New radiomics model at Shanghai Jiao Tong University improves height prediction for girls after menarche

Researchers in Shanghai develop a knee-based radiomics model that predicts final height in girls with superior accuracy compared to traditional bone age methods.

By: AXL Media

Published: Apr 29, 2026, 10:33 AM EDT

Source: Information for this report was sourced from the Chinese Academy of Sciences

New radiomics model at Shanghai Jiao Tong University improves height prediction for girls after menarche - article image
New radiomics model at Shanghai Jiao Tong University improves height prediction for girls after menarche - article image

Revolutionizing Stature Forecasting in Pediatric Care

A research team from the Shanghai Children’s Medical Center and Shanghai Jiao Tong University School of Medicine has introduced a groundbreaking method for predicting the final adult height of girls post menarche. While the onset of menstruation typically signals the final stage of physical growth, the remaining height gain varies significantly among individuals, often complicating clinical decisions regarding hormonal or orthopedic interventions. Traditionally, clinicians have relied on hand-wrist bone age via the Greulich-Pyle method, but this new study, published in the World Journal of Pediatrics, suggests that digital radiographs of the knee provide a more precise biological clock for lower limb growth.

The Shift from Hand-Wrist to Knee Imaging

The study’s focus on the knee, specifically the distal femur and proximal tibia, stems from the biological reality that these regions often fuse later than the small bones of the hand and wrist. Because final stature is heavily influenced by the extension of the lower limbs, imaging the knee offers a more direct observation of the body’s remaining growth potential. The researchers analyzed a cohort of 173 Chinese girls within three months of their first period, discovering that the distal femur preserves specific maturation signals that are frequently overlooked by standard clinical atlases.

Integrating Radiomics and Clinical Data

The "radiomics" approach involves extracting high-dimensional data from digital X-rays that are invisible to the naked eye, such as subtle patterns in texture, brightness, and shape. By combining these radiomic scores with the girl's height at menarche, her father’s height, and standard bone age, the team created a multi-variable prediction equation. This model successfully explained 73% of the variation in final adult height among the study participants, who gained an average of 8.94 cm after menarche. This integrated approach effectively mitigates the observer bias often found in manual bone age readings.

Categories

Topics

Related Coverage