Bibliogr. s. 70

Background: The aim of this work was to develop a method of volume determination utilizing the grey scale histogram of three-dimensional (3D) ultrasonic (US) image. Volumes are calculated as the product of single voxel volume and the area of histogram peak representing investigated object. The proposed solution performance was compared with two other methods. First utilizes two-dimensional crosssection areas on subsequent image layers while the second allows the volume determination on the basis of one-dimensional measurements. Material/Methods: The 3D US phantoms images were used to test the procedure. The usefulness of the method was also demonstrated on several clinical examples. The ultrasonic 3D images were collected, their histograms calculated and fitted with model curves allowing the volume calculations. The accuracy and precision was assessed and t-test was used for evaluating performance of all considered methods. Results: The accuracy (understood as the difference between real and measured volume) achieved in the proposed solution (3.6%) was the highest comparing to alternative methods (5.2% and 8.4%). Also the p-value (two-tailed t-test) was better in the case of the presented method. The performance increase was due to the elimination of subjective delineation of measured regions of interest. Conclusions: The presented procedure can be successfully used for volume assessment concerning its simplicity, accuracy and time consumption. It should be emphasized that the new method does not require image segmentation, unlike other methods in use. The procedure was tested with 3D US imaging but can be used successfully with any 3D imaging modality.