The 2020 American Thoracic Society (ATS), the Japanese Respiratory Society (JRS), and the Asociación Latinoamericana de Tórax (ALAT) HP clinical practice guidelines paved the way for the standardization of diagnosing HP on an international scale, which, among other benefits, is expected to prevent a misdiagnosis with other IP entities 11. Reporting institutional experiences on the impact of the new guideline is essential to highlight the strengths and potential pitfalls of the new guideline. An accumulation of such reports will inevitably prompt further modifications and ensure the suitable development and evolution of the guideline. Recently, it was reported that the 2020 HP guideline might induce an overdiagnosis of HP cases by MDD 11,15. The prior study found that pathological diagnosis holds considerable sway over the diagnosis reached through MDD 15. MDD is, in a sense, a method to find a compromise between different diagnostic modalities. In this process, it is possible that an erroneous conjecture from one modality may impact the final diagnosis and push it in the wrong direction. To improve the diagnosis, we must seek ways to eliminate errors within each modality.
In this study, 87 (40%) out of 217 cases had typical fHP or probable fHP according to the HP guideline; while fHP was not confirmed in the original diagnosis; it was altered in 54 cases (25%), according to the 2020 HP guideline 11. These results suggest that the diagnosis in nearly one-fourth of all cases may be changed under the new HP guideline. Among the 87 cases, a total of 39 cases were originally diagnosed as definite UIP or probable UIP according to the IPF guideline. Currently, it is unclear what therapy those 39 patients whose diagnoses changed from IPF to fHP in the new guideline may benefit from. This requires further clinical investigation.
We compared cases belonging to the HP + /HP + and HP-/HP + groups to understand the effect on cases whose diagnosis had changed to fHP using the HP guideline. We found significant differences in the lymphocyte fraction and the presence of bird antigens between them. Moreover, there was a statistically significant difference in the CD4/CD8 ratio in the BALF of the HP + /HP + and the indeterminate groups. Although no difference between the HP + /HP + and HP-/HP + groups was found, the results of the HP-/HP + group were similar to those of the indeterminate group. Although these are interesting data, as stated in the 2020 HP guideline, the clinical factors including BALF, serum data, and antigen exposure are useful for HP diagnosis. However, none of them are fully established as a diagnostic criteria by themselves. Further research is needed to address this issue.
Eight of the cases in the HP-/HP + group were determined to be typical fHP by the guideline because of the identification of ACF and granulomas. The granulomas in these cases were not specific for HP in the original diagnosis because they were found only in the pleura or in clusters of multinucleated giant cells without epithelioid cells. The HP guidelines describe the localization of granulomas as being more common in HP, such as in the peribronchial stroma and involved with PBM. However, the cited article indicated granulomas can also be observed in the pleura and airspace 16. Therefore, in the present study, the presence of multiple granulomas was considered positive, even if only in the air space. If granulomas are observed only in the pleura, they should not be regarded as HP granulomas.
Of the HP-/HP + group, 46 of the 54 patients had their diagnosis changed to probable fHP due to ACF. This result suggests that the ACF determination is essential in the diagnosis of HP. One setback may be that the HP guideline does not clearly state what level of ACF should be considered significant. Eight other cases were reclassified to typical fHP, and seven of them were originally diagnosed as definite (n = 3) or probable (n = 4) UIP (Figure, Supplemental Figs. 2,3). A review of their histopathology revealed granuloma or a cluster of giant cells, but they were present in the pleura and the airspace with concurrent smoking-related changes. The diagnosis of definite UIP was made because of the dense fibrosis with architectural destruction and subpleural/ paraseptal distribution and the clear presence of fibroblastic focus. Proper use of both the IPF and HP guidelines in such cases is challenging since those pathological findings are also features of fHP. It is not uncommon for a case judged probable fHP by the HP guidelines to be judged probable UIP by the IPF guidelines, and the diagnosis of one guideline cannot be used as a basis for rejecting the other. It will be very important to reach a consensus between the HP and IPF guidelines in the future. In addition, strong smoking-related pathological findings such as emphysema were observed in 25 of the 54 patients in the HP-/HP + group. It has been reported that smoking-related fibrotic lesions are found around the airways, and the original diagnosis was not considered a finding due to fHP, but rather idiopathic IP 20,21,22. As the HP guidelines state that ACF is not a specific finding for fHP, a pathologist should decide whether to consider these 25 cases as definite or probable UIP or probable fHP. This shows that the ACF misinterpretation should be kept in mind in cases of smoking-related pathological findings.
Only three cases were indeterminate for fHP according to the 2020 HP guideline,even though the original diagnosis was HP. As shown in Fig. 3A–D, two VATS biopsies, with three biopsy sites in both cases, showed end-stage honeycomb lung with granulomas, and it was impossible to determine the presence of ACF. For TBLC, samples were obtained from two sites; however, it was inadequate to identify ACF, probably owing to its limited size (Fig. 3E–F). Regarding the three HP + /indeterminate HP cases, as ACF was not observed, the patients were considered indeterminate for fHP according to the 2020 HP guidelines. However, all of these cases had multiple poorly formed granulomas in the stroma, and the original diagnosis was HP. The judgment of one criterion, ACF, is critical for the HP guideline; however, its specificity for fHP diagnosis is unclear. A previous study demonstrated that ACF had high sensitivity but low specificity for the diagnosis of HP 23. Tanizawa et al. compared UIP cases with and without ACF and found that, although cases with ACF were diagnosed significantly more frequently than fHP, there was no other significant difference between them, including genetic mutations 24. These reports show that the clinical relevance and reproducibility of ACF are unclear. These points need to be clarified in the future.
We examined whether the HP guideline could be applied to TBLC. Of the 65 TBLC specimens reviewed, 61 cases (94%) had quality sufficient for evaluation. The results were similar to those previously reported for TBLC specimen adequacy 25,26. We found that the proportion of cases identified as indeterminate for HP was highest in the TBLC group. This is because findings such as ACF and granuloma are rarely observed in TBLC without sampling a large section of lung tissue. Even if fHP is suspected pathologically, the judgment with the HP guideline is often indeterminate for fHP. This shows that the judgment of samples obtained using TBLC may underestimate HP detection using the current HP guideline. Other publications have similarly demonstrated that the frequency of HP in TBLC is lower than in VATS biopsy 27,28,29. Thus, based on our experience, simply applying the HP guideline in its current form to IP samples obtained with TBLC is not recommended.
There are several limitations to this study. First, the study comprised a purely pathological assessment of the effect of the guideline on the HP diagnosis, and MDD was not performed to reach a final clinical, radiological, and pathological consensus diagnosis. However, this was not within the scope of our study, which focused on the pathological domain, and it has been addressed in a separate study 15. Our data may show different trends after MDD input because it is a reconciliation process by radiologists and pulmonologists, and the actual effect of pathologic discrepancy may be diminished. Second, the study was retrospective in design, meaning we were restricted to the available data. Third, we used cases from a single center, which may have introduced some selection bias and composed of one MDD group. Fourth, the location of granuloma in the lung defined by HP guideline may be interpreted differently by different raters. In the future, it is necessary to examine whether the diagnosis of fHP differs depending on the location of granuloma. Nevertheless, this is the first study to address the impact of the newly introduced HP guideline in a large case series.
We confirmed that the pathological criteria of the 2020 HP guideline efficiently excluded most non-HP cases11. Concurrently, approximately one-fourth of all cases of fibrotic IP diagnostically changed from not HP to fHP. The significance of our results should be further evaluated and validated to clarify if such changes convey clinical and prognostic significance are an improvement or deterioration for the patients.