Cell blocks sequence like tissue; stained smears need their own rules

Design

Retrospective analysis of hybrid-capture NGS performance by specimen type

Population

10,900 specimens, including 455 cytology samples and 2,759 formalin-fixed surgical controls

Primary outcome

Sequencing quality (percentage of targets at ≥100× coverage) by preparation

Effect

Cell blocks ≥98% median coverage (like tissue); Papanicolaou smears ~74.7%, with more GC bias

Cytology is often the only sample available for genomic profiling in advanced cancer, yet quality thresholds for next-generation sequencing were built on formalin-fixed tissue. This study sequenced 10,900 specimens, including 455 consecutive cytology samples and 2,759 formalin-fixed surgical controls, across Diff-Quik and Papanicolaou smears, cell blocks and tissue.

Cell blocks performed like formalin-fixed tissue, with median target coverage at or above 98%. Stained smears yielded adequate DNA but less uniform coverage — median target coverage around 75% for Papanicolaou smears — with more GC bias. The differences came from the smear preparation and staining themselves, not from DNA input or tumour purity.

The practical message for a molecular lab is to treat smears as a distinct specimen type: a cell block can inherit tissue-based thresholds, but a smear needs specimen-specific validation and cytology-aware normalisation before its sequencing result is trusted. Applying FFPE thresholds unchanged to smears risks both false reassurance and lost calls.

• Cell blocks matched formalin-fixed tissue on sequencing quality (median target coverage ≥98%).
• Papanicolaou smears gave adequate DNA but lower coverage uniformity (median around 75%) and more GC bias.
• The gap was intrinsic to smear preparation and staining, not DNA input or tumour purity.
• Validate stained smears as their own specimen type rather than applying tissue-based thresholds unchanged.

Why it matters

It stops labs assuming every cytology sample is interchangeable for sequencing, and names the preparation that is.

Don't overread it

Smears still yield usable DNA — this argues for smear-specific validation, not for rejecting smears as a sequencing substrate.

The statistics, in plain English

Coverage uniformity is how evenly the sequencer reads the targeted regions; a drop to about 75% on smears versus 98% on cell blocks means parts of the panel are under-read, which can miss or distort variant calls even when total DNA looks adequate.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.