Robert Koprowski

Machine learning, medical diagnosis, and biomedical engineering research - commentary

A large number of papers are appearing in the biomedical engineering literature that describe the use of machine learning techniques to develop classifiers for detection or diagnosis of disease. However, the usefulness of this approach in developing clinically validated diagnostic techniques so far has been limited and the methods are prone to overfitting and other problems which may not be immediately apparent to the investigators. This commentary is intended to help sensitize investigators as well as readers and reviewers of papers to some potential pitfalls in the development of classifiers, and suggests steps that researchers can take to help avoid these problems. Building classifiers should be viewed not simply as an add-on statistical analysis, but as part and parcel of the experimental process. Validation of classifiers for diagnostic applications should be considered as part of a much larger process of establishing the clinical validity of the diagnostic technique.

Book review of "The Biomedical Engineering Handbook" fourth edition, edited by Joseph D. Bronzino, Donald R. Peterson

This article is a review of the book “The biomedical engineering handbook”, fourth edition: four volume set (ISBN 9781439825334, 254GBP, 5430 pages) edited by Joseph D. Bronzino and Donald R. Peterson published by the CRC Press Taylor & Francis group in 2015. The content of the book and its importance for biomedical engineering have been discussed in this invited review.

Headache in Children's Drawings:

Headache is a common health problem in childhood. Childrens drawings are helpful in the diagnosis of headache type. Children, especially younger ones, communicate better through pictures than verbally. The aim of the present study is to evaluate the usefulness of drawings of the childs headache in the diagnostic process carried out by a pediatrician and a pediatric neurologist. At the beginning of a visit in a neurological clinic, or on the first day of hospitalization, the child was asked, “Please draw your headache,” or “How do you feel your headache?” without any additional explanations or suggestions. Clinical diagnosis of headache type was made on the basis of the standard diagnostic evaluation. For the purpose of this study, childrens headaches were categorized as migraine, tension-type headache, or “the others.” One hundred twenty-four drawings of children with headaches were analyzed by 8 pediatricians and 8 pediatric neurologists. The analysts were unaware of the clinical history, age, sex, and diagnosis of the patients. The clinical diagnosis was considered the “gold standard” to which the headache drawing diagnosis was compared. There were 68 girls 5—18 years of age and 56 boys 7—18 years of age. Of the 124 children, 40 were clinically diagnosed with migraine (32.2%), 47 with tension-type headache (37.9%), and 37 (29.8%) as the others. Children with migraine most frequently draw sharp elements. Children with tension-type headache mainly drew compression elements and pressing elements. In the group of “the other” headaches, 21 children were diagnosed with somatoform disorders. The most frequent element in this groups drawings was a whirl in the head. Colors used most frequently were black and red, which signify severe pain. There was no difference in sensitivity of diagnoses between neurologists and pediatricians. Because the evaluation of drawings by children with headaches done both by pediatricians and pediatric neurologists was correct for approximately half of the children, the authors decided to prepare a set of test pictures, including characteristic presentations of pain. Preparing a ready set of test drawings may facilitate differentiation for the inexperienced doctors and encourage those children who refuse to draw.

Layers Recognition in Tomographic Eye Image Based on Random Contour Analysis

The paper presents an algorithm designed to detect layers of eye’s retina using an area analysis. This analysis has been implemented and tested in images obtained by means of Copernicus OCT (Optical Coherence Tomography). The algorithm created is an original approach to detect layers, contours and psuedoparallels. The approach presented is an expansion of approaches described in [1] and [2],[3] and enables identification and recognition of external limiting membranes, retina and others. The algorithm has been implemented in Matlab and C environment.

Fully automatic algorithm for the analysis of vessels in the angiographic image of the eye fundus

BackgroundThe available scientific literature contains descriptions of manual, semi-automated and automated methods for analysing angiographic images. The presented algorithms segment vessels calculating their tortuosity or number in a given area. We describe a statistical analysis of the inclination of the vessels in the fundus as related to their distance from the center of the optic disc.MethodsThe paper presents an automated method for analysing vessels which are found in angiographic images of the eye using a Matlab implemented algorithm. It performs filtration and convolution operations with suggested masks. The result is an image containing information on the location of vessels and their inclination angle in relation to the center of the optic disc. This is a new approach to the analysis of vessels whose usefulness has been confirmed in the diagnosis of hypertension.ResultsThe proposed algorithm analyzed and processed the images of the eye fundus using a classifier in the form of decision trees. It enabled the proper classification of healthy patients and those with hypertension. The result is a very good separation of healthy subjects from the hypertensive ones: sensitivity - 83%, specificity - 100%, accuracy - 96%. This confirms a practical usefulness of the proposed method.ConclusionsThis paper presents an algorithm for the automatic analysis of morphological parameters of the fundus vessels. Such an analysis is performed during fluorescein angiography of the eye. The presented algorithm automatically calculates the global statistical features connected with both tortuosity of vessels and their total area or their number.

EPR studies of free radicals decay and survival in gamma irradiated aminoglycoside antibiotics: sisomicin, tobramycin and paromomycin.

Radiation sterilization technology is more actively used now that any time because of its many advantages. Gamma radiation has high penetrating power, relatively low chemical reactivity and causes small temperature rise. But on the other hand radiosterilization can lead to radiolytic products appearing, in example free radicals. Free radicals in radiative sterilized sisomicin, tobramycin and paromomycin were studied by electron paramagnetic resonance (EPR) spectroscopy. Dose of gamma irradiation of 25kGy was used. Concentrations and properties of free radicals in irradiated antibiotics were studied. EPR spectra were recorded for samples stored in air and argon. For gamma irradiated antibiotics strong EPR lines were recorded. One- and two-exponential functions were fitted to experimental points during testing and researching of time influence of the antibiotics storage to studied parameters of EPR lines. Our study of free radicals in radiosterilized antibiotics indicates the need for characterization of medicinal substances prior to sterilization process using EPR values. We propose the concentration of free radicals and other spectroscopic parameters as useful factors to select the optimal type of sterilization for the individual drug. The important parameters are i.a. the τ time constants and K constants of exponential functions. Time constants τ give us information about the speed of free radicals concentration decrease in radiated medicinal substances. The constant K(0) shows the free radicals concentration in irradiated medicament after long time of storage.

Automatic analysis of selected choroidal diseases in OCT images of the eye fundus

IntroductionThis paper describes a method for automatic analysis of the choroid in OCT images of the eye fundus in ophthalmology. The problem of vascular lesions occurs e.g. in a large population of patients having diabetes or macular degeneration. Their correct diagnosis and quantitative assessment of the treatment progress are a critical part of the eye fundus diagnosis.Material and methodThe study analysed about 1’000 OCT images acquired using SOCT Copernicus (Optopol Tech. SA, Zawiercie, Poland). The proposed algorithm for image analysis enabled to analyse the texture of the choroid portion located beneath the RPE (Retinal Pigment Epithelium) layer. The analysis was performed using the profiled algorithm based on morphological analysis and texture analysis and a classifier in the form of decision trees.ResultsThe location of the centres of gravity of individual objects present in the image beneath the RPE layer proved to be important in the evaluation of different types of images. In addition, the value of the standard deviation and the number of objects in a scene were equally important. These features enabled classification of three different forms of the choroid that were related to retinal pathology: diabetic edema (the classification gave accuracy ACC1 = 0.73), ischemia of the inner retinal layers (ACC2 = 0.83) and scarring fibro vascular tissue (ACC3 = 0.69). For the cut decision tree the results were as follows: ACC1 = 0.76, ACC2 = 0.81, ACC3 = 0.68.ConclusionsThe created decision tree enabled to obtain satisfactory results of the classification of three types of choroidal imaging. In addition, it was shown that for the assumed characteristics and the developed classifier, the location of B-scan does not significantly affect the results. The image analysis method for texture analysis presented in the paper confirmed its usefulness in choroid imaging. Currently the application is further studied in the Clinical Department of Ophthalmology in the District Railway Hospital in Katowice, Medical University of Silesia, Poland.

The Cell Structures Segmentation

In this article we have presented an attempt to segmentation of cell structures images acquired while histological slides microscopic observation. The described algorithm of segmentation is also applicable in other matters, where the image segmentation is an important part.

Automatic ultrasound image analysis in Hashimoto's disease

The paper presents diagnostics of parenchyma echogenicity and organ dimensions in thyroid examinations in the case of Hashimotos disease using image processing methods. In the event of discovering focal changes within the thyroid, a method for their pathology evaluation was suggested. The detector proposed operates fully automatically; using the information on the image texture it detects an artery in the image, which fulfils the role of reference point, and based on it - detects the area of interest.

Identification of Layers in a Tomographic Image of an Eye Based on the Canny Edge Detection

In the paper we present an algorithm for the identification of retina layers using the Canny edge detection for images obtained with OCT (Optical Coherence Tomography) Copernicus. The developed algorithm is an extension of the approaches covered in [4] and allows the identification and detection of hyaline-retinal border layers, the retina and other. The algorithm was implemented in the Matlab environment and the C language.