Tissue Floaters and Contaminants in the Histology Laboratory Eric Platt, BS; Paul Sommer; Linda McDonald, MT, ASCP; Ana Bennett, MD; Jennifer Hunt, MD

● Context.—Anatomic pathology diagnosis is often based on morphologic features. In recent years, an appropriate increased attention to patient safety has led to an emphasis on improving maintenance of patient identity. Decreasing or eliminating cross-contamination from one specimen to another is an example of a patient identity issue for which process improvement can be initiated. Objective.—To quantify the presence of cross-contamination from histology water baths and the slide stainers. Design.—We assessed for the presence of contaminants in water baths at cutting stations and in linear stainer stain baths. We assessed the potential for tissue discohesion and carryover in tissue samples and we assessed the potential for carryover onto blank slides sent through the stainer.

Results.—In the 13 water baths examined (totalling 195 L of water), only one fragment of tissue was identified. The stain baths, however, contained abundant tissue contaminants, ranging in size from 2 to 3 cells to hundreds of cells. The first sets of xylenes and alcohols were the most heavily contaminated. Cross-contamination to blank slides occurred at a rate of 8%, with the highest frequency in the late afternoon. Conclusions.—Cross-contamination can present a significant challenge in the histology laboratory. Although the histotechnologists’ water baths are not heavily contaminated, the stainer baths do contain contaminating tissue fragments. Cross-contamination does occur onto blank slides in the experimental setting. (Arch Pathol Lab Med. 2009;133:973–978)

T

bad outcomes occur. Molecular approaches to resolving identity in tissue specimens have been reported to provide excellent results in most cases.1,2 One area that is a constant issue for the diagnostic pathologist is the possibility of tissue floaters and contaminants that are transferred to the glass slides during tissue processing. A retrospective review3 estimated that up to 3% of diagnostic slides in AP have tissue contaminants. These contaminants can occur at any step in the processing of tissues, although certain steps have been identified as having high potential for contamination. High-risk steps include the transfer of tissue to the glass slides in the water bath and traditional H&E staining procedures, which rely on dipping and dunking slides into sequential staining baths.3 These tissue floaters are particularly troublesome because they are often only on the glass slide and not in the block. This makes the molecular assessment quite difficult due to the floaters’ small nature and the fact that the DNA may be altered by the staining process. Therefore, approaches to minimize the possibility for contamination during the sectioning and staining process would be highly desirable in the AP laboratory. This study sought to quantify and assess the risk for tissue floaters and contaminants in the 2 areas that are the source for most floaters: the water bath and the traditional linear H&E stainer.

he production of histologic sections for diagnosis is a fundamental process that is critical for all diagnostic work in anatomic pathology (AP). Most clinical AP laboratories function in a similar fashion, from grossing of specimens in a designated space (gross room), to processing of tissues in processors, to embedding of tissues in paraffin wax, to cutting of sections on a microtome, to staining of glass slides for final review by a pathologist. It is critical that the pathologist be confident that the diagnostic material represented on that final hematoxylin-eosin (H&E)–stained slide truly represents the patient’s diagnostic material. And, most laboratories have established checks and balances within their processes to ensure the integrity of the tissue and the identity of the tissue throughout the flow of specimens. As every diagnostic pathologist realizes, however, this process is subject to errors and to possible mishaps at essentially every step, ranging from those that even occur before the laboratory (specimen identity issues) to those that occur within the laboratory. Frank specimen mix-ups have been the subject of reports in the literature but are also the subject of reports in the lay press, especially when Accepted for publication September 12, 2008. From the Department of Anatomic Pathology, The Cleveland Clinic, Cleveland, Ohio. The authors have no relevant financial interest in the products or companies described in this article. Presented as a platform presentation at the annual meeting of the United States and Canadian Academy of Pathology, Denver, Colo, March 2008. Reprints: Jennifer Hunt, MD, Department of Anatomic Pathology, L25, The Cleveland Clinic, 9500 Euclid Ave, Cleveland, OH (e-mail: [email protected]). Arch Pathol Lab Med—Vol 133, June 2009

MATERIALS AND METHODS General Methods For the purposes of this study, standard formalin fixation in 10% neutral buffered formalin and paraffin embedding of tissues was performed on all samples. The tissue processing was performed using standard techniques on traditional processors (LeiContaminants in Histology—Platt et al 973

ca Microsystems, Bannockburn, Ill; Sakura Finetek USA, Inc, Torrance, Calif). Typical microtome (Leica; Surgipath Medical Industries, Inc, Richmond, Ill; Microm International GmbH, Walldorf, Germany) and water bath setups (Leica; Baxter International, Inc, Deerfield, Ill) were used, again according to standard histology procedures. Slides were stained using a traditional linear stainer (also known as a ‘‘dip and dunk’’ stainer) (Leica). The linear stainer is a semiautomated machine. The slides are racked, with a maximum of 20 slides per rack, and dipped sequentially into each bath for the appropriate amount of time. They are then coverslipped on a separate automated coverslipper, and finally they are labeled by hand. Fluids that were collected and analyzed were processed on the ThinPrep machine (Hologic [formerly Cytyc], Marlborough, Mass). For the comparison staining, and particularly for assessment of contaminants, slides were stained in a continuous workflow, discrete slide stainer (Symphony, Ventana Medical Systems, Inc, Tucson, Ariz).

Water Bath Contamination Assessment The first experiment was designed to assess the number and type of contaminants in the water bath that histotechnologists use to float sections as they are cutting. The water in the bath was harvested in its entirety, approximately 1.5 L, and was broken down into aliquots of 50 mL of water. Each aliquot was spun down separately, using a centrifuge. The supernatant was discarded and the pellets were all recombined in ThinPrep solution and a ThinPrep slide was prepared for histologic review. This was performed at the end of the histotechnologists’ cutting day, which usually represented an 8-hour shift. The histotechnologists were asked to save the water from their baths if they did need to change it during the course of their day, and the same procedure was followed in that situation.

Linear Staining Bath Contamination The second experiment was designed to assess contaminants in the staining solutions of the traditional dip and dunk linear staining setup. There are 27 staining baths in the linear staining setup. They consist of the series of xylenes, alcohols, water, and specific stains shown in Table 1. Each bath contains about 250 mL of fluid. In our laboratory, the average number of blocks cut per day is 1092, and the average number of H&E slide stains that are processed is approximately 1637. The stains are generally changed once a day. Initial Assessment of Staining Baths. Initially, a pilot experiment was performed to assess overall potential for contaminants. In this experiment, 20 mL of fluid was extracted from each water bath at the end of 3 separate days, before the staining solutions were changed for the next day’s work. The fluid was spun down and ThinPrep slides were prepared and stained with the Symphony stainer. The slides were assessed for the presence, size, and type of tissue contaminants. Contaminants were defined as fragments that were more than 2 cells in size and contained at least one nucleus. Single keratinocytes were specifically excluded. Full Assessment of Staining Baths. The second experiment to assess for contaminants in the staining baths involved taking the entire contents of each stain bath on 1 day. At the end of the work day (approximately 5 PM), each bath was harvested. This was split into 50-mL aliquots, which were individually spun down and pelleted. The pellets were combined and ThinPrep slides were made. Slides were stained on the Symphony stainer and were assessed microscopically for the exact number, size, and type of tissue contaminants. Contaminants were defined as fragments that were more than 2 cells in size and contained at least one nucleus. Single keratinocytes were specifically excluded.

Cross-Contamination From Slide Pickup The final experiment was designed to determine whether contaminants in the stainer baths could be carried over to other slides during the process of staining. We also wanted to understand whether the contaminants in the water baths could be picked up on slides and be the source for tissue floaters. To assess 974 Arch Pathol Lab Med—Vol 133, June 2009

Table 1.

Bath No.

Staining Solutions and Contaminants in Staining Baths* Solution

No. of Contaminants per Slide (Entire Bath Contents)

1 Xylene 49 2 Xylene 33 3 Xylene 73 4 Xylene 41 5 Xylene 14 6 100% Alcohol 8 7 100% Alcohol 126 8 95% Alcohol 194 9 95% Alcohol 101 10 Water 1 11 Hematoxylin 5 12 Hematoxylin 2 13 Hematoxylin 2 14 Hematoxylin 0 15 Water 0 16 Define solution 4 17 Water 0 18 Bluing solution 20 19 Water 0 20 70% Alcohol 9 21 Eosin 1 22 100% Alcohol 2 23 100% Alcohol 4 24 100% Alcohol 6 25 Xylene 0 26 Xylene 1 27 Xylene 0 * This table demonstrates the solution in each of the staining baths on the linear stainer (see Bath No. and Solution columns). When the entire bath contents were analyzed for contaminants, different numbers of tissue fragments were seen in the lineup (see No. of Contaminants per Slide column). All stain solutions were purchased from Surgipath Medical Industries, Inc, Richmond, Ill.

this slide pickup during the staining process, we did 3 separate experiments. Contaminants Picked Up on Slides Prepared With Tissue Sections. The first experiment was to assess slides that had tissue on them for the presence of tissue floaters picked up during the staining process. Ten source tissues of different types were obtained from residual tissues from the gross room. These tissues included colon cancer, endometrial curettings, lung parenchyma, bone marrow with bone spicules, fibrous tissue, breast cancer, fibrous breast tissue, skin, thyroid, and gastrointestinal stromal tumor. The source tissues were cut into 3 different-sized fragments: approximately 2 mm in diameter, 4 mm in diameter, and 6 mm in diameter. These sets of differently sized fragments were embedded into 3 paraffin blocks, such that one block contained small fragments, one contained medium-sized fragments, and one contained large-sized fragments. Forty slides were cut from each block, using meticulous techniques; the histotechnologist was asked to change the water bath frequently and to minimize the potential for cross-contamination. These slides were stained, with 20 from each group stained in the linear stainer and 20 stained in the Symphony stainer. The slides were then assessed for external tissue contaminants and floaters, for lifting and discohesion of the known tissue fragments, and for movement of these discohesive fragments from one area of the slide to another. Contaminants Picked Up on Blank Slides Run Alone and Alternating With Tissue Sections. We then assessed the possibility of cross-contamination occurring in the staining baths, from adherence to slides as they are passed through the linear stainer. To do this, we sent 40 routine histology blank slides through the stainer at the end of an average load day, with no slides that contained tissues. For comparison, 40 slides were also sent through the Symphony stainer. Then, 200 charged slides Contaminants in Histology—Platt et al

were sent through the stainer in alternating positions with slides that had routine tissue on them. This latter experiment was performed at hourly intervals throughout the course of an average load day. All of the slides were then assessed microscopically for the presence of any tissue contaminants. Contaminants were defined as fragments that were more than 2 cells in size and contained at least one nucleus. Single keratinocytes were specifically excluded.

RESULTS Water Bath Contamination Assessment Thirteen water baths were analyzed, representing 13 different histotechnologists. The average number of blocks cut per histotechnologist on the routine cutting rotations is 15 to 20 per hour, or approximately 130 blocks on an average 8-hour shift. Of all the fluid spun down and analyzed microscopically, only one tissue fragment was identified. This was a large fragment with approximately 100 cells, and it consisted of lymph node tissue. There were many acellular contaminants in the water baths. These included keratin, fragments of paraffin, and minute specks of India ink. Linear Staining Bath Contamination Initial Assessment of Staining Baths. The initial assessment of the staining baths demonstrated tissue floaters and contaminants throughout the staining baths (Table 1). However, the highest density of contaminants was found in the first xylenes and alcohols. In these baths early in the staining lineup, the slides had between 12 and 30 different contaminating fragments on them. The contaminants ranged in size from 4 cells to more than 50 cells. The largest fragments were up to approximately 0.5 mm in diameter on the glass slide (Figure 1). Some of the contaminating fragments were morphologically malignant. Importantly, the contaminating fragments were extremely well preserved and could be easily recognized histologically. Abundant debris was also seen in the background of the slides. This included keratin, unrecognizable foreign debris, and minute specks of India ink. Full Assessment of Staining Baths. The day that was selected for the assessment was a moderately busy day in the histology laboratory, with approximately 1192 blocks processed and approximately 1734 H&E slides stained. The slides demonstrated contaminants that were again present in each staining bath but were concentrated in the early baths. Quantitation of the exact number of tissue contaminants was performed in this experiment (Figure 2; Table 1). The average number of contaminants per staining bath was 25.6 (median, 4; range, 0–194). The maximum number of contaminants was seen in the first set of alcohols (stain baths 6–9 in the stain lineup). These baths contained 8, 126, 194, and 101 contaminating fragments, respectively. In this complete sampling of the baths, a large variety of tissue types was seen, ranging from epithelial fragments, to lymphoid fragments, to stromal fragments. The size of the fragments was also quite variable. Many of the contaminants were around 10 to 30 cells. However, some were also quite large, measuring 0.5 to 1.0 mm on the glass slides. Again, in this experiment, some of the fragments showed morphologically malignant features. Cross-Contamination From Slide Pickup Contaminants Picked Up on Slides Prepared With Tissue Sections. In the experiment assessing whether conArch Pathol Lab Med—Vol 133, June 2009

Figure 1. Contaminants in the staining baths. A, This image shows a representative area on a slide prepared by spinning down the contents of an entire stain bath. Note the density of contaminating tissue fragments, the large sizes of these fragments, and the fact that the morphology is quite well preserved (hematoxylin-eosin, original magnification ⫻10). B, This image shows a higher power view of one of the contaminating fragments from a stain bath. The fragment is morphologically interpretable and sizable (hematoxylin-eosin, original magnification ⫻20).

taminants were picked up on prepared tissue sections, we found several interesting things. First, in tissue sections containing our known source tissues, we frequently saw displacement of fragments of the tissue across the slide. This was especially true of the more friable tissue types, such as the colon cancer. These displaced fragments were located at some distance from the source tissue on the glass slides. This appeared to be due to discohesion or lifting of the tissue fragments during staining. Second, the linear stainer had significantly more of these displaced tissue fragments than the Symphony stainer (45% of slides vs 22%, P ⫽ .007). Finally, 2 of the slides from the linear stainer (3% of the slides stained) had foreign tissue fragment floaters. In other words, these slides had tissue contaminants that did not match the type of tissues contained in the known source tissue block. None of the slides stained on the Symphony had foreign contaminants. Contaminants Picked Up on Blank Slides Run Alone and Alternating With Tissue Sections. We found that the 40 blank slides run through the linear stainer and the Contaminants in Histology—Platt et al 975

Figure 2. Graph of number of contaminating fragments from the stainer baths when the entire bath was evaluated. The x-axis has the stain bath identity and the y-axis has the overall number of contaminants on the ThinPrep slide that was prepared from the contents.

40 run through the Symphony stainer did not pick up contaminants. When the blank slides were alternated with tissue sections, and this was performed hourly, we did find 16 of the 200 blank charged slides harbored tissue floaters (Figure 3; Table 2). The number of contaminant fragments per slide ranged from 1 to 3 contaminants. The tissue contaminants were relatively small, with an average size of 13.9 cells (range, 4–50 cells). These contaminants were identified at various time points during the second half of the day. However, no contaminants were seen between 7 AM and 11 AM. The contaminants could be easily morphologically identified as specific cell types. The cell types include adipose and fibrous tissue, stromal tissue, skeletal muscle, and epithelial tissue (Figure 4). COMMENT One of the most serious issues faced in AP is misidentification of tissues, which includes mislabeled specimens, block identification problems, and tissue contaminants.3 In fact, patient identification errors in surgical pathology are the most rapidly growing category of malpractice claims involving pathologists.4 Most laboratories have elaborate identification processes to avoid specimen mix-ups, including regulations for how tissue samples and requisitions are matched up and checked during grossing, for how cross-checking of identity is done at each step, and for how the blocks and samples are labeled for accuracy. However, it is often difficult to eliminate all risk of tissue contaminants, especially given the hands-on nature of grossing, processing, embedding, sectioning, and staining of tissue sections. Floaters and contaminants can occur at 976 Arch Pathol Lab Med—Vol 133, June 2009

every step, including pickups at the gross bench during prosection, inside the processor, and at every step in the histology process. In one study, the percentage of diagnostic slides with tissue floaters or contaminants was estimated to be almost 3% through retrospective review.3 Up to 30% of these floaters consisted of abnormal or frankly malignant tissues. Many times these contaminants can be resolved at the histologic level, particularly when the tissue contaminant is derived from a completely different organ system. And, it has been shown that molecular analysis, using a DNA fingerprinting assay similar to those used in forensic analyses, can resolve most potential tissue floaters.1,2,5 However, in approximately 1% of cases, the tissue contaminant cannot be resolved, either histologically or through the use of molecular technology.2,3 Because of the diagnostic issues that contaminants can cause, the AP laboratory has a responsibility to reduce potential for error in every way possible. Process improvement and quality assurance initiatives should focus particularly on reducing the potential for false positives that can arise secondary to tissue floaters or contaminants.6 This study was aimed at further classifying the types of floaters and contaminants that occur during histology processing in the AP laboratory. In particular, we closely examined the potential for contaminants during the cutting of slides from the water bath and during the staining of slides on a traditional linear dip and dunk type of H&E stainer, which are considered to be the areas of most probable tissue floater contamination. Contaminants in Histology—Platt et al

Figure 3. Graph of the number of contaminants that were picked up on blank charged slides as they were sent through the linear stainer. The x-axis has the time of day and the y-axis has the percent of slides with contaminants.

Our results demonstrate that contaminants are present in the water bath for the microtome, although in very low concentrations. We postulated that because the tissue sections floated in the water bath are completely maintained within a thin sheet of paraffin, fragmentation and breaking off of tissue fragments that can cause floaters is less prevalent at this point in processing. That being said, there is still a very real potential for tissue floaters and contaminants from water bath contaminants. Further study of different time points in the day, different volumes of slides being cut, and during a series of different days may be

Table 2. Number and Type of Contaminants Picked Up on Blank Charged Slides Sent Through the Linear Staining Setup at Various Time Points

Time

7 8 9 10 11 12 1 2 3 4

AM AM AM AM AM PM PM PM PM PM

No. of Contaminant Fragments

Percentage of Slides With Contaminants

No. of Slides With Contaminants

0 0 0 0 1 3 1 3 8 4

0 0 0 0 5 15 5 15 25 10

0 0 0 0 1 3 1 3 5 2

Arch Pathol Lab Med—Vol 133, June 2009

useful in determining the overall risk of floaters from the water bath. In contrast, the potential for tissue contamination during the staining procedure may be much higher, because tissue is deparaffinized during the first steps in making an H&E stain. As the slides are dipped up and down into the staining baths, the deparaffinized tissue can fragment and small discohesive pieces can break free and lift from the slide. We demonstrated this by using tissue blocks with known tissue types within the blocks in serial sections. We noted frequent lifting and discohesion on these slides, especially when the native tissues were either friable (colon cancer), fragmented (endometrial curettings), or naturally more difficult to adhere to a slide (boney fragments). Knowing that discohesion occurs during the staining process, we were not surprised to also find high levels of contamination in the staining baths for the linear stainer. The contaminants were most abundant in the early baths of the staining lineup, particularly in the first set of xylenes and first set of alcohols. The latter baths (100% alcohol ⫻ 2 and 95% alcohol ⫻ 2) had a very high average number of contaminants, compared with all the other baths (107.3 fragments vs 11.4 fragments, averaged). However, tissue contaminants were found sporadically all the way through the staining lineup. The contaminating fragments ranged in size from a few cells all the way up to fragments of 0.5 to 1 mm. These larger fragments may Contaminants in Histology—Platt et al 977

Figure 4. Contaminants seen on glass slides that were sent through the linear stainer. These images show representative contaminants that were picked up onto a charged glass slide that was sent through the linear stainer with other tissue-containing slides (hematoxylin-eosin, original magnifications ⫻40 [A] and ⫻60 [B]).

ing through the staining lineup. Through a set of experiments using blank slides run either with or without other tissue slides, we found that cell fragments could indeed make their way onto blank slides and adhere. Up to 25% of the blank slides that were run through the lineup might contain contaminants, with this maximum number being reached at 3 PM on the study day. The highest number of contaminants on a single blank slide passed through the baths was 3. The fragments again ranged in size from 4 cells to 1 large fragment that contained approximately 100 cells. These results indicate that there is a risk for carryover from the stain baths to slides that are being passed through the linear stainer. There are several things that a laboratory can do to limit the risk of tissue contaminants and floaters from the water bath and a traditional linear stainer. First, meticulous cleaning of the microtome water bath and frequent clearing or changing of the water will almost certainly alleviate the rare contaminants that may pose a risk for carryover to subsequent sections being mounted. Second, because the stainer baths are a potential reservoir of tissue contaminants, changing the staining fluids may alleviate some of the potential for carryover from this source. And, as higher numbers of the contaminating fragments are localized to the first xylenes and alcohols, frequently changing these baths in particular may be useful. Finally, these experiments were performed on one specific linear stainer instrument. It is not known if they are transferrable to other machines that use similar technology. It is, however, unlikely that there is the same risk of cross-contamination on a newer instrument that uses a discrete slide staining process (Symphony, Ventana). In this continuous workflow instrument, the stain aliquots are used only once per slide and the potential for slide-to-slide or floater cross-contamination is theoretically nonexistent. Additional studies are needed to fully understand the spectrum of carryover and cross-contamination that occurs in routine histology processing and the impact that these can have on workflow. We gratefully acknowledge Ventana Medical Systems, Inc, for the placement of the Symphony instrument in our laboratory for the purposes of performing this study. References

contain hundreds of cells. Most of the contaminant fragments were very well preserved morphologically and some contained frankly malignant cells. It is highly likely that the contamination of the baths is dependent on the volume of slides being stained and the time point during the day that the samples are taken. Additionally, the results may be sporadic and different baths could be at risk depending on tissue type. Further experiments are needed to fully explore the reason for the variable presence of specific contaminants in the different baths. Lastly, we wanted to determine whether contaminants in the staining baths could adhere to slides that were pass-

978 Arch Pathol Lab Med—Vol 133, June 2009

1. Worsham MJ, Wolman SR, Zarbo RJ. Molecular approaches to identification of tissue contamination in pathology sections. J Mol Diagn. 2001;3:11–15. 2. Hunt JL, Swalsky P, Sasatomi E, Niehouse L, Bakker A, Finkelstein SD. A microdissection and molecular genotyping assay to confirm the identity of tissue floaters in paraffin-embedded tissue blocks. Arch Pathol Lab Med. 2003;127:213– 217. 3. Gephardt GN, Zarbo RJ. Extraneous tissue in surgical pathology: a College of American Pathologists Q-probes study of 275 laboratories. Arch Pathol Lab Med. 1996;120:1009–1014. 4. Troxel DB. Error in surgical pathology. Am J Surg Pathol. 2004;28:1092– 1095. 5. Venditti M, Hay RW, Kulaga A, Demetrick DJ. Diagnosis of ectopic tissue versus contamination by genetic routine surgical pathology specimen. Hum Pathol. 2007;38:378–382. 6. Renshaw AA, Gould EW. Measuring errors in surgical pathology in real-life practice. Am J Clin Pathol. 2007;127:144–152.

Contaminants in Histology—Platt et al