December 2010
Feature Story

William Check, PhD

Secret agents and double blinds are not usually part of education courses at CAP meetings. But in a way that describes the subject matter of the CAP ’10 course on infectious disease challenges in surgical pathology and cytopathology. Bobbi Pritt, MD, MSc, and Abdelmonem Elhosseiny, MD, presented five cases in which infectious organisms were found on tissue sections and cytology material but that only a vigilant pathologist proficient in the necessary tradecraft could see and identify. Interspersed were four quizzes that illustrated a twist: nonmicrobial entities mimicking infectious organisms in a way that could deceive even a vigilant pathologist who might not maintain a sufficiently high level of suspicion and ask the right questions.

“It is always important to keep infectious organisms in your differential diagnosis,” says Dr. Pritt, director of clinical virology and parasitology in the Division of Clinical Microbiology at the Mayo Clinic, Rochester, Minn. “Once you think you may have an infectious agent, you need to know what stains to do and how to interpret them.”

During the course the speakers used the five cases to show common challenges. The four quizzes, on the other hand, “were designed to break up the session a bit and inject some fun and show things people will not see all that commonly,” Dr. Pritt says. Once you take on the challenge of identifying infectious organisms, it is important to be aware of mimics. And some deceptive structures are not so rare. In the past few years Dr. Pritt has seen four cases of one type of crystal in urine that could be mistaken for a parasite egg. “We need to recognize them as artifacts as the first step in diagnosis,” she says.

Dr. Pritt is a pathologist with a subspecialty in microbiology and an interest in infectious organisms. After her pathology residency, she did a microbiology fellowship at Mayo, where she is now an attending physician. “As a pathologist, I can bring a unique perspective to the field of microbiology,” she says. Most important, as a pathologist Dr. Pritt can examine tissue sections. “In my division there are a large number of PhDs and MDs trained in clinical microbiology,” she says. “But I’m the only one with training in tissue pathology.” While most pathologists are not trained in microbiology, they can take advantage of the insights Dr. Pritt and Dr. Elhosseiny presented in their course.

In pathology at the University of Vermont Medical Center, infectious diseases are becoming more common, says Dr. Elhosseiny, who is a professor of pathology there. “We see them quite often, particularly in cytology.” That the pathologists can make a diagnosis quickly by visualization saves time (compared with microbiological testing), which means earlier treatment for the patient.

“This course combines Dr. Pritt’s expertise in microbiology with my expertise in cytology and surgical pathology,” Dr. Elhosseiny told CAP TODAY. That’s no coincidence. Dr. Pritt did her residency at the University of Vermont, where Dr. Elhosseiny was director of the residency program. After her fellowship at Mayo, Dr. Pritt visited the University of Vermont to give joint rounds with Dr. Elhosseiny, an experience that morphed into this first-time course.

In her opening remarks, Dr. Pritt called accurate and timely diagnosis of infectious diseases one of the most challenging tasks in medicine and noted it’s “typically the realm of the micro lab.” However, a pathologist’s ability to visualize an organism in tissue can contribute to a diagnosis when culture is not done, perhaps because there’s a lack of clinical suspicion of infection or the infecting agent is not culturable. In addition, Dr. Pritt said, even relatively rapid microbiological tests are time-consuming compared with looking at a slide, which in some instances can yield a diagnosis in minutes. Mycobacteria are an example. Then, too, anatomic pathology can offer valuable additional information, such as whether an organism appears to be colonizing a tissue or actually invading.

In the course, Drs. Pritt and Elhosseiny covered four classes of organisms: yeasts in tissue, filamentous bacteria, filamentous fungi, and worms in tissue, as well as one case in which an incidental finding of a protozoan was clinically significant.

In the first case, a 70-year-old man—a smoker with chronic obstructive pulmonary disease—had increasing shortness of breath. Chest x-ray showed lung opacity; a bronchoalveolar lavage specimen sent to cytopathology to look for lung cancer showed small oval structures inside mononuclear cells. These structures were positive on Gomori methenamine silver (GMS), a standard stain for fungi. Initially the suggestion was Histoplasma capsulatum; however, culture revealed Cryptococcus neoformans.

Dr. Pritt presented a three-step approach to identifying infectious organisms in tissue. First, list in your mind all potential organisms. In this case, where a yeast infection was suspected, there were many candidates, among them Candida albicans, C. glabrata, C. neoformans, H. capsulatum, Coccidioides immitis, and Pneumocystis jiroveci. Knowing the candidates, the pathologist can match characteristics of the yeasts to the case in hand. “Think about geography,” Dr. Pritt said. Histoplasma is found in the Mississippi and Ohio River valleys; C. immitis is found in the Southwest United States. Has your patient traveled to either of those places recently?

A further specification is that some yeasts, such as C. neoformans, are typically found in immunocompromised patients.

Know the appearance of each organism with various stains. H. capsulatum is practically invisible on H&E and shows a poorly staining cell wall on Gram stain. Despite its name, H. capsulatum does not have a true capsule. On Giemsa and H&E stains, H. capsulatum can show a faint halo, which is actually a “pseudocapsule.” This is demonstrated in Fig. 1A, where five yeasts of H. capsulatum are shown (arrows). Note the thin halo around each yeast, representing the poorly staining cell wall. However, when H. capsulatum yeast are stained with a cell wall stain such as GMS, the entire organism is stained black, with no residual halo or pseudocapsule (Fig. 1B,). This image also highlights the key morphologic features of H. capsulatum in tissue, including small (2 to 5 micron) uniform oval size and narrow-based budding.

C. neoformans is the only true encapsulated human yeast pathogen, Dr. Pritt emphasized. It has a polysaccharide capsule that often creates a large clear halo (arrows) on both H&E (Fig. 2A) and GMS (Fig. 2B). Occasionally, residual capsular material may also be seen.

When the organism is intracellular, it can also show an intracellular halo. In all stained preparations, the characteristic size variability and narrow-based budding can usually be identified.

Budding is an especially helpful feature for differentiating among similar appearing yeasts. For example, of the larger-sized yeasts, Blastomyces dermatitidis has broad-based budding, C. neoformans has narrow-based budding, and C. immitis has no budding. C. immitis forms what Dr. Pritt calls a “bag of marbles” appearance, where multiple endospores are contained within a mature spherule. Commonly, the whole bag of marbles may not be seen, but the “marbles” and often remnants of the “bag” are still present. Of course, deception appears in this criterion as well: Overlaying of separate structures, such as maturing spherules of C. immitis, can simulate budding. “Don’t be fooled into mistaking a bud,” Dr. Pritt cautions. A true budding yeast should have a well-demarcated attachment between the “mother” and “daughter” yeasts (Fig. 3).

After listing in your mind all potential organisms, the next step is to compare all the elements of their appearance against what you see on your slide—forms present, presence and type of budding, wall thickness, and hyphae/pseudohyphae.

Finally, use resources like special stains when appropriate. One stain that Dr. Pritt finds useful is mucicarmine for mucin, which can show the capsule of Cryptococcus and the cell wall of Blastomyces.

Also, Dr. Pritt advises, “Don’t be afraid to go to higher magnification. Use an oil immersion lens when dealing with microorganisms. They can be as small as 1 micron in diameter. To be definitive, you may need to examine the slide with a 100x objective and oil.” Even if you don’t use oil with your regular scope, she suggests dedicating one instrument in the lab to use whenever you see something you think might be an infectious agent. A microbiology consult may also be useful—there might be a culture against which to compare your presumptive identification.

Dr. Pritt concluded that there is a broad but defined group of yeasts that cause the majority of human infections; most can be differentiated by morphology and use of an objective ocular micrometer and special stains. A microbiology consult may be helpful in diagnosis. She shared an algorithm for distinguishing common yeast and yeast-like forms (Fig. 4).

In case 2, Dr. Elhosseiny described a 39-year-old woman with a history of Hodgkin’s lymphoma who had just completed extensive chemotherapy without a good clinical response. While awaiting bone marrow transplantation, she had signs of a serious infection, including respiratory failure and confusion. Chest x-ray and CT showed a pulmonary mass and diffuse smaller nodules in the lungs. Was this recurrence or infection? A fine-needle aspiration of the largest lung mass was urgently performed and examined in the radiology suite. With Giemsa staining under oil immersion, the paucicellular aspirate showed poorly staining, thin (1 micron) branched filamentous rods suggestive of a filamentous bacteria (Fig. 5).

Based on this morphologic appearance, a preliminary diagnosis was made and treatment was started. The following day, results from the special stains were available. Ziehl-Neelsen staining for acid-fast bacilli was negative, but a modified Kinyoun AFB stain (Fite stain) was positive (Fig. 6A). The bacteria were also nicely highlighted with GMS staining (Fig. 6B). Based on these staining patterns, a presumptive diagnosis of Nocardia was made.

Unfortunately, the woman died after three days due to the extent of her disease. Culture grew Nocardia spp. several days later. On autopsy, pulmonary sections showed diffuse lymphoma and abscesses consistent with recurrent tumor and infection. “In this case the key to diagnosis was a high index of suspicion,” Dr. Elhosseiny said, given that the organisms did not stain well with Giemsa. It was only with the appropriate suspicion and knowledge of this organism that the correct stain was ordered.

Differential diagnosis of filamentous fungi includes M. fortuitum (which is also weakly acid fast, filamentous, and occasionally branched); filamentous molds (which are much broader than filamentous bacteria); and Actinomyces (another filamentous bacteria that stains with GMS but is negative on modified AFB stain).

In summary, Dr. Elhosseiny said Nocardia is an opportunistic pathogen to which patients with alveolar proteinosis are particularly susceptible. Nocardia can have a mortality rate as high as 40 percent. It is often found in co-infections along with other pathogens in immunocompromised hosts. Culture confirmation is required but can take days to weeks; therefore, the ability of the pathologist to provide a rapid cytologic or histologic diagnosis is helpful in treatment decisions.

Turning to filamentous fungi, Dr. Pritt presented the case of a 45-year-old man with Hodgkin’s lymphoma that responded poorly to multiple chemotherapy regimens. He developed pneumonia with abscess after thoracic surgery; one month later he experienced shortness of breath, swelling, and erythema in the left chest wall. He was admitted with chest wall cellulitis and possible abscess formation. There was a black necrotic bulla in the center of the chest wall lesion. H&E staining of a skin punch biopsy was interpreted as probable Zygomycetes infection. Culture showed Rhizopus.

“The most common organisms I see in tissue are fungi,” Dr. Pritt said. Among the most dangerous fungal infections are those caused by the class Zygomycetes, called zygomycosis, especially its chief order Mucorales, which gives these infections their alternative name, mucormycosis. Zygomycetes are found widely in the environment and mostly infect immunocompromised hosts. Among the manifestations of zygomycosis the most common are rhinocerebral and pulmonary. “This organism can go from the sinuses into the eye and brain,” Dr. Pritt said. “You can’t diagnose it to genus level by histopathology alone; just call it zygomycosis.” What is most critical is that zygomycosis is a medical emergency, she said.

On slides, Zygomycetes show “classic” broad ribbon-like hyphae with collapsed forms—septations are rare. Vascular invasion is often present. Zygomycetes may not stain well with PAS or GMS but can usually be seen on H&E (Fig. 7).

Because zygomycosis is life-threatening and diagnosis affects therapy, it is important to differentiate it from other filamentous fungi. The easiest way to do this is by the broad irregular ribbon-like hyphae of Zygomycetes, as opposed to the narrow regularly septate hyphae of Aspergillus, Fusarium, and dematiaceous (pigmented) fungi.

Dr. Elhosseiny next showed a case of a 53-year-old man who underwent total hip replacement. Under the microscope, cyst-like structures containing coiled helminth (worm) larvae were seen in the synovial tissue and muscle. The spiral appearance of the worm, presence of a surrounding capsule, and location in muscle suggested the diagnosis of trichinosis (Fig. 8).

The major human reservoir for Trichinella spp. worldwide is swine; it can also be ingested in wild game meat such as bear. Trichinosis is not so common in the United States now that meat is often frozen before eating and increased regulations are placed on the pork industry. For diagnosis, Dr. Elhosseiny favors a muscle crush preparation, in which one can clearly see the three-dimensional coiled worms (Fig. 9).

Differential diagnosis of helminth larvae in tissue includes the dog heartworm, Dirofilaria, as well as Toxocara spp.

In the fifth case, a 57-year-old woman in otherwise good health had persistent diarrhea with nausea/vomiting. CT exam found a partially cystic mass in her pancreas. Ultrasound-guided endoscopic fine-needle aspiration of the mass was undertaken. Examination of the aspirate revealed Giardia. “To understand this case, you need to know how this type of endoscopy is done,” Dr. Elhosseiny said. “A knowledge of the procedure and anatomy is essential to correctly interpret the findings.” In this procedure, the endoscope passes through the duodenum to get to the pancreas. In the case at hand, while passing through the duodenum the endoscope picked up the Giardia organisms. So the main finding was pancreatic tumor, with giardiasis as an “incidental finding.” Treatment was affected, since the physician decided to treat the giardiasis before treating the tumor.

Several of the cautionary quizzes were striking. In one, PAS staining revealed filamentous-like structures in a brain biopsy. However, these were not fungi or bacteria, but simply blood vessels in necrotic tissue that were highlighted by the PAS stain. Keys included the overall pattern within the tissue, the presence of endothelial cells and red blood cells, and the size of the structures.

In a patient with colon cancer, a mesenteric lymph node contained darkly staining oval bodies on GMS (Fig. 10).

One might justifiably ask, Is this Histoplasma or Cryptococcus? In fact, it was neither. Rather they were Hamazaki-Wesenberg (H-W) bodies, which are found extracellularly in lymph node sinusoids. Investigation has shown that H-W bodies are oval giant lysosomes. They are positive on AFB stain and GMS and can appear to “bud.” In one study, H-W bodies were found in almost one-fourth of 359 superficial and deep lymph nodes from 46 necropsies, and their frequency was not related to age or disease (Tudway AJ. J Clin Pathol. 1979;32:52–55).

Equally startling was the case of a 50-year-old woman with a one-year history of bilateral panuveitis nonresponsive to steroid therapy. Vitreous fluid was submitted to “R/O lymphoproliferative disorder.” In the fluid were small, blue fusiform rods on Giemsa at 60× and 100× oil immersion (Fig. 11).

Were they bacteria? Toxoplasma? Fusarium? Again, they were not living organisms at all, but melanosomes released from the degenerating choroid layer and phagocytized by macrophages. They could be easily distinguished from microorganisms when compared directly to grampositive bacilli and Toxoplasma tachyzoites.