Jacek Kawa

Information Technologies in Biomedicine

As the medical information systems have been integrated in order to address the core of medicine, including patient care in ambulatory and in-patient setting, computer assisted diagnosis and treatment, telemedicine, and home care we are witnessing radical changes in the Information Technologies. This will continue in the years to come. This book presents a comprehensive study in this field and contains carefully selected articles contributed by experts of information technologies. It is an interdisciplinary collection of papers that have both a theoretical and applied dimension. In particular, it includes the following sections: - Image Processing and CAD, - Signal Processing, - Biotechnology, - Data Analysis, - Multimedia, - Biomechanics. This book is a great reference tool for scientists who deal with problems of designing and implementing information processing tools employed in systems that assist the clinicians in patient diagnosis and treatment.

Role of radiologists in CAD life-cycle

A modern CAD (computer-aided diagnosis) system development involves a multidisciplinary team whose members are experts in medical and technical fields. This study indicates the activities of medical experts at various stages of the CAD design, testing, and implementation. Those stages include a medical analysis of the diagnostic problem, data collection, image analysis, evaluation, and clinical verification. At each stage the physicians knowledge and experience are indispensable. The final implementation involves integration with the existing Picture Archiving and Communication System. The term CAD life-cycle describes an overall process of the design, testing, and implementation of a system that in its final form assists the radiologists in their daily clinical routine. Four CAD systems (applied to the bone age assessment, Multiple Sclerosis detection, lung nodule detection, and pneumothorax measurement) developed in our laboratory are given as examples of how consecutive stages are developed by the multidisciplinary team. Specific advantages of the CAD implementation that include the daily clinical routine as well as research and education activities are discussed.

Image Clustering with Median and Myriad Spatial Constraint Enhanced FCM

In the current study two approaches to the clustering problem have been tested. First, a sequential analysis of -ltering and fuzzy c-means (FCM) method is performed. Then, the standard FCM has been modi-ed by adding to the objective function a second term that formulates a spatial constraint. In both approaches mean, median, and myriad are implemented. The analysis has been performed on a synthetic image and clinical images.

Open architecture computer-aided diagnosis system

: In this study an approach to open architecture computer-aided diagnosis (CAD) is presented. The traditional goal of a CAD system, to assist the physicians in performing the diagnosis and treatment, has been extended. The platform also supports the system designer in developing a new CAD workflow by implementing general-purpose modules as well as problem-dependent procedures. A new CAD may require new procedures to be added, yet some of the already implemented functions can be employed. The CAD environment is subjected to the developmental process of three systems: the multiple sclerosis CAD, the lung nodule CAD and the pneumothorax CAD. Modules and procedures are briefly described and the CAD systems are evaluated. Results obtained during the CAD evaluation show prospective flexibility of the infrastructure. The trade-offs, well known to CAD designers, can easily be handled by the operators in a user-friendly manner by choosing various workflow paths.

Automated Fuzzy-Connectedness-Based Segmentation in Extraction of Multiple Sclerosis Lesions

In the current study, a fuzzy-connectedness-based approach to fine segmentation of demyelination lesions in Multiple Sclerosis is introduced as an enhancement to the existing ‘fast’ segmentation method. First a fuzzy connectedness relation is introduced, next a short overview of the ‘fast’ segmentation method is presented. Finally, a novel, automated segmentation approach is described. The combined method is applied to segmentation of clinical Magnetic Resonance FLAIR Images.

Radiological Atlas for Patient Specific Model Generation

The paper presents the development of a radiological atlas employed in an abdomen patient specific model verification.

Automatic brain tumour detection and neovasculature assessment with multiseries MRI analysis

In this paper a novel multi-stage automatic method for brain tumour detection and neovasculature assessment is presented. First, the brain symmetry is exploited to register the magnetic resonance (MR) series analysed. Then, the intracranial structures are found and the region of interest (ROI) is constrained within them to tumour and peritumoural areas using the Fluid Light Attenuation Inversion Recovery (FLAIR) series. Next, the contrast-enhanced lesions are detected on the basis of T1-weighted (T1W) differential images before and after contrast medium administration. Finally, their vascularisation is assessed based on the Regional Cerebral Blood Volume (RCBV) perfusion maps. The relative RCBV (rRCBV) map is calculated in relation to a healthy white matter, also found automatically, and visualised on the analysed series. Three main types of brain tumours, i.e. HG gliomas, metastases and meningiomas have been subjected to the analysis. The results of contrast enhanced lesions detection have been compared with manual delineations performed independently by two experts, yielding 64.84% sensitivity, 99.89% specificity and 71.83% Dice Similarity Coefficient (DSC) for twenty analysed studies of subjects with brain tumours diagnosed.

Spatial and dynamical handwriting analysis in mild cognitive impairment

Background and Objectives Standard clinical procedure of Mild Cognitive Impairment (MCI) assessment employs time-consuming tests of psychological evaluation and requires the involvement of specialists. The employment of quantitative methods proves to be superior to clinical judgment, yet reliable, fast and inexpensive tests are not available. This study was conducted as a first step towards the development of a diagnostic tool based on handwriting. Methods In this paper the handwriting sample of a group of 37 patients with MCI (mean age 76.1±5.8) and 37 healthy controls (mean age 74.8±5.7) was collected using a Livescribe Echo Pen while completing three tasks: (1) regular writing, (2) all-capital-letters writing, and (3) single letter multiply repeated. Parameters differentiating both groups were selected in each task. Results Subjects with confirmed MCI needed more time to complete task one (median 119.5s, IQR - interquartile range - 38.1 vs. 95.1s, IQR 29.2 in control and MCI group, p-value <0.05) and two (median 84.2s, IQR 49.2 and 53.7s, IQR 30.5 in control and MCI group) as their writing was significantly slower. These results were associated with a longer time to complete a single stroke of written text. The written text was also noticeably larger in the MCI group in all three tasks (e.g. median height of the text block in task 2 being 22.3mm, IQR 12.9 in MCI and 20.2mm, IQR 8.7 in control group). Moreover, the MCI group showed more variation in the dynamics of writing: longer pause between strokes in task 1 and 2. The all-capital-letters task produced most of the discriminating features. Conclusion Proposed handwriting features are significant in distinguishing MCI patients. Inclusion of quantitative handwriting analysis in psychological assessment may be a step forward towards a fast MCI diagnosis.

White matter segmentation from MR images in subjects with brain tumours

In this study an automatic White Matter (WM) detection method in Magnetic Resonance (MR) images is presented. The detected WM areas are intended to serve as reference areas for the Regional Cerebral Blood Volume (RCBV) perfusion maps analysis aimed at assessing brain tumour neovasculature. Two MR series, possessing the required WM to Gray Matter (GM) contrast, are analysed: T1-Weighted (T1W) and Fluid Attenuated Inversion Recovery (FLAIR). First, the FLAIR series is subjected to anisotropic diffusion filtering. Next, a two-dimensional histogram of the analysed series is calculated and clustered with the use of Kernelised Fuzzy C-Means (KFCM) clustering. Finally, the clustering results are used as WM seed points for the subsequent region growing, providing the WM masks. The methodology has been tested on 10 studies of subjects with brain tumours diagnosed and compared with the Golden Standard (GS) delineations performed by an expert physician. Three similarity measures have been calculated: sensitivity, specificity and the Dice Similarity Coefficient (DSC). Their values amounted to 67.86%, 97.55% and 69.98%, respectively.

Leg movement tracking in automatic video-based one-leg stance evaluation

Falls are a major risk in elder population. Early diagnosis is therefore an important step towards increasing the safety of elders. One of the common diagnostic tests is the Berg Balance Scale (BBS), consisting of 14 exercises arranged from the easiest (sitting-to-standing) to the most demanding (one-leg stance). In this study a novel approach to the automatic assessment of the time in which the patient can remain in the one-leg stance position without loosing balance is introduced. The data is collected using a regular video camera. No markers, special garments, or system calibration are required. Two groups are examined. The first group consists of 16 students: healthy, young adults (12 female, 4 male, avg. 20yrs±1). The second group consists of 50 elders (39 female, 11 male, avg. 78.8yrs±5.9). Data (short, one minute recordings) are collected in a controlled environment using a digital video recorder (50fps, 1920×1080) fixed to a tripod. Data are processed off-line. First, the region of interest is determined. Next, the Kanade-Lucas-Tomasi tracking is performed. Best tracks are selected based on the registered vertical movement and two tracks are obtained corresponding to the left and right leg movements. Tracks are later subjected to the sparse signal baseline estimation, denoising and thresholding to detect the raised leg. Results are compared frame-wise to the ground truth reference obtained in the manual processing procedure. Both legs are evaluated in the elder group (in all cases several attempts featuring both legs were registered), resulting in 89.18%±11.27% DICE, 93.07%±5.43% sensitivity and 96.94%±6.11% specificity values for both legs. The signal of a single leg is evaluated in the student group (in all cases only one attempt was needed to perform the full examination) resulting in 98.96%±1.2% DICE, 98.78%±1.65% sensitivity and 98.73%±2.69% specificity values. This is the first step towards a video-based system enabling the automatic assessment of the four last, most vital tasks of the Berg Balance Scale evaluation.