Gwilym S. Lodwick

Automated Radiographic Diagnosis via Feature Extraction and Classification of Cardiac Size and Shape Descriptors

One goal of digital processing of radiographic images is to provide the radiologist with quantitative measurements of human anatomy as well as an indication as to whether or not this anatomy is within normal limits. A computer algorithm is described, designed to automatically detect, extract quantitative measurements from, and diagnose the cardiac projection present in full-size anteriorview chest radiographs. A normal-abnormal diagnosis is demonstrated utilizing abnormal data from five classes of heart disease. In addition, normal-abnormal as well as normal-differential diagnoses are demonstrated for the rheumatic heart disease class. A feature extraction algorithm is developed using several ad hoc techniques, some of which were adapted from other feature extraction uses. The extracted features are classified into diagnostic classes using linear and quadratic discriminant functions. A concurrent study of physician diagnostic accuracy is also undertaken using the averaged diagnostic rates of ten radiologists on a representative subset of the radiographs used in the computer study.

Direct Measurement of the Effect of Film Miniaturization on Diagnostic Accuracy

A method for directly measuring the reduction in diagnostic accuracy incurred by altering original radiographs was applied to evaluate a 35 mm film system. A total of 4,290 readings of 2,145 proved cases were collected. Analysis of the results indicate that the 35 mm system reduced diagnostic accuracy by about 2.3% +/- 2.4% at a 95% confidence level. The same method can be used to evaluate other photographic reduction systems, television displays, enhancement schemes, or even the importance of clinical data in roentgenographic interpretation.

Towards computer analysis of pulmonary infiltration.

A feasibility study is described to provide quantitative texture measures to distinguish between normal lung, alveolar infiltrates and interstitial infiltrates. Advanced computer imaging technology and decision making processes were applied to distinguish between these textural patterns. The results, based on computer extracted quantitative measures, show an excellent separation of the three classes considered with 95% accuracy in the training phase and 90% accuracy in the testing phase.

The Computer Analysis and Diagnosis of Gastric Ulcers

THE VALUE of roentgenographic examination of the stomach in patients with upper gastrointestinal symptoms is well established (10, 12), and the accuracy of the roentgenologist in detecting all stomach abnormalities has been variously cited to be between 85 and 95 per cent (1, 3, 4, 7, 10). Even so, one of the most controversial differential diagnoses confronting a radiologist is that of whether a gastric ulcer is benign or malignant (15). Many roentgen signs have been described which are thought to be of value in making this differentiation, but the statistical significance of the various clinical findings and roentgenographic signs has never been established. The use of computers in analyzing data relating to medical disease entities is widely accepted. Computer diagnosis has exceeded the diagnostic accuracy of the trained radiologist in selected bone tumors (8). Radiographic manifestations of carcinoma of the lung have also been evaluated by computer technics (9). The statistical evaluation of 70 radiog...

Experience with MARS (Missouri Automated Radiology System)

Since April 1970, the Department of Radiology of the University of Missouri has used an on-line computer system to record, encode, and retrieve all radiological consultations. The System has greatly speeded up delivery of typed reports without slowing down the radiologists using it, and provides an excellent mechanism for retrieving case material for teaching and research. The major drawbacks to the System are computer reliability and cost. We hope to overcome these problems by employing a smaller computer system dedicated solely to operation of the department, and by the addition of programs to handle business operations.

Direct computer diagnosis of rheumatic heart disease.

The authors briefly describe their experiences in the computer screening of 14 × 17-inch chest radiographs and the identification of cardiac abnormalities. The computer identified the cardiac silhouette, measured its important parameters, and applied discriminant function analysis techniques to decide whether or not the heart was within normal limits; if not, the computer reported the likely type of abnormality within the spectrum of rheumatic valvular disease. With the use of the test library of normals and cases of rheumatic heart disease, the computer had a higher overall accuracy than the radiologists viewing the same library of proved cases.

COMPUTER ALGORITHMS FOR THE DETECTION OF BRAIN SCINTIGRAM ABNORMALITIES.

Abstract Algorithms for computer analysis of digitized brain scintigrams are described. A technique of variable thresholding and region enumeration is used to detect abnormalities which project clear of the calvarium, while a metric comparison detects distortions of the brain outline. The methods have been employed to process scintigrams digitized by means of an image dissector camera operating under control of an SEL-840A computer. It is felt that the algorithms will be equally applicable to brain scintigrams digitized by other means and that they may prove adaptable to the analysis of other organ system images.

Radiate—Updated and Redesigned for Multiple Cathode-Ray Tube Terminals

The concept of a computer system for on-line, real time synthesis and transmission of radiological consultations has been described (6, 7). Automatic coding and storing of the radiological report is a by-product of this system. Dictionaries for diagnostic examinations (240 entries), anatomic sites (1,007 entries), and diagnosis (1,354 entries) have been developed and expanded. Initially, a typewriter type terminal served to interface the computer with the radiologist. Now, a cathode-ray tube (CRT) with alphanumeric keyboard gives the radiologist the ability to seek, receive, acknowledge, or alter information from the computer memory or the established data base. During the past three years RADIATE has been broadened and tested on various computers on a research basis. Recently the system, though unchanged in basic concept, has been completely redesigned for the IBM 360 Series computers so that greater flexibility and faster information transaction could be achieved. The present RADIATE system provides for...

The new image of diagnostic radiology.

Diagnostic radiology has left the doldrums of the fifties and is now moving. It is beginning to present a vital image. This is not to say that we no longer have problems in this discipline; it would be more nearly correct to say that the problems have changed. Nearly twenty years ago the image of diagnostic radiology was that of a moribund discipline with little patient contact, with programs inadequately supported by the medical schools, and with most of the exciting new advances in radiology created and exploited within other disciplines whose interests overlap those of radiology. We have come a long way since then. We still have a long way to go. Diagnostic radiology has become revitalized, to the point where the increase of demand for a radiologists time approaches 25 per cent per year. Part of this change of status and increased demand is due to the radiologists greater participation in diagnostic procedures, as well as to an actual increase in demand for a greater variety of procedures. Radiology ...

General Purpose Digital Display System for Computed Tomography Images

The authors describe a microprocessor-based digital image display system utilizing a solid-state image memory with a high quality television-type monitor for display of computed tomography images. It is an alternative to a commercially available display system, and offers the potential for research in digital image display. It is now under clinical evaluation. Image input to the stand-alone display system is via 9-track magnetic tape common to almost all CT machines, and it may be programmed to accept digital tapes from any type of CT machine and display them in a common format.