—  SPECIALTY CONFERENCE  —

Infectious Disease Pathology

Case 1 - Pulmonary Cryptococcosis

James Kelly
Royal Jubilee Hospital
Victoria, BC


Click on each slide thumbnail image for an enlarged view
Clinical History
A 35 year old woman, 20 pack-year smoker, resident of Comox, Vancouver Island, presented with pleuritic chest pain. The pain came and went in multiple areas on both sides of the chest. There was no cough, hemoptysis or constitutional symptom and she felt well. Chest X-ray and CT showed multiple nodules in both lung fields resembling "metastases, infection or, less likely, infarcts" but no hilar or mediastinal adenopathy. An attempt at needle biopsy failed and a bronchoscopy with washings and transbronchial biopsies was negative. A video-assisted thoracoscopic biopsy yielded a wedge of lung containing a 2.3 cm nodule which had a yellowish cut surface. The surgeon divided the tissue sending some to histopathology and some to microbiology.


Case 1 - Figure 1 - Wedge resection of lung (low power), showing a necrotizing granuloma.

Case 1 - Figure 2 - Wedge resection of lung (intermedicate power), with yeast cells visible in the areas of necrotizing granuloma


Case 1 - Figure 3 - Wedge resection of lung (high power), encapsulated yeast in area of necrosis. The fungi are round or elliptical with thin eosinophilic walls and a surrounding capsule or fuzzy zone and large clear halos.

Case 1 - Figure 4 - Mucicarmine stain - Wedge resection of lung, stain shows the mucinous capsules surrounding narrow-necked budding yeast consistent with Cryptococcus.

Slide description
The section shows a necrotizing granuloma in which fungal yeast cells are readily visible. The fungi are round or elliptical, with thin eosinophilic walls, a surrounding capsule or fuzzy zone and large clear halos. Mucicarmine staining confirmed the mucinous capsules. These features are diagnostic of Cryptococcus. The granuloma shows palisaded histiocytes around the necrotic zone and organisms are present both within the macrophages and free in the necrotic zone. Outside the granuloma are abundant foamy histiocytes, lymphocytes and plasma cells. The necrotic zone retains the alveolar architecture and is not "caseous" or structureless.

Comment: This is a typical case of cryptococcal granuloma. I choose this case to illustrate the foamy macrophages because they can be so abundant as to be mistaken for clear cell carcinoma at frozen section. In our series (see below) foamy histiocytes were present in 11 of the 17 open lung biopsies and they were abundant in five cases (29%). When not abundant, foamy macrophages formed part of the palisaded granulomatous lining which was composed of eosinophilic epithelioid histiocytes in most cases.

Followup
Antifungal treatment resulted in rapid shrinkage of the pulmonary nodules and return to normality.

Histopathology of Pulmonary Cryptococcosis
The diagnosis of pulmonary cryptococcosis is made either by morphologic recognition of the organisms or by culture. FNA of lung is an effective diagnostic tool for cryptococcus. [1, 2] A variety of other specimen types are also used – endobronchial or transbronchial biopsies, open lung biopsies or bronchoalveolar washings. The host reaction depends on the immunologic competence of the host and the presence of a capsule around the organism. There may be no reaction at all in anergic subjects so that AIDS patients may show vast numbers of cryptococci in the interstitium with little or no inflammation. [3, 4] In immunocompetent patients most reports concern single cases and it has not been possible to study the time-course of the histological reaction to C. neoformans. A mixed suppurative and granulomatous reaction or a pure granulomatous reaction with varying necrosis is described. [3] Primary pulmonary cryptococcosis may be associated with granulomas in hilar lymph nodes. [5] Non-encapsulated strains elicit an initial strong suppurative response and later a suppurative granulomatous reaction.

The largest series reported to date, 36 patients, is an autopsy series from Johns Hopkins in the period 1936 to 1983, which is by its nature highly selected for severity, although not all deaths were due to the fungus. [6] The patients ranged in age from 2 to 89 years (mean, 49 years); all but three patients had underlying debilitating diseases and 23 patients had received steroids and/or chemotherapy. In 25 patients (69 per cent) cryptococcosis was a major factor contributing to death, through pulmonary disease in ten, systemic involvement in seven, and central nervous system disease in eight. In 15 patients (42 per cent) Cryptococcosis was diagnosed clinically.

Four morphologic patterns were described:

1.One or more peripheral pulmonary granulomas in seven patients (19 per cent).
2.Granulomatous pneumonia in nineteen patients (53 per cent), with intra-alveolar proliferating organisms and varying degrees of inflammatory response, which, when present, ranged from acute inflammation to diffuse intra-alveolar granulomas with giant cells.
3.In seven patients (19 per cent) organisms were present diffusely within alveolar capillaries and interstitial tissues, and reactions ranged from little or no inflammation with numerous organisms to few organisms with miliary granulomas.
4.In three patients (8 per cent) both intra-alveolar and intravascular organisms were present in massive numbers, and the primary route of infection was uncertain. The prominence of vascular invasion in these last two categories is of interest because of the tendency for Cryptococcus to disseminate to the meninges even in immunocompetent hosts.



A recent series of 11 cases of pulmonary cryptococcosis (7 immunocompromised, 4 immunocompetent) found that pulmonary nodules, either solitary or multiple, were the most common CT finding, present in 10/11 cases. [7] Discrete zones of ground glass attenuation on CT were seen surrounding nodules or adjacent to nodules and corresponded to airspace accumulation of foamy macrophages and edema fluid. [7]

Differential diagnosis

1.Coccidioidomycosis is a granulomatous inflammatory process caused by a dimorphic spherical yeast measuring 30-100 micrometres with a "double-walled" capsule within which are numerous endospores. Yeasts are seen in giant cells and necrotic tissue, and stain with GMS and PAS but not with mucicarmine.
2.Blastomycosis is caused by a dimorphic spherical yeast measuring 6-15 micrometres wide with a thick capsule and characteristic broad-based budding creating a dumbbell shape. The yeast is seen in giant cells and necrotic tissue, stains with GMS and PAS but negative or only weakly positive with mucicarmine. Mucosal surfaces may show pseudoepitheliomatous hyperplasia of the surface epithelium in reaction to the infection.
3.Histoplasmosis is a dimorphic oval-shaped, yeast measuring 3-5 micrometres in diameter with a single nucleus that buds via a narrow base. It has a surrounding clear space, is typically seen in clusters within macrophages and stains with GMS and PAS but not with mucicarmine.

The Recent Emergence of C. neoformans var gattii on Vancouver Island
There are two varieties of Cryptococcus, C. neoformans var. neoformans (CNVN) and C. neoformans var gattii (CNVG), and four serotypes (A, B, C, D). CNVN may be serotype A, D or AD whereas CNVG is serotype B or C. A new endemic focus of CNVG serotype B infection on Vancouver Island began around early 1999 and continues today. Previously, CNVGhas been geographically restricted to certain tropical and subtropical regions. It has been well-studied in Australia and Papua New Guinea where it more commonly causes disease in immunocompetent individuals than does CNVN. In Australia it is associated with eucalyptus trees. The strain of CNVG on Vancouver Island is similar to one of the Australian strains by DNA studies. The newly endemic region affects primarily the eastern coastal region of Vancouver Island, which includes the main populated areas on the island.

Seventy five human cases were confirmed from January 1999 up to July 2003. The rate in humans is approximately 2.7 cases per 100,000 population per annum, which is five times less than the rate of serious injury in motor vehicle accidents, but about 27 times the incidence rate observed in endemic areas in Australia. Many patients are smokers but a case-control study has not shown a statistical increase in risk attributable to smoking. Two thirds of patients have no evidence of underlying immunodeficiency, but about 25% were taking corticosteroids in doses exceeding 20 mg per day and altogether, about a third of patients have some form of immunodeficiency. The human cases to date have been predominantly respiratory but one case in four has meningitis. There have been fatalities.

The cause of the outbreak is unknown. Could it have been introduced by an infected exotic pet or plant? Is climatic change involved? It is unlikely we will ever know how the infection arrived here. Climatically, there has been a mean rise of winter temperatures at Nanaimo of two degrees over the past fifty years.

Targeted sampling of trees, air and soil from the vicinity of where affected patients live, has identified the organisms. Once an initial positive tree is identified, air and soil samples are taken outward in concentric circles. Positive trees have been identified from Victoria to Parksville. The positive trees are Douglas fir, Garry oak, Alder and Maple. Arbutus trees were first suspected as they are relatives of Eucalyptus trees which were implicated in former outbreaks of C. neoformans var gattii in Australia and Africa but these trees were initially negative although some have given positive samples since 2002. Known positive trees continue to be positive although not all replicate samples are positive in winter. Air concentrations are lower in the winter and begin to increase in April. The highest air counts have been at Rathtrevor beach in summer. Soil sample positives may decrease as air sample levels increase. Soil samples around an index tree may be high near and far from the tree. In Duncan and Victoria, certain trees have been positive for both serotypes A and B since September 2002.

The present outbreak affects animals, both domestic and wild, terrestrial and aquatic. Six Doll's porpoises that were found dead have proven positive over the past four years and 4 of these were found in the Strait of Georgia. Domestic cats and dogs, ferrets and other animals including such exotics as a tapir have also been affected. Animals present with head or neck nodules, lymphadenopathy, nasal and pneumonic signs. Diagnosis is made by cytology mainly. Some of the animals are being followed with Cryptococcus antigen titre during treatment with fluconazole. Like the humans, half of the animals had predisposing conditions that may have lowered their immunity. Many of the animals had traveled to Central Island (Parksville-Namaimo) but not all did.

Histopathology of the recent Vancouver Island cases
Before the present outbreak cryptococcal lung infection was rarely seen on Vancouver Island. In our histopathology files from 1999 to the present we have had 45 cases, 19 lung biopsies (17 wedge, two transbronchial), 2 autopsies and 24 cytologic specimens, (FNAs, sputa or bronchial washings). One autopsy was a fatal meningitis; the other was a cryptococcal pneumonia in a cirrhotic woman.

Lung biopsies were removed by thoracoscopic-assisted biopsy. Most were submitted for rapid diagnosis. They consisted of lung wedges containing one or more firm nodules with a brown or yellow, often mucoid appearance on cut surface. In many instances the diagnosis was established by examination of a scrape smear or imprint from the cut surface. In most cases cultures were not successful from these cases.

Histologically, all of the wedge-biopsied nodules showed a central necrotic zone or zones in which the ghosted outlines of alveoli were discernible. Organisms were usually readily visible in the necrotic zone, but not always numerous. The necrotic zone was delineated by macrophages, sometimes palisaded and eosinophilic but sometimes foamy. The organisms were usually confined to the necrotic zone. Most cases showed a palisade of eosinophilic macrophages surrounding the necrotic area, and a further thin layer containing a mixture of lymphocytes, plasma cells and fibroblasts. As mentioned above, foamy histiocytes were present in 11 of the 17 open lung biopsies and they were abundant in five cases (29%) forming collections outside the granuloma and filling the air spaces. At frozen section, the foamy macrophages were misdiagnosed once as clear cell carcinoma but raised the question of a clear cell carcinoma in the mind of the pathologist on other occasions particularly if the necrotic zone were not included in the section.

One case showed several tiny foci of necrosis and a mixed response of macrophages, eosinophils, lymphocytes and plasma cells. The fungi in that case were not localized to the centre of the lesion but scattered throughout the macrophages. One case of Cryptococcus infection occurred coincidentally with a bronchioloalveolar carcinoma.

On immunostaining with monoclonal antibodies provided by Dr Thomas Kozel, [8] seventeen histologically identified cases were confirmed as var. gattii at the time of writing.

Conclusion
There is a new epidemic focus of CNVG on Vancouver Island that presents as pneumonitis or meningitis, sometimes after a very short exposure during a visit to the island. Physicians in North America, and indeed worldwide, need to be aware of this risk in people who have traveled to the island.

References

  1. Silverman,J.F. & Johnsrude,I.S. Fine needle aspiration cytology of granulomatous cryptococcosis of the lung. Acta Cytol. 29, 157-161 (1985).
  2. Hsu,C.Y. Cytologic diagnosis of pulmonary cryptococcosis in immunocompetent hosts. Acta Cytol. 37, 667-672 (1993).
  3. Chandler,F.W. & Watts,J.C. Pathology of Infectious Diseases. Connor,D.H., Chandler,F.W., Schwartz,D.A., Manz,H.J. & Lack,E.E. (eds.), pp. 989-997 (Appleton and Lange, Stamford, CT,1997).
  4. Khardori,N., Butt,F. & Rolston,K.V. Pulmonary cryptococcosis in AIDS. Chest 93, 1319-1320 (1988).
  5. Salyer,W.R., Salyer,D.C. & BAKER,R.D. Primary complex of Cryptococcus and pulmonary lymph nodes. J Infect. Dis. 130, 74-77 (1974).
  6. McDonnell,J.M. & Hutchins,G.M. Pulmonary cryptococcosis. Hum. Pathol 16, 121-128 (1985).
  7. Zinck,S.E., Leung,A.N., Frost,M., Berry,G.J. & Muller,N.L. Pulmonary cryptococcosis: CT and pathologic findings. J Comput. Assist. Tomogr. 26, 330-334 (2002).
  8. Krockenberger,M.B. et al. An immunohistochemical method that differentiates Cryptococcus neoformans varieties and serotypes in formalin-fixed paraffin-embedded tissues. Med Mycol. 39, 523-533 (2001).