How many microns are you throwing away?

Frozen Section Quality · QA Protocol

How many micron are you throwing away?

Barry
September 11, 2026
Uncategorized

Every Mohs lab measures something. Turnaround time, stage count, slide volume, redo rate. Almost none of them measure the one number that sits upstream of all of those: how much tissue the lab destroys before it produces a single complete section.

Call it facing depth — the total thickness removed from the specimen block between first contact with the blade and the first section that shows the entire margin. It is trivially measurable with equipment you already own, it takes about thirty seconds per case, and it is the closest thing a Mohs lab has to a direct read on embedding quality.

This is a protocol for measuring it, charting it, and knowing what to do when the number is bad.

At a glance: measure trim passes × trim thickness · published benchmark 304 µm conventional vs 104 µm flattened · epidermal margin representation target 95% · section target 4–6 µm · chart monthly by tech and by site

1. What the number actually is

Facing depth is arithmetic, not estimation:

Facing depth (µm) = (trim passes × trim thickness setting) + (discarded sections × section thickness setting)

The cryostat already tells you both settings. A block trimmed at 30 µm for eleven passes, then sectioned at 5 µm with three sections discarded before a complete one appears, has consumed 330 + 15 = 345 µm. That is a third of a millimeter of margin gone before the diagnostic surface arrives.

Two definitions have to be fixed in writing before anyone starts counting, or the measurement will not be comparable between techs.

“First complete section” means: a single section containing an uninterrupted epidermal edge around the full circumference of the specimen, plus a continuous deep margin, with no gaps requiring the surgeon to infer coverage from an adjacent level. Not “good enough to read.” Complete.

“First contact” means: the first pass that removes tissue, not the first pass that removes embedding medium. Passes that cut only OCT do not count.

Write both definitions on a card and tape it to the cryostat. Ambiguity here is the main reason facing-depth audits fail.

2. Why this metric and not another

Because the published magnitude is large, and because it moves with technique.

A comparison of three tissue preparation methods reported in Virchows Archiv found that with a conventional chuck, specimens required an average of 304 µm of sectioning before the deep margin was cleared — and four out of five specimens still failed to contain 95% of the epidermal margin at that depth. A purpose-built flattening device brought the requirement to roughly 104 µm while achieving 95% epidermal margin representation.²

Read that as a ratio rather than as two absolute numbers. Preparation method accounted for roughly a threefold difference in tissue consumed, on the same kind of specimen, to reach the same diagnostic goal. That is not operator variance. That is method.

And the consumed tissue is not neutral. In a case-control review of recurrent tumors, findings associated with recurrence included visible remaining tumor (23%) and missing epidermal or dermal tissue (23%), alongside dense inflammation at a final margin where tumor had been present on prior slides (27%) — with tissue drop-out and cases requiring three or more stages emerging as significant factors.⁴ Tissue faced away is tissue that cannot be examined. When the epidermal edge is the part that goes missing, the audit trail reads as a lab failure, not a surgical one.

So: facing depth is a leading indicator for margin completeness, stage count, chair time, and a measurable share of the recurrences that come back to the practice years later. Turnaround time is a lagging indicator of everything except how fast people are moving.

3. The measurement protocol

Six steps. Nothing here requires new equipment.

Step 1 — Fix the settings. Trim thickness and section thickness are recorded per case, not assumed. If techs change trim thickness by feel mid-block, the arithmetic breaks. Standardize the trim setting across the lab and record any deviation.

Step 2 — Count from first tissue. The tech counts trim passes from the first pass that removes tissue. A hand tally counter on the cryostat shelf works better than memory.

Step 3 — Count discarded sections. Once at section thickness, count sections taken and discarded until the first complete one, as defined above.

Step 4 — Record the multiplication. Passes × trim setting, plus sections × section setting. One line on a sheet, or one column in a spreadsheet.

Step 5 — Record the four covariates. Facing depth is meaningless without them:

Anatomic site — nose, ear, eyelid, lip, trunk, scalp. Curvature drives difficulty.
Specimen size, longest dimension.
Tissue character — fatty, fibrous, cartilage-bearing, thin and friable.
Tech initials.

Step 6 — Record the failure mode when there is one. If the first complete section never arrives and the block is re-embedded, log it as a re-embed with a reason rather than as a facing depth. Re-embed rate is its own metric, and it is arguably the more damning one.

Thirty seconds per block. Run it on every stage for two weeks, then drop to a sampled audit — one full day per month, all cases, all techs — once you have a baseline.

4. What good looks like

Here is where honesty matters. The published figures above describe a specific study population under specific conditions. They are not a validated national benchmark, and no such benchmark exists. What follows is a working standard anchored on those figures, offered as a starting point for a lab that has never measured this. Your own baseline, once you have four weeks of data, is a better target than anything printed here.

Median facing depth	Reading	What it usually means
Under ~100 µm	Strong	Flattening is controlled and reproducible
~100–150 µm	Acceptable	Working range for most sites and specimen types
~150–300 µm	Investigate	Embedding technique is the likely driver, not the tissue
Over ~300 µm	Act	At or past the published conventional-chuck figure; method problem

Three cautions on reading that table.

Use the median, not the mean. A handful of cartilage-bearing ear cases or one 3 cm trunk specimen will drag an average upward and hide the ordinary work.

Segment before you conclude. A curved nasal ala and a flat trunk specimen are not the same measurement problem. Compare like to like, or the chart tells you nothing.

The spread is the finding. A lab with a median of 120 µm and a range of 80–160 µm is in control. A lab with a median of 120 µm and a range of 40–600 µm is not — it is averaging a good method and a bad one. Chart the spread, not just the center.

5. Charting it monthly

One chart, four series, reviewed at the same meeting every month:

Median facing depth, all cases — the headline trend line.
Median by tech — training signal. Persistent separation between two techs on comparable specimens is a technique difference, and technique differences are teachable.
Median by anatomic site group — tells you where to concentrate effort. The difficult sites are usually where recurrence risk is highest anyway.
Re-embed rate — blocks that never produced a complete section and had to be restarted.

Two derived numbers belong beside the chart. First, the 95th percentile, because the tail is where margin loss actually happens. Second, stages per case, plotted on the same timeline — if facing depth falls and stages per case falls with it, you have found something real rather than a measurement artifact.

Review it against the same three questions each month: Did the median move? Did the spread narrow? Did anything in the tail get explained?

6. Reading the chart

Pattern	Likely cause	First thing to check
High median, tight spread	Systematic method problem across the lab	The embedding procedure itself — everyone is doing the same thing
Normal median, long right tail	Specific case types defeating the method	Segment by site and tissue character
Split by tech, same specimen mix	Technique variance	Direct observation; one tech has a step the other lacks
Rising over months	Drift, or a change nobody logged	New chuck lot, new OCT, new freeze protocol, new hire
Falling median, rising re-embed rate	Techs stopping early	Verify the “complete section” definition is being applied

That last row is the one to watch for. A facing-depth program run without a hard definition of completeness creates an incentive to call an incomplete section complete — not maliciously, but reflexively. If the median drops while re-embed rate or stage count rises to meet it, the metric is being gamed. The written definition and the re-embed rate are the controls that prevent it.

Pattern	Likely cause	First thing to check
High median, tight spread	Systematic method problem across the lab	The embedding procedure itself — everyone is doing the same thing
Normal median, long right tail	Specific case types defeating the method	Segment by site and tissue character
Split by tech, same specimen mix	Technique variance	Direct observation; one tech has a step the other lacks
Rising over months	Drift, or a change nobody logged	New chuck lot, new OCT, new freeze protocol, new hire
Falling median, rising re-embed rate	Techs stopping early	Verify the “complete section” definition is being applied

7. When the number is bad

Work the list in this order. It runs cheapest-first, and the early items resolve more cases than lab folklore suggests.

1. The definition. Confirm everyone is measuring the same thing before changing anything. A meaningful share of first-audit surprises are definitional.

2. Flattening technique at the bench. Whether the epidermal edge, lateral margin, and deep margin are brought into one plane before freezing — rather than corrected afterward by cutting deeper — is the single largest lever. A specimen that goes onto the chuck already curved cannot be rescued at the microtome. It can only be consumed.

3. Freezing method and rate. Slow freezing produces ice-crystal artifact, which drives sections to be discarded for reasons unrelated to geometry and inflates the count. Rapid freezing — heat sink, isopentane, controlled spray — is the fix, and it is separate from the temperature dial.¹

4. Temperature by tissue type. A cryostat set once and left there is a standing source of discarded sections. Skin sections cleanly in the −25 to −30 °C range; cartilage shatters when very cold and prefers −16 to −22 °C; adipose needs longer freeze times and a deliberately thicker section.¹

5. Section thickness discipline. The working target is 4–6 µm, set by the roughly 8 µm thickness of epithelial cells. Cutting thick to get a section to hold together trades a geometry problem for an interpretive one: in the concordance literature, the most common category of disagreement between Mohs surgeons and dermatopathologists was the false positive, attributed largely to thick sections and to inflammation over-read as malignancy.¹³ Thick sections generate stages.

6. Blade condition. A damaged edge produces tears that look like tissue defects, and a tech chasing a “defect” that is actually a blade artifact will keep cutting.

7. Hardware. Chuck geometry and dedicated flattening equipment sit at the bottom of this list because they cost money, not because they matter least — the Virchows Archiv comparison puts method, hardware included, at the center of the difference.² Once steps 1 through 6 are clean and the median has not moved, the remaining variance is in the tools.

8. Making it stick

A number nobody owns stops being collected by week three. Three things keep it alive.

Put it in the competency file. Facing depth is one of the few Mohs histology skills that produces an objective, trendable number. A new tech’s curve over their first ninety days is a better competency record than a checklist signature.

Review it where decisions get made. If the chart only appears when something has gone wrong, it becomes a disciplinary instrument and the data quality collapses. Standing agenda item, good months included.

Report it upward in the surgeon’s units. Microns do not move a practice conversation. Stages per case, minutes of chair time, and re-excision rate do. Facing depth is the mechanism; those are the outcomes it predicts.

The short version

You cannot improve margin completeness by asking people to be more careful. You improve it by measuring how much tissue disappears before the first complete section, segmenting that number by tech and by site, and working a fixed list when it looks wrong.

The published difference between preparation methods is roughly threefold on the same specimens.² A lab that has never measured its own facing depth does not know which side of that ratio it is on.

Measure for two weeks. You will know by the end of the first one.

A note on sources

Every figure in this article is a paraphrased summary of findings reported in the cited sources. No text, tables, or figures from any source have been reproduced. Quantitative results are restated in our own words for commentary and educational purposes, with full attribution and a direct link to each original so readers can consult it firsthand.

Two clarifications are specific to this post. First, the 304 µm and 104 µm figures attributed to Virchows Archiv² are summarized from the published abstract, which sits behind a publisher paywall; we have not reproduced any portion of the article text, and readers who intend to rely on those numbers should obtain the paper directly from Springer.

Second, and more important for a protocol piece: the performance bands in Section 4 are our own working standard, not a published or validated benchmark. They are anchored on the figures above for orientation only. No professional body has issued a facing-depth standard for Mohs frozen sections, and we are not presenting one. Treat the bands as a starting scaffold and replace them with your own baseline as soon as you have four weeks of data. The measurement procedure in Section 3 is likewise a method we are proposing, not a cited protocol.

Citations to StatPearls¹ and to open-access articles hosted on PubMed Central³ are linked to the freely available versions. If you are a rights holder and believe any characterization here is inaccurate, contact us and we will correct it.

References

Ibrahim M, et al. Mohs Micrographic Surgery Interpretation of Technical Problems Impacting Slide Quality. StatPearls, NCBI Bookshelf. ↩↩↩↩
Kimyai-Asadi A, et al. Tissue preparation for Mohs’ frozen sections: a comparison of three techniques. Virchows Archiv, 2007. Paywalled — abstract only. ↩↩↩↩
Concordance rates among dermatopathologists and Mohs surgeons in frozen section Mohs slides: a systematic review. PubMed Central. ↩↩
Campbell T, et al. Surgeon error and slide quality during Mohs micrographic surgery: is there a relationship with tumor recurrence? Journal of the American Academy of Dermatology, 2013. ↩

Edit“How Many Microns Are You Throwing Away? A Facing-Depth QA Protocol”