An alternative approach to the current diagnostic guidelines for fibrotic interstitial lung disease
Fibrotic interstitial lung diseases (fILD) constitute a constellation of lung conditions including: idiopathic pulmonary fibrosis (IPF), chronic hypersensitivity pneumonitis (CHP), fibrosing non-specific pneumonitis (NSIP), ILD secondary to connective tissue disease or vasculitis, occupational lung disease and drug-induced lung fibrosis. fILD are a major health burden due to progressive fibrosis leading to end-stage lung disease.1 In addition, the diagnosis of fILD has challenged clinicians, radiologists and pathologists resulting in several reiterations of disease classification.2 Recently, two working groups—the Fleischner Society3 and the International Working Group (2017)4 have suggested changes to the previous American Thoracic Society (ATS)/European Respiratory Society (ERS)/Japanese Respiratory Society (JRS)/Latin American Thoracic Association (ALAT) guidelines of 2011 on the diagnosis of IPF.5 The International Working Group (2017) proposed that the terms ‘possible’ and ‘probable’ IPF not be used and the diagnosis of IPF be classified into high (>90% confidence) and low confidence (50–90%).4 In contrast, in December 2017, the Fleischner Society has published a white paper on the nomenclature for IPF—typical IPF, probable IPF, indeterminate IPF and features not consistent with IPF. Notably, the ATS/ERS/JRS/ALAT guidelines of 2011 are similar to the Fleischner guidelines but not in line with the International Working Group that encourages against the terms used in both classifications. This is emblematic of the difficulties finding a uniform approach to the diagnosis of fILD.
The major limitation of the classification proposals is the ambiguity of the diagnostic terms such as ‘possible’, ‘probable’ and ‘indeterminate’. Although this approach has flexibility in incorporating the wide-spectrum of fibrotic disorders under an umbrella diagnosis, it is unlikely to move the field forward since it is not possible to phenotype subjects around vague terms. This commentary proposes that we remove ambiguity and label conditions as directly and descriptively as possible. Although this is not without flaw it will help focus the diagnostic process on positive characteristics that have a direct relationship to clinical, radiological and pathological features thus improving patient outcomes.
THE SPECTRUM OF FIBROTIC fILD
Idiopathic pulmonary fibrosis
IPF has a prevalence of 60 per 100 000 with a 5-year mortality of 50%.6 This condition is characterized by a usual interstitial pneumonia (UIP) pattern on HRCT (Table 1) and pathology. The diagnostic criteria proposed by the ATS/ERS/JRS/ALAT guidelines of 2011 and the Fleischner guidelines are essentially the same for IPF (Table 1).
| Definite IPF | Possible IPF | Unclassifiable IPF | |
|---|---|---|---|
| ATS/ERS/JRS-2011 | Subpleural, basal predominance Reticular abnormality Honeycombing with or without traction bronchiectasis Absence of features listed as inconsistent with UIP pattern |
Same as definite IPF, Reticular abnormality No honeycombing |
Unclassifiable IPF: Diagnosis not made by clinical features or MDD. Inconsistent features between clinical, radiology and pathology. In patients with no biopsy MDD to make a diagnosis |
| Fleischner-2017 | Definite IPF: Basal predominant (occasionally diffuse), and subpleural predominant; heterogeneous. Honeycombing; reticular pattern with peripheral traction bronchiectasis or bronchiolectasis*; absence of alternative diagnosis | Same as IPF, Reticular abnormality No honeycombing |
Indeterminate IPF: Variable or diffuse. Evidence of fibrosis with some inconspicuous features suggestive of non-UIP pattern |
| Suggested new diagnosis | Definite IPF | Reticular predominant PF | Upper lobe predominant pulmonary fibrosis Inflammatory–pulmonary fibrosis |
- *Reticular pattern is superimposed on ground glass opacity, and in these cases it is usually fibrotic. ATS, American Thoracic Society; ERS, European Respiratory Society; IPF, idiopathic pulmonary fibrosis; JRS, Japanese Respiratory Society; MDD, multidisciplinary discussion; UIP, usual interstitial pneumonia.
- Marked fibrosis/architectural distortion
- Honeycombing subpleural/paraseptal distribution patchy involvement of lung parenchyma by fibrosis and fibroblast foci
- Absence of features against a diagnosis of UIP—upper or mid-lung predominance, peribronchovascular predominance, extensive ground glass abnormality, profuse micronodules (bilateral, predominantly upper lobes), discrete cysts (multiple, bilateral, away from areas of honeycombing), diffuse mosaic attenuation/air trapping and consolidation in bronchopulmonary segment(s)/lobe(s)5, 3
- Evidence of marked fibrosis/architectural distortion, honeycombing in a predominantly subpleural/paraseptal distribution
- Presence of patchy involvement of lung parenchyma by fibrosis
- Presence of fibroblast foci
- Absence of features that are not in keeping with a diagnosis of UIP. Hyaline membranes, organizing pneumonia, granulomas, marked interstitial inflammatory cell infiltrate away from honeycombing. Predominant airway centred changes. Other features suggestive of an alternate diagnosis.
Although the UIP pattern on both radiology and pathology is well characterized and attributed to IPF, it is a common feature found in a variety of different conditions characterized by fibrosis including CHP. Therefore, UIP should be recognized as a pattern of lung remodelling following an injury from a wide spectrum of disorders that may characterize a progressive phenotype and not a diagnosis in itself.
Possible (ATS/ERS/JRS/ALAT) or probable (Fleischner) IPF
The ‘possible’ IPF diagnosis of the ATS/ERS/JRS/ALAT criteria and the ‘probable’ IPF (Fleischner) guidelines describe the distinct entity of lung fibrosis with reticulation but no honeycombing on HRCT5.3
Since reticulation is the dominant feature of this condition this commentary suggests ‘reticular predominant pulmonary fibrosis’ in lieu of ‘possible’ or ‘probable’. This would describe the condition and enable a distinct phenotype to be identified.
Chronic hypersensitivity pneumonitis
A major source of uncertainty is the diagnosis of CHP (Table 2) since chronic stages of this disease lack typical clinical, radiological and pathological features.7 Mosaicism, air trapping and fibrosis on HRCT suggesting small airways disease may be the only features present since exposures, serology and pathology may be absent. As such there is a wide spectrum of presentations, both clinical and radiological that may be diagnosed as CHP without all the features of this condition leading to the exclusion of diseases that present in a similar way such as airway-centred lung fibrosis.8 As such, the author proposes, in the absence of diagnostic features of CHP but with mosaicism and fibrosis on HRCT the term ‘small airways predominant pulmonary fibrosis’ be used.
| Chronic hypersensitivity pneumonitis | Non-specific interstitial pneumonia | |
|---|---|---|
| Clinical | Exposures to precipitants | Progressive |
| Radiological | Mosaicism/air trapping Upper lobe involvement |
Subpleural, basal predominance. Ground-glass possible diffuse or peripheral. Rare honeycombing (∼5%) Reticular abnormality traction bronchiectasis |
| Serological | Positive precipitants | |
| BALF | Lymphocytosis | Non-specific |
| Pathology | Poorly formed granuloma | Temporal uniformity with contiguous involvement and inflammation and fibrosis (cellular or fibrosing). No architectural distortion. Rare honeycombing. Bronchiolectasis. Rare fibroblastic foci organizing pneumonia (usually <10%) |
| Suggested new diagnosis | Small airway-related pulmonary fibrosis (in the absence of exposure, serology or pathology) | Interstitial predominant inflammatory pulmonary fibrosis (when pathology is present) |
Non-specific interstitial pneumonia
The other diagnostic term that is misleading is NSIP since it has no relationship to the underlying condition. NSIP is characterized by specific features (Table 2) but despite this, many conditions that do not fit these criteria may also be diagnosed as NSIP. The condition is typically characterized by the presence of temporally homogenous interstitial inflammation and fibrosis on lung biopsy, the author suggests NSIP be called ‘interstitial predominant inflammatory pulmonary fibrosis’.
Unclassifiable/indeterminate IPF
‘Unclassifiable’ (ATS/ERS/JRS/ALAT 2011) denotes a radiological diagnosis frequently made in the absence of a lung biopsy whilst ‘indeterminate’ (Fleischner) demonstrates features of fibrosis and features of a non-UIP pattern.3 Both these categories incorporate several patterns of disease and phenotypes that do not have a biopsy to assist the indeterminate radiological diagnosis. Usually, the radiology has both features of fibrosis and inflammation. The author therefore suggests that the terms ‘indeterminate’ or ‘unclassifiable’ not be used and place all presentations with non-specific features on radiology as NSIP. If pathology is available then the condition can be more definitively diagnosed.
In the presence of idiopathic upper lobe fibrosis this should be described as ‘upper lobe dominant pulmonary fibrosis’ rather than ‘indeterminate’ or ‘unclassifiable’ disease.
THE NEW ALGORITHM CLASSIFYING fILD
This commentary suggests that vague terms should be removed from the diagnosis of fILD (Tables 1, 2). In the future, there will be several phenotypes that may be identified by unique molecular signatures. In moving closer to this aim, using existing diagnostic guidelines, that positively diagnose different types of fILD will reduce the need for terms such as ‘possible’, ‘probable’, ‘indeterminate’ and ‘unclassifiable’. This approach needs to be further investigated by studies evaluating the accuracy, confidence of the MDD in the diagnosis and improved phenotyping of fILD.
