Preliminary report on the validation of a grammar-based computer system for assessing skeletal maturity with the Tanner-Whitehouse 2 method

Automated bone age assessment: motivation, taxonomies, and challenges

Bone age assessment (BAA) of unknown people is one of the most important topics in clinical procedure for evaluation of biological maturity of children. BAA is performed usually by comparing an X-ray of left hand wrist with an atlas of known sample bones. Recently, BAA has gained remarkable ground from academia and medicine. Manual methods of BAA are time-consuming and prone to observer variability. This is a motivation for developing automated methods of BAA. However, there is considerable research on the automated assessment, much of which are still in the experimental stage. This survey provides taxonomy of automated BAA approaches and discusses the challenges. Finally, we present suggestions for future research.

[Assessment of carpal bone maturation by imaging: an alternative to bone age determination or a complementary study?]

Radiographic evaluation of skeletal maturation with or without automated systems relies on the comparison between the radiographic appearance of portions of a child's skeleton and the standardized appearance in a comparable population of children. Assessment of skeletal maturation using image analysis of wrist computerized-assisted tomography is a quantitative method allowing for measurements of the volume of each carpal bone, the axis of inertia, and the mean bone density. Morphological modifications of carpal bones are characterized by quantitative data relating to the length of the axis of inertia and the volume. Imaging is a potentially useful method for carpal bone assessment during growth.

Carpal bone maturation during childhood and adolescence: assessment by quantitative computed tomography. Preliminary results

The aim of our study was to measure the volume of each carpal bone during childhood and adolescence by image processing from computed tomography (CT) scans, and to analyze the relationship between the eight carpal bones. Thirteen CT scans were performed in nine normal prepubertal, peripubertal and post-pubertal children, six boys and three girls, aged 5-14 years. Each scan was processed in order to extract the carpal bones. The volume was computed for each bone. There was a significant correlation between carpal bone volume and age (0.55 < r < 0.79), and a very strong correlation between the volume of a given carpal bone and the volume of all the others, whatever the age (0.87 < r < 0.99, p < 0.01). Image processing is a potentially useful method for assessing bone maturation. The constant ratio between carpal bone volumes indicates that these bones interact with each other in wrist bone maturation.

Radiology's Achilles' heel: error and variation in the interpretation of the Röntgen image

The performance of the human eye and brain has failed to keep pace with the enormous technical progress in the first full century of radiology. Errors and variations in interpretation now represent the weakest aspect of clinical imaging. Those interpretations which differ from the consensus view of a panel of "experts" may be regarded as errors; where experts fail to achieve consensus, differing reports are regarded as "observer variation". Errors arise from poor technique, failures of perception, lack of knowledge and misjudgments. Observer variation is substantial and should be taken into account when different diagnostic methods are compared; in many cases the difference between observers outweighs the difference between techniques. Strategies for reducing error include attention to viewing conditions, training of the observers, availability of previous films and relevant clinical data, dual or multiple reporting, standardization of terminology and report format, and assistance from computers. Digital acquisition and display will probably not affect observer variation but the performance of radiologists, as measured by receiver operating characteristic (ROC) analysis, may be improved by computer-directed search for specific image features. Other current developments show that where image features can be comprehensively described, computer analysis can replace the perception function of the observer, whilst the function of interpretation can in some cases be performed better by artificial neural networks. However, computer-assisted diagnosis is still in its infancy and complete replacement of the human observer is as yet a remote possibility.