Ings

ROCHESTER,MINNESOTA

Vol. 65

FEBRUARY 1990

Clinical Implications of the Histopathologic Diagnosis of Pulmonary Lymphomatoid Granulomatosis

RICHARD J. PISANI, M.D., RICHARD A. DeREMEE, M.D., Division of Thoracic Diseases and Internal Medicine

We reviewed the epidemiologic, laboratory, roentgenographic, pulmonary function, and survival data from 28 patients who had a histologic diagnosis of lymphomatoid granulomatosis (LG) with involvement of the lungs. The mean age at the time of diagnosis was 51 years, and the male-to-female ratio was 3:2. Ten patients had other underlying diseases before LG was diagnosed. The most prominent symptoms were cough, dyspnea, fever, and rash, which were usually present for several months before diagnosis ofLG. Multiple nodules were detected on a chest roentgenogram in 68% of the patients. Immunoglobulin concentrations were abnormal in 8 of 12 patients studied. Although bronchoscopy established the diagnosis in approximately a third of the patients who underwent this procedure, open-lung biopsy was uniformly diagnostic. The median survival was 72 months, with follow-up through 12 years. In 11 patients, the original diagnosis ofLG was eventually changed to lymphoma. In five of these patients, the change in diagnosis was based on immunohistologic data obtained shortly after LG was discovered. Lymphoma diagnosed in this way was associated with a better prognosis than lymphoma diagnosed on the basis of conventional histopathologic findings. In three patients, solid tumors eventually developed. The diversity of clinical outcomes and frequent revisions of the diagnosis led us to consider the possibility that LG may also represent a histopathologic finding that occurs transiently in several disease processes.

Lymphomatoid granulomatosis (LG) was described by Liebow and associates! as a peculiar angiocentric and angiodestructive granulomatosis with large necrotizing nodules. Although

Individual reprints of this article are not available. Mayo Clin Proc 65:151-163, 1990

the lung is most frequently involved, LG is also commonly found in the upper respiratory tract, skin, kidneys, and central nervous system. Pulmonary infiltrates are histologically polymorphous and show variable degrees of cytologic atypia in the lymphoid cells. Vascular infiltration is characteristic and has led to the recent use of the term "angiocentric immunoprolifera151

152

PULMONARY LYMPHOMATOID GRANULOMATOSIS

tive lesions." Infrequent involvement of the hone marrow, spleen, and peripheral lymph nodes initially supported the distinction of LG from lymphoma. DeRemee and colleagues" have shown that the histopathologic changes are identical to those of a similar disorder, polymorphic reticulosis, which was first detected in the upper respiratory tract. Unfortunately, the pathologic distinction ofLG from lymphoma has been difficult, both conceptually and morphologically. 4 The advent of sophisticated immunohistologic and molecular biologic techniques has only partially clarified the histogenesis of these lesions and has suggested that they are T-cell proliferations, some of which have histologic and immunologic features of T-cell lymphomas." Numerous studies l ,3,6-9 have reviewed the clinical features of LG. Although the initial symptoms, organ involvement, chest roentgenographic findings, and laboratory data have been fairly consistent, the natural history of LG and its response to treatment remain variable from patient to patient and from study to study. Its frequent association with lymphoma has created controversy about whether it is actually a variant ofT-cell lymphoma. The fact that some patients have had complete responses or longterm remissions without therapy, however, makes it difficult to classify all cases of LG as "malignant." This clinical variability raises questions about the value of treating LG as a distinct clinical entity. Most of the pathologic features of LG may be seen in several lymphoproliferative disorders that affect the lung, including primary and secondary T- and B-cell non-Hodgkin's lymphomas, Hodgkin's disease, angioimmunoblastic lymphadenopathy (probably a T-cell lymphoma itself), and lymphoproliferations associated with immunodeficiencyfor example, acquired immunodeficiency syndrome (AIDS), the posttransplantation state, and Epstein-Barr virus. In this report, we review 28 consecutive cases of pulmonary LG from a single institution; the diagnosis was based on chest x-ray involvement and biopsy findings consistent with those initially described by Liebow and co-workers.'

Mayo Clin Proc, February 1990, Vo165

Patients who were eventually found to have alternative diagnoses were not excluded from the study because we wished to highlight the clinical and histologic heterogeneity of LG and address problems arising in management.

MATERIAL AND METHODS In a computer search of Mayo Clinic records for the period between 1977 and 1987, we identified 28 consecutive cases in which LG had been diagnosed on the basis of at least one pulmonary or extrapulmonary biopsy specimen. The diagnosis of LG was based on the following histopathologic findings: (1) a polymorphous lymphoid cell infiltrate with or without atypical cells and (2) angioinvasion. All available pathologic specimens were reviewed by one pathologist. Only patients with pulmonary involvement were selected for review; the study group included patients with compatible radiographic evidence of lung involvement if a biopsy specimen from another site was interpreted as LG (N = 4). One patient was excluded from analysis because, 1 month after a nasopharyngeal biopsy specimen demonstrated LG, an open-lung biopsy revealed Pneumocystis pneumonia without evidence of LG. The following clinical findings were assessed: specific symptoms and their duration, findings on physical examination, chest x-ray features, arterial blood gases, laboratory tests, pulmonary function tests, results of additional and subsequent biopsies, and survival data. Followup data were obtained from Mayo Clinic records and from personal contact with patients or their physicians (or both). Biopsy specimens from five patients whose diagnosis was made at the Mayo Clinic after 1983 were immunophenotyped for T- and B-cell markers in addition to routine histopathologic staining; on the basis ofthese studies, they were reclassified as T-cell lymphomas. These patients had minimal or no cellular staining with antibodies directed against B-cell surface markers (CD20 or CD22) and yet expressed T-cell (CD5) and helper T-cell (CD4) phenotype. A detailed description of the immunohistologic criteria used has been reported previously."

PULMONARY LYMPHOMATOID GRANULOMATOSIS

Mayo Clin Proc, February 1990, Vol 65

Survival data were plotted by using KaplanMeier estimates of the survival curve. Patients were divided into two groups: (1) those who eventually demonstrated evidence of a malignant lesion on conventional histopathologic evaluation and (2) those who had no obvious malignant involvement during the course of follow-up. The survival curves for these two groups were compared by using the log-rank test.

RESULTS

Clinical Features.-The clinical characteris-

tics of our patient population are shown in Table 1. The male-to-female ratio was 3:2. The mean age of the patients at the time of diagnosis ofLG was 51 years, but the ages varied considerably. Ten of our patients had 13 other underlying disease processes before LG was diagnosed (Table 2). The initial symptoms and their duration are listed in Table 3. Most patients complained of some combination of cough, dyspnea, fever, rash, anorexia, and weight loss for several months before eventually being diagnosed as having LG. In some patients, symptoms had been present for more than 2 years. Symptoms related to involvement of the nervous system were usually those of neuropathy. The findings on physical examination are summarized in Table 4. Despite the fact that all patients had lung involvement, clinical examination of the lungs revealed normal findings in two-thirds of the patients. The most frequent physical finding was rash. Several different skin lesions were noted in our patients with LG. Although many patients Table I.-Profile of Study Patients Diagnosed as Having Lymphomatoid Granulomatosis No. of patients Male Female Age (yr) of patients Mean± SD Range Smokers (mean pack-yr) Current Former

17 11

28

50.9 ± 14.5 24-79 23.7 9 6

153

Table 2.-Associated Disease Processes in 10 of 28 Study Patients Diagnosed as Having Lymphomatoid Granulomatosis Disease Autoimmune or infectious Crohn's disease Herpes ophthalmitis Juvenile rheumatoid arthritis Pancreatitis Resolving hepatitis B Hashimoto's thyroiditis Hematologic or lymphoid Myeloproliferative disorder Hodgkin's disease Plasmacytoma Thymoma

No. of cases 2 2 1 1 1 1

2 1

1

1

had erythematous maculopapular eruptions (N = 7), subcutaneous nodules (N = 5) and ulcers (N = 3) were also seen. The extremities were the most frequent site of involvement, but several patients had involvement of the trunk and head and neck areas. Most lesions were not painful or pruritic. Approximately a third of the patients had abnormalities on ear, nose, and throat examination. In six patients, physical examination disclosed evidence of involvement of the nervous system, including ataxia, visual field defects, trigeminal neuralgia, Bell's palsy, mononeuritis multiplex, and peripheral neuropathy. Laboratory Data.-Chest roentgenograms showed multiple nodules in 19 of the 28 patients (68%), and 12 patients (43%) had some degree of diffuse interstitial infiltrates (Table 5). Pulmonary function tests were performed in 15 patients, 4 of whom had normal findings. The most common abnormality noted was a reduction in lung volume consistent with restrictive disease (N 6). No one pattern was clearly associated with LG; obstruction was seen in three patients, and a mixed picture was noted in two. The arterial blood gases demonstrated similar variability; the mean values (± SD) were as follows: Pa0 2 , 66.5 ± 13.0 mm Hg; PaC0 2 , 33.6 ± 3.1 mm Hg; pH, 7.45 ± 0.04; and HC0 3 - , 21.2 ± 4.3 meq/liter. Most patients had an increased alveolar-arterial oxygen gradient and chronic respiratory alkalosis.

=

154

PULMONARY LYMPHOMATOID GRANULOMATOSIS

Table 3.-Initial Symptoms of Study Patients Diagnosed as Having Lymphomatoid Granulomatosis* Symptom

No. of patients

Cough Dyspnea Fever Rash Weight loss Malaise Arthritis or arthralgia Weakness or neuropathy Sinus or otorhinolaryngologic problem Peripheral adenopathy

Duration (mo) Mean Range

15 15 14 14 12 11 5

3.9 4.7 2.9 4.6 NA 4.3 NA

4

NA

4 1

NA NA

1-24 1-36 1-9 1-24 1-12

*NA = not available.

Laboratory data were not particularly helpful in establishing the diagnosis of LG. The mean erythrocyte sedimentation rate was increased (40.6 ± 39.0 mm in 1 hour), but the range of values was broad (4 to 144 mm in 1 hour). One finding that may be helpful in suggesting the diagnosis is an increased relative monocyte count. Eight patients had monocyte counts (reported on automated complete blood cell counts) greater than or equal to 12%. Concentrations of immunoglobulins A, G, and M were measured in 12 patients. IgA was decreased in two patients and increased in three. IgG was decreased in five patients and increased in three (the same patients who had elevated IgA). IgM was increased in two patients, one of whom had normal IgG and IgA and the other of whom had decreased IgA and IgG. Only 4 of 12

Table 4.-0rgan System Involvement (Detected Clinically or Radiographically) in 28 Study Patients Diagnosed as Having Lymphomatoid Granulomatosis Area of involvement Lung Skin Ear, nose, and throat Central nervous system Lymphatic system (adenopathy)

Patients No. % 28 15 8 6 3

100 54 29 21 11

Mayo

cue Proc, February 1990, Vol 65

patients studied had entirely normal immunoglobulin concentrations. Serum protein electrophoresis, which was done in 18 patients, revealed only 1 (6%) with monoclonal protein. This same patient had increased IgM associated with decreased IgG and IgA.

Histopathologic Diagnosis.-Fourteen

patients underwent bronchoscopy. Endobronchial abnormalities were noted in only two patients, and bronchial biopsy specimens were positive in one of these two. Transbronchiallung biopsy was done in 11 patients and was diagnostic ofLG in 3. Therefore, the diagnostic yield of bronchoscopy was 29% (4 of 14). Most diagnoses were made by open-lung biopsy, although ear, nose, and throat, skin, and brain biopsies established or confirmed involvement ofLG in a high percentage of the cases in which they were used (Table 6). Bone marrow biopsy was positive in only 1 of 19 cases. The characteristic histopathologic features oflow-grade and high-grade lesions of LG are shown in Figures 1 and 2, respectively. Treatment and Clinical Course.-Once LG had been diagnosed, the treatment was variable, depending on the degree of cytologic atypia, extent of involvement, severity of the illness, and clinical course. Most commonly, patients received either prednisone and cyclophosphamide (N = 7) or multiagent cytotoxic chemotherapy (N = 13). Most frequently, the chemotherapy consisted of doxorubicin (25 mg/m"), bleomycin (2 mg/m"), vinblastine (6 mg/m"), and dacarbazine (250 to 350 mg/m"). Two patients received no treatment. Letendre.'? recently summarized our institution's experience with aggres-

Table 5.-Findings on Chest Roentgenograms in 28 Study Patients Diagnosed as Having Lymphomatoid Granulomatosis Patients Finding Multiple nodules Diffuse infiltrate Localized infiltrate or mass Pleural effusion

No.

%

19 12 5 4

68 43 18 14

Mayo CUn Proc, February 1990, Vol 65

PULMONARY LYMPHOMATOID GRANULOMATOSIS

Table 6.-Biopsy Findings in Study Patients Diagnosed as Having Lymphomatoid Granulomatosis (LG) Biopsy site Bronchial Transbronchial Open lung Ear, nose, or throat Skin Central nervous system Bone marrow

Total no.

2 11 20 7

18 1 19

Positive for LG* No. % 1 3 20 6 8 1 'I

50 27

100 86 44 100 5

*See text for diagnostic histopathologic features.

sive cytotoxic chemotherapy for lymphomatoid granulomatosis. Because we thought that conventional histologic evidence of a malignant process at any time during a patient's course would predict a poor outcome, we divided patients into two groups for analysis of survival. Group 1 consisted of nine patients who, at some time after the pathologic diagnosis ofLG, were found to have evidence of a malignant lesion (lymphoma in six and solid tumor in three) based on conventional histopathologic studies. Group 2 consisted of 19 patients whose only diagnosis was LG; included were those who had either a conflicting diagnosis of nonmalignant disease or a diagnosis of

155

"lymphoma" based on immunohistologic findings. Survival data for each group are shown in Figure 3. Although no survival difference was noted through the first 6 months, the likelihood of survival was significantly less for group 1 after 6 months (P

Clinical implications of the histopathologic diagnosis of pulmonary lymphomatoid granulomatosis.

We reviewed the epidemiologic, laboratory, roentgenographic, pulmonary function, and survival data from 28 patients who had a histologic diagnosis of ...
3MB Sizes 0 Downloads 0 Views