay optimizes reading and provides consistent measurements.

Automated AI report

PANDA determines abnormal bone age findings by comparing to natural standard deviations accepted in the radiologic community. On the report, graphical elements are used to clearly highlight abnormal findings.

References

[1] D. G. King, D. M. Steventon, M. P. O’Sullivan, A. M. Cook, V. P. L. Hornsby, I. G. Jefferson, and P. R. King: Reproducibility of bone ages when performed by radiology registrars: an audit of Tanner and Whitehouse II versus Greulich and Pyle methods, The British Institute of Radiology, 2014.

[2] S.Serinellia, V.Panettab, P. Pasqualettib, D. Marchetti: Accuracy of three age determination X-ray methods on the left hand-wrist: A systematic review and meta-analysis, Legal Medicine, 2011. 

[3] Halabi, S. S., Prevedello, L. M., Kalpathy-Cramer, J., Mamonov, A. B., Bilbily, A., Cicero, M., … Flanders, A. E.: The RSNA Pediatric Bone Age Machine Learning Challenge, Radiology, 290(2), 2018. 498–503.

[4] Gaskin, C. M., Kahn, S. L., Bertozzi, J. C., & Bunch, P. M.: Skeletal Development of the Hand and Wrist. Oxford University Press, 2011.

[5] IB Lab Clinical Evaluation Study

[6] IB Lab Clinical Evaluation Study

Literature

  • Greulich, W. W., Pyle, S. I.: Radiographic atlas of skeletal development of the hand and wrist (2nd edition), Stanford Univ. Press, 1959. 
  • Gaskin, C. M., Kahn, S. L., Bertozzi, J. C., & Bunch, P. M.: Skeletal Development of the Hand and Wrist, Oxford University Press, 2011. 
  • Bayley, N., Pinneau, S. R.: Tables for predicting adult height from skeletal age: revised for use with the Greulich-Pyle hand standards, The Journal of Pediatrics, 40(4), 1952. 423–441.
  • Gertych et al: Bone Age Assessment of Children using a Digital Hand Atlas Comput Med Imaging Graph, 31(4-5), 2007. 322–331.
  • Larson DB, Chen MC, Lungren MP, et al.: Performance of a Deep-Learning Neural Network Model in Assessing Skeletal Maturity on Pediatric Hand Radiographs, Radiology 287(1), 2018. 313–322.