New AI Improves Paediatric Scoliosis Monitoring - EMJ

New Artificial Intelligence Improves Paediatric Scoliosis Monitoring

1 Mins

DATED scoliosis monitoring on X-rays is set to become a thing of the past, as a new artificial intelligence (AI) model has been developed to assist in predicting paediatric bone growth. Researchers, led by John Zech, New York University Langone Health, USA, have reported that their AI model provided improved predictions compared to the standard model used to predict scoliosis today.

Full-body biplanar slot scanning, a type of low-dose digital X-ray imaging, is the current method used to monitor scoliosis. Patients are generally imaged at 6-month intervals for a period of 1–5 years, allowing clinicians to estimate growth over time, and compare them with statistical averages in the Anderson-Green method. The Anderson-Green standards are, however, “based on a sample of only 100 children whose growth was evaluated more than 60 years ago, and who were not racially or ethnically diverse,” commented the authors.

In order to improve upon this process, the authors aimed to leverage AI by training a convolutional neural network model to automatically measure femorotibial length on X-ray images from a racially diverse set of paediatric patients. The study involved 1,874 examinations from 523 patients (aged ≤21 years). The convolutional neural network model was trained to segment the femur and tibia on the X-rays, and measure total leg, femoral, and tibial length, with results showing the model’s mean absolute error measurements were 0.25 cm for the femur, 0.27 cm for the tibia, and 0.33 cm for composite lower limbs. AI measurements were then used to develop femoral and tibial growth curves, which were compared to those made using the Anderson-Green method.

Zech and team found that the AI growth curves more accurately represented lower extremity growth in an external test set (n=154) than the Anderson-Green method. Growth reference ranges generated using the AI model showed higher coverage probability than those generated using the Anderson-Green method (86.7% versus 73.4%; p<0.001), the authors reported.

The researchers concluded: “The AI model offers an easily scalable method to obtain growth data from additional patients, which may allow the creation of stratified standards that more closely fit a child’s individual growth profile.”



Zech JR et al. Lower extremity growth according to AI automated femorotibial length measurement on slot-scanning radiographs in pediatric patients. Radiology. 2024;311(1):e231055.

Please rate the quality of this content

As you found this content interesting...

Follow us on social media!

We are sorry that this content was not interesting for you!

Let us improve this content!

Tell us how we can improve this content?