Morphologic Aspects of Interstitial Pneumonia With Autoimmune Features Open Access

The need for a better understanding and standardization of the entity ILD with autoimmune features, and the need for adequate management of patients with this entity resulted in the introduction of a new term, IPAF, in 2015, and in the formulation of a consensus-based classification proposal by several specialists.^15,16  The multidisciplinary consensus established the presence of 3 prerequisites for the definition of IPAF: evidence of ILD on HRCT and/or on surgical lung biopsy; clinical evaluation for the exclusion of known causes of ILD; and the absence of criteria that define a specific CTD type.¹⁵  The remainder of the classification was organized into 3 different domains: clinical, consisting of specific extrathoracic signs or symptoms; serologic, defined by the presence of specific autoantibodies; and morphologic, composed of characteristic changes on HRCT, histopathologic features on surgical lung biopsy, or multicompartment involvement.^15,17  To establish the diagnosis of IPAF, the patient needs to fulfill the 3 prerequisites and have at least 1 characteristic in at least 2 of the domains.^15,16 

Table 1 presents the classification criteria for IPAF. The clinical domain consists of signs and symptoms suggestive of autoimmunity, but these alone do not define CTD criteria. The serologic domain is formed by circulating autoantibodies frequently associated with CTDs, such as antinuclear antibody with titer equal to or greater than 1:320, or any titer with nucleolar or centromere staining pattern, preferably by indirect immunofluorescence assay; rheumatoid factor greater than 2 times the upper limit of normal; and other autoantibodies that are considered positive whenever their value exceeds the upper limit of normal. The morphologic domain consists of 3 categories thusly divided: ILD patterns observed on HRCT characterized by nonspecific interstitial pneumonia (NSIP; Figure 1, A and B), organizing pneumonia (OP; Figure 2, A), lymphoid interstitial pneumonia (Figure 3, A and B), or NSIP with OP; patterns of ILD or changes suggestive of autoimmunity in pulmonary surgical biopsy, that is, NSIP (Figure 1, C and D), OP (Figure 2, B), NSIP with OP, lymphoid interstitial pneumonia (Figure 3, C and D), and interstitial lymphoid aggregates with germinal centers and diffuse lymphoplasmacytic infiltration (Figure 4); and evidence of involvement of multiple thoracic compartments associated with ILD diagnosed by imaging exams, histopathologic findings, right heart catheterization, or pulmonary function tests.^15,18 

Subsequent studies that applied the IPAF concept have shown that these patients were more frequently female^14,16,19  and tended to be older than those with CTD-ILD.^14,16,20  In the serologic domain, antinuclear antibody was the most common finding,^14,16,19,21  and Raynaud phenomenon was the most frequent clinical feature.^14,16,19,21  Nonspecific interstitial pneumonia and usual interstitial pneumonia (UIP; Figure 5) were the most frequent patterns detected in the morphologic domain (Table 2).^14,19,21 

Nonspecific interstitial pneumonia is the pattern of ILD most commonly associated with CTD and other autoimmune disorders, and it may precede their diagnosis for years.^22,23  Kono and colleagues²³  evaluated the incidence of CTD in patients after an initial diagnosis of idiopathic NSIP. Of the 35 patients with idiopathic NSIP, 17.1% developed CTD during the follow-up period (5.5 ± 5.0 years). In addition, the authors observed that several patients in the idiopathic NSIP group as well as in the NSIP preceding CTD diagnosis group had already fulfilled the criteria for UCTD, LD-CTD, AIF-ILD, and IPAF at the time of initial NSIP diagnosis.²³  A study to characterize patients with LD-CTD demonstrated that NSIP was the most common morphologic pattern observed on HRCT (45%) and surgical lung biopsy (27%), and 15% of those patients prospectively met the criteria for a definitive diagnosis of a CTD.²⁴ 

In 2007, Kinder et al¹⁰  proposed that idiopathic NSIP is a manifestation of UCTD. In this study, among 18 patients classified as UCTD who underwent surgical pulmonary biopsy, 15 showed an NSIP pattern, in contrast to only 2 patients in the control group that was categorized as idiopathic interstitial pneumonia.¹⁰  In addition, a study by Suda et al²⁵  showed different prognoses in patients with NSIP associated with UCTD and patients with NSIP without criteria for UCTD. In their study, among 47 patients diagnosed with idiopathic NSIP who did not meet the criteria for a specific CTD, 22 had diagnostic parameters for UCTD. Comparison of survival curves during the course of 5 years showed that patients with UCTD-NSIP had a higher survival rate (100%) compared with idiopathic NSIP patients (58%).²⁵ 

Nonspecific interstitial pneumonia was the most common pattern identified on HRCT (57.1%) and on surgical lung biopsy (55%) in a retrospective cohort study of 56 patients who fulfilled classification criteria for IPAF. Most patients (98%) were treated with prednisone and/or immunosuppressive therapy. Baseline pulmonary function tests showed a mild restrictive pattern during follow-up (284.9 ± 143.3 weeks), and no decline in predicted forced vital capacity values and no deaths were observed.¹⁹  Similarly, NSIP was the most frequent pattern observed in a study of 57 IPAF patients (42.1% given diagnosis on HRCT and 31% among 16 patients who received diagnosis on biopsy), followed by UIP pattern (28% on HRCT and 19% on surgical lung biopsy). Diffuse lymphoplasmacytic infiltration and interstitial lymphoid aggregates were also observed on histologic examination (16% and 13%, respectively). Overall survival of IPAF patients at 1 year was not significantly different from overall survival of patients with idiopathic pulmonary fibrosis (IPF; 84% versus 95%, respectively). This may be partly explained by a significant proportion of UIP pattern among those with IPAF and also by the use of antifibrotic drugs in IPF patients. Considering only the IPAF group, smoking was a unique predictive factor associated with a higher risk of death. No difference in survival was observed when patients with UIP and NSIP pattern on HRCT were compared.²¹ 

Although NSIP is the pattern of ILD most often associated with autoimmunity, several other studies have reported UIP as the most prevalent pattern associated with UCTD. Vij et al¹²  performed a systematic evaluation of signs and symptoms, and serologic tests in patients with ILD in order to identify those patients who had evidence of an underlying autoimmune process, which they referred to as AIF-ILD. In this study, the predominant histopathologic pattern in patients with AIF-ILD was UIP (81%). The presence of a positive aldolase dosage (aldolase above the upper limit of normal) and increasing age were associated with worse survival, whereas high antinuclear antibody titers (1:1280 or higher) were associated with improved survival. There was no difference in the survival rate between patients with AIF-ILD and those with IPF.¹²  Similarly, the presence of autoantibodies in patients with a histologic diagnosis of UIP was not associated with an increase in survival, based on an analysis conducted by Song et al.²⁶  A subsequent study by Kim et al⁸  showed a different result, in which UIP was the most prevalent pattern associated with UCTD (44 of 105), followed by NSIP (29) and OP (9). In that study, the survival rate among patients with UCTD-NSIP was significantly higher compared with those with UCTD-UIP, but these patients had better survival when compared with patients with IPF.

Recently, Oldham et al¹⁴  demonstrated that among 144 patients who met the criteria for IPAF, 54.6% had a UIP pattern on HRCT and 73.5% had it on surgical lung biopsy. Survival analysis showed that the IPAF group had a worse outcome than the CTD-ILD group and had a marginally better outcome than that of the IPF group. After stratifying the IPAF group with respect to the presence or absence of UIP pattern, it was observed that patients with UIP pattern had similar survival compared with patients with IPF, and non–UIP pattern patients showed similar survival compared with patients with CTD-ILD.¹⁴  Although the diagnosis of UIP is already associated with poor prognosis in rheumatoid arthritis²⁷  and idiopathic interstitial pneumonia,²⁸  there are divergent data concerning the impact of this morphologic pattern in IPAF.

A systematic investigation of HRCT findings in patients with IPAF conducted by Chung and colleagues²⁹  showed that most patients with IPAF had UIP or possible UIP pattern on HRCT scan (65.4%; 89 of 136). A UIP or possible UIP pattern on HRCT was strongly correlated with UIP diagnosis on pathologic evaluation.²⁹  Survival analysis based on HRCT findings demonstrated that honeycombing and pulmonary artery enlargement were independently associated with poor survival. A smaller number of patients had a pattern inconsistent with UIP on HRCT (34.6%; 47 of 136), and the most common HRCT finding associated with this pattern was the presence of a significant degree of ground-glass opacity. Interestingly, 8% of the patients had an imaging pattern most consistent with hypersensitivity pneumonia.²⁹  It is known that hypersensitivity pneumonia may co-occur with autoimmune features or with a defined CTD, but the mechanisms underlying this association require further investigation.^30,31 

In summary, studies that evaluated survival in patients with IPAF and similar diseases indicate that multiple factors may be associated with prognosis in these patients, which include: (1) the prevalence of a UIP or NSIP pattern on HRCT and/or histology (a higher prevalence of a UIP pattern was associated with worse survival, whereas a higher frequency of an NSIP pattern was often associated with a better prognosis); (2) the use of prednisone and/or immunosuppressive drugs may have improved prognosis in IPAF patients; (3) the use of antifibrotic drugs may have improved survival in IPF patients; (4) a larger extent of honeycombing on HRCT determined a negative impact on survival; and (5) older age and a greater prevalence of smoking were associated with worse prognosis.

Although the studies demonstrate a relatively high frequency of UIP in patients with underlying autoimmune diseases, this morphologic pattern is not specific to IPAF. Therefore, for a patient to be categorized as IPAF, at least 1 aspect of the other 2 domains (clinical and serologic) or a morphologic characteristic different from UIP, such as interstitial lymphoid aggregates with germinal centers, diffuse lymphoplasmacytic infiltration, or multicompartment involvement, must be present.¹⁵ 

Usual interstitial pneumonia was the main histologic pattern found among 44 patients with LD-CTD (57%), followed by NSIP (30%). Additional morphologic findings suggestive of an autoimmune process have also been observed—that is, lymphoid aggregates with germinal centers, extensive pleuritis, prominent plasmacytic infiltration, and dense perivascular collagen. The most frequent finding was prominent plasmacytic infiltration, which was present in 72% of UIP and 85% of NSIP cases.³²  Histologic aspects, such as germinal centers and plasma cells, were also more frequently observed in patients with UIP and autoantibodies compared with patients with UIP without autoantibodies.²⁶  Interestingly, these morphologic aspects were not significantly different when comparing patients with IPF and autoantibody to CTD-UIP patients.²⁶ 

Lymphoid aggregates with germinal centers and diffuse lymphoplasmacytic infiltrate can emerge in the background of UIP and NSIP, and their presence, although not exclusive for IPAF or CTD-ILD, suggests the possibility of an underlying autoimmune process.³³  The reactive lymphoid hyperplasia can be observed within the fibrotic areas, in subpleural zones, along interlobular septa, and can be associated with bronchial walls.^33,34  A recent study by Adegunsoye et al³⁵  proposed a new methodologic approach to identify patients with histopathologic characteristics of IPAF. The authors used a quantitative evaluation for interstitial lymphoid aggregates with germinal centers (at least 3 germinal centers in a low-power field) and lymphoplasmacytic infiltration (presence of diffuse infiltrates of lymphocytes and at least 40 plasma cells in a high-power field) to identify patients fulfilling the morphologic criterion of IPAF. The use of this methodology associated with the histologic review of the cases and evaluation of characteristics of other domains allows patients with an initial diagnosis of IPF to be reclassified as IPAF.³⁵ 

The lungs are often involved in CTDs, and all compartments may be affected, including the interstitium, large and small airways, pleura, and pulmonary vasculature. A combination of patterns is often seen, and the association of more than 2 compartments is suggestive of an underlying autoimmune disease.⁴  In view of these aspects, the involvement of multiple compartments was included as one of the defining morphologic criteria of IPAF.¹⁵ 

The term “multicompartment involvement” refers to changes characterized by unexplained intrinsic airways disease (airflow obstruction, bronchiolitis, or bronchiectasis; Figure 6) evidenced by pulmonary function tests, HRCT, or histopathologic findings; unexplained pulmonary vasculopathy (precapillary pulmonary hypertension: group 1, pulmonary arterial hypertension; group 1′, pulmonary veno-occlusive disease; and group 3, pulmonary hypertension due to chronic lung disease and/or hypoxia; Figure 7) demonstrated mainly by cardiac hemodynamic assessment; and unexplained pleural or pericardial effusion or thickening detected by HRCT or ultrasound imaging or pleuritis on lung biopsy (Figure 8).¹⁵ 

Intrinsic airways disease was the most common multicompartment finding in IPAF patients observed in the study of Oldham et al,¹⁴  followed by pleural disease and pulmonary vasculopathy. Pulmonary hypertension was present in 10 of 57 IPAF patients (17.5%) studied by Ahmad et al.²¹  In a study by Adegunsoye et al,³⁵  pulmonary vasculopathy was also the most prevalent finding in 45 of 84 patients (53.6%) with IPAF. However, the prevalence of pulmonary vasculopathy in IPAF patients was not significantly different from the non-IPAF group (patients meeting criteria for IPF, idiopathic NSIP, cryptogenic organizing pneumonia, or UCTD who had not fulfilled criteria for IPAF).³⁵  In addition, patients with IPAF and pulmonary vasculopathy had worse survival rates compared with the IPAF group without such findings. The same analysis was performed among the non-IPAF patients, and the presence of pulmonary vasculopathy was not associated with worse survival rates in comparison with patients without pulmonary vasculopathy.³⁵ 

Whenever possible, ILDs should be discussed in a multidisciplinary meeting, with an expert panel of clinicians, radiologists, and pathologists.^16,36  The differential diagnosis of ILDs is extensive and challenging, and nonspecialized radiologists and pathologists may provide inaccurate diagnosis.³⁷  It is recommended that a detailed search for CTD be performed in all patients with ILD, and the participation of a rheumatologist is useful.¹⁶  A multidisciplinary approach to ILDs, including IPAF, increases the likelihood of obtaining an accurate diagnosis for optimal treatment and prognostic definition, reducing the frequency of unclassifiable diagnoses.^16,38  In current practice, a multidisciplinary discussion is necessary to determine whether a patient fulfills the criteria for IPAF; however, IPAF is a research statement,¹⁵  and for now this terminology should not be used in pathologic conclusions. In our practice, some commentaries with recommendations for the investigation of possible underlying autoimmune process are added to the pathologic reports.

Some patients with ILD exhibit suggestive but nondiagnostic clinical, serologic, and morphologic features of a definitive CTD. Different nomenclatures and criteria have been used to describe this condition: UCTD-ILD, LD-CTD, and AIF-ILD. The need for a single classification that would allow for prospective studies with uniform and comparable diagnostic parameters resulted in the introduction in 2015 of a new classification and terminology based on the consensus among several specialists: IPAF. Although the proposal is recent (2015), published studies using these criteria seem to identify a distinct subgroup of patients characterized primarily by female sex, younger age compared with individuals with IPF, and different prognoses. The studies show that NSIP and UIP are the most prevalent morphologic patterns and show discrepant results about the impact of the UIP pattern on survival. Prospective investigations are necessary to better define this subgroup and to determine the prognosis and appropriate clinical management of these patients.