The principle is a simple one: cytology is appropriate when you suspect a disease in which the tissue architecture is not important, and in which the lesion is more-or-less diffuse within the tissue. Tissue core biopsies (most often, tru cut biopsies) retrieve only tiny histologic samples. They are appropriate when you anticipate the histologic changes will be widespread within the tissue, and when you anticipate the need for at least some appreciation of architectural arrangement of the diseased tissue. Finally, larger traditional incisional or excisional biopsies are required when you anticipate the need for a broad geographic overview of the tissue organisation, to appreciate things like tumor invasion or the relative amounts of normal and abnormal tissue that may be critical to prognosis.

Cytology is fine for situations in which the change is likely to be diffuse within the tissue, and in which the critical diagnostic information is likely to be gained from examination of individual cell morphology or the detection of infectious agents, without regard for tissue architecture. These criteria apply to diseases like feline hepatic lipidosis, in which the fatty change affects virtually every hepatocyte. The enlarged kidney of a cat with suspected renal lymphoma would be another excellent candidate, since the diagnosis is made by seeing large numbers of monotypic lymphocytes without regard to their architectural arrangement or their relationship to pre-existent renal tubules. Cholangiohepatitis, on the other hand, has only multifocal accumulations of leukocytes that may be missed with fine needle aspiration, and the diagnosis requires not only the detection of the leukocytes but also the precise anatomic location in which they accumulate. The same limitation applies to aspiration of inflammatory skin disease, since in almost all cases the diagnosis requires not only an appreciation of what leukocytes are present, but exactly where they are in relationship to the hair follicles, adnexal glands, and epidermis. A disease like cirrhosis, in which there may be no leukocytes and in which the diagnosis requires the appreciation of fairly coarse architectural reorganisation, is a terrible candidate for cytologic diagnosis.

Tru cut biopsies are appropriate when you need at least some appreciation of tissue architecture, and when the architectural changes are likely to occur in a fashion that would be captured in small biopsies. In contrast to cytology aspirates, these samples will allow us to appreciate the location of leukocytes relative to the pre-existent tissue architecture, and they also allow us to appreciate how the individual cells (leukocytes or tumor cells) are behaving with respect to the pre-existent architecture and to one another.

This ability to evaluate tissue architecture becomes particularly critical in the diagnosis of malignancy, since destruction of pre-existent architecture is often a critical requirement to distinguish neoplasia from inflammation.

Invasion by atypical cells through normal tissue barriers is an observation of great prognostic importance, easily appreciated in well-selected tru cut biopsies but never with aspiration cytology. The greatest limitation of tru cut biopsies is that they do not allow appreciation of "coarse" architectural change. They allow only a keyhole view of a tumor that may or may not have tissue invasion in that specific location. They also may not allow appreciation of the variability in the degree of histologic aggressiveness within a neoplasm.

In situations other than neoplasia, the same strengths and weaknesses apply. Tru cut kidney biopsies, for example, are excellent for distinguishing glomerulonephritis from amyloidosis because, in both diseases, one expects virtually every glomerular tuft to be affected (the average tru cut biopsy captures 4-8 glomeruli). On the other hand, a disease like pyelonephritis typically creates random radial corticomedullary streaks of inflammation, and the lesion may be missed with a tru cut sample. Even if captured in the sample, the pathologist will not know whether the lesion reflects 10 percent or 90 percent of the kidney! Since the prognosis is directly correlated with the extent of the lesion, this is a critical weakness in the utility of renal tru cut biopsies. Unless you have remarkable skill in guiding such tru cut biopsy needles via ultrasound, tru cut biopsies are usually not suitable when you anticipate the lesions will be random and multifocal (like metastatic neoplasia, lesions of bacteremia, etc.).

Excisional or incisional surgical biopsies should be the gold standard against which all other biopsy techniques are measured. However, there are specific instances in which this most invasive of all biopsy techniques does not offer any substantial advantage over cytology or tru cut biopsy, and a few cases in which it may even be less accurate. Examples of the last category are not frequent; the one that comes to mind is gastrointestinal biopsy. One might imagine that full thickness biopsies would always be preferred, but in fact full thickness samples are limited in the number one can take, and by the fact that they are virtually always taken without knowing anything about the mucosa hidden by what usually is a normal tunica muscularis. Endoscopy, while limited to the most anterior and posterior regions of the gastrointestinal tract, has the advantage of direct visualization of suspected mucosal lesions, and the ability to harvest very large numbers of samples in a relatively short interval. For diseases that are notoriously patchy like eosinophilic gastroenterocolitis, endoscopic biopsy offers a substantial advantage.

There are several important circumstances in which the optimal result comes from using a combination of excisional biopsy and cytology, each bringing its different "talents" to the microscope. There will be many times when poorly melanotic melanoma, for example, is more easily confirmed with cytology than with histology. Anaplastic mast cell tumor will almost always have at least a few metachromatic granules visible with cytology, but trying to find those granules in a histology sample may be virtually impossible. Bone marrow is best evaluated by a combination of histology core and cytology. The histology core sample is superior for assessing overall cellularity and architectural changes, while the cytology/aspiration offers a more accurate assessment of what cell types are present. Cytology will often allow detection and identification of infectious agents with greater accuracy than histopathology.

Listed below are some "rules of thumb" about the relative reliability of cytology, tru cut/endoscopic biopsy, and excisional biopsies for the diagnosis of some of the most familiar diseases that you will attempt to confirm via cytology or histology. I warn you that all internists or oncologists may not share my opinions . . . but of course they usually do not have to interpret the samples!

¹N.D.: not usually done as a diagnostic procedure.