Pathology In Depth

Library › Ancillary testing on cytology specimens

Ancillary testing on cytology specimens

Last reviewed 2026-11-10 by a board-certified pathologist. 71 references.

Immunohistochemistry on Cytology Specimens: Validation, Panels for Carcinoma of Unknown Primary, and Smear Transfer

Definition. Immunohistochemistry on cytology specimens (immunocytochemistry) is the detection of antigens with labeled antibodies in cells obtained by fine-needle aspiration (FNA) or exfoliation. The substrates are formalin-fixed, paraffin-embedded (FFPE) cell blocks, direct smears, cytospins and liquid-based preparations. The preanalytic conditions of cytology specimens often differ from those of the positive control tissues (Fischer et al. 2014). The College of American Pathologists (CAP) guideline on analytic validation of immunohistochemical assays, updated in 2024, contains specific statements for cytology specimens fixed differently from the tissues used for initial validation (Goldsmith et al. 2024).

Related page: Flow Cytometry, FISH, Microbiology and Electron Microscopy on Cytology Specimens (flow cytometry, FISH, microbiology and electron microscopy) Related page: Molecular Testing and Predictive Biomarkers on Cytology Specimens (molecular testing, PD-L1 and predictive biomarkers) Related page: Digital Image Analysis and Artificial Intelligence in Cytology (digital image analysis and artificial intelligence) Related page: Cytology Specimen Types and Triage (specimen types, collection media and triage) Related page: Mesothelial vs Carcinoma Immunohistochemistry in Effusions (mesothelial and carcinoma markers in effusions)

Substrates for immunohistochemistry on cytology specimens

Substrate Fixation and processing Advantages Limitations and handling
FFPE cell block Formalin, or alcohol-based collection medium followed by formalin; paraffin embedding Serial sections for panels; same protocols and controls as histology when formalin-fixed Variable cellularity; low cellular yield is the leading cause of dissatisfaction (Crapanzano et al. 2014)
Direct smear, alcohol-fixed 95% ethanol, Papanicolaou stain Often the most cellular preparation; usable after destaining (Adiga et al. 2026) Requires validation separate from FFPE tissue (Goldsmith et al. 2024); a smear used for IHC is consumed
Direct smear, air-dried Romanowsky-type stain Prepared at rapid on-site evaluation Unsuitable for HER2 in the CAP survey (Fischer et al. 2014)
Cytospin Alcohol fixation or air drying Concentrates cells from fluids Alcohol-fixed, air-dried unstained cytospins keep for a month at 4°C and several months at −20°C (Jain et al. 2019)
Liquid-based preparation Methanol-based fixatives Uniform monolayer Methanol-based fixatives decrease staining intensity and cause false-negative results (Jain et al. 2019)
  • Cell block methods. More than 10 cell block methods were in use among 90 respondents to an American Society of Cytopathology (ASC) survey. No method was associated with greater satisfaction (Crapanzano et al. 2014).
  • Current practice. In a 2024 ASC survey of 294 respondents, 89% used FFPE cell blocks for IHC, 84% for molecular testing and 71% for in situ hybridization. For cell block preparation, 58% used a coagulating agent, and 67% used no fixative before formalin (Heymann et al. 2025).
  • Limits on ancillary testing. The two factors most often named as limiting ancillary testing on cytology specimens were low cellularity (49%) and lack of validation (23%) (Heymann et al. 2025).
  • Performance of cell blocks. Cell blocks of good quality and cellularity outperform other cytology preparations in reliability and versatility for ancillary testing (Nambirajan and Jain 2018).

Effect of alcohol-based fixatives on immunoreactivity

Alcohol-based and methanol-based collection media alter antigen preservation. The direction and size of the change differ by antibody and clone.

Antibody Fixative or pretreatment Effect compared with formalin Source
TTF-1, D2-40, CD20 CytoLyt fixation of normal tissues Absent or near absent expression Gruchy et al. 2015
p63, estrogen receptor, S100 protein, CD3, calretinin, chromogranin, synaptophysin CytoLyt fixation of normal tissues Reduced expression Gruchy et al. 2015
TTF-1 Hydrochloric acid decalcification Absent or near absent expression Gruchy et al. 2015
Ki-67 (MIB1 clone) CytoLyt-fixed cell blocks Mean labeling index 10% in CytoLyt versus 47% in formalin; mean loss 37%, up to 70% Buonocore et al. 2019
Ki-67 (30-9 clone) CytoLyt-fixed cell blocks Not inhibited; mean 83% in small cell carcinoma, against 60% with MIB1 Buonocore et al. 2019
INSM1 CytoLyt-fixed cell blocks Negative in 65% and equivocal in 30% of neuroendocrine neoplasms that showed more than 90% positivity in formalin-fixed cell blocks Garlin-Politis et al. 2024
PD-L1 PreservCyt and CytoLyt cell blocks of pleural effusions Moderate concordance with formalin at the ≥1% cutoff (κ 0.46–0.59); weaker staining intensity Mansour et al. 2026
  • Protocol modifications and MIB1. None of the modifications tested in 223 immunocytochemical reactions recovered MIB1 reactivity in CytoLyt-fixed material, except substitution of the 30-9 clone (Buonocore et al. 2019).
  • Diagnostic consequence for Ki-67. Markedly reduced MIB1 labeling is a potential error in the diagnosis of small cell carcinoma and in the grading of other tumors (Buonocore et al. 2019).
  • Diagnostic consequence for INSM1. Reduced INSM1 staining in CytoLyt-fixed cell blocks can produce a false-negative result for a neuroendocrine neoplasm (Garlin-Politis et al. 2024).
  • Alcohol-fixed smears over time. Prolonged alcohol fixation produced more variable ER, PR and HER2 staining and more false-negative results (Jain et al. 2019).
  • Epitope stability. Coverslipped alcohol-fixed smears retain epitopes for about 2 years. Unstained sections cut from FFPE blocks lose epitopes after 3 to 6 months at room temperature (Jain et al. 2019).
  • Validation on matched controls. The fixative data show the need to validate IHC protocols on control material treated in the same manner as the test material (Gruchy et al. 2015).

CAP 2024 requirements for validation of immunohistochemistry on cytology specimens

The 2014 CAP guideline contained 14 statements on initial validation and revalidation of IHC assays (Fitzgibbons et al. 2014). The 2024 update contains 15 statements and gives more specific guidance for cytology specimens (Goldsmith et al. 2024).

Previous rule (2014). Laboratories that regularly performed IHC on cytologic specimens processed differently from the validation tissues (for example alcohol-fixed cell blocks, air-dried smears or formalin-postfixed specimens) tested a sufficient number of such cases to ensure expected results, with the number of positive and negative cases set by the laboratory medical director (statement 7; expert consensus opinion). The 2014 concordance requirement of at least 90% excluded HER2, ER and PgR, for which separate validation guidelines existed (statement 2) (Fitzgibbons et al. 2014). The 2024 update replaced the cytology statement with a separate validation for each analyte and fixation method (statement 9) and a minimum of 10 positive and 10 negative cases, if possible (statement 10), and it set the concordance requirement uniformly at 90% for all IHC assays (Goldsmith et al. 2024).

Statement Content Strength
2 At least 90% overall concordance between the new assay and the comparator or expected results, for every assay used clinically Strong recommendation
3 Nonpredictive laboratory-developed assays: minimum 10 positive and 10 negative tissues Good practice statement
4 Laboratory-developed predictive assays: minimum 20 positive and 20 negative tissues Good practice statement
6 Predictive assays with distinct scoring schemes (for example, HER2 and PD-L1): each assay-scoring system combination validated with at least 20 positive and 20 negative tissues Strong recommendation
8 Validation tissues processed with the same fixative and processing methods as clinical cases, when possible Good practice statement
9 Cytology specimens not fixed like the initial validation tissues: separate validation for every new analyte and corresponding fixation method before clinical use Conditional recommendation
10 Minimum 10 positive and 10 negative cases for each cytology validation, if possible; more cases considered for predictive markers; if the minimum is not feasible, the rationale for using fewer cases is documented Good practice statement
11 IHC regularly done on decalcified tissues: a sufficient number of such tissues tested, the number set by the laboratory medical director Good practice statement
12 New antibody lot: at least 1 known positive and 1 known negative tissue Good practice statement
13 Change in antibody dilution, vendor (same clone) or incubation or retrieval time: at least 2 known positive and 2 known negative tissues Good practice statement
15 New antibody clone: full revalidation equivalent to initial validation Good practice statement

Source: Goldsmith et al. 2024.

  • Cytology specimens covered by statement 9. The guideline lists air-dried or alcohol-fixed smears, liquid-based preparations, alcohol-fixed cell blocks and specimens collected in alcohol or alternative fixative media and then postfixed in formalin (Goldsmith et al. 2024).
  • Formalin postfixation does not remove the requirement. A cell block collected in an alcohol-based medium and then postfixed in formalin is still listed as a cytology specimen that requires separate validation (Goldsmith et al. 2024).
  • Existing cytology assays. CAP states that statements 9 and 10 do not mandate revalidation of all IHC assays already performed on cytology specimens. Laboratories are expected to have some validation study for cytology specimens fixed differently from the original validation tissues (CAP guideline page, Goldsmith et al. 2024).

Cytology immunohistochemistry practice before the CAP 2024 update

  • CAP survey of 2009–2010. The survey findings on fixation platforms, mismatched controls and air-dried smears for HER2 are covered in Cytology Specimen Types and Triage. Most laboratories (155 of 184; 84%) did not alter antibody dilution or antigen retrieval for cytology IHC, and three had validated a nonformalin HER2 platform (Fischer et al. 2014).
  • Breast receptor testing. ER, PR and HER2 testing on breast cytology is covered in Invasive Breast Carcinoma of No Special Type and Invasive Lobular Carcinoma on FNA.
  • Cytology-specific controls. Nearly two-thirds of cytology laboratories in Europe and the United States use FFPE histology sections as controls for immunocytochemistry. Only 13% of 100 published articles specified cytology-specific controls (Jain et al. 2019).

Immunohistochemical panels for carcinoma of unknown primary on cytology specimens

Carcinoma of unknown primary (CUP) is often first diagnosed by FNA of a lymph node, liver lesion or serous effusion. The cell block is the usual substrate for the panel. A panel targeted by morphology and clinical features conserves material (Kandalaft and Gown 2016).

Clinical aims of the immunohistochemical workup of carcinoma of unknown primary

  • Treatable subsets first. Four primary sites with specific effective treatment and a better prognosis are investigated first: breast, ovary, prostate and thyroid (Selves et al. 2018).
  • Targeted therapy. A pulmonary or colorectal origin is also assessed because targeted therapies exist for these sites (Selves et al. 2018).
  • Residual unclassified cases. In approximately 5% of cases, IHC gives no definitive information, and an undifferentiated malignant neoplasm is reported without reference to its origin (Selves et al. 2018).

Lineage markers for an undifferentiated malignant neoplasm

  • Keratin. Keratin is expressed by some nonepithelial tumors with epithelial differentiation, such as synovial sarcoma, epithelioid sarcoma and desmoplastic small round cell tumor. Focal keratin expression has also been reported in angiosarcoma, melanoma, plasmacytoma, diffuse large B-cell lymphoma and anaplastic large cell lymphoma (Selves et al. 2018).
  • Vimentin. A vimentin-negative tumor is unlikely to be a sarcoma, lymphoma or melanoma, with the exception of alveolar soft part sarcoma (Selves et al. 2018).

CK7 and CK20 patterns in carcinoma of unknown primary

Pattern Typical tumors
CK7+/CK20− Breast, lung adenocarcinoma, endometrial, endocervical and serous ovarian carcinoma, cholangiocarcinoma, thyroid, salivary gland tumors, papillary renal cell carcinoma, mesothelioma, small cell lung carcinoma, a subset of urothelial carcinoma, pancreatic and gastric adenocarcinoma
CK7+/CK20+ Urothelial carcinoma, pancreatic adenocarcinoma, mucinous ovarian carcinoma, bladder adenocarcinoma, gastric adenocarcinoma, cholangiocarcinoma
CK7−/CK20+ Colorectal adenocarcinoma, Merkel cell carcinoma, gastric adenocarcinoma
CK7−/CK20− Prostate adenocarcinoma, clear cell renal cell carcinoma, hepatocellular carcinoma, adrenocortical carcinoma, nonseminomatous germ cell tumors, mesothelioma, small cell lung carcinoma, gastric adenocarcinoma

Source: Selves et al. 2018.

  • Overlap of patterns. Breast carcinomas are CK7+/CK20− in 86% to 94% of cases. The other patterns occur in decreasing order: CK7+/CK20+, CK7−/CK20+ and CK7−/CK20− (Selves et al. 2018).
  • Tumors in two patterns. Small cell lung carcinoma and mesothelioma appear in both the CK7+/CK20− and the CK7−/CK20− groups (Selves et al. 2018).
  • First markers after CK7+/CK20−. TTF-1, GATA3, PAX8 and WT1 are the markers most often used for a CK7+/CK20− CUP (Selves et al. 2018).

Site-specific markers and their limits

Marker Main use Limits
TTF-1 Lung adenocarcinoma; thyroid carcinoma About 50% positive in poorly differentiated lung adenocarcinoma, about 100% in well-differentiated tumors; positive in less than 5% of anaplastic thyroid carcinomas
GATA3 Breast carcinoma, ductal and lobular, including triple-negative tumors Positive in 80% to 90% of urothelial carcinomas
PAX8 Ovarian (79% to 99%), endometrial (84% to 93%) and renal cell carcinoma (90% to 100%); thyroid carcinoma Negative in breast, colon, prostate, liver and stomach adenocarcinoma in two large series
WT1 More than 90% of high-grade serous ovarian carcinomas Also expressed by benign and malignant mesothelial proliferations; cytoplasmic staining in breast carcinoma
CDX2 Intestinal differentiation Expressed in some gastric, esophageal, pancreatic, ovarian, urachal and sinonasal carcinomas; reduced in right-sided, poorly differentiated and mismatch repair–deficient colorectal carcinoma
SATB2 Lower gastrointestinal tract: more than 85% of colon adenocarcinomas and almost all appendiceal adenocarcinomas Expressed in almost half of bladder adenocarcinomas
PSA Prostatic adenocarcinoma Lost in 10% to 20% of metastases and after hormone therapy
Arginase-1 Hepatocellular carcinoma (more than 90%), including poorly differentiated tumors Less widely available

Source: Selves et al. 2018.

  • SATB2 in a CK7−/CK20+ tumor. SATB2 increases the predictive value for a colon origin from 93% to 99% in CK7−/CK20+ tumors. It is expressed in poorly differentiated colon carcinomas that are negative for CDX2 or CK20 (Selves et al. 2018).
  • CDX2-negative colorectal carcinoma. Colorectal carcinomas are negative for CDX2 in 15% to 20% of cases (Selves et al. 2018).
  • Napsin A in thyroid carcinoma. A minority of anaplastic and poorly differentiated thyroid carcinomas express napsin A. PAX8 separates these thyroglobulin-negative thyroid carcinomas from lung adenocarcinoma, which is PAX8 negative (Selves et al. 2018).
  • ERG in prostatic carcinoma. ERG protein, reflecting TMPRSS2::ERG, is detected in approximately 30% of prostatic carcinoma metastases (Selves et al. 2018).
  • Mesothelioma in a serous fluid. The mesothelial and carcinoma panels for effusions are covered in Mesothelial vs Carcinoma Immunohistochemistry in Effusions.

Immunohistochemistry on direct smears and smear transfer

When the cell block is acellular, the direct smear is the only source of tumor cells. A smear used for IHC or molecular testing is consumed and is no longer available as a diagnostic slide.

Destained Papanicolaou smears

  • Lung carcinoma subtyping. TTF-1 and p40 were applied to destained, 95% ethanol-fixed Papanicolaou smears of non-small cell carcinoma. TTF-1 on smears had a sensitivity of 97.2% and a positive predictive value of 100% against IHC on paraffin-embedded biopsies. p40 had a sensitivity and positive predictive value of 100%. Agreement between smear and biopsy was κ = 1.0 for both markers (Adiga et al. 2026).
  • ALK and PD-L1 on direct smears. Direct smears have also been used for ALK immunocytochemistry and for PD-L1 testing. For PD-L1, direct smears and small biopsies showed 97% concordance, against 82% for cell blocks, in one study summarized by the International Association for the Study of Lung Cancer (IASLC) Pathology Committee (Jain et al. 2019).

Cell transfer from a single smear

  • Principle. Cells from one smear are lifted in a mounting medium and divided among several slides. A battery of immunostains can then be applied to material from a single smear (Sherman et al. 1994).
  • Original series. In four effusions and five FNAs, cell transfer gave the expected immunostaining results for most antibodies, with low background staining and preserved morphology (Sherman et al. 1994).
  • Turnaround. Earlier transfer methods take about 3 to 40 hours, and the membrane with attached cells can stretch or tear when it is peeled off the slide. A nylon mesh method shortens the procedure without special equipment (Morito et al. 2021).

Sacrifice of smears for molecular testing

Choice of substrate for immunohistochemistry by clinical situation

Situation Preferred substrate Reason
Cellular cell block fixed in formalin Cell block sections Same protocols and controls as histology; serial sections for a panel (Nambirajan and Jain 2018)
Cell block collected in CytoLyt; Ki-67 required Cell block with the 30-9 clone, or a validated protocol MIB1 is inhibited by CytoLyt (Buonocore et al. 2019)
Cell block collected in CytoLyt; neuroendocrine marker required Formalin-fixed material, or a protocol validated on CytoLyt INSM1 is reduced in CytoLyt-fixed cell blocks (Garlin-Politis et al. 2024)
Acellular cell block, several smears with tumor Destained alcohol-fixed smear, validated for the analyte Smear immunocytochemistry agreed with biopsy IHC for TTF-1 and p40 (Adiga et al. 2026)
One smear with tumor, several markers needed Cell transfer to multiple slides Divides one smear among several stains (Sherman et al. 1994)
Smear needed for molecular testing Digital archiving, then scraping Preserves a record of the sacrificed slide; scraping maximizes DNA yield (Wang et al. 2026; Roy-Chowdhuri et al. 2016)

Flow Cytometry, Fluorescence In Situ Hybridization, Microbiology and Electron Microscopy on Cytology Specimens

Definition. Flow cytometry, fluorescence in situ hybridization (FISH), microbiologic studies and electron microscopy (EM) are ancillary tests applied to fine-needle aspiration (FNA) and exfoliative cytology material. Each test requires a different preparation: viable cells in suspension for flow cytometry, intact nuclei on a slide or section for FISH, unfixed material in a sterile container for culture and glutaraldehyde-fixed material for EM. The CAP guideline on thoracic small biopsy and cytology specimens includes flow cytometry, FISH, cytogenetics and microbiologic studies among the ancillary studies in its scope (Roy-Chowdhuri et al. 2020).

Related page: Immunohistochemistry on Cytology Specimens (immunohistochemistry and validation) Related page: Molecular Testing and Predictive Biomarkers on Cytology Specimens (molecular testing and predictive biomarkers) Related page: Lymph Node FNA Triage for Flow Cytometry, Cultures and Molecular Testing (allocation of lymph node FNA material) Related page: Cytology Specimen Types and Triage (collection media and triage) Related page: Granulomatous Lymphadenitis (Cytology) (mycobacterial and fungal lymphadenitis)

Requirements of each ancillary test

Test Material Handling Main limitation on cytology material
Flow cytometry Viable cells in suspension (needle rinse, fluid) Unfixed; RPMI 1640 medium, delivered within 2 hours in one series (Bandargal et al. 2024) Fragile large cells and scant neoplastic cells may not be detected (Ehinger et al. 2026); necrosis hinders analysis (Bandargal et al. 2024)
FISH Intact nuclei on smears, cytospins, liquid-based preparations or cell block sections Previously stained smears can be used Older archived slides fail more often (Nambirajan et al. 2022)
Culture Unfixed aspirate or fluid Sterile container at room temperature; promptly for bacteria, within 2 hours for fungi and mycobacteria, in the 2018 edition of the IDSA and ASM guide (Miller et al. 2018); the guide was updated in 2024 (Miller et al. 2024) Culture is an imperfect reference standard in paucibacillary specimens (Kohli et al. 2025)
Nucleic acid amplification for tuberculosis Aspirate or fluid Unfixed Sensitivity of 71.3% in lymph node aspirates for Xpert Ultra (Kohli et al. 2025)
Electron microscopy Needle rinse or fluid pellet Glutaraldehyde added promptly; isotonic medium; gentle handling Insufficient material in 19% of submitted cases (Etit et al. 2010)

Flow cytometry on fine-needle aspiration material

Diagnostic performance of FNA with flow cytometry for lymphoma

  • Wake Forest series. In 290 aspirates from 275 patients, FNA with flow cytometry had a sensitivity of 95% and a specificity of 85% for lymphoma when suggestive diagnoses were counted as positive. Specificity was 100% when only definitive diagnoses of lymphoma were counted (Meda et al. 2000).
  • Massachusetts General Hospital series. Flow cytometry was successful in 105 of 139 FNAs (75%). Among the cases with flow cytometry, 80 (77%) were true positive, 25 were indeterminate and none was false negative (Dong et al. 2001).
  • Subclassification. Initial subclassification was conclusive in 59% of positive cases. Review of morphology together with flow cytometry data raised it to 77% (Dong et al. 2001).
  • Subtype variation. Subclassification was 100% for small lymphocytic lymphoma/chronic lymphocytic leukemia, lymphoplasmacytic lymphoma, Burkitt lymphoma, mantle cell lymphoma and plasmacytoma. It was 33% for marginal zone lymphoma (Dong et al. 2001).
  • Hodgkin lymphoma. Hodgkin lymphoma was recognized as malignant in 7 of 16 cases (44%) (Dong et al. 2001).
  • Category assignment. With the algorithms of the WHO Reporting System for Lymph Node, Spleen and Thymus Cytopathology, more than 80% of lymphoid FNAs are placed in the benign or malignant category, particularly when flow cytometry is used (Ehinger et al. 2026).

Interpretation of flow cytometry results in the WHO lymph node system

Cytologic and flow cytometric findings Interpretation Source
Benign clinical and cytologic findings; small monotypic B-cell population (<0.5%) Atypical category may be considered; monoclonal B-cell lymphocytosis remains in the differential diagnosis Ehinger et al. 2026
Benign or atypical cytology; substantial monotypic B-cell population Suspicious for malignancy or malignant, even with low clinical suspicion Ehinger et al. 2026
Large atypical cells; no monotypic cells High-grade lymphoma usually categorized as malignant; fragile cells may not be detected Ehinger et al. 2026
Single, very large atypical cells (classic Hodgkin lymphoma pattern); polytypic flow cytometry A polytypic result does not exclude classic Hodgkin lymphoma; scant malignant cells may not be detected Ehinger et al. 2026
Monotonous small lymphocytes; no monotypic cells Inert nonspecific lymphadenitis can resemble low-grade lymphoma; flow cytometry prevents the error Ehinger et al. 2026
Plasma cell neoplasm or mastocytosis No monotypic lymphoid cells by design; aberrant mast cells expressing CD25, CD2 or CD30 are detected Ehinger et al. 2026

Allocation of lymph node FNA material and the limits of flow cytometry in large B-cell and classic Hodgkin lymphoma are also covered in Lymph Node FNA Triage for Flow Cytometry, Cultures and Molecular Testing.

  • Light-chain type. Lambda-monotypic B cells are more likely to be malignant than kappa-monotypic B cells (Ehinger et al. 2026).
  • Follicular lymphoma. Classic follicular lymphoma shows dimmer CD19, CD20 and CD38 expression than reactive hyperplasia (Ehinger et al. 2026).
  • Large B-cell lymphoma. Subtyping of large B-cell lymphoma on FNA is avoided unless large cells exceed 40% by both flow cytometry and cytomorphology. Follicular lymphoma grade 3A and 3B cannot be separated from diffuse large B-cell lymphoma on FNA (Ehinger et al. 2026).
  • Marginal zone lymphoma. A definitive diagnosis on FNA with flow cytometry is possible once lymphoplasmacytic lymphoma, mantle cell lymphoma, CD10-negative follicular lymphoma and large B-cell lymphoma have been excluded (Ehinger et al. 2026).
  • T-cell clonality. TRBC1 restriction analysis by flow cytometry or T-cell receptor gene rearrangement by PCR assesses T-cell clonality (Ehinger et al. 2026).
  • Lymphoblastic lymphoma. Flow cytometry separates B-lymphoblastic (CD19, CD79a, CD10, TdT positive) from T-lymphoblastic (CD3, CD7, CD10, TdT, CD99 positive) lymphoma (Ehinger et al. 2026).
  • Primary effusion lymphoma. The flow cytometric and diagnostic features of primary effusion lymphoma are covered in Lymphoma and Primary Effusion Lymphoma in Serous Effusions.

Fluorescence in situ hybridization on cytology specimens

Advantages and requirements of cytology material for FISH

  • Whole nuclei. Cytology specimens allow assessment of the entire nucleus. Signal loss from truncation artifact in tissue sections is avoided (Lozano et al. 2018).
  • Stained direct smears. FISH on stained direct smears is used for ALK and ROS1 rearrangements in non-small cell lung carcinoma (Lozano et al. 2018).
  • Evaluation criteria. FISH is the reference method for ALK rearrangement, and the same evaluation criteria apply in cytology as in histology (Savic and Bubendorf 2016).
  • Targeted evaluation. FISH is a morphology-based technique. A reliable result requires targeted evaluation of the cells in question, such as cancer cells or atypical cells (Savic and Bubendorf 2016).
  • Cell number. FISH for ROS1 requires 50 tumor cells (Jain et al. 2019).
  • ALK break-apart FISH scoring on the FDA-approved kit. The Vysis ALK Break Apart FISH Probe Kit is scored on 50 tumor nuclei. A nucleus is positive when orange and green signals are 2 or more signal diameters apart, or when a single orange signal is present without a corresponding green signal. A specimen is negative when fewer than 5 of 50 nuclei (<10%) are positive and positive when more than 25 of 50 (>50%) are positive. With 5 to 25 positive nuclei (10% to 50%), a second reader scores the slide, the two counts are combined over 100 nuclei, and the specimen is positive at 15% or more (≥15/100). The approved indication is 10% neutral buffered formalin-fixed, paraffin-embedded non-small cell lung carcinoma tissue, so FISH on cytology smears is a laboratory-validated modification (FDA PMA P110012, package insert).

Published performance of FISH on smears

  • ALK on Diff-Quik smears. Tumor-rich areas of Diff-Quik-stained smears were marked with a diamond-tipped scribe before FISH. Smear and cell block results agreed in 31 of 32 lung adenocarcinomas (97%). In the discordant case, the smear showed an ALK rearrangement, and RT-PCR of the cell block confirmed EML4::ALK, establishing a false-negative cell block FISH result (Betz et al. 2013).
  • Multiplex ALK and ROS1. Multiplex ALK/ROS1 FISH was performed on 27 non-cell block preparations from 20 patients. Results were obtained in 14 samples and agreed with FFPE FISH in 13. The discordant case showed ALK rearrangement on cytology, matching the ALK D5F3 companion diagnostic assay (Nambirajan et al. 2022).
  • Failures on archived slides. FISH failures occurred mainly in older archived slides, from overdigestion (5), hybridization failure (5) and excessive background fluorescence (3) (Nambirajan et al. 2022).

Applications of FISH on cytology specimens

Specimen Target Purpose Detailed coverage
Lung adenocarcinoma (FNA, effusion) ALK, ROS1 rearrangement; MET amplification Predictive testing; MET amplification is a mechanism of acquired resistance to EGFR inhibitors (Savic and Bubendorf 2016) Molecular Testing and Predictive Biomarkers on Cytology Specimens
Urine Chromosomes 3, 7, 17 and 9p21 Atypical urothelial cells equivocal for malignancy (Savic and Bubendorf 2016) Ancillary Tests, Indications and Management in Urine Cytology
Biliary brushing Multiprobe set Malignant biliary stricture Biliary adenocarcinoma on brushing and ancillary testing
Serous effusion CDKN2A (9p21) homozygous deletion Mesothelioma (Savic and Bubendorf 2016) Mesothelial vs Carcinoma Immunohistochemistry in Effusions
Soft tissue FNA Translocation-associated genes Classification of sarcomas Molecular and Ancillary Testing of Soft Tissue FNA Material
Lymph node FNA t(14;18), t(11;14), MYC Lymphoma subtyping Lymph Node FNA Triage for Flow Cytometry, Cultures and Molecular Testing

Microbiologic studies on cytology specimens

Collection of aspirates for mycobacterial culture

  • CAP thoracic guideline. Statements 15 (EBUS-TBNA of granulomatous nodes with suspected tuberculosis) and 16 (pleural fluid for extrapulmonary tuberculosis) are covered in Cytology Specimen Types and Triage.
  • Mycobacterial culture in superficial nodes. In 400 superficial lymph node aspirates inoculated on Löwenstein-Jensen medium, 180 (45%) showed tuberculous lymphadenitis. The authors concluded that culture increases diagnostic accuracy beyond cytology and Ziehl-Neelsen staining (Mitra et al. 2017). Allocation of lymph node material for culture is covered in Lymph Node FNA Triage for Flow Cytometry, Cultures and Molecular Testing.
  • Confirmation of mycobacterial infection. Acid-fast bacilli stain, PCR, culture and the tuberculin skin test are used for suspected mycobacterial infection. PCR, culture and drug sensitivity results confirm the infection (Al-Abbadi et al. 2020).

Nucleic acid amplification tests for tuberculosis on cytology specimens

Specimen Test Sensitivity (95% CI) Specificity (95% CI) Reference standard
Lymph node aspirate (6 studies) Xpert MTB/RIF Ultra 71.3% (64.3–77.4) 97.4% (82.3–99.7) Composite
Pleural fluid (13 studies) Xpert MTB/RIF Ultra 74.0% (60.8–83.9) 88.1% (78.8–93.6) Microbiological
Cerebrospinal fluid (16 studies) Xpert MTB/RIF Ultra 88.2% (83.7–91.6) 96.0% (86.8–98.9) Microbiological
Rifampicin resistance (13 studies) Xpert MTB/RIF Ultra 100.0% (93.4–100.0) 99.4% (92.1–100.0) Phenotypic or genotypic susceptibility testing

Source: Kohli et al. 2025 (Cochrane review informing a WHO policy update).

  • Interpretation of a negative result. Sensitivity varies across extrapulmonary specimens, while specificity is high for most specimens (Kohli et al. 2025). A negative NAAT on a lymph node aspirate therefore does not exclude tuberculosis.

Electron microscopy on cytology specimens

Processing of FNA material for electron microscopy

  • Glutaraldehyde timing. Poor preservation was traced to delay before glutaraldehyde was added to a needle rinse held in a physiologic transport medium (Ravinsky et al. 1993).
  • Transport medium and handling. A tendency toward hypotonicity of a commercial transport medium and vigorous handling of the needle rinse also impaired preservation (Ravinsky et al. 1993).
  • Scant material. Fewer smears and filtration of the needle rinse to capture tissue fragments increased the material available for EM (Ravinsky et al. 1993).
  • Yield. With adequately cellular, well-preserved material, EM contributed clinically valuable information in 48% of cases (Ravinsky et al. 1993).

Diagnostic contribution of electron microscopy in a cytopathology practice

Measure Result in 178 cytology cases (131 FNAs, 47 exfoliative)
Frequency of EM requests 178 of 149,006 nongynecologic cytology cases, 1993–2006
Additional diagnostic information 32%
Confirmation of the cytologic diagnosis 48%
Insufficient material 19%
Discrepant result 1%
FNA applications: epithelial versus mesenchymal tumor 45.6%
FNA applications: nonneoplastic processes (alveolar proteinosis, amyloidosis) 15.7%
FNA applications: microorganisms 12.2%

Source: Etit et al. 2010.

Ultrastructure of mesothelioma and adenocarcinoma in effusions

Feature Mesothelioma Adenocarcinoma Source
Microvillus length-to-diameter ratio 19.1 ± 7.0 9.1 ± 2.2 Sakuma et al. 1999
Microvilli Longer, thinner and more complex Shorter Warhol et al. 1982; Sakuma et al. 1999
Microvillus rootlets Absent Present in 2 of 10 cases Warhol et al. 1982
Lamellar inclusion bodies Absent Present in 2 of 10 cases Warhol et al. 1982
Intermediate filaments More abundant Less abundant Warhol et al. 1982; Sakuma et al. 1999
Long junctions (desmosomes over 1 μm) Present in 8 of 12 cases Not reported Sakuma et al. 1999
  • Reactive mesothelial cells. Reactive mesothelial cells had a microvillus length-to-diameter ratio of 9.2 ± 2.4, the same range as adenocarcinoma, so a high ratio separates mesothelioma from both (Sakuma et al. 1999).
  • Use in effusions. EM was applied infrequently to exfoliative specimens to separate mesothelioma from adenocarcinoma, and it was useful in that setting (Etit et al. 2010).
  • Current first-line markers. Immunohistochemical separation of mesothelial cells from carcinoma, and of benign from malignant mesothelial cells, is covered in Mesothelial vs Carcinoma Immunohistochemistry in Effusions.

Choice of ancillary test by cytologic question

Cytologic question First test Reason
Monomorphous small lymphoid cells in a node Flow cytometry Detects monotypic B or T cells and separates small B-cell lymphomas (Dong et al. 2001; Ehinger et al. 2026)
Scattered very large atypical cells in a mixed background Cell block immunostains; excisional biopsy A polytypic flow cytometry result does not exclude classic Hodgkin lymphoma (Ehinger et al. 2026)
Lung adenocarcinoma with exhausted cell block FISH or other molecular test on stained smears Smear and cell block FISH agree (Betz et al. 2013)
Necrotizing granulomas in a thoracic node Mycobacterial smear and culture; tuberculosis PCR CAP thoracic guideline, statement 15 (Roy-Chowdhuri et al. 2020)
Effusion with epithelioid cells, mesothelioma versus adenocarcinoma, immunostains inconclusive EM on a glutaraldehyde-fixed pellet Microvillus length-to-diameter ratio and rootlets (Sakuma et al. 1999; Warhol et al. 1982)

ancillary test handling on cytology material lookup

Molecular Testing and Predictive Biomarkers on Cytology Specimens, Including Supernatant Cell-Free DNA

Definition. Molecular testing on cytology specimens is the detection of gene mutations, fusions, amplifications and other genomic alterations in DNA or RNA extracted from cell blocks, direct smears, liquid-based preparations or the cell-free fraction of cytology fluids. Predictive biomarkers on cytology specimens also include protein markers assessed by immunohistochemistry, such as ALK, ROS1 and PD-L1. The CAP/IASLC/AMP lung cancer molecular testing guideline of 2018 accepts cell blocks and other cytologic preparations as suitable specimens for biomarker molecular testing (Lindeman et al. 2018). Cytology samples are similar to surgical samples in identifying clinically relevant genomic alterations, with success rates up to 93% after process optimization (Kim et al. 2025).

Related page: Immunohistochemistry on Cytology Specimens (validation of immunohistochemistry on cytology specimens) Related page: Flow Cytometry, FISH, Microbiology and Electron Microscopy on Cytology Specimens (FISH on smears and cell blocks) Related page: Metastatic Lung Adenocarcinoma in Serous Effusions (lung test menu, PD-L1 tumor proportion score) Related page: Invasive Breast Carcinoma of No Special Type and Invasive Lobular Carcinoma on FNA (ER, PR and HER2 on breast cytology) Related page: Biliary adenocarcinoma on brushing and ancillary testing (biliary FISH and sequencing) Related page: Molecular and Ancillary Testing of Soft Tissue FNA Material (soft tissue molecular testing) Related page: Pancreatic cyst fluid chemistry and molecular testing (cyst fluid and supernatant DNA in pancreatic lesions)

CAP/IASLC/AMP 2018 statements on cytology specimens and tumor content

Topic 2013 statement 2018 statement
Cytology specimens Cytologic samples suitable for EGFR and ALK testing, cell blocks preferred over smears Cell blocks or other cytologic preparations are suitable specimens for lung cancer biomarker molecular testing
Assay sensitivity and tumor content EGFR methods able to detect mutations in specimens with at least 50% cancer cells; more sensitive tests detecting mutations with as little as 10% cancer cells strongly encouraged Assays able to detect alterations in specimens with as little as 20% cancer cells, in house or at a reference laboratory (expert consensus opinion)
Total EGFR immunohistochemistry Not recommended for selection of EGFR TKI therapy Not used to select patients for EGFR-targeted TKI therapy (strong recommendation)

Source: Lindeman et al. 2018, summary of recommendations (strength labels are given for the 2018 assay-sensitivity and total EGFR statements).

A revision of the 2018 guideline was posted for public comment from August 22 to September 12, 2024 (CAP newsroom). No revised guideline had been published at the time of review, and the 2018 statements remain current.

Performance of cytology preparations for next-generation sequencing

Cell blocks and supernatant cell-free DNA in a large institutional series

In 4,871 prospectively sequenced clinical cytology samples tested with a large next-generation sequencing (NGS) panel, the following results were reported (Kim et al. 2025).

Measure Cell block Supernatant cell-free DNA
Successful testing 81% (3,616 of 4,457) 71% (190 of 268), tested only when the cell block was unsuitable
Median total DNA yield 427.5 ng 182.2 ng
At least one somatic alteration in sequenced cases 93.9% 92.6%
Low-level cross-contamination 4.7% of cases Negligible
  • Overall success. Of all cytology samples, 81% (3,806 of 4,725) were tested successfully. Supernatant testing used as a rescue raised the success rate of the cytologic procedures from 77% to 81% (Kim et al. 2025).
  • Causes of failure. In descending order, failures were caused by DNA yield below the sequencing cutoff (11.3%), very scant tumor of less than 10% (4.6%), low sequencing coverage (1.8%), sample-level DNA contamination (1.6%) and low DNA quality (0.1%) (Kim et al. 2025).
  • Optimization steps. Pretreatment of pelleted cells with 95% ethanol before HistoGel increased pellet density. Mineral oil deparaffinization reduced nucleic acid losses. In patients undergoing EBUS-TBNA for lung cancer, the change in deparaffinization alone raised sequencing success from 76.3% to 93% (Kim et al. 2025).
  • Internal and external cell blocks. Success rates for internal cell blocks reached 92% to 93%, against 79% to 82% for cell blocks from outside laboratories, in the last 2 years of the series (Kim et al. 2025).
  • Minimum DNA input. The minimum DNA input for cell blocks was lowered from 50 ng to 30 ng without a change in sequencing coverage (Kim et al. 2025).
  • Cross-contamination. Batching and pooling in tissue processors, shared embedding and cutting equipment and paper wrapping or HistoGel in cell block processing are points at which low-level cross-contamination can occur. Contamination checks are part of quality control for cytology samples (Kim et al. 2025).
  • Surgical comparison. Clinically relevant contamination occurred in 0.81% (5 of 619) of surgical samples, and the contaminated surgical samples were minute biopsies with low tumor purity (Kim et al. 2025).

Molecular testing on direct smears

  • Use of smears. Direct smears are suitable for molecular testing, and NGS can be performed on very small amounts of DNA from cytology samples (Lozano et al. 2018).
  • DNA extraction from slides. Scraping yields more DNA than cell lifting, and nonfrosted slides yield more DNA than fully frosted slides (Roy-Chowdhuri et al. 2016).
  • RNA fusion testing. An anchored multiplex PCR–based RNA fusion assay on 14 sacrificed smears obtained adequate nucleic acid in all cases and detected fusions in 64% (Wang et al. 2026).
  • Smear sacrifice and documentation are covered in Immunohistochemistry on Cytology Specimens.

Supernatant cell-free DNA from cytology specimens

Supernatant cell-free DNA (ScfDNA) is DNA recovered from the fluid that remains after centrifugation of an FNA rinse, effusion or liquid-based collection medium. This material is usually discarded (Hannigan et al. 2019).

Published performance of supernatant cell-free DNA

Study Specimens Main result
Hannigan et al. 2019 150 lung carcinoma FNA supernatants All yielded adequate DNA; NGS successful in 104 of 116 (90%); 100% concordance with concurrent FNA and 96% with concurrent core biopsy; 84% concordance with plasma for actionable driver mutations
Kim et al. 2025 268 supernatants tested after the cell block was unsuitable Successful in 71%; variant allele frequencies similar to the paired cell block
Thakur et al. 2024 20 pleural effusion supernatants from lung adenocarcinoma Concordant EGFR mutations in 10 of 11 positive cases; one additional EGFR T790M mutation; NGS found additional EGFR mutations missed by real-time PCR in 2 of 10
Gutierrez Amezcua et al. 2026 57 aggressive thyroid carcinomas (51 anaplastic) Cartridge-based BRAF p.V600E PCR successful in 90% of supernatant samples; 100% agreement with NGS or droplet digital PCR; median turnaround 2.86 hours, against 46.7 hours for BRAF V600E immunocytochemistry
Roy-Chowdhuri et al. 2026 22 FNA supernatants, rapid NGS assay 100% sensitivity, specificity and accuracy for single-nucleotide variants and insertions or deletions (variant allele frequency ≥0.5%) and for fusions; specimen stability up to day 7

Selection and interpretation of supernatant samples

  • Selection by the matched preparation. In one laboratory, the supernatant was considered suitable when the corresponding ThinPrep slide showed tumor cells at 10% or more by visual inspection (Kim et al. 2025).
  • Surrogate for tumor fraction. Variant allele frequencies in the supernatant were similar to those in the matched cell block. The cell block or cytopreparation therefore serves as a surrogate for the tumor-derived proportion of supernatant DNA (Kim et al. 2025).
  • DNA quality. Supernatant DNA is not exposed to formalin. Its integrity allows lower DNA inputs when the tumor proportion is suitable (Kim et al. 2025).
  • Quantity and success. In effusion supernatants, success of mutation detection did not correlate with cell-free DNA quantity or with tumor fraction in the sediment (Thakur et al. 2024).
  • Turnaround. Supernatant testing requires no histologic processing or preextraction, which gives the shortest turnaround among the preparations compared (Gutierrez Amezcua et al. 2026). Conventional tissue-based NGS usually takes several weeks (Roy-Chowdhuri et al. 2026).

Supernatant cell-free DNA compared with plasma cell-free DNA

Feature Cytology supernatant Plasma
Source Fluid from the sampled lesion after centrifugation Circulating blood
Guideline status in lung cancer (Lindeman et al. 2018) Covered by the acceptance of cytologic preparations; validation required Insufficient evidence for diagnosis of primary lung adenocarcinoma (no recommendation); may be used for EGFR mutations when tissue is limited (recommendation)
Tumor content check Paired cell block or cytopreparation (Kim et al. 2025) Not available from the same procedure
Concordance for actionable mutations (Hannigan et al. 2019) 100% with concurrent FNA, 96% with concurrent core biopsy Not reported against tissue; supernatant–plasma concordance 84%

Cell-free DNA and circulating tumor cells in the CAP/IASLC/AMP 2018 guideline

Statement Content Strength
13 Laboratories testing for EGFR T790M at secondary resistance to EGFR TKIs use assays able to detect T790M in as little as 5% of viable cells Recommendation
14 Insufficient evidence for or against routine ALK mutation testing at progression on an ALK TKI No recommendation
15 Insufficient evidence for circulating cell-free DNA methods in the diagnosis of primary lung adenocarcinoma No recommendation
16 When tissue is limited or insufficient, physicians may use a cell-free DNA assay to identify EGFR mutations Recommendation
17 Cell-free DNA may be used to identify EGFR T790M at progression on an EGFR TKI; the tumor sample is tested if the plasma result is negative Expert consensus opinion
18 Insufficient evidence for circulating tumor cell molecular analysis for diagnosis, mutation identification or T790M detection No recommendation

Source: Lindeman et al. 2018.

  • Plasma first or tissue first. The 2021 IASLC liquid biopsy statement addressed the "plasma first" and "tissue first" approaches to molecular testing in advanced non-small cell lung carcinoma, after FDA approval of two circulating tumor DNA assays (Rolfo et al. 2021).

Predictive biomarkers in lung carcinoma on cytology specimens

Genes and methods in the CAP/IASLC/AMP 2018 guideline

  • ROS1 scope. ROS1 testing is performed on all patients with advanced-stage lung adenocarcinoma, irrespective of clinical characteristics (statement 1; strong recommendation) (Lindeman et al. 2018).
  • Required and panel genes. The genes required by the guideline, the genes included in larger panels and the preference for multiplexed panels are covered in Metastatic Lung Adenocarcinoma in Serous Effusions.
  • Discordant results. Unexpected, discordant, equivocal or low-confidence results are confirmed or resolved with an alternative method or sample (Lindeman et al. 2018).
  • Non-adenocarcinoma histology. Molecular testing may be used in other histologies when clinical features indicate a higher probability of an oncogenic driver (Lindeman et al. 2018).

Immunohistochemistry as a predictive test on cytology specimens

Marker Guideline position Published cytology data
ALK IHC is an equivalent alternative to FISH (recommendation) (Lindeman et al. 2018) Cell block studies with the 5A4 or D5F3 clones: sensitivity 100%, specificity 83% to 100% against FISH; studies without the D5F3 Ventana assay: sensitivity 66% to 86% (Jain et al. 2019)
ROS1 IHC may be used as a screening test in advanced-stage lung adenocarcinoma; positive results confirmed by a molecular or cytogenetic method (statement 2; expert consensus opinion) (Lindeman et al. 2018) Scoring at ≥20 tumor cells; positivity commonly defined as H-score 100–150 or ≥2+ cytoplasmic staining in 50% to 75% of tumor cells (Jain et al. 2019)
EGFR, mutation-specific antibodies Not addressed as a replacement for mutation testing Specificity 96% to 99%; sensitivity about 16% to 30% on cytology when only strong membranous staining is accepted (Jain et al. 2019)
  • ALK FISH on the FDA-approved kit. The Vysis ALK Break Apart FISH assay is positive when more than 25 of 50 tumor nuclei (>50%) show split or isolated 3′ (orange) signals, or, after an equivocal first count of 5 to 25 of 50 (10% to 50%), when 15% or more of 100 nuclei scored by two readers are positive. The approved specimen is formalin-fixed, paraffin-embedded tissue (FDA PMA P110012, package insert). Scoring details are given in Flow Cytometry, FISH, Microbiology and Electron Microscopy on Cytology Specimens.
  • Antibody-drug conjugate targets. HER2, MET, TROP2 and HER3 have gained therapeutic relevance as antibody-drug conjugate targets in lung carcinoma, and MTAP deficiency can be assessed by IHC on cell blocks (Osama et al. 2026).

PD-L1 testing on cytology specimens

Evidence Result Source
Proportion of cytology specimens evaluable 92.0% of 709 specimens in 7 publications Gosney et al. 2020
Agreement with histology at the ≥50% cutoff (9 studies) Overall 89.7% (n = 428); positive 72.0% (n = 218); negative 95.0% (n = 258) Gosney et al. 2020
Agreement at the ≥1% cutoff Similar to the ≥50% cutoff Gosney et al. 2020
22C3 on cell blocks (94 cytology, 138 surgical) No significant difference in tumor proportion score; about 35% with TPS ≥50% in both groups; similar response to pembrolizumab Torous et al. 2018
22C3 and 28-8 clones on cell blocks 95% concordance Jain et al. 2019
Direct smears and cell blocks against biopsies 97% for smears and small biopsies, 82% for cell blocks, in one study Jain et al. 2019
  • Regulatory status. PD-L1 immunohistochemical testing was approved for FFPE surgical pathology specimens (Torous et al. 2018). Clinical trial–validated assays were validated with histology specimens, although about 30% to 40% of patients with advanced non-small cell lung carcinoma have only cytology specimens (Gosney et al. 2020).
  • Validation. CAP validation requirements for PD-L1 assays on cytology specimens (statements 6 and 9) are covered in Immunohistochemistry on Cytology Specimens.
  • Laboratory-developed tests. Many laboratories use laboratory-developed PD-L1 tests, and the specimen type, assay validation and external quality assurance are part of appropriate implementation (Lantuejoul et al. 2020).
  • Fixative effects on PD-L1 staining in cell blocks are covered in Immunohistochemistry on Cytology Specimens. The tumor proportion score and its exclusions are covered in Metastatic Lung Adenocarcinoma in Serous Effusions.

Choice of cytology material for molecular testing

Situation Preferred material Reason
Cellular cell block with adequate tumor content Cell block sections Highest success rate in the largest series (Kim et al. 2025)
Cell block exhausted or acellular; tumor on smears Scraped direct smear Smears are suitable for DNA and RNA testing (Lozano et al. 2018; Wang et al. 2026)
Cell block unsuitable; residual liquid-based medium with tumor on the matched preparation Supernatant cell-free DNA Rescue testing successful in 71% (Kim et al. 2025)
Need for a same-day single-gene result (for example BRAF p.V600E in anaplastic thyroid carcinoma) Supernatant cell-free DNA on a rapid platform Median turnaround 2.86 hours (Gutierrez Amezcua et al. 2026)
Progression on an EGFR TKI; plasma negative for T790M New tumor sample Statement 17 (Lindeman et al. 2018)

Digital Image Analysis and Artificial Intelligence in Cytology

Definition. Digital cytology is the capture, storage, transmission and interpretation of cytology slides as digital images, most often as whole slide images (WSI). Digital image analysis applies software to these images to measure features such as nuclear staining or to locate objects of interest. Artificial intelligence (AI) in cytology uses machine learning algorithms, most often deep learning, to detect, classify or triage cells. The American Society of Cytopathology (ASC), with the International Academy of Cytology and the Digital Pathology Association, published a two-part white paper in 2024 on digital cytology implementation and on AI in cytology, with best practice recommendations and legal considerations (Kim et al. 2024a; Kim et al. 2024b).

Related page: Immunohistochemistry on Cytology Specimens (immunohistochemistry substrates and validation) Related page: Molecular Testing and Predictive Biomarkers on Cytology Specimens (PD-L1 and predictive biomarkers) Related page: The Bethesda System for Reporting Cervical Cytology (cervical cytology categories) Related page: Cervical Cancer Screening, Specimen Collection and Quality Assurance in Cervical Cytology (screening and quality assurance) Related page: The Paris System for Reporting Urinary Cytology (urine cytology categories) Related page: Ancillary Tests, Indications and Management in Urine Cytology (urine biomarkers and artificial intelligence in urine cytology)

Technical problems of whole slide imaging in cytology

Focusing problems in three-dimensional cytology preparations

  • Three-dimensional preparations. WSI is used less in cytology than in histology because cytology preparations are three-dimensional. Capturing z-stack information increases file size and creates usability problems in viewing (Capitanio et al. 2018).
  • Sources of focusing difficulty. Three-dimensional cell clusters, variable thickness and overlapping cells increase the demands on focusing, scanning time, image storage and infrastructure (Laohawetwanit et al. 2026).
  • Z-stacking. Z-stacking scans each slide at multiple focal planes, and the viewer moves between the planes during review. Each additional layer multiplies the file size and lengthens scanning. There is no consensus on the added value of z-stacking for clinical use (Mistry et al. 2026).
  • Liquid-based preparations. ThinPrep and other liquid-based preparations may reduce the problem by concentrating cells in a smaller area with fewer clusters (Kim et al. 2022).
  • Deferral to glass. Cases with three-dimensional clusters or other features that are difficult to visualize digitally can be deferred to glass slides (Mistry et al. 2026).

Measured effects of z-stacking and magnification

Study Material Finding
Kim et al. 2022 30 urine ThinPrep slides (22 high-grade urothelial carcinoma) Scan time and file size 6.27 minutes and 0.827 GB without z-stack, against 14.06 minutes and 2.650 GB with a 3-layer z-stack; concordance with the original report κ 0.606–1.0 without and 0.533–1.0 with z-stack; minimal diagnostic benefit
Hanna et al. 2017 10 ThinPrep Papanicolaou tests and 20 nongynecologic cases Region-of-interest z-stacks permitted fine focusing of thick smears and clusters; screening took 5.5 times (panoramic system) and 1.8 times (single-plane WSI) as long as glass; no difference in diagnostic concordance
Ahmed et al. 2025 98 thyroid FNAs, 463 slides scanned at ×20 Mean scan time 5.48 minutes; z-stacking requested in 23% and ×40 in 14% of reviews; intrareader agreement κ 0.86–0.95
Satturwar et al. 2026 100 thyroid ThinPrep FNAs with AI assistance A 7-layer z-stack did not improve concordance over single-layer WSI; review time with AI assistance was shorter on the z-stacked images

Validation of whole slide imaging for cytology

Guidelines for validation of whole slide imaging in cytology

  • Update of 2022. The CAP guideline on validating WSI systems for diagnostic purposes updated the 2013 guideline of 12 statements. It offers 3 strong recommendations and 9 good practice statements (Evans et al. 2022).
  • Cytology excluded. The systematic review covered frozen sections, anatomic pathology specimens and hematopathology cases. Cytology cases were excluded (Evans et al. 2022).
  • Core requirements reaffirmed. The update reaffirms a validation set of at least 60 cases, intraobserver diagnostic concordance between WSI and glass slides and a 2-week washout period between modalities. Overall WSI–glass concordance below 95% is a cause for concern, and all discordances are reconciled (Evans et al. 2022).
  • Cytology-specific guidance. The ASC Digital Cytology Task Force white paper offers best practice recommendations for incorporating digital cytology into practice (Kim et al. 2024a).
  • Training period in the United Kingdom. Royal College of Pathologists guidance requires a strictly monitored period of double reporting with glass and digital slides before digital reporting (Mistry et al. 2026).
  • Quality assurance programs. Several national cytology quality assurance programs include digital cytology (Mistry et al. 2026).

Published validation studies of digital cytology

Study Design Concordance
Mistry et al. 2026 88 cases, gynecologic and nongynecologic, smears, ThinPreps, cytospins and cell blocks; scanned at 40× equivalent (0.23 μm/pixel) 95.1% after exclusion of cases with discrepancy on glass; 89.5% to 89.8% with all cases included; κ 0.86
Underwood et al. 2026 1,748 Papanicolaou tests, Genius Digital Diagnostics System against the ThinPrep Imaging System 98.5% (1,722 of 1,748); 95% concordance required for validation
Ahmed et al. 2025 98 thyroid FNAs for rapid on-site evaluation, five cytopathologists, 2-week washout Intrareader κ 0.86–0.95; interreader κ 0.48 for WSI and 0.56 for light microscopy
Hanna et al. 2017 30 slides, glass against two digital modalities No statistical difference in diagnostic concordance
  • Sources of discordance. In the Mistry study, false-negative digital diagnoses included missed nonnecrotizing granulomas in a lymph node, adenocarcinoma cells in cerebrospinal fluid and bronchial brushings and squamous cell carcinoma in sputum (Mistry et al. 2026).
  • Unsatisfactory and negative Papanicolaou tests. Of the 26 discordant Papanicolaou tests, 20 were discordances between unsatisfactory and negative for intraepithelial lesion or malignancy (NILM), attributed to cellularity quantification by the digital system (Underwood et al. 2026).
  • Thyroid categories. WSI concordance for thyroid FNA was highest for Bethesda categories I and VI and lowest for categories III to V (Ahmed et al. 2025).
  • Interpretation time. Interpretation with WSI took significantly longer than with light microscopy in the thyroid rapid on-site evaluation study (Ahmed et al. 2025).
  • Effect of excluding discrepant cases. Exclusion of cases that lack diagnostic consensus on glass may inflate concordance estimates, because some excluded cases are the more difficult ones (Mistry et al. 2026).
  • Range of specimen types. Most earlier validation studies were restricted to a single specimen or preparation type, with a median sample size of 22 (Mistry et al. 2026).
  • Detection by the digital system. One low-cellularity LSIL with rare diagnostic cells in the Genius image tiles had been interpreted as NILM on the ThinPrep Imaging System (Underwood et al. 2026).
  • Future preparations. Expansion of digital cytopathology may increase the use of liquid-based preparations relative to direct smears. Computational approaches such as semantic or extended focusing have also been proposed (Mistry et al. 2026).

Telecytology and remote rapid on-site evaluation

  • Rapid on-site evaluation by WSI. WSI is a feasible alternative to light microscopy for rapid on-site evaluation of thyroid FNAs, with high intrareader agreement and comparable interreader agreement (Ahmed et al. 2025).
  • Uses of telecytology. Telecytology has shown high diagnostic concordance and improves access to expert assessment, particularly for rapid on-site evaluation in remote or resource-limited settings (Laohawetwanit et al. 2026).
  • Applications of digital cytology. Digital cytology is used for primary diagnosis, remote rapid on-site evaluation, consultation, quality assurance, education, archiving and research (Laohawetwanit et al. 2026).

Computer-assisted screening of cervical cytology

System Preparation Evidence Source
ThinPrep Imaging System ThinPrep 10,742 specimens at 4 sites; more sensitive than manual screening for ASC-US+; equivalent for LSIL+ and HSIL+; higher specificity for HSIL+; daily screening rates doubled Biscotti et al. 2005
BD FocalPoint GS Imaging System SurePath 12,313 slides; sensitivity for HSIL+ increased by 19.6% and for LSIL+ by 9.8%, with small decreases in specificity; ASC-US+ not different Wilbur et al. 2009
Genius Digital Diagnostics System with the Genius Cervical AI algorithm ThinPrep FDA De Novo request granted January 31, 2024; classified as a digital cervical cytology slide imaging system with an AI algorithm (21 CFR 864.3900, product code QYV) FDA De Novo DEN210035
  • Workflow effect of the Genius system. Hands-on time fell from more than 5 hours with the ThinPrep Imaging System workflow to just over 1 hour. Review time per case fell from 138.0 to 70.1 seconds, with similar agreement with the sign-out diagnosis (Murphy et al. 2025).
  • Location-guided screening. The FocalPoint GS system combined computer-assisted, field-of-view screening with device-directed quality control rescreening (Wilbur et al. 2009).
  • Cervical cytology reporting categories are covered in The Bethesda System for Reporting Cervical Cytology.

Artificial intelligence in cervical cytology

  • Training and design. A supervised deep learning algorithm was trained on 188,542 digital cytology images. In 2,145 women referred from organized screening, two liquid-based slides were prepared from each sample, one for AI-assisted reading and one for manual reading (Bao et al. 2020).
  • Detection of CIN2+. AI-assisted reading detected 92.6% of CIN2 and 96.1% of CIN3+ (Bao et al. 2020).
  • Comparison with skilled cytologists. AI-assisted reading had equivalent sensitivity (relative sensitivity 1.01) and higher specificity (relative specificity 1.26) (Bao et al. 2020).
  • Comparison with less experienced readers. Against cytology doctors from primary hospitals, AI-assisted reading had higher sensitivity (1.12) and specificity (1.36) (Bao et al. 2020).
  • HPV-positive women. In HPV-positive women, AI-assisted reading improved specificity for CIN1 or less without loss of sensitivity (Bao et al. 2020).
  • Population. The study population was a referral population. The authors called for studies in the general population (Bao et al. 2020).

Artificial intelligence in nongynecologic cytology

Specimen System or method Result Source
Urine (200 ThinPrep slides) AIxURO, AI-assisted digital urine cytology on two scanners Sensitivity 85.0% and 88.3% against 79.3% for microscopy; specificity 85.7% and 82.3% against 94.3%; positive predictive value 85.6% and 83.3% against 93.3%; median review time reduced by 66% to 78% Lajara et al. 2026
Thyroid FNA (908 cases with surgical follow-up) Machine learning algorithm on WSI Sensitivity 92.0% and specificity 90.5% for malignancy; area under the curve 0.932, against 0.931 for the cytopathologist Elliott Range et al. 2020
Thyroid FNA (100 ThinPrep slides) AIxTHY on single-layer and 7-layer WSI Sensitivity for Bethesda III or higher rose from 60% to 78% and 79%; specificity fell from more false-positive indeterminate calls; review time reduced by 40% to 65% Satturwar et al. 2026
  • Trade-off in urine cytology. The sensitivity gain with AIxURO came with lower specificity and positive predictive value. The AI modalities reduced the number of atypical-or-worse cases classified as negative for high-grade urothelial carcinoma (Lajara et al. 2026).
  • Biopsy-confirmed urine cases. In the 98 urine cases (49%) with biopsy confirmation, AIxURO showed higher sensitivity and negative predictive value than microscopy (Lajara et al. 2026).
  • Combined reading in thyroid FNA. The combination of the algorithm with the cytopathologist's diagnosis performed better than either alone, and the authors proposed the algorithm as an adjunct for the indeterminate categories (Elliott Range et al. 2020).
  • Urine and thyroid reporting categories are covered in The Paris System for Reporting Urinary Cytology and The Bethesda System for Reporting Thyroid Cytopathology. Other AI systems for urine cytology are covered in Ancillary Tests, Indications and Management in Urine Cytology.

Quantitative image analysis of immunostains on cytology material

Application Finding Source
HER2 immunohistochemistry in breast cancer Quantitative image analysis is validated before implementation, followed by maintenance and ongoing quality control and quality assurance; performance and reporting are supervised by pathologists with expertise in image analysis Bui et al. 2019
Ki-67 grading of pancreatic neuroendocrine neoplasms on EUS-FNA cell blocks (58 cases, >100 tumor cells) Manual count graded correctly in 69% (κ 0.44); digital image analysis in 55% (κ 0.19); manual count more accurate Jin et al. 2016
PD-L1 tumor proportion score in non-small cell lung carcinoma (51 SP263 cases) Agreement with the median pathologist score at the 50% cutoff: fair for one commercial algorithm (κ 0.354), substantial for another (κ 0.672) Plass et al. 2025
  • Causes of Ki-67 discordance on cell blocks. Nontumor cell contamination and insufficient sampling were the main reasons for discordance with surgical grading (Jin et al. 2016).
  • Grade 2 tumors. Ki-67 grading on cell blocks was least reliable for grade 2 tumors (Jin et al. 2016).
  • Pathologist agreement on PD-L1. Interobserver agreement among pathologists was moderate at a TPS below 1% (κ 0.558) and almost perfect at 50% or more (κ 0.873). Intraobserver agreement ranged from κ 0.726 to 1.0 (Plass et al. 2025).
  • Structure of the CAP HER2 image analysis guideline. The guideline contains 11 recommendations: 7 based on CAP laboratory accreditation requirements and 4 based on expert consensus opinion (Bui et al. 2019).
  • Fixative effects on Ki-67 in CytoLyt-fixed cell blocks are covered in Immunohistochemistry on Cytology Specimens.

Limits on adoption of digital cytology and artificial intelligence

Barrier Description Source
Image acquisition Three-dimensional clusters, variable thickness and overlapping cells Laohawetwanit et al. 2026
Storage and scan time File size multiplies with each z-stack layer Mistry et al. 2026; Kim et al. 2022
Regulation and liability Regulatory and medicolegal concerns Laohawetwanit et al. 2026; Kim et al. 2024b
Validation Validation and training are required; the CAP WSI guideline excluded cytology Laohawetwanit et al. 2026; Evans et al. 2022
Specificity of AI triage Sensitivity gains with lower specificity in urine and thyroid cytology Lajara et al. 2026; Satturwar et al. 2026

References

  1. Adiga DSA, V S HK, Bhatt AS, Patel S, Shetty DD. Analysis of a two-marker immunohistochemistry panel (TTF-1 and p40) for distinguishing lung adenocarcinoma from squamous cell carcinoma on destained direct cytologic smears. Ann Diagn Pathol. 2026;82:152617. doi:10.1016/j.anndiagpath.2026.152617. PMID 41713158
  2. Ahmed MS, Klippel-Almaraz D, Amin SE, et al. Utility of whole-slide imaging for rapid evaluation of thyroid FNA: a multireader prospective study. Cancer Cytopathol. 2025;133(9):e70046. doi:10.1002/cncy.70046. PMID 40889104
  3. Al-Abbadi MA, Barroca H, Bode-Lesniewska B, et al. A proposal for the performance, classification, and reporting of lymph node fine-needle aspiration cytopathology: the Sydney System. Acta Cytol. 2020;64(4):306–322. doi:10.1159/000506497. PMID 32454496
  4. Bandargal S, Florianova L, Dmitrienko S, et al. Fine needle aspirate flow cytometry's ancillary utility in diagnosing non-Hodgkin lymphoma in the head and neck. J Otolaryngol Head Neck Surg. 2024;53:19160216241296127. doi:10.1177/19160216241296127. PMID 39511788
  5. Bao H, Bi H, Zhang X, et al. Artificial intelligence-assisted cytology for detection of cervical intraepithelial neoplasia or invasive cancer: a multicenter, clinical-based, observational study. Gynecol Oncol. 2020;159(1):171–178. doi:10.1016/j.ygyno.2020.07.099. PMID 32814641
  6. Betz BL, Dixon CA, Weigelin HC, Knoepp SM, Roh MH. The use of stained cytologic direct smears for ALK gene rearrangement analysis of lung adenocarcinoma. Cancer Cytopathol. 2013;121(9):489–499. doi:10.1002/cncy.21286. PMID 23536384
  7. Biscotti CV, Dawson AE, Dziura B, et al. Assisted primary screening using the automated ThinPrep Imaging System. Am J Clin Pathol. 2005;123(2):281–287. PMID 15842055
  8. Bui MM, Riben MW, Allison KH, et al. Quantitative image analysis of human epidermal growth factor receptor 2 immunohistochemistry for breast cancer: guideline from the College of American Pathologists. Arch Pathol Lab Med. 2019;143(10):1180–1195. doi:10.5858/arpa.2018-0378-CP. PMID 30645156
  9. Buonocore DJ, Konno F, Jungbluth AA, et al. CytoLyt fixation significantly inhibits MIB1 immunoreactivity whereas alternative Ki-67 clone 30-9 is not susceptible to the inhibition: critical diagnostic implications. Cancer Cytopathol. 2019;127(10):643–649. doi:10.1002/cncy.22170. PMID 31398281
  10. Capitanio A, Dina RE, Treanor D. Digital cytology: a short review of technical and methodological approaches and applications. Cytopathology. 2018;29(4):317–325. doi:10.1111/cyt.12554. PMID 29665178
  11. College of American Pathologists. Principles of analytic validation of immunohistochemical assays (guideline page and frequently asked questions). https://cap.org/protocols-and-guidelines/cap-guidelines/current-cap-guidelines/principles-of-analytic-validation-of-immunohistochemical-assays
  12. Crapanzano JP, Heymann JJ, Monaco S, Nassar A, Saqi A. The state of cell block variation and satisfaction in the era of molecular diagnostics and personalized medicine. Cytojournal. 2014;11:7. doi:10.4103/1742-6413.129187. PMID 24799951
  13. Dong HY, Harris NL, Preffer FI, Pitman MB. Fine-needle aspiration biopsy in the diagnosis and classification of primary and recurrent lymphoma: a retrospective analysis of the utility of cytomorphology and flow cytometry. Mod Pathol. 2001;14(5):472–481. doi:10.1038/modpathol.3880336. PMID 11353059
  14. Ehinger M, Calaminici M, Cozzolino I, Zeppa P, Field AS. Practical implications of the World Health Organization Reporting System for Lymph Node, Spleen, and Thymus Cytopathology: categories and ancillary testing for subtyping of hematolymphoid tumors on FNA biopsy cytopathology using a pattern-based approach. Cancer Cytopathol. 2026;134(1):e70066. doi:10.1002/cncy.70066. PMID 41432088
  15. Elliott Range DD, Dov D, Kovalsky SZ, Henao R, Carin L, Cohen J. Application of a machine learning algorithm to predict malignancy in thyroid cytopathology. Cancer Cytopathol. 2020;128(4):287–295. doi:10.1002/cncy.22238. PMID 32012493
  16. Etit D, Selig M, Faquin WC, et al. The utility of electron microscopy in cytopathology. Ultrastruct Pathol. 2010;34(5):252–259. doi:10.3109/01913121003775114. PMID 20568988
  17. Evans AJ, Brown RW, Bui MM, et al. Validating whole slide imaging systems for diagnostic purposes in pathology. Arch Pathol Lab Med. 2022;146(4):440–450. doi:10.5858/arpa.2020-0723-CP. PMID 34003251
  18. Fischer AH, Schwartz MR, Moriarty AT, et al. Immunohistochemistry practices of cytopathology laboratories: a survey of participants in the College of American Pathologists Nongynecologic Cytopathology Education Program. Arch Pathol Lab Med. 2014;138(9):1167–1172. doi:10.5858/arpa.2013-0259-CP. PMID 24840035
  19. Fitzgibbons PL, Bradley LA, Fatheree LA, et al. Principles of analytic validation of immunohistochemical assays: guideline from the College of American Pathologists Pathology and Laboratory Quality Center. Arch Pathol Lab Med. 2014;138(11):1432–1443. doi:10.5858/arpa.2013-0610-CP. PMID 24646069
  20. Garlin-Politis M, Upadhyay Baskota S, Picon S, et al. CytoLyt fixation impedes insulinoma-associated protein 1 (INSM1) immunoreactivity compared to formalin fixation. J Am Soc Cytopathol. 2024;13(3):213–218. doi:10.1016/j.jasc.2024.02.002. PMID 38575468
  21. Goldsmith JD, Troxell ML, Roy-Chowdhuri S, et al. Principles of analytic validation of immunohistochemical assays: guideline update. Arch Pathol Lab Med. 2024;148(6):e111–e153. doi:10.5858/arpa.2023-0483-CP. PMID 38391878
  22. Gosney JR, Boothman AM, Ratcliffe M, Kerr KM. Cytology for PD-L1 testing: a systematic review. Lung Cancer. 2020;141:101–106. doi:10.1016/j.lungcan.2020.01.010. PMID 32007657
  23. Gruchy JR, Barnes PJ, Dakin Haché KA. CytoLyt fixation and decalcification pretreatments alter antigenicity in normal tissues compared with standard formalin fixation. Appl Immunohistochem Mol Morphol. 2015;23(4):297–302. doi:10.1097/PAI.0000000000000082. PMID 25265430
  24. Gutierrez Amezcua JM, Arcila ME, Ng DL, et al. Ultrarapid BRAF mutation detection on supernatant cell-free DNA obtained by FNA: an accurate and expedient method for BRAF assessment in aggressive thyroid carcinomas. Cancer Cytopathol. 2026;134(5):e70098. doi:10.1002/cncy.70098. PMID 42018428
  25. Hanna MG, Monaco SE, Cuda J, Xing J, Ahmed I, Pantanowitz L. Comparison of glass slides and various digital-slide modalities for cytopathology screening and interpretation. Cancer Cytopathol. 2017;125(9):701–709. doi:10.1002/cncy.21880. PMID 28558124
  26. Hannigan B, Ye W, Mehrotra M, et al. Liquid biopsy assay for lung carcinoma using centrifuged supernatants from fine-needle aspiration specimens. Ann Oncol. 2019;30(6):963–969. doi:10.1093/annonc/mdz102. PMID 30887015
  27. Heymann JJ, Pineda CM, Booth CN, et al. Tools, techniques, and challenges in preparing cytology specimens for ancillary studies: results of the ASC Optimizing Cytology and Small Biopsy Specimen Processing for Ancillary Studies task force survey. J Am Soc Cytopathol. 2025;14(1):55–63. doi:10.1016/j.jasc.2024.10.001. PMID 39547835
  28. Jain D, Nambirajan A, Borczuk A, et al. Immunocytochemistry for predictive biomarker testing in lung cancer cytology. Cancer Cytopathol. 2019;127(5):325–339. doi:10.1002/cncy.22137. PMID 31050216
  29. Jin M, Roth R, Gayetsky V, Niederberger N, Lehman A, Wakely PE Jr. Grading pancreatic neuroendocrine neoplasms by Ki-67 staining on cytology cell blocks: manual count and digital image analysis of 58 cases. J Am Soc Cytopathol. 2016;5(5):286–295. doi:10.1016/j.jasc.2016.03.002. PMID 31042505
  30. Kandalaft PL, Gown AM. Practical applications in immunohistochemistry: carcinomas of unknown primary site. Arch Pathol Lab Med. 2016;140(6):508–523. doi:10.5858/arpa.2015-0173-CP. PMID 26457625
  31. Kim D, Burkhardt R, Alperstein SA, et al. Evaluating the role of Z-stack to improve the morphologic evaluation of urine cytology whole slide images for high-grade urothelial carcinoma: results and review of a pilot study. Cancer Cytopathol. 2022;130(8):630–639. doi:10.1002/cncy.22595. PMID 35584402
  32. Kim D, Sundling KE, Virk R, et al. Digital cytology part 1: digital cytology implementation for practice: a concept paper with review and recommendations from the American Society of Cytopathology Digital Cytology Task Force. J Am Soc Cytopathol. 2024;13(2):86–96. doi:10.1016/j.jasc.2023.11.006. PMID 38158316
  33. Kim D, Sundling KE, Virk R, et al. Digital cytology part 2: artificial intelligence in cytology: a concept paper with review and recommendations from the American Society of Cytopathology Digital Cytology Task Force. J Am Soc Cytopathol. 2024;13(2):97–110. doi:10.1016/j.jasc.2023.11.005. PMID 38158317
  34. Kim D, Vanderbilt CM, Yang SR, et al. Maximizing the clinical utility and performance of cytology samples for comprehensive genetic profiling. Nat Commun. 2025;16(1):116. doi:10.1038/s41467-024-55456-8. PMID 39747849
  35. Kohli M, Inbaraj LR, Salomon A, et al. Low-complexity automated nucleic acid amplification tests for extrapulmonary tuberculosis and rifampicin resistance in adults and adolescents. Cochrane Database Syst Rev. 2025;8(8):CD012768. doi:10.1002/14651858.CD012768.pub4. PMID 40757508
  36. Lajara S, Cuda J, Geisler DL, et al. Enhancing urothelial carcinoma diagnosis with artificial intelligence-integrated urine cytology: biopsy-validated accuracy and efficiency gain. Cancer Cytopathol. 2026;134(6):e70110. doi:10.1002/cncy.70110. PMID 42144873
  37. Lantuejoul S, Sound-Tsao M, Cooper WA, et al. PD-L1 testing for lung cancer in 2019: perspective from the IASLC Pathology Committee. J Thorac Oncol. 2020;15(4):499–519. doi:10.1016/j.jtho.2019.12.107. PMID 31870882
  38. Laohawetwanit T, Chong Y, Yoshikawa AL, Bychkov A. Cytology 2.0: digital cytopathology, telepathology, and AI implementation. Semin Diagn Pathol. 2026;43(6):151064. doi:10.1016/j.semdp.2026.151064. PMID 42962376
  39. Lindeman NI, Cagle PT, Aisner DL, et al. Updated molecular testing guideline for the selection of lung cancer patients for treatment with targeted tyrosine kinase inhibitors: guideline from the College of American Pathologists, the International Association for the Study of Lung Cancer, and the Association for Molecular Pathology. Arch Pathol Lab Med. 2018;142(3):321–346. doi:10.5858/arpa.2017-0388-CP. PMID 29355391. Also published in J Mol Diagn. 2018;20(2):129–159. doi:10.1016/j.jmoldx.2017.11.004. PMID 29398453
  40. Lozano MD, Echeveste JI, Abengozar M, et al. Cytology smears in the era of molecular biomarkers in non-small cell lung cancer: doing more with less. Arch Pathol Lab Med. 2018;142(3):291–298. doi:10.5858/arpa.2017-0208-RA. PMID 29494220
  41. Mansour MSI, Pettersson L, Pedersen R, et al. Choice of fixative affects programmed death-ligand 1 expression in cell blocks from pleural effusions with metastatic pulmonary adenocarcinoma. Cancer Cytopathol. 2026;134(3):e70078. doi:10.1002/cncy.70078. PMID 41663326
  42. Meda BA, Buss DH, Woodruff RD, et al. Diagnosis and subclassification of primary and recurrent lymphoma: the usefulness and limitations of combined fine-needle aspiration cytomorphology and flow cytometry. Am J Clin Pathol. 2000;113(5):688–699. doi:10.1309/0Q7F-QTGM-6DPD-TLGY. PMID 10800402
  43. Miller JM, Binnicker MJ, Campbell S, et al. A guide to utilization of the microbiology laboratory for diagnosis of infectious diseases: 2018 update (2018 edition) by the Infectious Diseases Society of America and the American Society for Microbiology. Clin Infect Dis. 2018;67(6):e1–e94. doi:10.1093/cid/ciy381. PMID 29955859
  44. Miller JM, Binnicker MJ, Campbell S, et al. Guide to utilization of the microbiology laboratory for diagnosis of infectious diseases: 2024 update by the Infectious Diseases Society of America (IDSA) and the American Society for Microbiology (ASM). Clin Infect Dis. 2024 (online ahead of print). doi:10.1093/cid/ciae104. PMID 38442248
  45. Mistry T, Hunter H, Oukrif D, et al. Validation of digital cytology for primary diagnosis across a range of specimen types. Cytopathology. 2026;37(3):222–231. doi:10.1111/cyt.70063. PMID 41808272
  46. Mitra SK, Misra RK, Rai P. Cytomorphological patterns of tubercular lymphadenitis and its comparison with Ziehl-Neelsen staining and culture in eastern UP (Gorakhpur region): cytological study of 400 cases. J Cytol. 2017;34(3):139–143. doi:10.4103/JOC.JOC_207_15. PMID 28701826
  47. Morito S, Nitanda T, Tsukamoto R, et al. Rapid cell transfer by means of nylon mesh to improve cellular diagnosis: the role of immunocytochemistry. Acta Cytol. 2021;65(5):424–429. doi:10.1159/000516555. PMID 34098550
  48. Murphy KM, Weatherhead K, Chenault C, et al. Impact of the Genius Digital Diagnostics System on workflow and accuracy compared with the ThinPrep Imaging System for review of ThinPrep Papanicolaou tests. Am J Clin Pathol. 2025;164(5):746–751. doi:10.1093/ajcp/aqaf099. PMID 40994090
  49. Nambirajan A, Jain D. Cell blocks in cytopathology: an update. Cytopathology. 2018;29(6):505–524. doi:10.1111/cyt.12627. PMID 30153355
  50. Nambirajan A, Rana D, Samant K, Prabakaran A, Malik P, Jain D. Multiplex fluorescence in situ hybridization testing for anaplastic lymphoma kinase and c-ros oncogene 1 gene rearrangements on cytology smears in lung adenocarcinomas: comparative study with formalin-fixed paraffin-embedded sections. J Am Soc Cytopathol. 2022;11(3):154–164. doi:10.1016/j.jasc.2022.01.001. PMID 35148960
  51. Osama MA, Kamaluddin Z, Bubendorf L, Jain D. Predictive immunohistochemical biomarkers for antibody-drug conjugate therapy in pulmonary cytology. Acta Cytol. 2026 (online ahead of print). doi:10.1159/000553423. PMID 42406652
  52. Plass M, Olteanu GE, Dacic S, et al. Comparative performance of PD-L1 scoring by pathologists and AI algorithms. Histopathology. 2025;87(1):90–100. doi:10.1111/his.15432. PMID 39961605
  53. Ravinsky E, Quinonez GE, Paraskevas M, Diocee MS, Kutryk EM. Processing fine needle aspiration biopsies for electron microscopic examination: experience implementing a procedure. Acta Cytol. 1993;37(5):661–666. PMID 8362575
  54. Rolfo C, Mack P, Scagliotti GV, et al. Liquid biopsy for advanced NSCLC: a consensus statement from the International Association for the Study of Lung Cancer. J Thorac Oncol. 2021;16(10):1647–1662. doi:10.1016/j.jtho.2021.06.017. PMID 34246791
  55. Roy-Chowdhuri S, Agarwal N, Gulraiz F, et al. Analytical validation of a rapid next generation sequencing-based liquid biopsy assay using fine needle aspiration supernatants for clinical use. J Am Soc Cytopathol. 2026 (online ahead of print). doi:10.1016/j.jasc.2026.05.003. PMID 42309846
  56. Roy-Chowdhuri S, Chow CW, Kane MK, et al. Optimizing the DNA yield for molecular analysis from cytologic preparations. Cancer Cytopathol. 2016;124(4):254–260. doi:10.1002/cncy.21664. PMID 26630358
  57. Roy-Chowdhuri S, Dacic S, Ghofrani M, et al. Collection and handling of thoracic small biopsy and cytology specimens for ancillary studies: guideline from the College of American Pathologists in collaboration with the American College of Chest Physicians, Association for Molecular Pathology, American Society of Cytopathology, American Thoracic Society, Pulmonary Pathology Society, Papanicolaou Society of Cytopathology, Society of Interventional Radiology, and Society of Thoracic Radiology. Arch Pathol Lab Med. 2020;144(8):933–958. doi:10.5858/arpa.2020-0119-CP. PMID 32401054
  58. Sakuma N, Kamei T, Ishihara T. Ultrastructure of pleural mesothelioma and pulmonary adenocarcinoma in malignant effusions as compared with reactive mesothelial cells. Acta Cytol. 1999;43(5):777–785. doi:10.1159/000331291. PMID 10518130
  59. Satturwar S, Li Z, Yang CS, et al. Artificial intelligence-assisted digital thyroid FNA cytology: improved agreement and sensitivity for higher-risk Bethesda categories with enhanced screening efficiency. Cancer Cytopathol. 2026;134(10):e70149. doi:10.1002/cncy.70149. PMID 42779199
  60. Savic S, Bubendorf L. Common fluorescence in situ hybridization applications in cytology. Arch Pathol Lab Med. 2016;140(12):1323–1330. doi:10.5858/arpa.2016-0202-RA. PMID 27479335
  61. Selves J, Long-Mira E, Mathieu MC, Rochaix P, Ilié M. Immunohistochemistry for diagnosis of metastatic carcinomas of unknown primary site. Cancers (Basel). 2018;10(4):108. doi:10.3390/cancers10040108. PMID 29621151
  62. Sewell WA, Al-Abbadi MA, Geddie WR, et al. The power of flow cytometry in the cytopathologic diagnosis of lymphoid lesions. Acta Cytol. 2026 (online ahead of print). doi:10.1159/000550290. PMID 41505355
  63. Sherman ME, Jimenez-Joseph D, Gangi MD, Rojas-Corona RR. Immunostaining of small cytologic specimens: facilitation with cell transfer. Acta Cytol. 1994;38(1):18–22. PMID 8291350
  64. Thakur S, Rathor A, Jain S, et al. Pleural effusion supernatant: a reliable resource for cell-free DNA in molecular testing of lung cancer. J Am Soc Cytopathol. 2024;13(4):291–302. doi:10.1016/j.jasc.2024.03.006. PMID 38704351
  65. Torous VF, Rangachari D, Gallant BP, et al. PD-L1 testing using the clone 22C3 pharmDx kit for selection of patients with non-small cell lung cancer to receive immune checkpoint inhibitor therapy: are cytology cell blocks a viable option? J Am Soc Cytopathol. 2018;7(3):133–141. doi:10.1016/j.jasc.2018.02.003. PMID 29713584
  66. Underwood DL, Chaari RR, Webb S, Booth CN, Doxtader EE. Evaluation of the Hologic Genius Digital Diagnostics System for clinical use in Papanicolaou test interpretation: a validation study of 1748 cases. J Am Soc Cytopathol. 2026;15(5):378–384. doi:10.1016/j.jasc.2026.05.004. PMID 42309847
  67. US Food and Drug Administration. De Novo classification request DEN210035: Genius Digital Diagnostics System with the Genius Cervical AI algorithm (Hologic, Inc.). Decision date January 31, 2024. https://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfpmn/denovo.cfm?id=DEN210035
  68. US Food and Drug Administration. Premarket approval P110012: Vysis ALK Break Apart FISH Probe Kit (Abbott Molecular), package insert 30-608495/R1, August 2011. https://www.accessdata.fda.gov/cdrh_docs/pdf11/p110012c.pdf
  69. Wang CI, Zhang ML, Madrigal E, Faquin WC, Chebib I. Diagnostic utility of anchored multiplex polymerase chain reaction-based fusion assay on sacrificed cytology slides. Cancer Cytopathol. 2026;134(6):e70105. doi:10.1002/cncy.70105. PMID 42119039
  70. Warhol MJ, Hickey WF, Corson JM. Malignant mesothelioma: ultrastructural distinction from adenocarcinoma. Am J Surg Pathol. 1982;6(4):307–314. PMID 6287872
  71. Wilbur DC, Black-Schaffer WS, Luff RD, et al. The Becton Dickinson FocalPoint GS Imaging System: clinical trials demonstrate significantly improved sensitivity for the detection of important cervical lesions. Am J Clin Pathol. 2009;132(5):767–775. doi:10.1309/AJCP8VE7AWBZCVQT. PMID 19846820