A 2-year follow-up MRI study for the evaluation of an age estimation method based on knee bone development.

Age estimation is an actual topic in the area of forensic medicine with a special focus on the age limits of 16 and 18 years. Current research on this topic relies on retrospective data of inhomogeneous populations relating to sex, age range, and socioeconomic status. In this work, we present a 2-year follow-up study for the evaluation of an age estimation method on a prospective magnetic resonance imaging (MRI) knee data collective of a homogeneous population. The study includes 40 male subjects from northern Germany aged 14 to 21 years. Three MRI examinations were evenly acquired within 2 years for each subject. As a first evaluation, a three-stage system was used to assess the ossification status of the knee (I:"open", II:"partially ossified", III:"fully ossified"). Three raters assessed the growth plate of the distal femur, proximal tibia, and proximal fibula based on central 2D slices. A good inter-rater agreement was attained (κ = 0.84). All subjects younger than 18 years were rated as stage I and had a cumulative knee score (SKJ) ≤ 5. Based on the follow-up datasets, new parameters quantifying the intra-individual ossification process were calculated. The results of this follow-up analysis show a different start, end, and speed of each growth plate's maturation as well as an ossification peak for individuals at the age of 16. The generated MRI database provides new insights into the ossification process over time and serves as a basis for further evaluations of age estimation methods.

Automated segmentation of the knee for age assessment in 3D MR images using convolutional neural networks.

Age assessment is used to estimate the chronological age of an individual who lacks legal documentation. Recent studies indicate that the ossification degree of the growth plates in the knee joint correlates with chronological age of adolescents and young adults. To verify this hypothesis, a high number of datasets need to be analysed. An approach which enables an automated detection and analysis of the bone structures may be necessary to handle large datasets. The purpose of this study was to develop a fully automatic 2D knee segmentation based on 3D MR images using convolutional neural networks. A total of 76 datasets were available and divided into a training set (74%), a validation set (13%) and a test set (13%). Multiple preprocessing steps were applied to correct image intensity values and to reduce the image size. Image augmentation was employed to virtually increase the dataset size for training. The proposed architecture for the segmentation task resembles the encoder-decoder model type used for the U-Net. The trained network achieved a dice similarity coefficient score of 98% compared to the manual segmentations and an intersection over union of 96%. The precision and recall of the model were balanced, and the error was only 1.2%. No overfitting was observed during training. As a proof of concept, the predicted segmentations were used for the age estimation of 145 subjects. Initial results show the potential of this approach attaining a mean absolute error of 0.48 ± 0.32 years for a test set of 14 subjects. The proposed automated segmentation can contribute to faster, reproducible and potentially more reliable age estimation in the future.