Accurate diagnosis of interstitial lung disease (ILD) is crucial for management and prognosis but can be challenging even for experienced clinicians. Expert multidisciplinary discussion (MDD) is considered the reference standard for ILD diagnosis; however, there remain concerns regarding lack of validation studies and relative limited information on the impact of MDD in real-life clinical practice. The goal of this study was to assess the effect of MDD in providing a specific ILD diagnosis, changing the diagnosis provided upon referral, and to determine how often and in which way MDD altered management.
Interstitial lung diseases (ILD) are a heterogenous group of diseases with varied causes, treatment, and prognoses. Accurate diagnosis of ILD is crucial for management decision and prognosis; however, it can be challenging due to the diverse manifestations and lack of defined diagnostic criteria for some ILDs. Patients with ILD require a dynamic multidisciplinary assessment for integration of clinical, radiological and, at times, pathological findings. The importance of multidisciplinary discussion (MDD) has been emphasized since the publication of the American Thoracic Society (ATS) and European Thoracic Society (ERS) consensus statement on the classification of idiopathic interstitial pneumonias in 2002 [
The goal of this study was to determine the impact of MDD in the diagnosis and management of patients with ILD. Specifically, the study assessed the ability of MDD to provide a specific diagnosis, the proportion of cases in which MDD altered the final diagnosis, and the concordance between pre-MDD and MDD diagnoses. The study also assessed how often and in which way MDD altered the management plan.
This retrospective observational study was conducted at The Ottawa Hospital, a tertiary academic center with expertise in ILD, and was approved by the institutional ethics review board; informed patient consent was waived. Cases were retrieved from the ILD-MDD database from January 1st, 2013, to May 31st, 2018. This prospectively maintained database of cases of ILD reviewed in face-to-face weekly/biweekly meetings. The MDD team consists of respirologists, a registered nurse, a thoracic pathologist, and a thoracic radiologist all with expertise in ILD who review clinical information, imaging studies and pathology slides when available, of patients seen at the ILD clinic or referred from the community. Inclusion criteria for the study consisted of patients with suspected or confirmed fibrotic ILD with available recent CT scan of the chest, database entry containing pertinent clinical and pathological information, and documented MDD consensus diagnosis. Patients with confirmed diagnosis of collagen vascular disease, lack of documented consensus diagnosis, or with no CT scan of the chest were excluded.
Information regarding patients’ demographics, origin of referral, diagnosis at the time of referral (pre-MDD diagnosis), management before referral if any, consensus diagnosis (MDD diagnosis), and management following MDD were obtained from the ILD-MDD database and, when not available, extracted from patient’s electronic records. Origin of referral was documented according to medical specialty and type of practice (non-MDD academic physician or community-based physician). Pre-MDD diagnoses were based on clinical diagnosis provided by referring physician and/or HRCT diagnosis when available. Specific pre-MDD diagnoses were considered valid and recorded when in accordance to the current consensus classification for idiopathic interstitial pneumonias [
MDD consensus diagnosis strictly followed current consensus guidelines for diagnosis and management of IIP (idiopathic interstitial pneumonias) and IPF [
Pre-MDD diagnoses were compared to MDD consensus diagnoses, change in final diagnosis was recorded in percentage (%), and discordant cases were documented. Relevant change in patient’s management following MDD diagnosis was recorded, including initiation or change in pharmacological treatment, termination of medical treatment, referral to surgical lung biopsy, and institution of nonpharmacological management.
During the study time, 172 cases were reviewed in the ILD-MDD meeting. Patients with confirmed diagnosis of collagen tissue disease (CTD, 19 cases) and long-term ILD patients followed by the MDD respirologist and reviewed for instructive purposes (14 cases) were excluded from the analysis; 13 cases were excluded for missing data. A total of 126 cases were included in the final analysis (79 men, 47 women, 36–93 years, mean 70 years, median 69). Patients were referred to ILD-MDD meetings by physicians of various specialties, most commonly community family physicians and respirologists. Table
Origin of referral.
Specialty | Number of cases |
---|---|
General practice | 64 |
Respirology | 34 |
Internal medicine | 6 |
Cardiology | 5 |
Emergency | 3 |
Surgery | 3 |
Intensive care unit | 1 |
Rheumatology | 1 |
Oncology | 1 |
Not specified | 8 |
A specific pre-MDD ILD diagnosis was provided in 59% (74/126) of patients, most commonly IPF (24/74, 32%) and HP (20/74, 27%); the remainder of the patients (52/126, 41%) were referred without a specific diagnosis (ILD-NOS).
When available, the pre-MDD diagnosis was based on a clinical diagnosis made by the referring physician in 45 patients; in the remainder 81 cases, pre-MDD diagnosis was based on initial HRCT diagnosis which in some cases corresponded to the clinical diagnosis provided by the referring physician. A summary of pre-MDD diagnoses is provided in Table
Pre-MDD and MDD consensus diagnoses.
Diagnosis | Pre-MDD, | MDD, |
---|---|---|
ILD-NOS | 52 (41%) | 0 (0%) |
HP | 20 (16%) | 21 (16.6%) |
IPF | 24 (19%) | 34 (27.0%) |
SR-ILD | 5 (4.0%) | 10 (8.0%) |
NSIP | 6 (4.7%) | 6 (4.7%) |
OP | 3 (2.4%) | 2 (1.6%) |
Drug toxicity | 4 (3.2%) | 4 (3.2%) |
Asbestosis | 1 (0.8%) | 1 (0.8%) |
Sarcoidosis | 4 (3.2%) | 1 (0.8%) |
LIP | 1 (0.8%) | 1 (0.8%) |
GL-ILD | 1 (0.8%) | 1 (0.8%) |
Pulmonary ossification | 1 (0.8%) | 1 (0.8%) |
Aspiration | 1 (0.8%) | 1 (0.8%) |
HP/IPF | 2 (1.6%) | 0 (0%) |
DIPNECH | 0 (0%) | 1 (0.8%) |
ILA | 0 (0%) | 4 (3.2%) |
Cystic disease | 1 (0.8%) | 0 (0%) |
IPAF | 0 (0%) | 5 (4.0%) |
Unclassifiable ILD | 0 (0%) | 27 (21.4%) |
No ILD | 0 (0%) | 6 (4.7%) |
A specific MDD diagnosis was provided in 62% (78/126) of cases; 12% (15/126) had a provisional diagnosis established at initial discussion sufficient to direct management. ILD was deemed unclassifiable at initial discussion in 21% (27/126) patients. In 6 patients, there was no clinical and radiological evidence of ILD after MDD with parenchymal changes attributed to atelectasis or poor technique. A summary of post-MDD diagnoses is provided in Table
Provisional diagnoses and effect on management.
Provisional diagnosis ( | Effect on management ( |
---|---|
HP = 8 | Initiation pharmacological treatment (4) |
No change (4) | |
IPF = 4 | No change (4) |
NSIP = 3 | No change (2) |
Initiation pharmacological treatment (1) |
HP = hypersensitivity pneumonitis; IPF = idiopathic pulmonary fibrosis; NSIP = nonspecific interstitial pneumonia.
The overall agreement for specific pre-MDD and MDD diagnosis was 41% (52/126). MDD altered diagnosis provided upon referral in 37% (47/126) of patients, including 22 specific MDD diagnoses for ILD-NOS. For specific ILD diagnoses, of the 24 patients referred with IPF diagnosis, 20 (80%) were confirmed at initial MDD discussion; one was diagnosed as smoking-related ILD (SR-ILD) and three were unclassifiable (two later diagnosed as UIP/IPF after lung biopsy). Of the 20 patients referred as chronic HP, 12 (60%) were confirmed, including 4 working diagnoses; one patient was diagnosed as SR-ILD, and one was found to have no ILD. Six pre-MDD HP were deemed unclassifiable, and one was later diagnosed as UIP/IPF following lung biopsy (Figure
A 72-year old man referred to MDD meeting with a diagnosis of sarcoidosis. HRCT pattern favored chronic HP; however, the case was deemed unclassifiable after multidisciplinary discussion as no exposure was elicited and due to the high clinical probability of IPF. The patient was referred to surgical lung biopsy that demonstrated a usual interstitial pneumonia (UIP) pattern and led to a diagnosis of IPF. The patient was subsequently started on antifibrotic treatment.
Concordant Pre-MDD and MDD diagnoses.
Concordant diagnoses | % | |
---|---|---|
HP | 12 | 23 |
IPF | 19 | 37 |
SR-ILD | 4 | 7 |
NSIP | 5 | 9 |
OP | 2 | 4 |
Drug toxicity | 3 | 6 |
Asbestosis | 1 | 2 |
Sarcoidosis | 2 | 4 |
LIP | 1 | 2 |
GL-ILD | 1 | 2 |
Pulmonary ossification | 1 | 2 |
Aspiration | 1 | 2 |
MDD led to change in management in 39% (50/126) of patients, notably initiation of medical therapy (30/50), lung biopsy (6/50), and change in medical treatment (4/50), including one patient with no ILD after MDD who had been started on immunosuppression and discontinued after consensus discussion. A total of 21 patients continued medical treatment initiated before MDD (5 with antifibrotic therapy and 16 with immunosuppressant). In the remainder of cases, patients were referred to smoking-cessation program and further rheumatological investigation, and one was referred to lung transplant. Amongst cases with concordant pre-MDD and MDD diagnoses, management was altered in 46% (24/52), either initiation or change in medical treatment. Consensus discussion endorsed pharmacological therapy instituted before referral in 41% (52/126) cases. In 21% (27/126) of patients, MDD recommendation was continued clinical follow up along with referral to smoking-cessation program in eight cases.
Lung biopsy was indicated in 22 (17%) of patients but performed in only 6 cases; 3 patients declined biopsy, and 13 were deemed not eligible to biopsy due to advanced disease. Of the 6 unclassifiable cases with lung biopsy, 4 were diagnosed with UIP/IPF, one with NSIP, and one with UIP pattern and areas of organising pneumonia.
Multidisciplinary discussion established a specific ILD diagnosis in 62% of our patients. When compared to diagnosis upon referral, MDD provided a new diagnosis or altered preexisting diagnosis in 37% of the cases. Our results, corroborating previous studies that showed change in diagnosis following MDD in 40%–53% of cases [
A substantial proportion of patients were referred to MDD for evaluation of ILD not specified. For those with a specific pre-MDD diagnosis, concordance rate with consensus MDD diagnosis was 41%. The highest concordance rate was for the diagnosis of IPF (81%), the commonest diagnosis provided by referring physicians. In the seminal study by Flaherty et al. [
Strict adherence to diagnostic criteria maximizes diagnostic certainty but in clinical practice may result in a substantial number of unclassifiable cases, potentially leading to delay in diagnosis and treatment. In current practice, when diagnostic criteria are not fully met for a confident ILD diagnosis, expert assessment may be able to provide a leading diagnosis that is more likely than not based on clinical judgement. For such cases, an International Working Group has recently proposed the term “provisional” diagnosis. In our cohort, 12% of the patients did not fulfill diagnostic criteria and received a provisional diagnosis considered adequate to directed management. Including these cases, the percentage of patients with ILD diagnoses post-MDD raised from 62% to 73%.
Cases that cannot be given a confident or provisional diagnosis even after comprehensive MDD are considered unclassifiable ILD [
MDD altered management in 39% of our patients, mostly initiation of medical therapy. In 6 cases, MDD led to change or termination of medical treatment deemed inappropriate, including one patient receiving immunosuppressant with no clinical or radiological evidence of ILD upon MDD. While comprising a relatively small proportion of patients, the implication of this in an individual case cannot be underestimated. It often affected management even when diagnosis remained the same and for cases deemed unclassifiable after initial MDD. Interestingly, even for cases with concordant pre-MDD and MDD diagnoses, management was altered in a considerable proportion (46%) of patients emphasizing the central role of expert opinion not only for diagnosis but for institution of appropriate treatment, remarkably antifibrotic drugs for IPF and immunosuppressive/anti-inflammatory therapy for non-IPF ILD. Change or initiation of nonpharmacologic management was common and included referral to smoking-cessation programs and supportive treatment (supplemental oxygen and pulmonary rehabilitation).
This study has several limitations, notably the inherent limitations of a retrospective single-center design, including selection and referral bias. The study included only one tertiary center; thus, MDD diagnosis was not validated and not assessed for inter-MDD agreement. Due to the nature of MDD diagnosis, while our study assessed changes in diagnosis, it cannot assess the accuracy of MDD diagnosis. Another limitation is the relative small number of patients with surgical biopsy which otherwise might have increased the rate of specific diagnoses. Finally, there was no longitudinal follow up to evaluate disease behavior and change in clinical or radiological patterns that might have altered diagnosis and management. Thus, the effect of MDD in outcome was not assessed. Nonetheless, we believe our study strengthens the importance of MDD in patient diagnosis and management decisions with key real-life implications including biopsy referral, initiation, and change of therapies.
In summary, MDD provided a specific diagnosis discordant with pre-MDD diagnosis in a significant proportion of cases and was particularly valuable in the diagnosis of non-IPF ILD. MDD altered management in the majority of patients and had relevant impact on management even in cases with concordant pre-MDD diagnosis.
The data used to support the findings of this study are restricted by The Ottawa Hospital Research Ethics Board in order to protect patients’ privacy. Data are available from the corresponding author for researchers who meet the criteria for access to confidential data.
Ghofran Ageely, Nha Voduc, and Saly Zahra declare that they have no conflicts of interest. Carolina Souza received consultant fees and honorarium from Pfizer, Boehringer–Ingelheim, AstraZeneca, and Hoffmann-La Roche, an educational grant from Boehringer–Ingelheim, is a member of the advisory board of AstraZeneca and Boehringer–Ingelheim. Kaissa De Boer received consultant fees and honorarium from Boehringer–Ingelheim and Hoffmann-La Roche and chair funding support from Boehringer–Ingelheim. Marcio Gomes received educational grants from Pfizer, BMS, Merck, AstraZeneca, Roche, Eli Lilly, and Boehringer–Ingelheim and honoraria from Amgen, AstraZeneca, Merck, and Roche.