5.3 Frozen Sections

5.3 Frozen Sections

Quick Answer: Frozen IHC starts from unfixed or briefly slide-postfixed tissue in OCT, not from a processed paraffin block. Snap-freezing limits ice-crystal holes; slow freezing in a warm cryostat does not. Frozen slides are often postfixed in cold acetone or in formalin. Acetone protocols usually skip or shorten HIER because epitopes were never formalin-masked. Lipid-rich tissues keep fat that paraffin processing would have extracted. Intraoperative frozen IHC is a speed-and-triage test. Research frozen IHC can be colder and slower but is still not FFPE. Compare frozen and paraffin results as different methods for the same antigen, not as copies.

This OpenExamPrep section is independent teaching on frozen-section specimen handling as it appears under Specimen Handling on the published QIHC topic outline. It does not claim ASCP or manufacturer approval of a cryostat brand or OCT lot. Chapter 4 covers paraffin microtomy. Chapter 6 covers full HIER chemistry. Here the point is why frozen tissue often should not receive that full HIER program.

Intraoperative versus research frozen IHC

Intraoperative frozen IHC exists to answer a question before the patient leaves the room or before the specimen is submitted entirely to processing: cytokeratin on a sentinel node, a rapid pituitary hormone stain, Hirschsprung-related stains in some pediatric workflows, or a limited lymphoma panel on a node that must be triaged for flow cytometry. Time is measured in minutes. Retrieval, if any, is abbreviated. Controls may be a previously frozen positive that you cut now. Morphology is cryostat morphology: acceptable for is there keratin-positive tumor in this margin, not for a 0–3+ HER2 score.

Research or scheduled frozen IHC (enzyme histochemistry, some muscle and nerve panels, antigens known to fail in paraffin) can use colder isopentane snaps, −80 C archives, and overnight drying. It is still frozen tissue. It still needs its own validation. It is not an excuse to skip controls.

Do not use intraoperative leftover OCT blocks as if they were predictive FFPE assays the next day without a frozen-method file. A 12-minute cytokeratin cocktail does not become a companion diagnostic because the chuck is still in the cryostat.

OCT compound

OCT supports the tissue on the chuck. Apply it in the cryostat, orient, freeze the mound, then face and cut (often about 4–6 µm for IHC; thicker cuts increase background and overlap). OCT is hydrophilic: it must not dry into a sugary crust on an uncovered block. Label the chuck; unlabeled frozen tissue is not recoverable by gross color.

Never melt OCT-infiltrated tissue into formalin and pretend the epitopes had a documented 6–72 hour NBF time. That hybrid has no ischemia clock and a freeze-thaw. OCT is also not a transport medium for DIF; Michel's/Zeus handling is Section 5.2. If tissue for frozen IHC must travel, it is usually snap-frozen, shipped on dry ice, and kept frozen—not soaked in OCT at room temperature.

Snap freeze versus slow freeze

Ice is the enemy of membrane continuity.

• Slow freeze (tissue placed on a room-temperature chuck, parked in a −20 C cryostat, or thrown onto dry ice as a large wet piece): water coalesces into large crystals. Crystals leave holes, shred epithelium, and break glomerular capillaries. Swiss-cheese skin cannot support a linear IgG call or a membrane stain.
• Snap freeze: small crystals. Practical methods include isopentane (2-methylbutane) cooled in liquid nitrogen. Liquid nitrogen alone can vapor-insulate the surface and freeze more slowly than a cold isopentane bath. Embed in OCT first or freeze and then OCT-mount, per SOP, but do not thaw to re-embed.

A large fatty breast excision will not snap-freeze like a 3 mm kidney cortex. Slice tissue thin. Wipe surface saline that forms ice armor. Repeat freeze-thaw cycles after the first snap are how a beautiful block becomes unreadable overnight.

Postfixation of frozen slides

Once cut, frozen sections are still unfixed protein films until you postfix. Common slide postfixatives:

Acetone is often used cold (refrigerated or −20 C) for a short, validated time. Over-long acetone can over-extract and flatten morphology. Formalin postfix of a frozen slide is not equivalent to overnight NBF of a core: penetration is from one cut surface, time is minutes, and a 6 µm section is not a 3 mm biopsy.

Air-dry briefly so sections stick, then postfix. Dripping-wet slides wash off in acetone. Charged slides help; they do not replace postfixation.

Why HIER may be omitted or shortened

HIER exists mainly to reverse formalin methylene bridges and to reopen epitopes after paraffin processing. Frozen tissue that never saw NBF does not have that mask. Full pressure-cooker high-pH retrieval on an acetone-fixed frozen slide commonly lifts the section, cooks remaining OCT residue into background, and destroys morphology that was already ice-fragile.

If the frozen slide was formalin-postfixed, a short retrieval (lower temperature, shorter time, milder pH) may be needed for a clone that was titered on FFPE. Copying the automated stainer's FFPE HER2 retrieval onto a frozen intraoperative slide is a common blank or washed-off result.

Enzyme retrieval (trypsin, proteinase K) is even harsher on frozen sections than on alcohol smears. Use only if that frozen method was optimized. Do not treat retrieval as a volume knob that always increases signal.

Lipid-rich tissue

Paraffin processing extracts lipids with alcohols and xylene. Frozen sections keep lipid. That is why Oil Red O and some enzyme histochemistry must be frozen. For IHC:

• Adipose, brain, adrenal cortex, and fatty breast can repel aqueous reagents, trap antibody, and show dirty background.
• Lipid droplets can autofluoresce if the same block is used for IF, or can look like holes if they wash out later.
• Acetone postfix extracts some lipid, which can help reagent access or can collapse the tissue.
• Do not expect a fatty frozen margin to look like an FFPE H&E of the same excision.

Wipe excess fat from the chuck face if it smears across the knife and carries tissue fragments onto the next slide. Contamination is a frozen-section specialty: yesterday's thyroid on today's parathyroid.

Same antigen: frozen versus FFPE

A clone can be brilliant on frozen and dull on FFPE, or the reverse. Concordance is a validation study, not a hope. Controls must match the method: a paraffin on-slide control does not prove the frozen run worked. If a frozen cytokeratin cocktail is used intraoperatively, the permanent FFPE cytokeratin the next day is still required for the medical record when that is the laboratory's practice. They are sequential methods, not duplicates.

Worked example. Intraoperative sentinel node: a frozen CK AE1/AE3 highlights a 1 mm cluster in 12 minutes, no HIER, acetone postfix. The next-day FFPE block is retrieved at high pH and stained with the same cocktail plus a more specific keratin. The frozen result triaged the case; the FFPE result is the archived method. Do not confirm HER2 on the leftover frozen OCT from that node using the tissue HER2 0–3+ rules.

High-yield traps

• Thawing and refreezing an OCT block to reorient it.
• Using a warm cryostat and calling the holes negative membrane staining.
• Applying full FFPE HIER to acetone-fixed frozen slides.
• Reporting predictive cutoffs from leftover intraoperative tissue.
• Cutting fatty tissue thick, then interpreting lipid holes as absence of antigen.