A 60-cell cell block and eight requested predictive tests
A 62-year-old man with a 30-pack-year smoking history presents with a 3.2 cm right upper lobe mass, bulky subcarinal adenopathy and a single 8 mm brain lesion. EBUS-TBNA of the subcarinal node yields one H&E-stained slide and a 1 mm cell block containing roughly 60 viable tumour cells in a single ribbon. The H&E shows a poorly differentiated carcinoma with focal squamous eddies and rare mucin vacuoles. The treating oncologist has requested TTF-1, p40, PD-L1 (22C3), EGFR, ALK, ROS1 and a 50-gene DNA/RNA next-generation sequencing panel, and wants results within seven working days for a tumour board. The specimen was fixed in 10% neutral buffered formalin for nine days, and the block is nearly exhausted.
1. Before any further sections are cut, how do you allocate the remaining tissue among the requested tests? Give your sequence with reasons, and state what you would defer or decline.
Read the worked answer
Governing principle: the block is a finite, non-renewable resource. Each 4 µm section removes cells that cannot be recovered, so cutting must be planned as a single batch rather than serially as requests arrive.
Step 1 - protect the diagnosis. Review the H&E and confirm that the cell block really shows carcinoma; if the diagnosis rests on the smear alone, say so.
Step 2 - tier the requests by therapeutic consequence and tissue cost.
- Tier 1 (one section each, same run): p40 and TTF-1, ideally as a two-colour duplex on a single slide; PD-L1 22C3 on one charged slide.
- Tier 2: EGFR by real-time PCR, which needs one section with ≥20% tumour cellularity - cheaper in tissue than a 50-gene panel.
- Tier 3: NGS. Sixty cells fixed for nine days gives low DNA yield and fragmented RNA; failure probability is high and a failed run consumes the block without producing data.
Step 3 - stop cutting while reserve remains. Keep at least 5 unstained sections plus the block for future needs, and record the number of sections consumed.
Step 4 - communicate. Tell oncology in writing that the specimen is at its limit, that NGS has been deferred rather than refused, and that a repeat biopsy or liquid biopsy should be planned now rather than after a failed run. Documenting 'insufficient material for molecular testing - recommend re-biopsy' is a clinically actionable result, not a clerical note.
2. The tumour is TTF-1 negative and diffusely p40 positive, with rare mucin vacuoles. What is your diagnosis, and how does it change the way you handle the molecular requests?
Read the worked answer
Diagnosis: non-small cell carcinoma with squamous differentiation. p40 is a highly specific marker of squamous lineage (more specific than p63, which also stains adenocarcinomas and lymphomas), and diffuse strong positivity in a poorly differentiated tumour supports squamous cell carcinoma.
Why the mucin matters: the presence of mucin vacuoles is incompatible with a pure squamous cell carcinoma in the WHO classification. The honest line is 'non-small cell carcinoma, favour squamous cell carcinoma; focal mucin present, adenosquamous carcinoma cannot be excluded on this sample.' That sentence is not hedging - it changes testing.
Consequences:
- For a genuinely pure squamous carcinoma, routine EGFR/ALK/ROS1 testing is not standard, and reflex testing is often omitted. On a small, non-definitive sample that omission is unsafe, because an adenosquamous or adeno-component would be treatment-relevant, and the patient has a brain metastasis where a driver mutation would alter first-line therapy.
- PD-L1 testing is indicated regardless of lineage.
- NGS remains the most informative single test (it covers EGFR, KRAS, BRAF, MET, ERBB2 and resistance mechanisms) if tissue allows.
Reporting: state the diagnostic uncertainty explicitly and name the tests performed, so that a later negative molecular result is not misread as a negative tumour.
3. NGS fails for low input. PD-L1 returns a tumour proportion score of 40%. How do you report this, and what options remain?
Read the worked answer
Report content:
- PD-L1 TPS 40% with clone 22C3 and the assay platform named; note that only ~60 tumour cells were assessable, which makes the score an estimate rather than a precise value.
- A pre-analytical caveat: nine days in formalin exceeds the 24–72 h range used in most validation studies and can reduce immunoreactivity. Flag it rather than silently reporting a possibly falsely low score; the laboratory should have documented fixative and fixation time at accession.
- NGS must be reported as not performed / failed, never as 'negative' or 'wild type'. Distinguishing 'no mutation detected' from 'no result' is a patient-safety issue that drives inappropriate first-line immunotherapy.
What remains available:
- EGFR mutation analysis by real-time PCR on one spare section.
- ALK and ROS1 immunohistochemistry as a screening surrogate for rearrangement; positive cases then need confirmation by FISH or RNA-based testing.
- Repeat sampling: re-biopsy of the nodal station, or plasma circulating tumour DNA, which is validated for EGFR at least and avoids further exhausting the block.
Communication: give the tumour board a clear statement of what is known, what is unknown, and what the next specimen should be used for. The most common error in this scenario is treating an uninformative test as a reassuring one.