Pancreatic Tumors. Группа авторов

FNA using ThinPrep is inferior to that of conventional smears. However, ThinPrep® remains a viable substitute for smear preparation for practices in which a cytologist is not available on-site to prepare the slides. At our institution, we receive pancreatic FNA specimens submitted directly into CytoLyt, and have found them to have a good diagnostic yield. Aspirates form cysts have been identified to have background mucin and the epithelium is well preserved.

      Cell Blocks

      A good cell block preparation can be utilized for additional ancillary studies such as immunohistochemical stains, fluorescence in situ hybridization, and other molecular studies necessary for the diagnosis and treatment of the patients. Ancillary tests that are validated for formalin-fixed specimens may have to be revalidated for specimens that are fixed in alcohol. Alcohol-fixed specimens may have altered antigenicity which may render some of the tests falsely negative.

Different cell block preparation methods utilize different collection mediums for transportation and fixation of the specimen collected, for example methanol is used as a fixative for the automated Cellient cell block technique [19]. Other methods of cell block preparation are the clot and scrape method, BBC cell block fixative method, plasma thrombin cell block preparation, collodion bag cell block procedure, Shandon cytoblock method, HistoGel method, Tissue coagulum clot method, and Formalin or alcohol vapor method [19]. The clot and scrape method and formalin/alcohol vapor method are both inexpensive techniques but are reported to have variable quality and cellularity. The collodion method, although time consuming, has a good cellular yield, making it ideal for samples with scant cellularity. Collodion is admixed with an ether-based solvent; it must be handled in a laminar flow hood and stored in small volumes in a flameproof enclosure [20]. The Cellient automated system is time consuming and expensive but is great for small/scanty samples and provides a crisp cellular architecture. As it is fully automated, there is little chance of cross-contamination. It utilizes methanol as the fixative which could interfere with certain immunohistochemical stains. However, in a separate protocol formalin can be used as the fixative. The instrument can accommodate only one specimen block at a time, so high-volume labs may require multiple instruments. The HistoGel method is tedious but helps in the concentration of cells and provides good cellular preservation. The plasma thrombin method is an inexpensive and simple method and provides a clean background. The disadvantages of this technique include an inability to control clot formation and irregular distribution of cells within the pellet. Some thromboplastin agents contain epithelial cells that may interfere with the diagnosis and with the interpretation of ancillary studies [21].

      Processing of Cytology Samples from Pancreatic Cystic Lesions

The goal of pancreatic cyst fluid evaluation is to differentiate mucin from non-mucinous lesions, and to identify cysts carrying a high-risk malignancy [22, 23]. Hence, both symptomatic and asymptomatic pancreatic cysts are actively evaluated to determine the nature of the cyst. It is extremely crucial to examine the cyst fluid aspirated to make this distinction and to rule out high-grade dysplasia or adenocarcinoma associated with the cyst.

      Pancreatic cyst fluids need to be submitted fresh and undiluted. The pancreatic cyst fluid needs to be triaged for biochemical analysis, to include carcinoembryonic antigen (CEA) and amylase, and for molecular analysis, in addition to cytology. CEA and amylase can be performed from the supernatant, which is preferred, as it preserves the cells in the specimen for cytological evaluation.

Chai et al. [24] developed a volume-based protocol using minimal pancreatic cyst fluid volumes for biochemical, molecular, and cytology analysis. The addition of molecular analysis to the traditional CEA and cytological examination of the pancreatic cyst fluid has led to a reappraisal of how these samples are triaged and the comparative roles of these different modalities [25, 26]. Ideally, if the cyst fluid volume is sufficient it is split into neat (unaltered fluid), smears, cell block, and supernatant fluid. Chai et al. [24] proposed the following pancreatic cyst fluid triage based on volume:

Fig. 3. A volume-dependent protocol for processing of aspirated pancreatic cyst fluid [28](reproduced with permission from Cambridge University Press).

      1.Volume ≤1 mL

      (a)Centrifuge the specimen for generating supernatant fluid and deposit (pellet)

      (b)Deposit to process as a cell block for cytology review and KRAS testing

      (c)Supernatant fluid (0.5 mL) to biochemistry for CEA

      (d)Remaining supernatant fluid frozen and stored

      2.Volume >1 mL

      (a)Neat fluid (0.5 mL) to biochemistry for CEA

      (b)Centrifuge specimen to generate deposit and supernatant fluid

      (c)Deposit to process as a cell block and 2 wet-fixed smears

      (d)Supernatant and any excess neat fluid frozen and stored

      (e)KRAS mutation testing on cell block material

      (f)For difficult cases KRAS retesting can be performed on stored supernatant and/or neat fluid

      A limitation of this proposal is that it only includes CEA measurements, and uses neat fluid for CEA measurements if the volume is greater than 1 mL. Also, KRAS is the only mutation evaluated. Current analyses include a greater number of targets.

A protocol proposed by Dr. Pitman calls for only using the supernatant for CEA and amylase, and preserving the neat fluid for cytology and molecular testing (Fig. 3). This protocol is also volume dependent.

      1.Volume <0.5 mL

      (a)Divide in equal halves for molecular testing and CEA level

      2.Volume ≥0.5 mL

      (a)Vortex the specimen and remove 0.3 mL for molecular assessment

      (b)Centrifuge the remaining fluid; send the supernatant for CEA and amylase levels, and the remainder for cytology

      The ideal protocol for cyst fluid processing may be laboratory dependent. The CEA and amylase may be processed from the supernatant. Therefore, it is not necessary to send the neat fluid. Recent studies have shown DNA retrieval from residual cytological material, so it may be possible to perform a limited mutational panel on the cyst fluid supernatant.

Cytology is the most sensitive test for determining malignancy in pancreatic cysts [27]. However, cytology is hampered by low cellularity and the presence of degenerative changes. The presence or absence of background mucin and high-grade dysplasia should be reported. Finally, it is a combination of CT/MRI findings, molecular test results, CEA and amylase levels, and cytological features present on slides and/or cell block sections that determine the final diagnosis rendered and subsequent surgical/non-surgical management of the patients.