Cytopathology: Adequacy, Categories, and the Two-Error Model
Adequacy, categories, and the two-error model
Cytology questions rest on two ideas: adequacy and risk stratification. Learn the reporting systems by name and structure-Bethesda cervical (NILM, ASC-US, LSIL, ASC-H, HSIL, AGC, SCC, AIS), Bethesda thyroid (categories I–VI), Milan for salivary gland, and Paris for urine (negative, AUC, SHGUC, HGUC). Know adequacy thresholds: thyroid FNA generally requires at least six groups of ten well-visualized follicular cells, with stricter criteria when only microfollicles are present. Cervical cytology adequacy concerns obscuring factors; transformation-zone sampling is a quality indicator, not a mandatory adequacy element.
Separate sampling error from interpretive error. A negative smear in a patient later found to have cancer is often a sampling problem rather than a screening failure. Know that HPV genotyping triages ASC-US and that p16/Ki-67 dual stain improves specificity in triage.
Traps: calling a sparsely cellular specimen “negative”; overusing “atypical”; mistaking reactive mesothelial cells for adenocarcinoma; confusing small-cell carcinoma with lymphocytes; and forgetting that a cell block may be the only material available for ancillary studies.
Rule of thumb: state adequacy first, then category, then implied risk. If a specimen is inadequate, that is the report-not “negative.”
Worked example: pleural fluid with a few atypical epithelioid clusters. A two-step panel resolves most cases-mesothelial markers (calretinin, WT1, CK5/6, D2-40) versus epithelial markers (MOC-31, Ber-EP4, CEA, B72.3), with site-specific markers such as TTF-1, PAX8, or CDX2 added as needed. If results are discordant, report the ambiguity and recommend biopsy rather than forcing a diagnosis.