Bone age assessment: automated techniques coming of age?

Bone age determination from hand radiographs is one of the oldest radiographic procedures. The first atlas was published by Poland in 1898, and to date the Greulich Pyle atlas, although it dates from 1959, is still the most commonly used method. Bone age rating is time-consuming, suffers from an unsatisfactorily high rater variability, and therefore already 25 years ago it was proposed to replace the manual rating by an automated, computerized method, a field nowadays referred to as computer-aided diagnosis (CAD). The pursuit of this goal reached a first stage of accomplishment in 1992-1996 with the presentation of several systems. However, they had limited clinical value, and efforts in CAD research were increasingly focused on lesion detection for cancer screening. It was only in 2008 that a fully-automated bone age method was presented, which appears to be clinically acceptable. In this paper we consider the requirements that should be met by an automated bone age method and review the state of the art. Integration in PACS and saving time are important factors for radiologists. But it is the validation of the methods which poses the greatest challenge, because there is no gold standard for bone age rating, and the direct comparison to manual rating is therefore not sufficient for demonstrating that manual rating can be replaced by automated rating. One needs additional studies assessing the precision of a method and its accuracy when used for adult height prediction, which serves as an objective.

Accuracy and self-validation of automated bone age determination.

The BoneXpert method for automated determination of bone age from hand X-rays was introduced in 2009 and is currently running in over 200 hospitals. The aim of this work is to present version 3 of the method and validate its accuracy and self-validation mechanism that automatically rejects an image if it is at risk of being analysed incorrectly. The training set included 14,036 images from the 2017 Radiological Society of North America (RSNA) Bone Age Challenge, 1642 images of normal Dutch and Californian children, and 8250 images from Tübingen from patients with Short Stature, Congenital Adrenal Hyperplasia and Precocious Puberty. The study resulted in a cross-validated root mean square (RMS) error in the Tübingen images of 0.62 y, compared to 0.72 y in the previous version. The RMS error on the RSNA test set of 200 images was 0.45 y relative to the average of six manual ratings. The self-validation mechanism rejected 0.4% of the RSNA images. 121 outliers among the self-validated images of the Tübingen study were rerated, resulting in 6 cases where BoneXpert deviated more than 1.5 years from the average of the three re-ratings, compared to 72 such cases for the original manual ratings. The accuracy of BoneXpert is clearly better than the accuracy of a single manual rating. The self-validation mechanism rejected very few images, typically with abnormal anatomy, and among the accepted images, there were 12 times fewer severe bone age errors than in manual ratings, suggesting that BoneXpert could be safer than manual rating.

Metacarpal thickness, width, length and medullary diameter in children--reference curves from the First Zürich Longitudinal Study.

SUMMARY: Metacarpal thickness (T), width (W), length (L) and medullary diameter (M) were measured in 3,121 X-rays from 231 healthy Caucasian children aged 3 to 19 years and analysed for bone age, age, height, weight and gender-related characteristics, showing highly differentiated growth patterns with prepubertal dips. Reference data for the four metacarpal measures are presented. INTRODUCTION: The aim of the study was to create and explore a reference database for metacarpal T, W, L and M in children. METHODS: Three thousand one hundred twenty-one left-hand X-rays (1,661 from boys) from 231 healthy Caucasian subjects (119 boys) aged 3 to 19 years were analysed by BoneXpert, a programme for automatic analysis of hand X-rays and bone age (BA; in years). RESULTS: In boys, growth of T, W and L shows a prepubertal decrease from BA 7 to 13 and then accelerates again. In girls, the same is seen only for T starting from BA 8 to 11, whereas W and L grow at a declining rate. M shows steady growth until BA 10.5 in girls and BA 13.5 in boys and then grows smaller in both. W is greater in boys from BA 6 onwards, while L is greater in girls from BA 9 to 13 and T from BA 11 to 14. BA is reflected best by L until start of puberty and by T and L thereafter. CONCLUSION: T, W, L and M show highly differentiated growth patterns. These reference data provide a basis for further research into skeletal development and the management of hormone therapies in children.

A paediatric bone index derived by automated radiogrammetry.

SUMMARY: Hand radiographs are obtained routinely to determine bone age of children. This paper presents a method that determines a Paediatric Bone Index automatically from such radiographs. The Paediatric Bone Index is designed to have minimal relative standard deviation (7.5%), and the precision is determined to be 1.42%. INTRODUCTION: We present a computerised method to determine bone mass of children based on hand radiographs, including a reference database for normal Caucasian children. METHODS: Normal Danish subjects (1,867), of ages 7-17, and 531 normal Dutch subjects of ages 5-19 were included. Historically, three different indices of bone mass have been used in radiogrammetry all based on A = piTW(1 - T/W), where T is the cortical thickness and W the bone width. The indices are the metacarpal index A/W(2), DXR-BMD = A/W, and Exton-Smith's index A/(WL), where L is the length of the bone. These indices are compared with new indices of the form A/(W(a) L(b)), and it is argued that the preferred index has minimal SD relative to the mean value at each bone age and sex. Finally, longitudinal series of X-rays of 20 Japanese children are used to derive the precision of the measurements. RESULTS: The preferred index is A/(W(1.33) L(0.33)), which is named the Paediatric Bone Index, PBI. It has mean relative SD 7.5% and precision 1.42%. CONCLUSIONS: As part of the BoneXpert method for automated bone age determination, our method facilitates retrospective research studies involving validation of the proposed index against fracture incidence and adult bone mineral density.

A review of Bayesian neural networks with an application to near infrared spectroscopy.

MacKay's (1992) Bayesian framework for backpropagation is a practical and powerful means to improve the generalization ability of neural networks. It is based on a Gaussian approximation to the posterior weight distribution. The framework is extended, reviewed, and demonstrated in a pedagogical way. The notation is simplified using the ordinary weight decay parameter, and a detailed and explicit procedure for adjusting several weight decay parameters is given. Bayesian backprop is applied in the prediction of fat content in minced meat from near infrared spectra. It outperforms "early stopping" as well as quadratic regression. The evidence of a committee of differently trained networks is computed, and the corresponding improved generalization is verified. The error bars on the predictions of the fat content are computed. There are three contributors: The random noise, the uncertainty in the weights, and the deviation among the committee members. The Bayesian framework is compared to Moody's GPE (1992). Finally, MacKay and Neal's automatic relevance determination, in which the weight decay parameters depend on the input number, is applied to the data with improved results.

In-line image analysis in the slaughter industry, illustrated by Beef Carcass Classification.

This paper describes a complete framework for the quantitative quality control of biological objects using computer vision. The techniques are described in the context of BCC-2, the second generation Beef Carcass Classification centre, which has operated as a prototype since March 1995. Installed in the slaughterline, BCC-2 analyses one half, dehided carcass. BCC-2 determines the visual properties: conformation, fatness and fat colour as well as objective quantities such as the percent of saleable meat and the cross sectional area of the rib eye. BCC-2 measures geometry and colour quantitatively. A procedure for maintaining the same calibration over time for several BCC-2 units has been developed. BCC-2 is built from a few inexpensive components: A frame that positions the half carcass in the slaughterline, a camera, two PC's, and a terminal. In addition, two slide projectors project stripes of light onto the carcass at an angle to the camera to provide information about the three-dimensional shape. The biological variation of the carcasses requires the use of advanced information processing techniques: traditional pattern recognition, principal component analysis, and neural networks. BCC-2 is adaptive, i.e. it is trained by examples, and BCC-2 is robust in the sense that it classifies all carcasses except the ones most damaged in the slaughter process.

Diversity in auxology: between theory and practice. Proceedings of the 18th Aschauer Soiree, 13th November 2010.

Auxology has developed from mere describing child and adolescent growth into a vivid and interdisciplinary research area encompassing human biologists, physicians, social scientists, economists and biostatisticians. The meeting illustrated the diversity in auxology, with the various social, medical, biological and biostatistical aspects in studies on child growth and development.

Computational radiology in skeletal radiography.

Recent years have brought rapid developments in computational image analysis in musculo-skeletal radiology. Meanwhile the algorithms have reached a maturity that makes initial clinical use feasible. Applications range from joint space measurement to erosion quantification, and from fracture detection to the assessment of alignment angles. Current results of computational image analysis in radiography are very promising, but some fundamental issues remain to be clarified, among which the definition of the optimal trade off between automatization and operator-dependency, the integration of these tools into clinical work flow and last not least the proof of incremental clinical benefit of these methods.

Estimation of bone mineral density by digital X-ray radiogrammetry: theoretical background and clinical testing.

A new automated radiogrammetric method to estimate bone mineral density (BMD) from a single radiograph of the hand and forearm is described. Five regions of interest in radius, ulna and the three middle metacarpal bones are identified and approximately 1800 geometrical measurements from these bones are used to obtain a BMD estimate of the distal forearm, referred to as BMDDXR (from digital X-ray radiogrammetry, DXR). The measured dimensions for each bone are the cortical thickness and the outer width, in combination with an stimate of the cortical porosity. The short-term in vivo precision of BMDDXR was observed to be 0.60% in a clinical study of 24 women and the in vitro variation over 12 different radiological clinics was found to be 1% of the young normal BMDDXR level. In a cohort of 416 women BMDDXR was found to be closely correlated with BMD at the distal forearm measured by dual-energy X-ray absoptiometry (r = 0.86, p < 0.0001) and also with BMD at the spine, total hip and femoral neck (r = 0.62, 0.69 and 0.73, respectively, p<0.0001 for all). The annual decline was estimated from the cohort to be 1.05% in the age group 55-65 years. Relative to this age-related loss, the reported short-term precision allows for monitoring intervals of 1.0 years and 1.6 years in order to detect expected age-related changes with a confidence of 80% and 95%, respectively. It is concluded that the DXR method offers a BMD estimate with a good correlation with distal forearm BMD, a low variation between geographical sites and a precision that potentially allows for relatively short observation intervals.