Fluorescence in situ hybridization (FISH) is a useful technique for analyzing nucleic acid sequences at the individual cell level. However, the application of FISH to solid tumor specimens is currently performed either on disaggregated cells, on touch preparations or on thin (4-6 mm) thick sections. Analysis of disaggregated cells suffers from the fact that the cellular organization of the tissue is lost. Touch preparations are biased in favor of those cells which most easily detach from the specimen, and are thus not representative of the cell population. Thin sections do not contain intact nuclei thus the number of FISH signals per nucleus may be underestimated. A solution to these limitations is to use thicker (20 mm) sections, because they contain intact nuclei from which FISH signals can be accurately enumerated. FISH labeling techniques for 20 mm sections followed by confocal imaging has already been reported1 and recently we reported 3D image display and interactive image analysis algorithms that make it possible to enumerate FISH signals in individual nuclei of thick sections 2.

The analysis algorithms for FISH signal enumeration first made maximum-intensity projection images through the acquired confocal images at different angles. When these projections were viewed sequentially, the original, confocal 3D image appeared semi- transparent and rotating. The analyst interactively marked FISH signals in the projection images with the computer's mouse. After marking all signals, several images of each pair of FISH signals closer than an analyst-defined distance (typically 1.5 nuclear diameters) were displayed. The most useful of these images was a 2D slice cut from the confocal image, where the x-axis of the slice was the line in the (x,y) plane of the confocal image joining the two signals and the y-axis of the slice was the z (depth) axis of the confocal image. From this slice, it was generally possible to tell if the signals were in the same or different nuclei and if the nuclei had been truncated by the tissue sectioning. The analyst told the algorithm (via mouse interaction with a menu) if the signals were in the same nucleus, different nuclei, or if one or both signals should be rejected. After consideration of all signal pairs, the algorithm overlaid the number of FISH signals in each intact nucleus on a projection image.

These algorithms have been used to enumerate chromosome specific FISH probes in breast tumors, because changes in chromosome copy number (aneusomy) are frequent in breast cancers and are potential prognostic markers. As a control, we enumerated the signals in normal skin sections labeled with a probe for the centromere of chromosome 8, where two signals per intact nucleus would be expected. For the control, over 80% of the intact nuclei reported two signals. This result was compared with the result that would have been obtained from a thin section of the same specimen by extracting the central slices from the acquired confocal image that represented a 4 mm section. Less that 40% of these nuclei reported two signals and the remaining nuclei reported one or zero signals.

In conclusion, these enumeration algorithms enabled analysis of small lesions in intact tissue sections that could not be analyzed by other molecular techniques and permitted the direct correlation of FISH information with tissue histology and cytology.

___________

This work was supported by a Biomedical Engineering research grant from the Whitaker Foundation, and by the Director, Office of Energy Research, Office of Basic Energy Sciences, of the U.S. Department of Energy under Contract No. DE-AC03- 76SF00098.