Context.—Spontaneous pneumothorax can be idiopathic (primary), or it can occur in association with an underlying predisposing condition (secondary). Spontaneous pneumothorax may be a harbinger of an undiagnosed clinical condition, which may be associated with serious systemic abnormalities, making early recognition and diagnosis important. The pulmonary pathology of some of these disorders has not been fully elucidated.
Objective.—To review cases of pneumothorax in the hope of identifying pathologic features that might correlate to specific clinical syndromes.
Design.—The pathology computer files at 3 hospitals were searched for all cases of spontaneous pneumothorax, primary and secondary, regardless of etiology during a 11-year period. Ninety-two cases were retrieved. Each of the cases was evaluated for reactive eosinophilic pleuritis, elastosis, pleural fibrosis, emphysema, intra-alveolar macrophages, cholesterol clefts, vasculopathy, and intraparenchymal or intrapleural cysts. Clinical information regarding asthma and smoking history, site of the pneumothorax, family history, radiographic findings, predisposing conditions, recurrence, age, and sex were extracted from the medical records.
Results.—In 11 patients (12% of all the patients with spontaneous pneumothorax), a distinctive pattern of pleural fibrosis with islands of fibroblastic foci within a myxoid stroma was noted at the pleural-parenchymal interface or leading edge. These lesions correlated with a select subset of patients, consisting predominantly of young men.
Conclusions.—Our review identified a distinct pattern of pneumothorax–associated fibroblastic lesions in a subset of cases of spontaneous pneumothorax. Whether this is related to the pathogenesis of the pneumothorax remains to be elucidated.
Spontaneous pneumothorax (SP) is a well-recognized pulmonary complication of a wide variety of disorders, ranging from endometriosis to metastatic malignancy. It has been separated into primary and secondary types, where the secondary cases are associated with a clear etiology. In primary or idiopathic cases of SP, a clear etiology is not apparent.1 Pathologic examination of cases is important because SP may be the initial presentation in a variety of conditions.2–4 Some of those diseases are well described and have clearly recognizable pathologic features, such as lymphangioleiomyomatosis or malignancy. Others are just recently being recognized, both clinically and pathologically, for example Birt-Hogg-Dubé syndrome.5,6 For other pneumothorax-associated syndromes, clearly identifiable pathologic features have been elusive, despite careful analysis. For example, a recent article evaluating known cases of Marfan syndrome7 describes distal acinar (paraseptal) emphysema as a feature of Marfan syndrome. However, that is also a feature believed to represent a nonspecific reaction to a pneumothorax.7,8 Identifying the underlying disorder may carry important clinical ramifications, exemplified by the development of renal tumors in Birt-Hogg-Dubé syndrome and aortic root dissection in patients with Marfan syndrome. With this in mind, we conducted a multicenter, 10-year, retrospective review of all cases of SP with the hope of identifying and correlating histopathologic patterns with specific clinical entities. During this review, we identified a clinicopathologic entity recognizable based on its histologic features. Early reports of SP8 seem to include a subset of similar cases that were identified based on clinical features, not histopathology. This report describes the histologic entity, which we have termed pneumothorax-associated fibroblastic lesion (PAFL).
MATERIALS AND METHODS
Cases of SP were retrieved from the pathology files of 3 institutions (St Agnes Hospital, Baltimore, Maryland; Northwest Hospital, Randallstown, Maryland; and Sinai Hospital, Baltimore) for a 11-year period (2001–2011). The pathology computer files were searched for cases with the term spontaneous pneumothorax listed anywhere in the report. All cases of SP were reviewed, regardless of clinical impression of etiology (primary or secondary). Using as a guideline the morphologic features found in previous reviews of SP,8–12 each case was evaluated for reactive eosinophilic pleuritis, elastosis, pleural fibrosis, emphysema, intra-alveolar macrophages, cholesterol clefts, vasculopathy, intraparenchymal or intrapleural cysts, and underlying disorders, such as lymphangioleiomyomatosis, infection, malignancy, or endometriosis. In addition, the presence or absence of fibroblastic foci, a feature not identified in prior reviews of cases of pneumothorax, was added. The following clinical data were also extracted from the medical records when available: asthma and smoking history, site of the pneumothorax, family history, radiographic findings, predisposing conditions, recurrence, age, and sex of the patient. Elastic stains (Verhoeff-van Gieson) were performed on 5 cases. Ninety-two cases were retrieved (the results from this large review will be reported elsewhere). The studies were performed with the approval of institutional review boards of each of the hospitals.
RESULTS
Eleven patients (12% of all patients with SP in the study) were found to have a distinctive pattern of pleural fibrosis with islands of fibroblastic foci within a myxoid stroma at the pleural-parenchymal interface or leading edge. Often, these lesions exhibited a wedge-shaped configuration, with the broad base at the pleural surface and the apex toward the lung parenchyma. In one patient (9%), serial lung biopsies from separate episodes of pneumothorax were available for review (biopsies 2 and 8 in Table 1).
Clinical Characteristics of Spontaneous Pneumothorax With Pneumothorax-Associated Fibroblastic Lesion

The clinical and radiologic findings of our patients are described in Table 1. Almost all were male. Ten of the patients (91%) were younger than 25 years old (range, 16–24); one (9%) was 53. The average age in the larger, 92-patient group was 54, and the average in the PAFL subset was 20. In comparison to the main study group, this age difference is statistically significant (P < .05). The 53-year-old patient was a nonsmoker. Of the others, 4 (36%) were nonsmokers, 5 (45%) were smokers, and the smoking status was unknown for 1 patient (9%). Three (27%) had a history of asthma; 3 (27%) did not. The asthma history of the others (5 of 11; 45%) was unknown. In all patients (100%), the upper lobes or apex was involved by the pneumothorax. One (9%) also had lower lobe involvement.
On gross examination, 4 of the resected lung specimens (36%) had recognizable blebs or bullae. The largest bulla was in a 16-year-old, nonsmoking patient and measured 9 cm at greatest dimension. Other bullae/blebs measured up to 2 cm in greatest dimension. No masses were identified.
Histopathologic review revealed a distinctive lesion with a zonal pattern characterized by dense, collagenous fibrosis at the pleural-subpleural surface and younger, more-active, fibroblastic-like foci at the pleural-parenchymal interface (Figure 1, A through C). The fibroblastic-like foci are key to recognition of the entity. The fibroblastic foci were scattered along the leading edge (Figure 1, D). Less fibrotic, subpleural lesions with a “stuck-on” appearance could be identified (Figure 2, A). The lesions were multifocal and extended varying lengths along the pleural surface, separated by intervening normal pleura (Figure 1, A). As indicated above, the lesions often assumed a triangular or pyramidal configuration that was quite striking on low power. They sometimes extended down pleural septa but did not seem to disproportionately involve the septa. The amount of subpleural surface involved varied by patient, and the lesions ranged in size up to one ×10 field, along the inner pleural surface. The zonal pattern of dense, peripheral fibrosis with more immature, reactive changes toward the parenchymal surface was suggestive of temporal heterogeneity. The band of fibrosis in closest proximity to the pulmonary parenchyma often showed a parallel nuclear orientation of both the collagen bands and nuclei to the border with the adjacent parenchyma (Figure 2, B). In some patients, the fibrosis extended superficially into the lung parenchyma, tracking along the alveolar or interlobular septa. These areas of fibrosis typically extended only a few millimeters into the lung and were never associated with more-extensive interstitial pulmonary fibrosis. Intrapleural cysts up to 5 mm were noted in the dense pleural fibrosis. The cysts were lined by reactive hobnail cells, representing either hyperplastic type 2 pneumocytes or mesothelial cells and were, therefore, different from blebs (intrapleural, unlined collections of air). Occasional blebs were also noted. In some of these cysts, the wall showed the same reactive, fibroblastic appearance as the leading edges of the lesion showed (Figure 2, C). Vascular proliferation was noted in the dense pleural and subpleural fibrosis. The adjacent alveoli ranged from normal to showing mild, type-2 pneumocyte hyperplasia. Elastic stains demonstrated disruption of the normal pleural elastic tissue (Figure 2, D). This fibrotic, zonal lesion was absent in the other cases (n = 81) reviewed for the larger study, which served as a random, unbiased internal control.
Histopathologic features of pneumothorax-associated fibroblastic lesion (PAFL). A, Wedge-shaped area of zonation demonstrated by peripheral collagenous fibrosis and intervening normal pleura (brackets indicate 1 PAFL). B and C, Higher-power view illustrating the zonation and fibroblastic foci (arrowheads) at the leading edge. D, Fibroblastic focus (hematoxylin-eosin, original magnifications ×10 [A], ×20 [B and C], ×40 [D].
Histopathologic features of pneumothorax-associated fibroblastic lesion (PAFL). A, Wedge-shaped area of zonation demonstrated by peripheral collagenous fibrosis and intervening normal pleura (brackets indicate 1 PAFL). B and C, Higher-power view illustrating the zonation and fibroblastic foci (arrowheads) at the leading edge. D, Fibroblastic focus (hematoxylin-eosin, original magnifications ×10 [A], ×20 [B and C], ×40 [D].
A, Stuck-on appearance of a small pneumothorax-associated fibroblastic lesion. B, Parallel fibrosis. C, Intrapleural microcyst with myxoid fibrosis in the wall. Note the laminar collagen and fibroblasts parallel to the pleural surface. D, Disruption of elastic tissue (hematoxylin-eosin, original magnifications ×20 [A] and ×40 [B and C]; Verhoeff-van Gieson stain, original magnification ×20 [D]).
A, Stuck-on appearance of a small pneumothorax-associated fibroblastic lesion. B, Parallel fibrosis. C, Intrapleural microcyst with myxoid fibrosis in the wall. Note the laminar collagen and fibroblasts parallel to the pleural surface. D, Disruption of elastic tissue (hematoxylin-eosin, original magnifications ×20 [A] and ×40 [B and C]; Verhoeff-van Gieson stain, original magnification ×20 [D]).
Additional findings in individual patients included eosinophilic vasculitis; reactive eosinophilic pleuritis; intra-alveolar macrophages; hemorrhage; patchy, subpleural and peribronchiolar, chronic inflammation; and atelectasis. The presence of these findings was variable from case to case. Areas of patchy, subpleural airspace enlargement or distal acinar emphysema were also present (see Table 2 for a list of the additional major pathologic features).
Fibroblastic foci were not specific for this lesion. Six additional cases (6 of 81; 7.4%) contained fibroblastic foci, but they lacked the wedge-shaped PAFL lesions. In these 6 cases without true PAFL lesions, the fibroblastic foci could be ascribed to other conditions: 2 patients (33%) had usual interstitial pneumonia, 2 (33%) had occupational lung disease (pneumoconiosis and mixed dust pneumoconiosis from coal work) and 2 (33%) were heavy smokers with advanced emphysema. One of the latter patients had also been treated for lung cancer with chemotherapy and radiotherapy. Because true PAFL lesions were absent in these cases, they were not included in the analysis.
COMMENT
Pneumothorax is associated with a wide variety of disorders, ranging from lymphangioleiomyomatosis and inherited disorders of collagen synthesis to pulmonary endometriosis. Given the diversity of conditions associated with pneumothorax, it is most likely that several different mechanisms are responsible for a common result. With that in mind, we hoped that careful review of these cases would reveal morphologic features that segregate cases of pneumothorax into relatively homogeneous clinical categories. Our review identified a distinct pattern of pneumothorax-associated fibroblastic lesions in a subset of cases of SP.
We think there is a substantial overlap between patients with the histologic entity that we have described and the cases reported by Lichter and Gwynne8 in their pioneering study of SP. Of note, Lichter and Gwynne8 selected cases based on clinical features (young age, absence of disease) and not pathologic findings. They then examined the pathology and described emphysema and cyst formation, atelectasis, fibrosis with dense subpleural scars, chronic inflammation, pigment deposition, alveolar cell proliferation, bronchial lesions with peribronchial fibrosis, and vascular lesions. Our cases also demonstrated these features. In addition, we were struck by the presence of fibroblastic foci, a feature not identified in their analysis. It is the recognition of fibroblastic foci in conjunction with the subpleural zonal pattern of fibrosis that is distinctive in this subset of patients and that can aid in the recognition of a specific clinicopathologic entity. As mentioned above, fibroblastic foci are not specific for any one entity but must be a part of the entire morphologic lesion. The fibroblastic foci play a similar role in their use in recognizing usual interstitial pneumonia, where again, they are not pathognomonic for usual interstitial pneumonia but are a well-recognized, morphologic feature.13,14
Although the key features of PAFL (zonation, fibroblastic foci, and wedge-shaped configuration) were not recognized by Lichter and Gwynne,8 one of their photographs is suggestive of a PAFL lesion. Of course they, like others, were struck by the clinicopathologic presentation of the cases and hypothesized a “nonspecific,” infectious etiology abetted by “some local inherent predisposition.”
Our findings are of interest because we took a different approach in our study of patients with pneumothorax. We deliberately included all patients in our larger morphologic review, with a theory that, although some had underlying conditions that could account for the development of pneumothorax, others might have unique, histologic lesions not yet recognized. Following review of our pathologic findings, we noted that a subset of these patients12 had an interesting, distinctive, subpleural lesion, with a zonation pattern and fibroblastic-like foci at the leading edge. Analysis of this group revealed it to be composed predominantly, but not solely, of young men. This is a similar group to those identified on clinical grounds by Lichter and Gwynne,8 who also identified 2 young women in their group.
The patchy pleural distribution and presence in only 12% of all cases we examined (D.A.B., K.S., C.D.G., unpublished data, 2012) argue against this being a nonspecific, reactive process. It is possible that these foci represent sites of old disease because, like most patients biopsied for pneumothorax, 8 out of 11 of our patients (73%) have had recurrent pneumothorax (Table 1). However, we did encounter cases in our larger review (D.A.B., K.S., C.D.G., unpublished data, 2012) that were recurrent but did not have PAFL, and in the one case where we have tissue from both the initial pneumothorax and the recurrence (biopsies 2 and 8, Table 1), PAFL are present in both. The cases from the larger review serve as negative controls; that is, the patients experienced SP but did not exhibit PAFL. Similarly, although PAFL is most common in the young, it is not present in the pathology specimens of every young person with a SP. In the larger study of pneumothorax cases, there were 20 patients (21.7%) younger than 26 years, and PAFL was only identified in 10 of these (50%). The PAFL lesions are not described in any systematic fashion in the pathology specimens of lymphangioleiomyomatosis, Birt-Hogg-Dubé,15,16 Marfan syndrome, or Langerhans cell histiocytosis,17 entities with high rates of pneumothorax. This suggests that PAFL is integral to the development of a specific type of pneumothorax, rather than a nonspecific reactive effect.
The pathogenesis of SP is unknown. Currently, pneumothoraces are believed to arise from a combination of emphysema-like changes, which are blebs and bullae, and abnormalities in the structure of the lung called pleural porosity.18 As most cases of primary pneumothorax are found to have emphysema-like changes and many diseases that lead to secondary pneumothorax are associated with cystic changes in the lung, it has been postulated that these cysts rupture and lead to alveolar and pleural air leakage.18–20 However, fluorescein studies in patients with primary SP show leakage of fluorescent dye away from the emphysema-like changes.18,21 In addition, 6% to 15% of patients without pneumothorax have been found to have blebs on thoracoscopy, so blebs alone are not sufficient for pneumothorax.19,22,23 It is theorized that the pleura is structurally abnormal, leading holes or pores to form, allowing the leakage of air, so-called pleural porosity.18,20 Whether the pleura is abnormal in the same way in each condition is unclear. It is possible that PAFL is a result of either a primary structural abnormality of the pleura or an abnormal healing process.
In summary, we present a unique histologic lesion arising in a subset of predominantly male patients presenting with primary SP. This lesion occurs in the apices of the lung and in young patients who have no, or relatively short, smoking histories. We note a strong, but not complete, overlap between patients carrying our histologic lesion and the clinically defined set of patients first identified by Lichter and Gwynne.8 The etiology of this lesion remains uncertain. Although this may be a florid, reactive-reparative process, it is also possible that these patients have an underlying abnormality that predisposes them to pneumothorax. We look forward to independent confirmation of our findings and, with a histologic entity in hand, further evaluation of a possible etiology.
References
Author notes
The authors have no relevant financial interest in the products or companies described in this article.
Competing Interests
Presented in part at the annual meeting of the United States and Canadian Academy of Pathology, Vancouver, British Columbia, Canada, March 21, 2012.