Rapid on-site evaluation (ROSE) involves a cytology staff member, either a (cyto)-pathologist or cytotechnologist, providing real-time evaluation of the specimen during the procedure and communicating the findings to the clinician (e.g., pulmonologist, interventional radiologist, surgeon, or endosonographer). Correlation with the history and imaging provides an integrated approach to the diagnosis (Table 3.1).

Traditionally, providing a preliminary assessment was sufficient. Now, in light of the added importance and responsibility for the cytologist to properly triage the specimen, it is intuitive that the individual performing ROSE has experience in cytomorphology interpretation, be knowledgeable about ancillary tests necessary for various diagnoses, ensures that sufficient tissue is allocated for potential ancillary tests (e.g., immunohistochemistry and/or molecular testing for carcinomas, microbiological cultures in cases of inflammation or granulomas, or flow cytometry for lymphoma assessment), and be informed about methods of specimen fixation/preservation required for each test. Proper decision-making and specimen submission at the time of the procedure can directly affect patient care. Recent data strongly support that aspirates performed with ROSE optimize the use of the aspiration procedure [27, 28] by improving diagnostic yield [29] and aiding in the procurement of additional material for potential necessary special studies [29–40].

There is also value added to a well-prepared specimen—one without poor preservation or preparation—which expedites final examination. When the technical quality of a specimen is maximized, final interpretation is not compromised (Fig. 3.2). High-quality preparations limit intradepartmental consults and time consumed interpreting slides with limited cellularity, poor preparation, or artifacts and shift focus on diagnostic dilemmas. In cases where the same pathologist performs ROSE and issues the final report, time may not be “wasted” [13] as on-site assessment provides an opportunity to obtain the clinical history and preview of a fraction of the slides.

ROSE of lung nodule aspirates is effective at discriminating malignant versus nonmalignant lesions, such as sarcoidosis [41], and having a cytopathologist present is associated with a significantly greater accuracy in diagnosis of malignancy [32–34, 37, 39]. ROSE increases sensitivity and specificity [42] and the predictive value of a negative test. Moreover, it increases the likelihood of attaining a higher rate of positive diagnoses [34], when present, and decreases unsatisfactory results [37]. While the average reported non-diagnostic rate of FNAs is 20 % when on-site evaluation is not utilized [27], a diagnostic rate of 98 % has been achieved in a study of over 5,600 FNAs with ROSE [27]. Diagnoses rendered at time of ROSE are highly accurate and usually concordant with the final interpretation [43, 44]. In the same study with >5,600 FNAs, the on-site interpretation was in agreement with the final diagnosis in 85 % of cases; there was disagreement in 2.7 % of cases and deferment to the final diagnosis in 12 %.

The high concordance of on-site immediate interpretation and final diagnosis allows for reliable decision-making that can enhance efficiency and cost-effectiveness of patient care [27]. With a preliminary diagnosis, clinical decisions can be expedited [27, 45–49] and scheduling for additional imaging, appointment with a specialist, or surgical procedure can be initiated. The procedure can be terminated as soon as diagnostic material is obtained. By doing so, fewer needle passes [34, 50, 51] than those predetermined in the absence of ROSE [51] are performed at fewer sites, thereby decreasing the time a patient is under sedation, reducing the complication rate [52], and increasing the safety of the procedure. Next, ROSE results in a decrease in the number of repeat [37] or additional procedures necessary for a diagnosis [50], which has downstream effects. Assuming non-diagnostic FNAs are repeated, direct institutional charges can be significant. The diagnostic value and cost-effectiveness of ROSE was evaluated in 5,688 FNAs [27]. In the absence of ROSE, the estimated calculated cost of performing repeat FNAs of non-diagnostic specimens resulted in additional direct institutional charges of $2,022,626 over 5 years. These savings are significant despite the professional service fee of on-site interpretation [27]. Also, this does not take into account indirect costs to the patient, such as time off work and prolonged hospital stays. ROSE has downstream effects on the facility as well. Specifically, with shorter procedures times and fewer re-biopsies, utilization of procedure rooms and imaging facilities [53] is optimized.

ROSE is not always feasible, either by a pathologist or cytotechnologist, due to limited resources of the facility. For a pathologist, ROSE results in time consumption, workflow disruption, and little reimbursement, i.e., high cost for time spent at the procedure [29, 53, 54] (Table 3.2). From a pathologist’s economic standpoint, time is spent better signing out cases in the office than performing immediate, on-site evaluation of FNA specimens [53]. Compared with routine surgical pathology, compensation schedules for intraprocedural FNA consultations by pathologists are insufficient [40, 53]. Moreover, relative value units for signing out cases in the office outweigh those generated during a single ROSE. The time and overall benefit are not calculated in the measurement of productivity. In light of this suboptimal compensation and often prolonged procedure times [54], some institutions defer ROSE to cytotechnologists or forego ROSE and place the entire specimen in a liquid-based cytological container [55]. Despite the extra expense for the on-site cytology service [56], reports document that on-site analysis, specifically for transbronchial FNA, is cost-effective [56, 57].

Though infrequent [49], another drawback of ROSE includes diagnostic misinterpretation, both false positive and false negative [58, 59]. In the majority of cases with differing on-site immediate diagnosis and final diagnosis, the reason is sampling error, i.e., only normal-appearing tissue evident on Diff-Quik stained smears and diagnostic cells present in the remaining specimen. In one report, the diagnostic yield (determining tissue adequacy) of ROSE in cases with high suspicion for lung cancer was 77.1 % (101 of 131) but increased to 93.9 % upon final cytological analysis following review of the remaining material processed in the laboratory [60]. This suggests that the lesional tissue was sampled but not identified during ROSE; training in procurement and triage can improve the on-site yield in such instances. Additional reasons for discrepancy include scant cellularity, need for intradepartmental consultation or ancillary studies, intraobserver variability, and challenging diagnosis [49].

Laboratories may find it challenging to provide ROSE due to high volume, insufficient staffing, time constraints, and distance issues. Telecytology (TC), a relatively recent means to view slides remotely in real time, addresses these issues while allowing the pathologist to be involved in the evaluation process. Telepathology (TP) has been shown to be a valuable tool for frozen section diagnosis with high diagnostic accuracy [71–74]. While TC still lags behind TP in validation and use [75], it has value in off-site assessment of multiple organs [36, 76–83] and lymph nodes sampled under EUS, EBUS, and CT guidance [36, 78, 82] without significant difference in agreement between TC and ROSE [83]. Concordance between TC and conventional ROSE has also been reported in adequacy assessment, assignment of preliminary diagnostic categories (non-diagnostic, benign, atypical, suspicious, malignant), correlation with final diagnoses [36, 77–81], and average number of passes performed [78]. Although sample size was small, one study showed an 89 % sensitivity of both ROSE and TC; however, ROSE showed a higher specificity compared to TC (93 % versus 87 %, respectively) [83].

There are three main available methods of image transmission for TC and TP including (1) static image capture, (2) dynamic live video (real time), and (3) whole slide image (WSI) [75, 84]. For both the static and dynamic systems, a team approach that involves a skilled on-site operator experienced in cytology is essential. Briefly, static image capture, consisting of a microscope with an attached digital camera and internet access for file transmission, offers low cost, wide applicability, and access. Limitations include selective and narrow scope of image review and inability to view on multifocal planes [75, 84]. With the dynamic system, comprising a microscope with an attached camera capable of generating video, the operator at the originating site controls the microscope, and the pathologist situated remotely views the slide passively [84]. Robotic systems are also available with capability of complete and unbiased slide review as well changes in magnification and focusing by the interpreting pathologist. WSI captures the entire slide content and permits slide review in a manner similar to that of a glass slide under a microscope [75, 84]. Disadvantages include increased time for entire slide scan and suboptimal depth of field views [75, 84].

TC has been shown to be a more effective use of a pathologist’s time on the clinical service, with a reported average time of 30 min saved per procedure [78]. Although the actual time spent reviewing a case via TC versus conventional on-site assessment may be similar and not statistically significant [36, 78], the potential time saved in a given day for a pathologist working on a busy FNA service can be significantly reduced by eliminating time spent traveling to and from the procedure site and downtime either during prepping or in between passes. In a high-volume FNA service, the time saved for cytopathologists could potentially be measured in hours.

Loss of personal contact and communication between the cytopathologist and clinician/radiologist performing the FNA is a noteworthy concern. Although this issue is not addressed in much of the TC literature, one study noted that clinician satisfaction has not been negatively impacted by using TC over in-person assessment and that no issues related to TC have been identified by them [78]. Others specifically report that even with TC, interaction with the endoscopist is not sacrificed and that TC allows for mass communication/teaching beyond any distance and advances communication so that pathologists can reach out to more clinicians, thereby improving patient care [81].

TC also allows the cytopathologist to actively participate in making decisions about proper triage for ancillary studies [78], including IHC, flow cytometry, molecular testing, and cultures; however, successful triaging requires that the on-site operator, whether a cytotechnologist, cytology fellow, or pathology resident, be knowledgeable of various preservation media and technically skilled in the preparation of slides that allows for necessary material to be saved for such ancillary tests. Comparison of rates of adequate material for ancillary tests, specifically IHC and molecular testing for lung cancer, in FNAs with in-person ROSE versus TC ROSE has not yet been addressed and is an issue that is worth investigating.

Several fixatives and transport media are available for cytology specimens. For EGFR molecular testing, guidelines provided by the College of American Pathologists, the International Association for the Study of Lung Cancer, and the Association for Molecular Pathology recommend using formalin-fixed (10 % neutral buffered), alcohol-fixed (70 % ethanol), fresh, or frozen specimens following validation [11]. Cytology specimens initially placed in other solutions (i.e., saline, RPMI) can be transferred to formalin or CytoLyt subsequently. Fixation for 6 to 12 hours is recommended for small specimens [11].

As stated above and discussed in the following section on cell blocks, formalin-fixed paraffin-embedded (FFPE) tissue is currently recommended for ancillary tests such as IHC stains and molecular assays [2, 11, 85, 86]. Typically, formalin is the fixative in which IHC stains and most molecular assays are optimized and validated, and FFPE material has potential to yield multiple tissue sections that are suitable for these ancillary studies [38, 87–90].

Liquid-based preparations (LBPs) offer a standardized alternative to conventional smear preparation and fixation for FNAs and exfoliative specimens, including respiratory specimens and effusions. The current FDA-approved liquid-based preparations (LBPs) in use are ThinPrep®/CytoLyt (Cytyc Corp, Marlborough, MA) and SurePath™ (Becton Dickinson) [91]. The specimen is placed in an alcohol-based solution (e.g., CytoLyt), which serves as a fixative and transport medium. An automated system in the laboratory creates slides, which are stained with Papanicolaou stain, with an even distribution of cells and without loss of cells. Additionally, the process and solution minimize contamination by blood, inflammation, and debris.

There are several additional advantages of LBPs (Table 3.4). They preserve cells at room temperature in the fixative solution for extended periods of time (3 weeks in CytoRich and 3 months in PreservCyt) [91]. CytoLyt solution appears to be superior to CytoRich Red in terms of the yield of suitable DNA [92]. Cell blocks for ancillary tests [55] can be prepared from the unused cells in solution. Even without cell blocks, ancillary tests, such as IHC [93–95] and molecular analysis [92, 96–100], can be performed on LBPs. In fact, the percentage of malignant cells on FNA specimens assessed on ThinPrep® has been found to be significantly high [101]. EGFR and KRAS mutation analyses have been successfully carried on lung carcinomas preserved in CytoLyt [96–98] and have shown comparable results between FNA samples and histological tissues [98]. Potential advantages of utilizing LBP for molecular techniques include (1) homogeneous cell distribution, which is helpful in determining percentage of malignant cells in a specimen dominated by benign or normal cells, and (2) preparation of multiple slides, which spares the original routinely stained slides/smears [92].