Contamination of histologic sections with “floaters” is occasionally encountered both in cytology and surgical pathology. The presence of extraneous cells and/or tissue fragments can be quite vexing and creates an unexpected diagnostic dilemma. Failure to recognize extraneous tissue can lead to unnecessary treatments and/or repeat procedures. The ideal approach would be recognize the extraneous tissue fragment and discount it; however, the burden of proof belongs to the pathologist and it is often difficult to prove given the multiple variables involved with specimen preparation. Another problematic issue is the series of questions that arise after identification of these extraneous tissue fragments, including the following: where did this tissue come from and how should it be proven, is the procedure for tissue processing creating contamination, and how should this be documented? Here we aim to review and provide a road map f or addressing some of these quality issues that cytopathology laboratories can face.

In a 1994 Q-Probes study conducted by the College of American Pathologists (CAP), which included 275 laboratories and 321,757 slides in a prospective study arm, the authors identified “floaters” at a rate of 0.6% with more than half (59.5%) identified near the tissue being evaluated. A retrospective study arm was also included by the authors (including 262 laboratories and 57,083 slides) and a rate of 2.9% was identified, with more than half (53.2%) of the tissue being identified distant from the tissue being evaluated. The higher incidence in the retrospective study was not surprising since it involved a re-review of slides specifically targeting the identification of “floaters”. These tissue fragments were non-neoplastic in most cases (87.3% prospective, 94% retrospective); nevertheless, when pathologists were surveyed, the “floaters” did cause a severe degree of difficulty in 0.4% in the prospective study and 0.1% in the retrospective study. Only 6.1% of laboratories had written protocols for documentation of “floaters”. Methods included labeling the slide as “floater” (55%), comment in report (37%) and documented in a log book (6%). 1 In a more recent 2011 study performed by Associated Regional and University Pathologists (ARUP) laboratories, a tissue contaminant rate of 0.01% was identified in a review of 521,661 slides (125,000 cases) reviewed for QA purposes over a 64 month period. In a prospective review of 1,000 slides, a tissue contamination rate of 1.2% was identified. 2 The 17 year gap between these two studies gives an indication that “floaters” are a persistent problem despite advancing technologies. Additionally, they tend to be recognized and considered more problematic when it is close to the tissue being evaluated. Methods for assessing “floaters” vary but include histologic characteristics such as the proximity of the tissue fragment (e.g., distant fragments are more likely to be extraneous tissue), the plane of focus (e.g., “floaters” may be out of the plane of focus) and presence in the paraffin block (e.g., “floaters” may be in a single section and not present in the paraffin block, or present in the block and therefore on multiple sections). 1,2 The presence in the block, however, may be more of an indication at which point in the process the tissue appeared (e.g., “floaters” in the block may be an indication that tissue contamination arose prior to or during paraffin embedding). This will be discussed in further detail below.

The potential for cross contamination has been recognized both by CLIA ’88 and CAP which require written policies for the prevention of cross contamination. 3 When investigating “floaters”, one of the earliest steps involves communication between the pathologist, the laboratory and the clinician. Reviewing the clinical history, determination of what site was biopsied and how the procedure was performed can be helpful. 4 Electronic medical records have made this endeavor considerably easier. Next, communication with individuals responsible for processing the specimen is vital in determining which other specimens may have been processed by the individual, whether any abnormalities were noted during processing (e.g., friable tissue fragments, instrument and surface cleanliness, distractions during processing etc.) and whether any of the procedures used in processing deviated from the standard process established by the laboratory. Review of the gross descr iption may also be useful in attempting to match tissue fragments. 4 In cytology, this is less relevant due to the small size of FNA samples. In the Q-probes study, possible sources of “floaters” identified in specimen processing included: gloves, dissecting surfaces, instruments, containers, saws, ink and fixation solutions. 1 Additionally important for cytology laboratories is the use of on-site or laboratory staining solutions which bypass the histology laboratory. In such cases, cells identified on the smears which are not identified in the cell block sections may help isolate the issue to the staining solutions used by cytology laboratories.

Other potential points of tissue contamination involve the histology laboratory and include embedding tissue in paraffin, cutting tissue sections with the microtome, transfer of tissue to the glass slide in a water bath, and staining H&E tissue sections. 5 In a review of potential histology laboratory contaminations performed by Platt et al, they identified tissue contamination in 1 of 13 water baths analyzed with a higher incidence of acellular contamination (e.g., keratin, paraffin fragments, and ink). In assessment of linear staining baths at the end of a “moderately busy day”, the highest density of contamination was found in the initial xylene and alcohol baths, and ranged between 8 to 101 contaminating fragments with sizes ranging between 10 cells and 1.0 mm tissue fragments. In examination of water baths for tissue pick-up, they identified a “floater” rate of 8% when blank slides were alternated with tissue sections. This phenomenon seem ed to be emphasized in the later parts of the day. 5 In a separate study by Cahill et al which included 69 histopathology laboratories in six different countries, charged blank slides were included with the last run of the linear H&E stainers and 57% of the laboratories tested exhibited “floaters” which were assumed to have migrated to the blank slides from the stainer baths. Additionally, the contents of the first reagent bath as well as the 95% and 100% alcohol reagent baths were evaluated for extraneous tissue. None of the stainer baths analyzed were free of contaminates and the number ranged from 3 to 3,018 contaminates. 6 These studies highlight that stainer baths and fixation solutions can be a significant source of tissue contamination. Therefore, cytology laboratories should establish protocols for which specimens should be stained first (i.e. cases that are cellular or positive for malignant cells are stained last to minimize carry over) and st aining solution filtration practices (i.e. once daily or between processing of cellular or hypocellular specimens).

The histologic characteristics of the potential “floater” may also be useful. For example, in the retrospective review of the ARUP study of extraneous tissue mentioned previously, 25% of the floaters were found to lie above the plane of the patient tissue and 43% of floaters were present in the tissue block and identified on multiple levels. Fifty-seven percent were only present on one tissue section level. 2 Once a tissue fragment is determined to be likely extraneous, the next step may involve definitive characterization of the origin. Examination of specimens which were processed in the cytology or histology laboratories at or around the same time may identify a morphologic match. Additionally, immunohistochemistry looking for differences in staining characteristics may be useful, but it is important to remember that tissue is frequently only present on 1 or 2 tissue sections and may not be present on deeper levels. 2,4 The paucity of tissue may make fur ther molecular characterization of tissue origin impossible. 4 Immunohistochemistry for the ABO blood group antigens has been described, but is considered unreliable due to a high incidence of non-specific staining and sub-optimal antibodies. Additionally, this testing relies on differences in blood types between patients, which may not always be the case. 7

Molecular techniques, applied in forensic practices for identity testing, can also be used to confirm if a piece of tissue is extraneous or “foreign”. Florescent in situ hybridization (FISH) performed on 5 um sections utilizing probes to the centromere DNA portions of the X and Y chromosomes have been used to determine if a fragment of tissue has a different sex chromosome complement; however, this methodology only works if the tissues in question comes from a patient of a different gender. 8 Another methodology is a polymerase chain reaction (PCR) to analyze short tandem repeats (STRs). STRs represent areas of DNA where a sequence of nucleotides is repeated in tandem and is considered a genetic polymorphism due to a lack of phenotypic impact. While the number of repeats at a single locus may not be specific for a particular individual, examining STRs at multiple loci allows the construction of a genetic fingerprint of an individual. STRs exist throughout th e human genome and can be targeted with commercially available PCR kits. These kits provide primers for a subset of commonly targeted loci and the amplification test results indicate the number of repeated nucleotide sequences at a specific target. This is generally performed on DNA extracted from both tissue known to originate from the patient (another surgical specimen or patient blood) and tissue which is suspected to represent contamination. These tests require very little DNA, but there is potential for biologic or instrumentation artifact. For example, if a mutation exists in the DNA sequence adjacent to the STR of interest in the primer annealing region, the PCR primer may fail to anneal and amplify giving a false result. Other potential artifacts include stutter products, split peaks due to incomplete adenylation, triallelic patterns, instrument voltage spikes, dye blobs and bleed through of dye colors. 9,4

Once a “floater” is identified, the next step is to identify and address the source of the contamination. This is often achieved through quality improvement (QI) initiatives. Lean practices, which evolved from the Toyota Production System, have become an important driver in healthcare QI initiatives; however, the idea of Lean is often viewed speculatively as too time intensive, costly and soft science at best. Lean should not be viewed as a specific entity, but instead as a set of tools and ideas. These can be used to identify problems and help improve the quality of the service delivered to patients. Ideally it would also minimize cost and fully utilize the skills of the individuals working in the system. Lean respects individuals and their humanity with understanding that individuals will make mistakes. 10,11,12 One instrument in the Lean toolbox which is frequently utilized in the investigation of “floaters” is Root Cause Analysis (RCA). RC A identifies the unwanted results and follows a line of direct physical causes back through the process chain with clear identification of individuals who own the steps in the process. Solutions can then be generated when the root cause is clearly understood. 12 Different systems for RCA have been proposed, but probably the easiest methodology is the “5 Why’s”. In this approach, when a problem is identified the process owner should ask a series of “why?” questions to attempt to identify the root cause itself. 11 (For example: Why was the “floater” on the slide? The stainer bath was contaminated in the laboratory. Why was the stainer bath contaminated? The solutions were not filtered. Why were the solutions not filtered?…etc.) Five is not meant to be a strict rule and it may take more or less to get to the root cause of the issue. When more than one potential root cause is identified, additional tools such as a fishb one diagram can be useful in investigating causes. A fishbone diagram uses broad categories such as “management” or “equipment” and re-examines the problem within the context of these categories. 11

Another methodology of RCA, which has been described and examined in the context of “floaters” by Dimenstein, is Eindhoven’s original taxonomy of Root Cause Analysis which breaks down failures into three broad categories: Organizational, Technical and Human. In a review of RCA using this methodology, Dimenstein identified some of the broad components which can be used to address these categories. Organizational failures can be addressed by including process guidelines for “floater” prevention as well as protocols for documenting, investigating and following up on these incidences. Technical includes many of the items which have been previously mentioned (e.g., grossing trays, filtration systems, gloves, instruments, water baths, paraffin embedding etc). Human failures can be mitigated by making sure there is sufficient education and individuals are following proper handling techniques. 13 At our institution, a tool which has been implement ed for RCA is the “Situation Background Assessment Recommendation” (SBAR) methodology. The SBAR identifies a problem and then users (i.e. the involved individuals including technicians, administrative assistants, pathologists etc.) create a timeline of events and explanations which led to the event. SBARs are reviewed by an administrative team and follow-up recommendations are made to address the issue.

Whatever the approach to RCA, a few words of caution should be remembered. First, RCA is simply a means to identify a root cause and solutions need to be generated by the process owners and tailored to the needs of the system. Second, implementation of a QI initiative is not the end of Lean thinking but rather the beginning. The process and implemented solution should be continually re-evaluated to make sure it is addressing the issue. Third, after the initial implementation of a QI initiative, mistakes are likely to occur as individuals adjust to a new process. Post-implementation “slip-ups” should not be taken as a failure but a learning opportunity. Fourth, Lean and RCA are not punitive and should not single out a single individual as this is often counter-productive and may lead to concealing flaws. “Why?” questions can seem accusatory and improvement should emphasize group participation. One principle that is re-iterated in Lean practice i s Dr. W Edwards Deming’s 94/6 rule which states that 94% of the problems probably lies in the system and 6% with the person. 12 Deming’s principle is advice that those who endeavor into Lean thinking would be wise to remember.

The risk of “floaters” can never be totally eliminated and the process of investigating and documenting floaters is largely dependent on the protocols established by the laboratory. There are many potentially contaminating sources in pathology and in investigating the causes of these it is important to remember that it is not a process of assigning blame. The goal should be to improve the process and the services that the laboratory offers. It is a continuous process of improvement and all must be involved. Cytopathology is unique in that there are a variety of different cell preparations and different stains utilized, which may make it difficult to determine the source of a “floater”, but having a “quality” laboratory involves being aware of such issues and methods for discovering and addressing “floaters” when they occur.