Correlation of histology and acoustic parameters of liver tissue on a microscopic scale.

The correlation of several acoustic parameters with histological features was investigated in healthy White New Zealander rabbit liver (n = 10). Thin sections (10 microns) were studied by means of a light microscope in combination with a digital image processing system. Adjacent thick sections (250 microns) were studied by means of a custom-designed acoustic microscope. Markers of black silk suture material, that could be identified both optically and acoustically, enabled the spatial correlation of both imaging modalities. Acoustic images were reconstructed from the velocity of ultrasound, the attenuation at the central frequency (30 MHz), the attenuation spectral slope, the backscatter spectral slope and the backscatter at the central frequency. The measurements comprised the frequency range from 10 to 50 MHz, yielding a resolution of approximately 50 microns. From the thin sections (10 microns) the local histological composition was obtained by digital segmentation techniques. The features that were segmented are: the collagen rich fibrous tissue content, the area lumina, the area interstitial spaces, the number density of nuclei and the area parenchymal tissue. Correlation techniques revealed that the main feature responsible for attenuation is collagen. There was a fair correlation between area lumina and attenuation, but this was caused by a high correlation between the collagen that surrounds the lumina, and attenuation. No correlation was found between any histological feature and backscatter parameters or velocity.

Quantitative immunohistologic study of lip biopsies. Evaluation of diagnostic and prognostic value in Sjögren's syndrome.

In a group of 45 patients with Sjögren's syndrome (SS) and 80 controls the high specificity (95%) and sensitivity (100%) of a recently proposed bivariate quantitative immunohistologic (QIH) criterion for SS, based on percentages of IgA and IgG-containing plasma cells in labial salivary gland (LSG) tissue, was confirmed. The best univariate QIH criterion for discrimination between LSG biopsies of SS patients and controls appeared to be based on the percentage of IgA containing plasma cells, and had a specificity of 99% and a sensitivity of 96%. A criterion based only on the percentages of IgM-containing plasma cells, proposed in another recent study, resulted in a high number (31%) of false negatives. Interobserver reproducibility of QIH diagnoses was excellent. Moreover it was demonstrated that accuracy, precision and the interobserver reproducibility of plasma cell counting depends on the choice of tissue fixation and immunohistologic staining procedure. The combination of formol sublimate fixation and peroxidase anti-peroxidase procedure appeared to be the best combination for QIH examination. Furthermore, in 2 SS patients systemic monoclonal IgM/kappa gammopathy was preceded by high predominance of IgM and kappa containing plasma cells in the plasma cellular infiltrate of the LSG tissue.

Effects of tissue processing techniques in acoustical (1.2 GHz) and light microscopy.

In this study the influence of various tissue processing and staining techniques on the acoustical properties of liver tissue was investigated. A qualitative study was performed using ultrasound attenuation as the imaged parameter of a combined optical/acoustical microscope with a 1.2 GHz transducer. Images were made of three sets of adjacent liver sections (6 microns in thickness) which were prepared in ten different ways: fixed by alcohol or formalin; stained by hematoxylin-eosin (HE), toluidine blue (TB) or non-stained; sectioned by a cryostat or by a paraffin microtome. It was concluded that the images obtained from cryostat sections were of much higher quality than those from paraffin sections. Images obtained from sections that were sectioned while embedded in paraffin displayed no detail at all. No consistent effect was noticed with respect to staining by HE or TB. Alcohol fixed sections gave more detailed images than formalin fixed sections. Formalin fixation in combination with cryostat sectioning yielded many cytoplasmic vacuoles.

A new method to correlate acoustic spectroscopic microscopy (30 MHz) and light microscopy.

A powerful new method is used to investigate the correlation between light microscopic and acoustic properties of biological tissues. Specimens of liver were sectioned into successive slices, 250 micrometers and 10 micrometers thick. The thick sections were investigated acoustically, the thin sections by means of light microscopy. Markers that could be detected and located, both optically and acoustically, were used to find and reconstruct corresponding regions in the acoustic and optical sections (2.5 x 2.5 mm). Parameter images were reconstructed from the sections investigated acoustically. The acoustic parameters were attenuation at 30 MHz, the slope of the attenuation spectrum (between 10 and 50 MHz), backscattering at 30 MHz, the slope of the backscattering spectrum (between 10 and 50 MHz) and the local ultrasound velocity. Acoustic images were obtained in the frequency range from 10 to 50 MHz, yielding a lateral resolution of about 50 micrometers. The sections for light microscopy were stained according to the Goldner trichrome staining technique. The histological composition was determined quantitatively, using digital image segmentation techniques. The percentage of collagen-rich fibrous tissue, luminal structure and interstitial spaces, and the number of nuclei were calculated for regions of 250 x 250 micrometers. These histological features were correlated with the acoustic parameters obtained from the corresponding regions in adjacent sections. It was thus possible to find the histological components responsible for acoustic parameters.

Influence of fixation and immunohistological technique on accuracy, precision and inter-observer reproducibility of plasma cell counting.

Recently two highly sensitive and specific diagnostic criteria for Sjögren's syndrome based on percentages of IgA-, IgG-, and IgM-containing plasma cells measured in immunohistologically stained labial salivary gland tissue have been described. The reliability of such a criterion is dependent on the accuracy, precision and inter-observer reproducibility in plasma cell counting. The present study evaluates the effect of tissue fixation and immunohistological procedures on the aforementioned factors. Immunoglobulin (Ig)-containing plasma cells in sections of lamellated submandibular salivary gland tissue, alternately fixed in a 4% buffered formol solution or formol-sublimate solution and stained with an indirect immunoperoxidase and unlabelled peroxidase anti-peroxidase (PAP) method respectively, were enumerated by three independent observers. Relative numbers of Ig-containing plasma cells appeared to be less sensitive for systematic errors due to tissue fixation and immunohistological procedure than absolute numbers of Ig-containing plasma cells. The best inter-observer reproducibility of plasma cell counts was obtained in sections from formol sublimate-fixed specimens stained according to the PAP procedure.