Feasibility of automating the micronucleus assay.

The results of a feasibility study on the automation of the micronucleus assay in whole blood cultures of human lymphocytes are reported. The assay requires determination of the number of lymphocytes with micronuclei among the proliferating population. Using an in-house-assembled image analysis system, a prototype software package was developed that addressed two problems: micronuclei identification and discrimination of nonproliferating cells from proliferating lymphocytes (the only ones that can give rise to micronuclei). The results of manual verification of automated micronucleus scoring showed that 70% of all digitized micronuclei were extracted from the images and 90% of them were correctly classified and paired with a parent nucleus by an "affinity function". The discrimination between proliferating and nonproliferating cells was carried out by linear discriminant analysis of simple nuclear features extracted from Feulgen-stained cells. Among the Feulgen-stained nuclei that were identified by autoradiography as proliferating or not, 85% were correctly classified by a six-feature discriminant function.

A modified harlequin stain for cell kinetics.

Lymphocytes cultured with 4 X 10(-4) M bromodeoxyuridine (BrdU) and stained by the harlequin procedure show wide variations in the colour and texture of their interphase nuclei. We demonstrate that we are able, by inspection, to identify four sets of nuclei: small dark nuclei (D) in cells that have not transformed in response to mitogenic stimulation; red nuclei (R) in transformed cells that have not synthesized DNA in culture; speckled nuclei (S)--red speckles on a bluish background--in cells that are synthesizing DNA for the first time in culture; blue nuclei (B) in cells that have completed one or more cycles of DNA synthesis. By using this modified harlequin stain to score the fraction of proliferating cells, and using the modified or conventional harlequin stain to determine the relative frequencies of mitotic cells in the first, second, or later divisions, we are able to estimate the fraction of lymphocytes in the original blood sample that can divide in culture. In addition, we show that the technique lends itself to automation. For this we have digitized black and white video images of the cells and extracted features based on the distribution of optical densities in the nuclei. Discriminant analysis carried out on these descriptors resulted in the correct classification of the three major groups R, S and B with an accuracy of 80.3%, 58.5% and 85.1%, respectively.

Attenuation correction in echocardiography.

A commercial two-dimensional echo unit, modified to permit digital acquisition (2.2 MHz) of echo signals prior to time-gain-compensation (TGC), was ratiometrically calibrated. Simulating the TGC in software, we demonstrated improvements in image quality as compared to conventional video methods. A path-dependent attenuation correction (PDAC) algorithm, utilizing the gray level image statistics to assign one of three attenuation coefficients (for chest, myocardium or blood) to each image pixel, was then developed. Using it on left ventricular (LV) short axis images obtained in nine healthy closed chest dogs, we demonstrated that the backscatter intensity at end diastole (ED) was 2.0 +/- 0.5 dB (mean +/- SEM) higher (p less than 0.01) than at end systole (ES), in circumferential segments of the myocardium oriented anteriorly and posteriorly, relative to the transducer. In seven of the dogs, subsequently occluded for one hour in the anterior descending branch of the left coronary artery (LAD), this normal phasic myocardial backscatter (PMB) variation decreased or was reversed in ischemic segments, relative to preocclusion values, suggesting utility of the method for sequential study of myocardial ischemia and its treatment.

DNA content of micronuclei in human lymphocytes.

We have calculated the distribution of DNA contents in micronuclei (MN) induced by ionizing radiation in human lymphocytes on two assumptions: the MN arise from acentric chromosome fragments (ACF), and the ACF result from the random breakage and rejoining of chromosomes. Measurements show that about 80 per cent of MN have a DNA content in the range of 0.5-6 per cent of the G1 nucleus. This group is consistent with the model and shows little dependence on radiation dose over the dose range of 0.5-4 Gy, or on lymphocyte culture time, varying from 48 to 76 hours. The MN with DNA content from 6 to 20 per cent of the G1 nucleus are probably the result both of spindle defects and of DNA synthesis in MN.

An improved micronuclear assay in lymphocytes.

We have found that by increasing the bromodeoxyuridine (BUdR) concentration to 4 X 10(-4) M in cultures and using the conventional harlequin staining procedure we can distinguish the proliferating lymphocytes by their blue color from the non-proliferating cells which have red nuclei. The micronuclei (MN) are also blue and are associated with the proliferating cells. The high BUdR concentration does not appear to affect the yield of MN. By scoring the MN only in the (blue) proliferating cells the MN assay is made less sensitive to perturbations in the lymphocyte cultures caused by radiation or by variations in the culture time on PHA concentration. Moreover, the frequency of cells with MN approximately doubles and thus comes into better agreement with the frequency of cells with chromosome aberrations. A better agreement yet can be achieved by correcting the assay for the effect of cell division.

Gray scale of ultrasound instruments: evaluation using a test object.

B-mode scans of castor oil in a wedge-shaped container provide gray level maps. The relation between the gray levels and the amounts of absorbing medium can be used to measure the effects of electronic gain, transducer characteristics, and the overall gray level performance of different ultrasound systems. We report results from four units and ten transducers for varying scanning depths.