Introduction to the Practical Algorithmic Pattern Recognition Approach to Fine Needle Aspiration Biopsy

FNAB Pattern Recognition

Numerous texts over the past 50 years have described the attributes of fine needle aspiration biopsy (FNAB). The purpose of this text is not an attempt to replace these salient works in diagnostic cytopathology of FNAB lesions. This book serves as a “field guide” to the diagnosis of lesions by FNAB, incorporating a practical algorithmic pattern recognition method to achieve this goal. This algorithmic technique is similar to the process that an ornithologist would use to speciate a particular bird, a botanist would use to speciate a particular plant, or an ichthyologist would use to speciate a particular fish.

Renshaw previously described a pattern recognition approach for the diagnosis of FNAB lesions that concentrated on a differential diagnosis (DD) derived from a dominant particular cell type or background pattern (e.g., a DD from a dominant element of small blue cells, monotonous cells, pleomorphic cells, biphasic cell population, hypocellular background, or background dominant).

We describe a diagnostic, step-by-step algorithmic approach to FNAB that entails identifying key single cell, tissue fragment, and background elements that comprise a particular direct smear pattern. This algorithmic method is best achieved by first examining a slide at low magnification . Once a particular pattern of a lesion within a particular organ system is recognized as presented within this book, one can refer to a limited list of known entities that share that particular smear pattern. The microscopist can then use unique cytomorphologic characteristics identified at low and high magnification to distinguish a particular entity from others on that limited list, followed by appropriate ancillary tests to confirm the specific diagnosis in almost all cases. This algorithmic process often yields a specific cytologic diagnosis or offers a limited number of possibilities within the DD. It should be kept in mind that some lesions may fall into more than one direct smear pattern category, each with its own DD list.

We define “low magnification” as use of the ×2, ×4, ×10, or ×20 objectives and “high magnification” as the ×40 or ×60 objectives. At low magnification, the entire slide should be scanned and the overall “background environment” that exists on that slide assessed. The microscopist should not use high magnification at this pattern recognition stage except to briefly confirm significant details, so as to avoid making the mistake of “missing the forest for the trees.” At low power the overall cellularity and the presence or absence of extracellular substances, such as mucin, myoepithelial-derived stromal material, or cystic and necrotic debris, are assessed. Once the overall cellularity and background are assessed, the microscopist then focuses on what particular types of cells are present on the slide, as well as on the interaction of the cell types and/or stroma in tissue fragments. Is there only one specific cell type present or multiple cell types? Do these particular cells or a mixture of cells form tissue fragments or are they dissociated from each other? Do these cells and/or tissue fragments actually make up a particular architectural structure, a papillary or glandular formation, a two-dimensional flat sheet, or three-dimensional tissue fragment perhaps? Are there neoplastic cells closely associated with blood vessels in an angiocentric or coronal distribution? Are neoplastic cells intimately associated with a particular type of extracellular stromal material? What is the significance of these multicellular interactions in the diagnosis?

The diagnostic process is a step-by-step process: a low-power assessment of patterns followed by a high-power assessment of cell types, dispersed cells, and tissue fragments. This algorithmic exercise yields a specific diagnosis or at least a DD. If a specific diagnosis or DD cannot be established and if the low- and high-power assessments are discrepant, then the microscopist must go back to the low-power pattern recognition and similarly the high-power diagnostic features and reassess until the low- and high-power findings coalesce into a specific diagnosis or a list of DD.

Deconstructive Cytology

This text also introduces the concept of “deconstructive cytology,” a helpful exercise that emphasizes the recognition of the most dominant or significant basic key elements or details that comprise a particular smear pattern on a slide. Once the key elements are observed and categorized, these elements serve as building blocks in the construction of a simple sketch or cartoon within one’s mind eye that serves as a schematic rendition of a particular pattern. This process forces the observer to consciously commit to the identification of important key elements that create a unique pattern. By engaging in the creation of a certain cartoon that represents a recognized pattern on a smear slide, a cartoon pattern that is associated with a limited number of pathologic entities, one is more able to commit these various patterns to memory. Consequently, it will be easier to recognize lesions that are associated with a particular pattern when one encounters a similar case in the future. Cartoon renditions of all smear patterns introduced by the contributors of the various organ system chapters within this book are found on the plate pages in the text. These cartoons serve as a visual training guide to the reader that emphasizes the most important key details or elements that describe a particular smear pattern.