Why Cytology Labs Run Random Re-Screens — and Why Most Still Track Them on Paper

Every cytology case that gets signed out as negative carries an implicit claim: a trained professional reviewed this material, found nothing actionable, and the patient can be managed accordingly.

That claim rests entirely on human attention — and human attention is variable. It varies by screener experience, by time of day, by workload, by the monotony of reviewing slide after slide of normal material. The literature on screening fatigue in cytology is substantial. The consequence when attention lapses is not a billing error or a turnaround delay. It is a missed diagnosis.

Random re-screening exists precisely because this variability is predictable, measurable, and — in a functioning laboratory quality system — manageable. The question most labs are still working through is not whether to re-screen. Regulation settled that years ago. The question is how: how to select cases without bias, document results without manual error, track concordance without end-of-year reconstruction, and act on discordance before it becomes a pattern.

Why Re-Screening Is a Regulatory Requirement, Not a Best Practice

The foundation is regulatory. In the United States, the Clinical Laboratory Improvement Amendments (CLIA) — codified at 42 CFR §493.1274 — require that laboratories performing gynecological cytology rescreen at least 10 percent of cases interpreted as negative by each cytotechnologist per year. That 10 percent must include a random sample of negative cases, as well as targeted rescreening from high-risk patient populations defined by the laboratory director. The rescreening must be performed by a qualified supervisor or cytotechnologist — not the original screener — and the results must be documented and retained (1).

For laboratories seeking accreditation from the College of American Pathologists (CAP), the Cytopathology Checklist extends and operationalises this requirement. Accredited labs must maintain a formal rescreening program that covers random selection, documentation of results, concordance tracking, and procedures for handling discordant cases. Laboratories must also demonstrate that statistical records are maintained — including the total number of negatives rescreened and the number of cases reclassified following rescreen (2).

Beyond gynecological cytology, ISO 15189:2022 — the international standard governing quality and competence in medical laboratories — provides the broader framework. Clause 7.4 requires that laboratories implement internal quality control procedures for all examination processes, with documented evidence that results are monitored for accuracy and reproducibility over time. For cytology, this means that non-gynecological specimens — sputum, urine, fine needle aspirations, effusions, cerebrospinal fluid — are not exempt from the requirement to have a systematic, documented quality monitoring program. ISO 15189 does not prescribe a specific percentage, but it requires that the laboratory define and justify its quality control parameters based on the risk profile of the examination (3).

The logic across all of these frameworks is the same: a single screener's interpretation of a negative case cannot be self-validating. An independent second look on a statistically meaningful sample is the mechanism by which the system checks itself.

How Most Cytology Labs Currently Do It

The manual workflow for cytology rescreening is almost universally the same, regardless of laboratory size or accreditation status.

At the end of each screening session — or at batch close — the cytotechnologist or supervisor produces a list of all cases interpreted as negative. In most facilities, this list is extracted from the LIS or maintained in a spreadsheet. A random number generator — typically Excel's RAND() function, occasionally a printed random number table — is applied to the list, and the top 10 percent are identified. Those case numbers are written into a log or a second spreadsheet column. The physical slides for those cases are retrieved from storage. A qualified re-screener reviews them. The result — concordant or discordant — is entered manually. At year-end, concordance rates are calculated and reported as part of the annual quality review.

The problems with this workflow are not primarily about the difficulty of the task. They are about what the manual approach systematically fails to guarantee.

Selection is rarely truly random. When random selection is performed manually or with ad hoc tools, there is consistent opportunity for inadvertent bias — selecting cases from easier-to-retrieve storage locations, drifting toward certain screeners, or applying inconsistent high-risk targeting criteria. Random number generation that is not seeded and logged produces a selection that cannot be audited.

Documentation is a reconstruction rather than a record. When results are entered after the fact into a spreadsheet maintained separately from the LIS, the audit trail is fragmented. If a regulatory inspection or accreditation survey asks for documentation of the selection method for a specific case, many labs cannot produce it. The log exists; the methodology that produced it is gone.

Discordance detection is retrospective. In a manual system, a pattern of discordant re-screens by a specific screener may not be visible until the annual summary is compiled. By that time, the pattern has persisted for months.

What New Hardware Is Doing to the Problem

The most significant hardware-level development in cytology quality control is the emergence of AI-assisted digital imaging systems that can perform autonomous or semi-autonomous re-screening at scale.

The BD FocalPoint GS Imaging System — one of the earlier FDA-cleared automated cytology platforms — uses a robotic microscope and automated image analysis to rank slides by priority, flagging the fields most likely to contain abnormal cells. This approach shifts the re-screening burden from purely manual review to targeted review of algorithmically identified high-risk regions.

Hologic's Genius Digital Diagnostics System represents the current generation of this technology. It is the first FDA-cleared digital cytology system to combine volumetric imaging with a deep-learning AI that reviews every cell on a ThinPrep preparation, identifies the most diagnostically relevant images, and presents a ranked gallery to the cytologist or pathologist. Studies published in PMC have validated the system's performance against conventional manual screening, with demonstrated non-inferiority in sensitivity for high-grade lesions (4).

Further still, recent work published in Cancer Cytopathology (2025) demonstrated the feasibility of 100 percent quality control review using whole-slide imaging and the Techcyte SureView AI system — replacing the statistical sampling requirement with comprehensive AI-based re-screening of every case (5). If that approach achieves broader regulatory acceptance, the 10 percent rule becomes a floor rather than a ceiling.

The aspiration is clear: a future in which every slide that enters the laboratory also undergoes automated AI review, and the concordance between human screener and AI becomes the quality control mechanism — continuous, objective, and fully documented.

Most cytology labs are not there yet. The capital cost of whole-slide digital imaging infrastructure remains significant, and regulatory frameworks have not fully adapted to AI-based QC equivalence. The majority of labs still operate in the world of partial automation at best and paper-based tracking at worst.

How SlidePath Handles Random Re-Screening

SlidePath's approach to cytology re-screening is built directly into the LIMS workflow — not bolted on as a separate module or managed through an adjacent spreadsheet.

When a cytologist closes a screening batch, the process begins at the box level. The cytologist selects the box from within SlidePath and initiates the randomizer. The system draws a statistically valid random selection from all cases in the box that were interpreted as negative, applying the laboratory's configured parameters: minimum 10 percent of total negatives, stratified by screener to ensure per-screener compliance, with optional weighting for high-risk case flags defined by the laboratory director.

The selected cases are automatically queued for second review within the same system. A qualified re-screener — not the original cytotechnologist — is assigned the queue. The second review is performed and the result recorded directly in SlidePath, linked to the original case and the original interpretation.

SlidePath then compares the re-screen result against the original and evaluates whether the pair falls within the laboratory's defined concordance bracket. Concordance thresholds are configurable: the lab sets the acceptable discordance rate, and SlidePath tracks actual performance against that benchmark in real time. Cases that fall outside the bracket — where the re-screen result represents a clinically significant reclassification of the original negative — are automatically flagged for supervisor review. The supervisor receives the flag, the original result, the re-screen result, and the full case history. No manual review of logs is required to surface the discordance.

The audit trail is complete and automatic. Every case selected, every re-screen performed, every concordance determination, and every flag generated is timestamped and attributed to a named user. At any point, the laboratory can generate a rescreening compliance report for a given period, screener, or case type — without reconstruction and without manual calculation.

For CAP, CLIA, or ISO 15189 surveys, the documentation is already there.

The Compliance Case for Doing This in Software

Random re-screening is not optional, and the consequences of doing it poorly compound over time. A laboratory that cannot demonstrate that its 10 percent selection was genuinely random, that its re-screen results were systematically compared against originals, and that discordant cases were acted on is not merely at risk of an accreditation finding. It is operating a quality system that is failing to catch its own errors.

The technology to close this gap does not require a whole-slide imaging platform or an AI system. It requires a LIMS that treats rescreening as a core workflow — one that selects randomly, documents automatically, measures concordance in real time, and surfaces problems before they become patterns.

See how SlidePath's randomizer works in your laboratory environment. Book a workflow session with the SlidePath team and we'll show you the re-screening module running against your actual case volume — from box selection to concordance report in a single session.

Book a free workflow assessment →

Sources

1. U.S. Department of Health and Human Services. Clinical Laboratory Improvement Amendments (CLIA). 42 CFR §493.1274 — Standard: Cytology. Code of Federal Regulations. Available at: https://www.law.cornell.edu/cfr/text/42/493.1274

1. College of American Pathologists. Cytopathology Accreditation Checklist. CAP Laboratory Accreditation Program; 2024. [Checklist criteria for random rescreening program, concordance documentation, and discordant case review.]

1. International Organization for Standardization. ISO 15189:2022 — Medical Laboratories: Requirements for Quality and Competence. Geneva: ISO; 2022. [Clause 7.4: Examination processes — internal quality control.]

1. Pantanelli S, et al. Validation of AI-assisted ThinPrep Pap test screening using the Genius Digital Diagnostics System. PMC / published in peer-reviewed cytopathology journal, 2024. Available at: https://pmc.ncbi.nlm.nih.gov/articles/PMC11304920/

1. Rivera Rolon MV, et al. Implementing 100% quality control in a cervical cytology workflow using whole slide images and artificial intelligence provided by the Techcyte SureView System. Cancer Cytopathology. 2025. doi:10.1002/cncy.70019