Table 2 General guidelines for FISH deletion assays using FFPE tissue sections

FISH test validation

It should be performed according to the most appropriate Standards and Guidelines for Laboratories [22–24, 40, 41], and for monitoring and reporting data [42]. There are a number of reviews that address clinical applications of FISH [43, 44].

Standard controls

The laboratory should periodically check assay performance (including control probes) as part of quality monitoring. Monitoring FISH testing over time to assess adverse technical trends is also recommended.

Analytical standards

Assessment of several normal metaphase cells should be considered for validation that the correct probe was used for the study:

● In typical analytic validations, the FISH probe is hybridized to metaphase and interphase cells from peripheral blood cultures of five karyotypically normal control males.

● For each specimen, the number of FISH signals in at least 50 consecutive interphase cells is recorded, and then the hybridization sites in 20 metaphase cells are identified by banded chromosome morphology.

● The analytic sensitivity and specificity for metaphase cells, and the percentage of nuclei that meet the signal pattern criteria for normal cells are calculated as described [23].

● This evaluation also ensures that there are no background signals or cross-hybridizations to related genes that could be misinterpreted in interphase FISH tissue section analyses.

Establishment of cutoff values

The cutoff levels to be used to identify a sample as deleted should be established as part of the FISH test validation for the laboratory.

● The cutoff value used is established by analyzing a reference panel of histological tissue sections from normal healthy cases.

● The use of suitable normal control tissue with similar sized nucleus to the target tissue being analyzed can help to establish the expected percentage of signal losses due to signal truncation artefacts. In this context, setting up a normal database for each probe being used in the laboratory is suggested.

● The laboratory’s cutoff database should address each type of target tissue and it should identify the thickness of samples used for FISH (the same thickness should be maintained for all specimen testing). Wiktor et al. [23] published an excellent clinical FISH validation approach, which describes a cutoff method to establish an analytic sensitivity with a 95% confidence level. The ECA guidelines also discuss various approaches to establishing robust cut-offs [40].

● Monitoring and revising existing cutoff values should also be considered as probes used and test approaches change with time.

Positive and negative controls

If possible run positive and negative controls. For FISH deletion assays these will be samples with known homozygous, hemizygous and undeleted copies of the tumor suppressor gene of interest.

Tissue section thickness

The standard thickness of unstained tissue sections is between 5 - 6 microns. The tissue sections must be applied to a positively charged or silanized slide to minimize the incidence of tissue sections falling off the slides during processing.

Tissue sections less than 5 - 6 microns will result in an increased truncation of nuclei affecting established cutoff values, and a greater thickness will result in poor probe hybridization.

Hematoxylin–eosin slides

An hematoxylin–eosin slide must accompany unstained slides for FISH test with the diseased area circled in either pen or felt marker. Designated areas should be representative of viable tumor regions, with necrosis, hemorrhage, poor fixation, or histological artifacts being excluded.

General considerations for FFPE FISH analysis

Use optimal filter set for the deletion assay probe combination and check there is no bleed-through between different filters.

Review slides for hybridization performance. There should be >85% efficiency with minimal non-specific noise.

Pre-screen the tumor area selected by the pathologist using an adjacent hematoxylin and eosin section map for the following features:

● The area is tumor rich

● Nuclei have a regular shape and uniform DAPI staining

● Nuclei do not have evidence of digestion damage such as ‘doughnut- like’ appearance with empty epicenters

● Nuclei should not be covered by a cloudy typically yellowish layer or obscured by auto-fluorescent structures.

● Nuclei have hybridization signals with uniform intensity and similar patterns of granularity.

Preparations not meeting these criteria should not be used for signal enumeration.

Ensure that the entire selected area of tumor has been pre-screened carefully before selecting nuclei to score. Sometimes a small area containing a clonal deletion may be missed without this pre-screen.

Only examine nuclei that are distinct and ideally separated from each other. Select cells in which the borders of individual nuclei can be clearly distinguished. Avoid scoring nuclei that are crowded, overlapping, or distorted.

When selecting nuclei focus up and down on the z-axis and make sure the entire volume of the chosen nucleus is present inside the section and that the FISH signals at all focal planes are enumerated. A bias in distribution to the upper or lower face of the section may indicate truncation.

Signals may be either bright and compact oval shapes, split into two smaller but connected dots, or a stringy diffuse shape. Pay attention to the signal intensity.

The probes flanking the tumor suppressor gene can help distinguish between truncation losses and ‘real’ interstitial deletions. Sometime the deletion may be larger and include one (or both) flanking probe sets.

Scoring of probe signals

All scores should be entered onto score sheets in an unbiased manner. A routine FISH evaluation should be scored by two technologists.

All scores should be entered onto score sheets together with comments that may be relevant concerning heterogeneity, signal quality etc.

Score appropriate number of nuclei according to the Standards and Guidelines for Laboratories. When inconsistent results are obtained a third reader is required or additional nuclei should be scored based on the laboratory director’s guidance.

Be aware of the possibility of clonality of deletions (such as mixture of hemi-and homozygous deletion). All clones should be evident once appropriate number of nuclei has been scored.

Sometimes in complex cases there is more than one type of clone:

● Each clone should be scored individually (score appropriate number of nuclei for each clone) and the location of the clone marked on the hematoxylin and eosin section map.

● In such complex tissue where there is more than one type of aberration, each clone should be scored individually (ideally scoring 100 cells for each).

Once completed the scoring, re-scan the marked tumor area to ensure nothing has been missed.

Typical scoring results for tumor suppressor gene FISH assays will describe the % of normal cells, the frequency of homozygous and/or hemizygous deletions or monosomies. In addition, there may be a percentage of cells with ploidy alterations or gains of the chromosome.

The criteria for scoring deletion FISH should in general be developed for the tumor suppressor gene of interest after taking into account the previous experience of the laboratory and using data from the literature from other groups performing similar assays.

Quality control

Ensure that signals from all probes are present in normal surrounding tissue adjacent to tumor areas on all slides to confirm successful FISH hybridization.