8.2 In Situ Hybridization (FISH, CISH, SISH)
Key Takeaways
- DNA probes report gene copy number or rearrangements in the nucleus; RNA probes report transcripts; immunohistochemistry reports protein—CISH is not a protein stain just because it is brown.
- FISH uses fluorescent probes and a fluorescence microscope; CISH uses hapten probes plus enzyme-chromogen detection on a brightfield microscope; SISH deposits silver for stable black nuclear dots, also brightfield.
- ASCO/CAP 2018 dual-probe HER2 ISH groups combine HER2/CEP17 ratio and mean HER2 copies per cell: group 1 (ratio ≥2.0 and copies ≥4.0) is ISH-positive; group 5 (ratio <2.0 and copies <4.0) is ISH-negative; groups 2–4 need additional workup, not a dumped full algorithm.
- Protease (often pepsin) opens chromatin for the probe; overdigestion yields hollow, moth-eaten nuclei, lift-off, and lost signals; underdigestion yields autofluorescence or weak hybridization.
- ISH slides still need invasive-tumor morphology, documented pretreatments, and controls; a technologist does not reclassify a group 4 tumor as group 1 by counting only the brightest nuclei.
8.2 In Situ Hybridization (FISH, CISH, SISH)
Quick Answer: ISH hybridizes a labeled nucleic-acid probe to DNA or RNA in the tissue. FISH is fluorescent and needs a fluorescence microscope. CISH is chromogenic and brightfield. SISH deposits silver and is also brightfield. HER2 dual-probe assays use both a HER2/CEP17 ratio and mean HER2 copy number (ASCO/CAP 2018 groups). Protease opens the nucleus; pepsin overdigestion hollows nuclei and loses signal. CISH is gene (or transcript) detection, not a substitute protein IHC.
This OpenExamPrep section is independent teaching on in situ hybridization covering published QIHC staining topic areas, including FISH, CISH, and SISH. It is not an ASCP publication and does not claim Board, ASCO, or CAP approval. HER2 group numbers below come from published ASCO/CAP breast HER2 testing guidance (2018 focused update), taught at technologist level—not as a pathologist's full sign-out algorithm.
What ISH is doing on an FFPE slide
Immunohistochemistry binds antibody to protein. In situ hybridization binds a probe to a complementary nucleic-acid sequence that is still in the cell. The section stays on glass, so you can see which nuclei (or which cytoplasmic RNA clouds) carry the target—unlike a tube PCR that destroys tissue architecture.
DNA probes detect nuclear gene sequences: amplification (extra copies), deletion, or a rearrangement when two differently labeled probes split or fuse. HER2 dual-probe assays are the QIHC-relevant example: one probe to the HER2 (ERBB2) locus on 17q, one probe to a chromosome 17 centromeric sequence (CEP17) as a copy-number reference.
RNA probes detect transcripts. RNA ISH can look more like a protein stain in distribution (cytoplasmic or nuclear RNA), but the target is still nucleic acid. RNA is more labile than DNA: RNase-free technique, prompt fixation, and avoidance of RNA-destroying pretreatments matter. A tumor can show extra HER2 DNA copies with modest protein, or (less often for HER2) protein expression without classic amplification—the assays are not interchangeable.
Probes may be labeled directly with a fluorophore (classic FISH) or with a hapten (digoxigenin, DNP, biotin) that an antibody-enzyme step later visualizes (CISH and most SISH). Hapten detection is why CISH looks like IHC under the brightfield microscope even though the first binding event was hybridization, not antibody-to-protein.
FISH, CISH, and SISH as staining methods
| Method | Probe label / detection | Microscope | Morphology while scoring | Signal permanence | Typical QIHC use |
|---|---|---|---|---|---|
| FISH | Fluorophore on DNA (or RNA) probes; DAPI nuclear counterstain | Fluorescence, dark room, oil immersion, filter cubes | Nuclear outlines by DAPI; tissue architecture is harder | Fluorescent signals fade; store dark, limited re-review | HER2 dual-probe, other gene copy or break-apart assays |
| CISH | Hapten probe → antibody-HRP or AP → chromogen (often DAB brown or a red AP chromogen) | Brightfield | Hematoxylin counterstain; invading glands remain visible | Chromogen is generally stable | HER2 and other copy-number assays when fluorescence is unavailable |
| SISH | Hapten probe → HRP → silver deposition (black nuclear dots) | Brightfield | Morphology retained; silver is dense and fine | Silver is highly stable | Automated dual HER2/CEP17 silver ISH as a platform example |
FISH remains the reference hybridization method in many HER2 concordance studies. It requires a fluorescence microscope, trained scoring, and attention to fade, overlapping nuclei, and crush. Dual-color FISH (HER2 one color, CEP17 another) is how ratio and copy number are both obtained.
CISH was developed so a laboratory can read gene copies on the same brightfield microscope used for IHC. You see brown or red dots in nuclei plus hematoxylin. You do not need a fluorescence room. You also do not get multi-fluorophore flexibility as easily as FISH.
SISH is metallographic: HRP catalyzes silver deposition at the probe site. Dots are black, high contrast, and archival. Dual SISH (HER2 one silver or chromogen color, CEP17 another, depending on the kit) is an example of automated brightfield HER2 ISH—not the only valid platform.
Do not treat CISH or SISH as "IHC with extra steps." The first specific step is still hybridization after denaturation. Antibody in CISH/SISH is detecting the hapten on the probe, not HER2 protein.
HER2 dual-probe FISH groups at technologist level
ASCO/CAP breast HER2 guidance (2018 focused update) classifies dual-probe ISH using two numbers together: the HER2/CEP17 ratio and the average HER2 signals per cell. Learn that both exist. Do not memorize every concurrent-IHC recount branch as if this were a pathology boards dump.
| Group | HER2/CEP17 ratio | Mean HER2 copies per cell | Technologist-level meaning |
|---|---|---|---|
| 1 | ≥ 2.0 | ≥ 4.0 | ISH-positive pattern |
| 2 | ≥ 2.0 | < 4.0 | Unusual; not a standalone "positive"—additional workup (concurrent IHC) |
| 3 | < 2.0 | ≥ 6.0 | High copy number without ratio ≥ 2.0; additional workup |
| 4 | < 2.0 | ≥ 4.0 and < 6.0 | Former "equivocal" window; additional workup |
| 5 | < 2.0 | < 4.0 | ISH-negative pattern |
Group 1 is the classic amplified dual-probe result: ratio at least 2.0 and at least four HER2 signals per cell on average. Group 5 is the classic non-amplified result: ratio below 2.0 and fewer than four HER2 copies per cell. Groups 2–4 are uncommon patterns that ASCO/CAP directed into concurrent IHC and, when IHC is 2+, a blinded ISH recount—not into a technologist forcing the count into group 1 by selecting only the brightest nuclei.
Single-probe HER2 ISH (no CEP17) uses copy-number cutoffs (classically ≥ 6.0 copies as positive and 4.0 to < 6.0 as a historically equivocal window). Dual-probe is the usual clinical assay. HER2-low is an IHC reporting construct (1+ or 2+ with ISH-negative), not a sixth FISH group. IHC 0 / 1+ / 2+ / 3+ protein scoring is a different assay (preliminary screening and predictive-marker chapters).
Technologist duties that actually affect the group: hybridize only invasive carcinoma when that is the requested target (do not score DCIS as if it were the invasive tumor), count non-overlapping nuclei with intact borders, document average signals and ratio, run on-slide controls, and escalate heterogeneous or unscorable slides rather than averaging crush artifact into a false ratio.
Protease pretreatment and pepsin overdigestion
FFPE nuclei are protein-packed. After dewax and, typically, a heat or acid pretreatment specified by the kit, a protease (often pepsin; sometimes proteinase K) nicks the protein mesh so the probe can reach DNA or RNA. Digestion is a timed, temperature-controlled enzyme step. Lots differ in activity. A validated 10 minutes at 37 °C is not "until it looks done."
Underdigestion: probe cannot reach target. FISH shows autofluorescence, weak or absent dots, and a protein haze. CISH/SISH shows pale or no nuclear dots. The temptation is to blame the probe or the hybridizer.
Overdigestion (classic pepsin over-run): nuclei look hollow, moth-eaten, or empty; nuclear borders dissolve; tissue lifts; DNA (or RNA) is lost; remaining signals are few, faint, or outside ruined nuclei. You cannot score a group 1 versus group 5 on a digested ghost. Shorten time, lower temperature, or dilute pepsin per the SOP; do not "rescue" by counting whatever dots remain at the edge of a hole.
Morphology check after digestion, before expensive probe, is a QIHC-practical habit: nuclei should be crisp enough to count, not fried and not still rubbery. ISH protease is not the same recipe as IHC enzyme-induced epitope retrieval, even if both use pepsin or proteinase K. Do not borrow an IHC EIER time for a FISH kit.
CISH versus IHC: gene versus protein
A brown nuclear CISH dot is hybridized probe, usually DNA copy number. A brown HER2 IHC stain is protein on the membrane (complete intense circumferential staining in the 3+ definition). They can disagree: amplification without much protein, protein without amplification, or a 2+ IHC sent to ISH exactly because protein was equivocal. That disagreement is why laboratories run both, not why CISH "replaces" IHC.
Practical traps:
- Reading CISH as if it were HER2 IHC 3+ membrane scoring.
- Calling SISH silver dust "necrosis pigment" without checking nuclear localization.
- Using an IHC polymer detection kit as if it were a CISH detection kit without the hybridization/denaturation program.
- Treating a faded FISH slide as a negative group 5.
Worked example
A breast core with HER2 IHC 2+ goes to dual-probe FISH. Mean HER2 copies are 5.2, CEP17 copies 4.8, ratio 1.08. That is group 4 (ratio < 2.0, copies ≥ 4.0 and < 6.0), not group 1. The technologist documents the counts and does not re-score only the two brightest nuclei to push the ratio over 2.0. Concurrent IHC workup is the published pathologist-integrated next step. A parallel CISH on the same block, if used, is still a gene assay; a 3+ membrane IHC on a different block would still be protein. If pepsin was left on twice as long and nuclei are hollow, the correct action is a new section and a correct digestion—not a group assignment on destroyed DNA.
What is the essential difference between HER2 immunohistochemistry and HER2 chromogenic in situ hybridization on the same FFPE breast tumor?
After pepsin pretreatment for HER2 FISH, nuclei look hollow and moth-eaten, tissue is lifting, and fluorescent signals are few or absent. What is the most likely cause?