An update on touch preparations of small biopsies

Abstract

Touch preparations (TPs) are being increasingly utilized in the era of personalized medicine. They fill a gap in cytopathology practice by providing a method to perform rapid onsite evaluation of small tissue samples such as core needle biopsies. However, there is a paucity of literature about how best to perform and interpret a TP. A high-quality TP can provide excellent diagnostic accuracy and good concordance with core needle biopsy histopathology findings. Although many of the cytomorphologic features of TPs overlap with fine needle aspirate smears, TP cytology is unique and differs from conventional smears in many aspects. It is important for cytologists to recognize these features, as well as potential pitfalls and artifacts in order to avoid misinterpretation. Core depletion of tumor cells is a notable drawback if TPs are performed too aggressively. TP slides are also valuable for ancillary testing because they often contain a cellular and pure population of whole tumor cells. This paper reviews all of the aspects of TPs including their clinical utility, proper slide preparation techniques, distinctive cytomorphologic characteristics, limitations, and potential pitfalls.

Introduction

Definitive subtyping of malignancies including their molecular characterization is expected in pathology. In the current era of personalized medicine, there has been an increased demand for minimally invasive procedures, including small core needle biopsies (CNBs). These CNBs are not only utilized to render accurate diagnoses, but also to triage specimens at the time of procurement for ancillary testing. As a result, several cytology practices have been called upon during rapid onsite evaluation (ROSE) to accordingly perform touch preparations (TPs) on these small biopsies to confirm their adequacy, provide a preliminary interpretation, and appropriately triage samples.^1,2 Compared with frozen sections, TPs are rapid and easy to perform, do not require dedicated equipment, avoid freezing tissue artifact, and, under optimal conditions do not potentially deplete the entire sample thereby precluding subsequent ancillary testing.³ In a study by Li et al., onsite TPs of CNBs is useful in reducing the number of cores needed to obtain diagnostic material.⁴

The TP technique was first introduced into the medical literature in 1927 when researchers from London (Dudgeon and Patrick) developed a simple method of rapid imprint using freshly cut specimens.⁵ They studied 200 consecutive specimens and found reliable diagnostic accuracy of their imprints when compared with formalin-fixed paraffin-embedded tissue sections. Since then, TPs have been widely utilized for many organ systems.^6,7

Overall, the sensitivity and specificity of TPs vary for different organs.^8–11 The factors responsible for obtaining good TP results include the type and gauge of needle used, operator skill, the organ and lesion being sampled, and, most importantly, the technique used to make a TP slide. TPs are particularly helpful for immediately evaluating CNBs and other small specimens (eg, endoscopic or bronchoscopic biopsies) that may be too small to smear without sacrificing the entire sample. Table 1 summarizes the advantages and disadvantages of TPs. Although the cytomorphology seen on TPs generally resembles that observed on conventional direct smears, there are some differences that cytologists need to recognize. Some important differences seen with TPs include the presence of more tumor cell—enriched slides with less blood contamination (Fig. 1), intact architecture,¹² less background material (eg, lymphoglandular bodies), scant to absent endothelial wrapping in hepatocellular carcinoma, and different artifacts (eg, streaking versus smearing, false clustering of cells) compared with aspirate smears. Table 2 summarizes key cytomorphological differences between TPs, aspirate smears, and core biopsy histopathology.

Table 1.

Advantages and disadvantages of touch preparations.

Advantages

Disadvantages

• Minimally invasive, rapid, and easy to perform • Requires inexpensive equipment • Less dependent on operator skill than FNA • Useful for ROSE of small core needle biopsies • Better preserved architecture compared to FNA • Facilitates immediate diagnosis and triage of specimens • Fewer artifacts (smearing, freezing, formalin fixation) • Complements concomitant nondiagnostic aspirates • Complements frozen section during intraoperative consultations • Offers cytomorphologic features not visualized on tissue sections • Additional source of material to perform ancillary studies • Unique applications (eg, for difficult-to-perform frozen sections)

• Greater complication risk with CNB compared with FNA, particularly with vascular lesions • Unfamiliarity with properly preparing and evaluating TPs • Limited utility for fibrous, cystic, and necrotic lesions • Challenging cytomorphologic differences compared with aspirate smears (eg, unexpected clustering of tumor cells) • Damage (eg, fragmentation) of core with forceful TP • Loss of tumor cells in remaining core biopsy, limiting ancillary testing • Limited evaluation of architecture compared with tissue sections • Absence of morphologic clues (eg, missing background material) • Artifacts due to improper handling (eg, streaking)

Figure 1.

A, Core needle biopsy of a retroperitoneal lymph node showing predominantly dense fibrous tissue with only a small focus of metastatic urothelial carcinoma (hematoxylin and eosin stain, 400×). B, A touch preparation of this core biopsy shows a hypercellular slide with abundant discohesive, malignant urothelial cells characterized by spindle to cercariform shapes and containing bland ovoid nuclei with an increased nuclear:cytoplasmic ratio (Diff-Quik stain, 100×).

Table 2.

Cytomorphological differences between touch preparations, aspirate smears, and core biopsy.

Cytomorphology	Touch preparation	Aspirate smear	Core biopsy
Overall architecture	Disrupted	Disrupted	Intact
Cellularity	Variable	Variable	Variable
Cohesive cell groups	Present without surrounding stroma	Present without surrounding stroma	Present with surrounding stroma
Discohesive cells	May show artifactual clustering	Present	Infiltrative cells
Heterogeneity	Limited	High, parallels number of passes	Limited
Stroma	None to scant	None to scant	Usually present
Cystic fluid	Unsuitable	Easy to smear	Useful to identify lining epithelium
Background material	Variable, may be patchy or absent	Variable	Differenta
Blood dilution	Limited	Often present	Usually absent
Artifacts	Streaking, thick areas, and stripped cells	Air-drying, crushed cells and obscuring material	Freezing or formalin artifact (eg, cellular contraction)

^aCertain background material present in cytology slides (eg, tigroid appearance) may not be found in core biopsy sections.

Given the increasing role of cytopathologists in the assessment of TPs of small biopsies, and the fact that some cytologists may not be that familiar with these cytology preparations, this review addresses the utility and accuracy of TPs, the various techniques of preparing a good TP, and highlights potential problems and pitfalls that may be encountered, as well as which ancillary tests are most suitable in this setting.

Utility of touch preparations

In general, biopsies of most specimen types can undergo TP evaluation. TPs are suitable for small biopsies such as CNB. When performed on cellular, nonsclerotic lesions, TPs can yield diagnostic slides. TPs are not optimal for cystic lesions, however. To the best of our knowledge, the application of TPs instead of smears specifically for fine-needle biopsy samples [eg, SharkCore (Medtronic Corporation, Minneapolis, MN), EchoTip ProCore (Cook Medical Incorporation, Bloomington, IN), Acquire (Boston Scientific, Marlborough, MA)] used during endobronchial ultrasound or endoscopic ultrasound procedures has not been well studied.¹³

Laboratories may utilize a combination of fine-needle aspiration (FNA) and TP for procuring adequate specimens. In this setting, FNA is typically performed first to confirm that the proceduralist is in a good location of the targeted lesion (eg, viable lesional cells are present), to minimize tissue trauma and bleeding, and to utilize the aspirate for appropriate ancillary studies (eg, flow cytometry) that is sometimes more suitable than a CNB. Some laboratories, especially academic medical centers or subspecialized private practices, separate their FNA and TPs from the CNB material, allowing the CNB on which a TP was performed onsite by the cytology division to go to the subspecialist surgical pathologist for evaluation and final sign out. Others, however, believe it is important to keep these specimens together for logistic purposes, to limit duplication of ancillary testing, avoid double billing, and minimize potential discrepant diagnoses. Like FNA, TPs can be used for ROSE of small biopsies to confirm tissue adequacy, render an immediate diagnosis, and triage the specimen for ancillary studies including procurement of specimens for clinical trials.^14–17 A positive diagnosis using a TP avoids the need for extra passes. Usually only a few lesional cells are needed on a TP to render a helpful interpretation on site, thereby Although a TP can be performed on any tissue type, they are particularly helpful when evaluating hematolymphoid neoplasms, certain soft tissue tumors like small round blue cell tumors, and bone specimens.

Hematolymphoid neoplasms

When evaluating a specimen for a possible lymphoproliferative disorder, it is extremely helpful to be able to assess cellular cytomorphology (eg, monotonous versus polymorphous lymphoid population, small versus large lymphoid cells). Apart from lymphoma, TPs are also helpful to rapidly evaluate other hematolymphoid neoplasms (eg, leukemia, plasma cell neoplasms) to confirm the diagnosis and appropriately triage material for flow cytometry.^18,19 Unstained TP slides can be easily collected for fluorescent in situ hybridization (FISH) studies, which are often more desirable than tissue sections because they avoid truncation artifact of chromosomes. We recommend preparing at least 3 unstained TP slides for possible FISH studies. Lymphoglandular bodies are commonly seen in FNA smears of benign and malignant lymphoid specimens. These cytoplasmic fragments may not always be observed with TPs, however, because, unlike smears, the lymphoid cells may not have been disrupted if they get gently imprinted directly on to a glass slide²⁰ (Fig. 2). It is also important to be aware that when a TP is performed in cases of Hodgkin lymphoma the Reed-Sternberg and Reed-Sternberg-like cells may cluster together (Fig. 3), which could mimic carcinoma.

Figure 2.

Touch preparation from a core needle biopsy of a lymph node involved by follicular lymphoma. Note the intact lymphoma cells that were imprinted on the slide and absence of background lymphoglandular bodies (Diff-Quik stain, 400×).

Figure 3.

Touch preparation from a core needle biopsy of a lymph node involved by classical Hodgkin lymphoma showing a loose cluster of Reed-Sternberg cells (black arrow) (Diff-Quik stain, 200×).

Soft tissue tumors

Although FNA certainly plays a helpful role in the workup of soft tissue tumors, it is well known that they are often unable to subtype many of these tumors because of the need to evaluate certain tissue architecture features (eg, zonal patterns, vasculature).^21,22 Also, grading of sarcomas is often not reliable on FNA samples alone. Further, certain soft tissue tumors are best evaluated by CNB because they are paucicellular (eg, myxoma), constituted mostly of blood vessels (eg, hemangioma), contain predominantly dense fibrosis (eg, sclerosing epithelioid fibrosarcoma), or have marked ossification (eg, myositis ossificans).²³ Like FNA smears, TPs are likely to be easier to interpret when soft tissue tumors are hypercellular and when tumor cells easily dissociate from their surrounding stroma. For this reason, small round blue cell tumors typically yield good TPs. A characteristic finding in Ewing sarcoma FNA smears is a background tigroid appearance. With TPs, however, because there is typically less disruption of cells, there is less glycogen present in the background. Furthermore, if there is no smearing action performed with a TP then this glycogen-rich background does not form characteristic tigroid stripes and may lead to a misinterpretation (Fig. 4).

Figure 4.

A, Touch preparation from a core needle biopsy of Ewing sarcoma showing a bubbly glycogen-rich background among small blue tumor cells (Diff-Quik stain, 400×). Without any smearing action this background does not form tigroid stripes (black arrow) that are typically seen in fine-needle aspiration smears, as shown in B (Diff-Quik stain, 200×).

Bone specimens

FNA of bone lesions often yields aspirates of low cellularity, especially in cases of sclerotic metastases. Compared with CNB, some investigators have reported that FNA of bone tumors can be more inconclusive.²⁴ In this setting, TPs may be helpful, even if only a few lesional cells are dislodged from a dense CNB onto a glass slide during preparation.^15,23 TPs are particularly helpful to diagnose bone metastases (eg, carcinoma) and hematolymphoid neoplasms (eg, myeloma) involving the bone marrow, whereas subtyping sarcomas and bone tumors may be more difficult.^15,25 When interpreting a bone TP it is important not to misinterpret normal bone marrow elements (eg, megakaryocytes) or reactive bone (eg, osteoblasts) as malignancy.

Accuracy and performance

The accuracy of TPs has been reported to be high (78% to 96%) for a variety of specimens, showing a general concordance of 80% to 92% with concurrent CNB histopathology diagnoses.^1,16,26,27 ROSE evaluation of TPs can thereby help reduce the number of cores required for adequate diagnostic material. Paulose et al reported that in their clinical practice, TPs prepared from 90 lung CNBs demonstrated a sensitivity, specificity, positive predictive value, and negative predictive value of 89%, 100%, 100%, and 68%, respectively.²⁸ Not surprisingly, a combination of FNA with TP of a CNB often maximizes the diagnostic yield (eg, up to 95% accuracy).¹ The performance of TPs is particularly high for epithelial-type malignancies, especially those that are hypercellular and where the tumor cells easily imprint onto slides. Apart from being used for ROSE, TPs also provide tumor-enriched samples that offer an excellent source of high-quality tumor DNA for ancillary studies (eg, molecular analysis).

Touch preparation technique

Very few articles address the actual preparation technique of TPs. Proper technique is indeed necessary to produce a good impression or imprint of cells from a tissue sample onto a glass slide that appropriately preserves cytomorphologic architecture without creating artifacts or excessive loss of diagnostic material. General guidelines for proper technique are summarized in Table 3. TPs may be performed by cytotechnologists, pathologists, trainees, or even proceduralists (eg, radiologists). Samples suitable for TPs need not only include small CNBs, but can be performed with any portion of antemortem or postmortem tissue.

Table 3.

Guidelines for optimal preparation of touch preparations.

• Touch the core along the short axis of the slide to minimize movement

• For most cases touching a core biopsy once or twice is enough

• Do not drag tissue core biopsies for >1 cm along the slide

• Gently handle the core biopsy to minimize tissue manipulation

• For delicate cores that may easily fragment, the touch and pick method is best

• Use uncoated slides to limit tissue sticking to slides and thus fragmenting cores

In general, TPs should not be performed on frosted slides. The TP technique simply involves touching a CNB on to a glass slide (Fig. 5). This can be performed using several methods.^29–31 With the “drag method”, a piece of tissue or core biopsy is placed onto a glass slide and gently dragged along the surface of the slide (eg, gently using the tip of a needle, wooden applicator stick, or the edge of a second glass slide). With the very similar “roll method” a tissue core is lightly rolled on a glass slide (gently using the tip of a needle, etc). This is best suited for solid firm cores that can be easily rolled (eg, bone cores). An alternate roll method can be performed with softer tissue cores by sticking it to a plastic needle sheath and then rolling it gently on the slide. If the core sticks to this sheath, however, precious cells may be lost. Hence, after the TP is made the plastic needle sheath is directly dipped into the formalin container and stirred gently to dissociate the tissue core. This helps minimize core manipulation. The “touch and pick method” uses a needle or a stick to gently touch a tissue biopsy on to the surface of a clean glass slide, then pick it up again without dragging it, and gently allowing the tissue to next touch another area of the slide. This should not be performed too many times because such manipulation can damage the core, cause streaking artifact, and may needlessly cause too many cells to be removed from the tissue core. Lightly wetting a slide beforehand can be considered to prevent tissue cores from firmly sticking to a slide, making it easier to move them around. Finally, there is the “sandwich” method, where a second glass slide is gently placed on top of the CNB to create an imprint. Forceps should be avoided because they may crush cells. Slides prepared by any of the aforementioned methods can either be air dried or wet-fixed (eg, immediate immersion in 95% ethanol). For air-dried specimens Romanowsky-type stains such as a Diff-Quik can be performed. For wet-fixed slides a Papanicolaou stain can be used. A toluidine blue stain may also be used, which requires only one drop of stain.

Figure 5.

Left, The “touch and pick method” of performing a touch preparation from an intact core needle biopsy is shown. Right, This slide shows that multiple touch imprints of a core needle biopsy were made (Diff-Quik stain). Too many touches like this are to be avoided as this may be associated with excessive loss of cells from the core, introduction of artifacts, and damage of the tissue core.

Problems and pitfalls

TP artifacts may distort or obscure cytomorphology and thereby create a diagnostic dilemma. Typical artifacts that may be seen with TPs include streaking (Fig. 6), crushed cells, and thick clusters that make it hard to visualize individual cells.²⁹ Streaking of cells occurs when CNBs are dragged, sometimes even slightly, along the slide. Streaking is especially common with lymphoid lesions and small cell carcinoma. Certain cytologic features characteristically found in FNA samples created with direct smears, such as lymphoglandular bodies in lymphoid lesions or a tigroid background in glycogen-rich neoplasms (eg, seminoma, Ewing sarcoma), may not be evident with TPs. Over-manipulation of friable tissue cores can cause them to fragment, which is a major problem with necrotic lesions. CNBs from lesions with abundant myxoid or mucoid material may also cause fragmentation or excess loss of material from the core.

Figure 6.

Touch preparation from a core needle biopsy of a lymph node involved by mantle cell lymphoma showing streaking artifact. Note the absence of lymphoglandular bodies in the background. (Diff-Quik stain, 200×).

One of the major criticisms of TPs is the potential “loss” of most/all of the cellular material on the slide, resulting in a lack of remaining tumor cells within the tissue for subsequent permanent tissue section evaluation (Fig. 7) and/or inadequate material for ancillary testing. This typically occurs with vigorous manipulation (eg, excessive dragging) of the CNB when performing a TP (Fig. 8). To demonstrate this drawback, Rekhtman et al designed an experiment in which ex vivo CNBs were used to prepare TPs that were dragged along a glass slide for different lengths.³² Although the overall cellularity present on their TP slides relative to the corresponding CNBs increased respectively (eg, 19% for imprint only, 33% for 1-cm drag, 41% for 2-cm drag, and 46% with full-slide drag), this also resulted in a subsequent proportional loss of DNA content from the CNBs (15% for imprint only, 36% for 1-cm drag, and 50% for 2-cm drag).³² Of note, there was no difference in the diagnostic accuracy for less versus more extensive TPs. In a study by Tong et al that evaluated 1100 CNB cases with associated TPs the authors found that 48 cases (4.4%) had fewer tumor cells in the TP than the corresponding CNB, whereas 36 cases (3.3%) had more tumor cells in the TP when compared with the CNB.³³ These investigators further demonstrated that the 3 most common anatomic sites subject to differences in cellularity were lung, bone, and soft tissue.³³ Biopsy needle gauge also had an influence on the depletion of tumor cells in this study. A 20-G needle was used in all of the lung CNB cases that lacked tumor cells, whereas none of the lung biopsies that were performed with an 18-G needle were completely depleted of tumor cells.³³ Therefore, the recommendation is to avoid forceful TPs. To mitigate the drawback from depletion of cells, an additional untouched core should be requested to ensure that adequate tumor cells are available for ancillary testing.

Figure 7.

Examples of core depletion after aggressive touch preparations. Left, This touch preparation (top image) shows a hypercellular slide (Diff-Quik stain, 100×) whereas the corresponding core needle biopsy shows only scant remaining mantle cell lymphoma (hematoxylin and eosin stain, 40×). Right, This touch preparation (top image) of a metastatic colonic adenocarcinoma shows abundant tumor cells imprinted on the slide (Diff-Quik stain, 100×) while the corresponding core needle biopsy of the liver shows absence of metastatic adenocarcinoma (hematoxylin and eosin stain, 40×).

Figure 8.

Aggressive touch preparation where the core needle biopsy was unnecessarily dragged across the full length of the slide.

Ancillary testing

Subtyping malignancies by means of ancillary testing (eg, immunohistochemistry, FISH, or molecular tests) is part of everyday cytopathology practice. These ancillary tests can be performed on a variety of cytology samples including liquid-based collections, smears, and cell blocks. More recently, CNBs are being increasingly procured because they provide larger tissue fragments for testing than cell blocks with scant aspirated material. However, the drawback is that CNBs may contain many normal and inflammatory cells as well as abundant desmoplastic stroma, which lowers the overall tumor cellularity. As a result, tissue cores often require laborious microdissection to select a pure population of neoplastic cells for molecular testing. Because unstained TP slides performed during ROSE often contain abundant intact neoplastic cells, they serve as potential alternatives to the traditional material used for ancillary testing.

TP slides have not been widely used for immunocytochemistry. However, TP slides can be used to perform FISH and molecular genetic testing.^34,35 Several reports have demonstrated the value of utilizing TPs for FISH studies to diagnose lymphoma,³⁶ carcinoma,³⁷ and even detect congenital abnormalities at autopsy.³⁸ Unstained TP slides are particularly advantageous for FISH as they contain the whole (ie, unsectioned) nucleus of cells imprinted on the slide, which avoids truncation artifact and formalin fixation shrinkage artifact. Truncation artifact is the underestimation of copy numbers that is caused by incomplete DNA complements within transected nuclei typically seen in tissue sections.³⁹ High-quality TP slides made at the time of onsite evaluation are very important, however, because with too many overlapping cells it may be difficult to reliably interpret distinct nuclear FISH signals. One study showed similar results for TP cytology and formalin-fixed paraffin-embedded tissue when performing next generation sequencing of lung tumors.⁴⁰ Cytology laboratories should, accordingly, have a standard procedure in place for suitable preparation, storage, and use of unstained TPs.

Related applications

Other uses of TPs include postmortem examination for rapid assessment, testicular biopsy to evaluate for spermatogenesis during male infertility workup, and biobanking collection.^41–43 Imprints are also widely performed during intraoperative consultations when receiving large specimens. To make an imprint a clean, dry glass slide should be pressed gently on to a freshly cut large specimen for a few seconds. Imprint cytology is especially beneficial for the intraoperative evaluation of neuropathology cases^44–46 and bone tumor resection margins.³ The utility of intraoperative imprints of sentinel lymph nodes has been studied for a variety of cancers such as breast carcinoma, melanoma, and cervical carcinoma. It is often easier to perform cytology imprints on fatty sentinel lymph nodes than try to cut a good frozen section. A study by Motomuro et al showed that imprint cytology for the intraoperative diagnosis of sentinel node metastases in breast cancer achieved high sensitivity (84.6%), specificity (96.6%) and accuracy (94.1%).⁴⁷ Nejc et al similarly showed 94% accuracy of intraoperative imprint touch cytology of sentinel nodes in skin melanoma patients.⁴⁸ In general, metastatic carcinoma cells tend to cluster and are hence easily recognized on imprints, even at low magnification against a background with lymphohistiocytic clusters. Similar to frozen sections, however, micrometastases and lobular breast carcinoma may be difficult to diagnose by imprint alone.

Intraoperative imprint cytology of thyroid nodules is another instance that may be more helpful than frozen sections,^49,50 especially for cases that were previously diagnosed as suspicious for papillary thyroid carcinoma (PTC). This can help guide the appropriate surgical procedure such as whether the surgeon should perform a total thyroidectomy for PTC instead of just a lobectomy. It is easier to diagnose PTC by means of cytology versus frozen section because of the ability to visualize characteristic nuclear features (eg, grooves, pseudoinclusions). In a study involving 84 patients, Anila and Krishna showed that imprint cytology of thyroid lesions yielded a diagnostic sensitivity of 75%, specificity of 100%, positive predictive value of 100%, and negative predictive value of 98.74%.⁵¹ Of note, intraoperative TPs have also been used for the identification of parathyroid glands.^52,53

Intraoperative imprint cytology is unsuitable for lesions in which the cells do not dissociate well (eg, dense fibrosis or bone tumors⁵⁴), to evaluate depth of tumor invasion,⁵⁵ distinguish in situ carcinoma from invasive carcinoma, distinguish adenoma from carcinoma, diagnose well-differentiated neoplasms,^54,56 evaluate certain inflammatory diseases, and for evaluating joint tissue during intraoperative consultation of arthroplasty revisions for determining neutrophil counts.

Conclusion

Touch preparation cytology is unique and differs from conventional FNA cytology in many aspects. TPs fill a gap in cytopathology practice by providing a method to perform ROSE of small CNBs. Moreover, TP slides have proven to be extremely valuable given that they often contain a cellular and pure population of whole tumor cells for ancillary testing. Not surprisingly, the evaluation of TPs is becoming a new trend in the era of personalized medicine. However, the benefits of TPs are highly dependent on well-prepared slides. Therefore, awareness of different TP techniques is essential. It is also important for cytologists to recognize potential pitfalls and artifacts associated with TP cytology in order to avoid misinterpretation of these preparations. Finally, apart from their value with immediate diagnostic interpretations and ancillary studies, TPs are also valuable in research studies, clinical trials, and biobanking.