Unpaid scientific collaborator & advisor with Veracyte, Inc.

Diagnosis and Classification of Idiopathic Interstitial Pneumonias: Role of Histopathology in the Golden Age of Consensus Jeffrey L. Myers, M.D. A. James French Professor of Diagnostic Pathology Vice Chair for Clinical Affairs & Quality Director, MLabs University of Michigan, Ann Arbor, MI myerjeff@med.umich.edu Unpaid scientific collaborator & advisor with Veracyte, Inc. Diagnosis and Classification of Idiopathic Interstitial Pneumonias: Role of Histopathology in the Golden Age of Consensus At the conclusion of this talk those who were actively engaged will be able to, diagnose histologically classic usual interstitial pneumonia (UIP) in surgical lung biopsies, diagnose histologically classic nonspecific interstitial pneumonia (NSIP) in surgical lung biopsies, articulate the significance of a histologic diagnosis of UIP or NSIP in the context of multidisciplinary review

Diffuse Parenchymal Lung Diseases DPLD of known cause or association (eg, drugs, CTD, occupational/environmental) Other radiologically distinct DPLDs (eg, LCH, LAM) Granulomatous DPLD (eg, HP, sarcoidosis) Idiopathic interstitial pneumonias clinical & radiology pathology Idiopathic Interstitial Pneumonias Pathology Based Classification 2016 chronic fibrosing IPs smokingrelated IPs acute/ subacute IPs CPR Diagnosis idiopathic pulmonary fibrosis (IPF) idiopathic nonspecific IP (NSIP) respiratory bronchiolitis interstitial lung disease (RBILD) desquamative IP (DIP) cryptogenic organizing pneumonia acute IP (AIP) Morphologic Pattern usual IP (UIP) nonspecific IP (NSIP) respiratory bronchiolitis desquamative IP (DIP) organizing pneumonia diffuse alveolar damage (DAD) Travis et al. Am J Resp Crit Care Med 2013; 188: 733 Idiopathic Interstitial Pneumonias Evidence Based Guidelines for Diagnosis & Management a specific form of chronic, progressive fibrosing interstitial pneumonia of unknown cause, occurring in older adults, limited to the lungs, and associated with the histopathologic and/or radiologic pattern of UIP Raghu et al. Am J Respir Crit Care Med 2011; 183: 788-824.

Diagnostic Process Diagnostic Process Usual Interstitial Pneumonia (UIP) Histologic Criteria fibrosis heterogeneity (variegated pattern) geographic heterogeneity temporal heterogeneity

temporal heterogeneity acute chronic fibroblast focus Multiple microscopic foci of injury occurring over time pro-collagen Focal fibroblast proliferation (fibroblast foci) collagen IV Recurrent microscopic injury Alveolar collapse, collagen deposition, architectural distortion Progressive clinical course Death Fibroblast Foci & The Pathogenesis of UIP

Usual Interstitial Pneumonia (UIP) Histologic Criteria fibrosis heterogeneity (variegated pattern) geographic heterogeneity temporal heterogeneity architectural distortion honeycomb change fibrotic scarring peripheral/subpleural accentuation fibroblast foci geographic heterogeneity temporal heterogeneity architectural distortion peripheral distribution fibroblast foci geographic heterogeneity TBBx? temporal heterogeneity architectural distortion peripheral distribution

Usual Interstitial Pneumonia Role for Transbronchial Lung Biopsies? 7 (32%) diagnostic of UIP (ie, patchwork fibrosis + ff and/or HCC) 22 patients with UIP/IPF HRCT = UIP (4) SLBx = UIP (19) explants = UIP (2) 2 (9%) consistent with UIP (ie, fibrosis + ff and/or HCC) Berbescu et al. CHEST 2006; 129: 1126-31 13 (59%) nonspecific (9) or insufficient (4) Usual Interstitial Pneumonia Role for Transbronchial Lung Biopsies? patchwork fibrosis + honeycomb change 3/32 (9.4%) TBBx diagnostic of UIP The primary and main role of transbronchial biopsy in diagnosis of usual interstitial pneumonia is thought not to confirm a diagnosis, but to exclude other infiltrative diseases, such as malignancy, sarcoidosis or infections.

MultiDisciplinary Discussion MultiDisciplinary Discussion Bell Sympoisum: Idiopathic Interstitial Pneumonias - Myers Pathologists 1 & 2 Respir Res 2012; 13: 96 TBB criteria for UIP patchy fibrosis fibroblast foci honeycomb change 1 = compatible with UIP even modest amount of patchy interstitial fibrosis, fibroblast foci, honeycomb changes detected on TBB can be highly predictive of a UIP pattern. Conversely, the absence of UIP histopathologic criteria on TBB does not rule out UIP. Sensitivity Specificity PPV NPV Pathologist 1* 30% (12/40) 100% (24/24) 100% (12/12) 46% (24/52) Pathologist 2* 30% (12/40) 92% (22/24) 86% (12/14) 55% (22/40) *kappa coefficient of agreement 0.61 (95% CI 0.31-0.91) Usual Interstitial Pneumonia Role for Transbronchial Lung Biopsies? History & HRCT review TBBx results SLBx? Final diagnosis Independent HRCT review (3 radiologists) Independent TBBx review (4 pathologists) n = 33 (30 UM, 3 UCLA) HRCT, TBBx, SLBx for ILD UIP 2 patchwork fibrosis, HNCB, FF Sheth et al. CHEST (2016), doi: 10.1016/j.chest2016.09.028. Usual Interstitial Pneumonia Role for Transbronchial Lung Biopsies? History & HRCT review Independent HRCT review (3 radiologists) SLBx results Final diagnosis Independent SLBx review (1 pathologist) 5 months after TBBx MDD no TBBx results included order of review shuffled Sheth et al. CHEST (2016), doi: 10.1016/j.chest2016.09.028.

Usual Interstitial Pneumonia Role for Transbronchial Lung Biopsies? All Patients n=33 No SLB n=10 Undecided n=2 Yes SLB n=21 Definite HRCT n=4 No Definite HRCT n=6 No Definite HRCT n=2 No Definite HRCT TBB & SLB Agree n=4 TBB & SLB Agree n=6 TBB & SLB Agree n=2 TBB & SLB Agree n=7 TBB & SLB Incongruent n=14 Sheth et al. CHEST (2016), doi: 10.1016/j.chest2016.09.028. Usual Interstitial Pneumonia Role for Transbronchial Lung Biopsies? TBB specimens n=33 Adequate n=27 Inadequate n=6 Diagnostic n=13 Nondiagnostic n=14 UIP n=9 OP n=1 HP n=3 Diagnostic Performance for UIP All Cases Inadequate excluded Sensitivity 33% (5/15) 50% (5/10) Specificity* 78% (14/18) 76% (13/17) PPV* 56% (5/9) 56% (5/9) NPV 58% (14/24) 72% (13/18) *4 false positives include 2 fibrotic HP with UIP-like features Sheth et al. CHEST (2016), doi: 10.1016/j.chest2016.09.028. Diffuse Parenchymal Lung Diseases Role for Transbronchial Biopsy DPLD Sensitivity Specificity Comment Histologically Distinct DPLDs LCH variable yield Granulomatous DPLDs LAM HP sarcoidosis highest diagnostic value Idiopathic Interstitial Pneumonias UIP/IPF lowest diagnostic non-uip/ipf value

Diffuse Parenchymal Lung Diseases Role for Transbronchial Biopsy DPLD Sensitivity Specificity Comment Diagnostic value of TBB in idiopathic interstitial pneumonias is limited to a small subset of patients with UIP, but can never provide specific support for diagnoses of NSIP, RBILD, or DIP. Idiopathic Interstitial Pneumonias UIP/IPF lowest diagnostic non-uip/ipf value Diagnostic Process Usual Interstitial Pneumonia (UIP)/IPF Clinical Features older adults (mean age 59 years) men > women ( 2:1)

Idiopathic Pulmonary Fibrosis (IPF) Incidence/Prevalence in Medicare Beneficiaries (2000-2011) All patients Age groups 66-69 70-74 75-79 80 Female Male Unadjusted incidence 93.7 (91.9-95.4) 63.3 (61.1-65.6) 95.2 (91.7-98.9) 118.7 (114.2-123.3) 129.8 (125.3-134.5) 86.1 (84.0-88.3) 104.8 (102.0-107.7) Adjusted incidence rate 1.00 (reference) 1.53 (1.45-1.61) 1.92 (1.82-2.02) 2.14 (2.04-2.26) 1.00 (reference) 1.29 (1.24-1.33) Raghu et al. Lancet Respir Med 2014; 2: 566-72 p value -- <.0001 <.0001 <.0001 -- <.0001 Usual Interstitial Pneumonia (UIP)/IPF Clinical Features older adults (mean age 59 years) men > women ( 2:1) insidious onset dyspnea, dry cough PFTs: restriction, DLco, hypoxemia Idiopathic Interstitial Pneumonias Evidence Based Guidelines for Diagnosis & Management a specific form of chronic, progressive fibrosing interstitial pneumonia of unknown cause, occurring in older adults, limited to the lungs, and associated with the histopathologic and/or radiologic pattern of UIP Raghu et al. Am J Respir Crit Care Med 2011; 183: 788-824.

Survival (%) Bell Sympoisum: Idiopathic Interstitial Pneumonias - Myers Usual Interstitial Pneumonia (UIP)/IPF Radiological Findings subpleural, basilar reticulations honeycombing ± traction br ectasis minimal ground glass Usual Interstitial Pneumonia (UIP)/IPF Clinical Features older adults (mean age 59 years) men > women ( 2:1) insidious onset dyspnea, dry cough PFTs: restriction, DLco, hypoxemia poor prognosis (mortality 65%) (mean survival 3 years) Usual Interstitial Pneumonia (UIP)/IPF Survival After Diagnosis 100 80 60 mean age 58 years M:F 2:1 40 20 0 2.8 yrs median survival 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 Bjoraker et al.: AJRCCM 1998 Years following diagnosis UIP

Usual Interstitial Pneumonia (UIP)/IPF Treatment Lung Transplantation Patients listed 98 Waiting list 89 Removed from waiting list 9 Elective 57 High Urgency* 32 Survival After Lung Txplt median survival 120 months (10 years) Dead 22 Actively waiting 8 Transplanted 52 Dead 7 *HU criteria ICU in txplt center Acute life-threatening Imminent need for controlled ventilation OR Assisted/controlled ventilation ten Klooster et al. Lung 2015; 193: 919-26 Usual Interstitial Pneumonia (UIP)/IPF Treatment Anti-fibrotic Agents $96K Patients receiving nintedanib had significant reductions in the rate of decline in forced vital capacity (FVC) at 1 year. trend toward a reduced rate of death that mirrored the reduced rate of decline in lung function. pirfenidone met the primary end point by showing significant reduction in the 1-year rate of decline in FVC may also reduce mortality in patients with idiopathic pulmonary fibrosis $94K The studies provide little insight into the use of these drugs in patients with more severe disease (FVC <50% of the predicted value) or with an acute disease exacerbation. Hunninghake. A new hope for idiopathic pulmonary fibrosis. NEJM 2014 Sixty-four cases of interstitial pneumonia were identified that could not be classified into one of the three main categories of idiopathic interstitial pneumonia. Nonspecific interstitial pneumonia must be separated from the three main forms of idiopathic interstitial pneumonia because of better prognosis and different treatment options.

It should not be considered a specific disease, however, because it may have varying etiologies...; less often it may reflect a nonrepresentative biopsy of another process. Nonspecific Interstitial Pneumonia Histologic Criteria chronic interstitial pneumonia with ( fibrotic NSIP ) or without ( cellular NSIP ) fibrosis temporally uniform without architectural distortion lacking other features (e.g. granulomas, pigmented [ smoker s ] alveolar histiocytes) to allow more specific classification Katzenstein & Fiorelli. Am J Surg Pathol 1994; 18: 136 cellular NSIP

fibrotic NSIP fibrotic NSIP NSIP UIP

Diagnosis and Classification of Idiopathic Interstitial Pneumonias: Role of Histopathology in the Golden Age of Consensus At the conclusion of this talk those who were actively engaged will be able to, diagnose histologically classic usual interstitial pneumonia (UIP) in surgical lung biopsies, diagnose histologically classic nonspecific interstitial pneumonia (NSIP) in surgical lung biopsies, articulate the significance of a histologic diagnosis of UIP or NSIP in the context of multidisciplinary review Idiopathic Nonspecific Interstitial Pneumonia Definition The final diagnosis in the 67 [of 193] cases was established when, the surgical lung biopsy showed a NSIP pattern (cellular or fibrosing); the HRCT showed a pattern consistent with NSIP and not diagnostic of other entities such as UIP or chronic hypersensitivity pneumonitis; and there were no clinical features of another chronic ILD, such as collagen vascular disease, drug, or inhaled antigen exposure at the time of diagnosis. Travis et al. Am J Resp Crit Care Med 2008; 177: 1338 Idiopathic Nonspecific Interstitial Pneumonia Definition The final diagnosis in the 67 [of 193] cases was established when, the surgical lung biopsy showed a NSIP pattern (cellular or fibrosing); the HRCT showed a pattern consistent with NSIP and not diagnostic of other entities such as UIP or chronic hypersensitivity pneumonitis; and there were no clinical features of another chronic ILD, such as collagen vascular disease, drug, or inhaled antigen exposure at the time of diagnosis. Travis et al. Am J Resp Crit Care Med 2008; 177: 1338

Histopathologic Variability in UIP/IPF 109 patients with UIP or NSIP who underwent biopsy of 2 lobes concordant UIP (UIP + UIP) 51 NSIP (NSIP + NSIP) 30 discordant UIP (UIP + NSIP) 28 Flaherty et al. Am J Respir Crit Care Med 2001; 164: 1722 Histopathologic Variability in UIP/IPF 1.0 0.8 NSIP Cumulative proportion surviving 0.6 0.4 Discordant UIP 0.2 Concordant UIP 0.0 0 1 2 3 4 5 6 7 8 9 Years Flaherty et al. Am J Respir Crit Care Med 2001; 164: 1722 Histopathologic Variability in UIP/IPF concordant UIP concordant NSIP discordant UIP Flaherty et al. 2001 UIP is a specific diagnosis that defines natural history key learning point NSIP is not and is always SLBx necessary a diagnosis but of insufficient exclusion for diagnosis of NSIP Monaghan et al. 2004 discordant UIP concordant NSIP concordant UIP

NSIP-like Areas Common in UIP Histologic Study of Biopsy and Explant Specimens areas resembling nonspecific interstitial pneumonia (NSIP-like areas) are present in the majority of UIP cases in both biopsy and explant specimens, and they are extensive in some. Katzenstein et al. Am J Surg Pathol 2002; 26: 1567 Idiopathic Nonspecific Interstitial Pneumonia Diagnosis of Exclusion SLBx = UIP NSIP HP fibrotic/ healed LCH late/persistent/ resolving DAD? systemic CVD? non-diagnostic/ non-classifiable lesions SRIF? NSIP Idiopathic Nonspecific Interstitial Pneumonia Diagnosis of Exclusion our CRP consensus review revealed diagnoses of hypersensitivity pneumonitis, organizing pneumonia, and UIP in cases in which lung biopsies showed a definite or probable NSIP pattern. Travis et al. Am J Resp Crit Care Med 2008; 177: 1338

Idiopathic Nonspecific Interstitial Pneumonia Clinical Features women > men (2:1) mean age 52 yrs 69% never smokers N = 67 5 yrs 82.3% 10 yrs 73.2% Travis et al. Am J Resp Crit Care Med 2008; 177: 1338 Usual Interstitial Pneumonia/IPF Role of HRCT in Separating from NSIP Histologic Diagnosis HRCT Dx UIP NSIP Total UIP* 27 0 27 *includes definite and probable diagnoses data from Flaherty et al. Thorax 2003; 58: 143 Usual Interstitial Pneumonia/IPF Role of HRCT in Separating from NSIP Histologic Diagnosis HRCT Dx UIP NSIP Total UIP* 27 0 27 Indeterminate 20 5 25 *includes definite and probable diagnoses data from Flaherty et al. Thorax 2003; 58: 143

Radiological Honeycombing Bell Sympoisum: Idiopathic Interstitial Pneumonias - Myers Usual Interstitial Pneumonia/IPF Role of HRCT in Separating from NSIP Histologic Diagnosis HRCT Dx UIP NSIP Total UIP* 27 0 27 Indeterminate 20 5 25 NSIP* 26 (59%) 18 44 *includes definite and probable diagnoses data from Flaherty et al. Thorax 2003; 58: 143 Honeycombing in UIP Histopathology Radiology Microscopic Honeycombing Absent Rare Present None 8 18 33 Probable 0 2 7 Definite 8 12 28 CT/SLBx concordance 50% 49% Chung et al. CT scan findings of probable usual interstitial pneumonitis have a high predictive value for histologic usual interstitial pneumonitis. CHEST 2015; 147: 450 Usual Interstitial Pneumonia/IPF Role of HRCT in Separating from NSIP Histologic Diagnosis HRCT Dx UIP confident/ UIP* 27 accurate HRCT diagnosis of UIP Indeterminate 20 in 27 (37%) of 73 NSIP* 26 (59%) patients *includes definite and probable diagnoses data from Flaherty et al. Thorax 2003; 58: 143

AJRCCM 2004; 170: 904-10 n = 58 consecutive patients independent pathology review before step 1 weekend in Michigan AJRCCM 2004; 170: 904-10 Frequency of UIP/IPF diagnosis Clinician A Clinician B Clinician C Step 1 28 17 24 κ 0.75 Step 2 28 19 24 κ 0.86 κ 0.73 κ 0.75 Step 3 24 23 24 κ 0.80 κ 0.92 Frequency of UIP/IPF diagnosis Clinician A Clinician B Clinician C Step 1 28 17 24 Step 2 28 19 24 Step 3 24 23 24 SLBx results κ 0.75 κ 0.86 AJRCCM 2004; 170: 904-10 Step 4 32 29 29 κ 0.92 κ 0.73 κ 0.75 κ 0.89 Step 5 31 29 30 κ 0.80 κ 0.92 κ 0.60 κ 0.55 κ 0.50 κ 0.79

Consensus Path Dx Radiologist A AJRCCM 2004; 170: 904-10 Frequency of UIP/IPF diagnosis Radiologist B Step 1 30 15 15 κ 0.92 Step 2 30 15 15 κ 0.95 Step 3 30 15 16 κ 0.90 κ 0.98 Frequency of UIP/IPF diagnosis Consensus Path Dx Radiologist A Radiologist B Step 1 30 15 15 Step 2 30 15 15 Step 3 30 15 16 SLBx results AJRCCM 2004; 170: 904-10 κ 0.92 κ 0.95 Step 4 30 32 30 Step 5 30 30 30 κ 0.90 κ 0.98 κ 0.20 κ 0.16 κ 0.84 κ 0.88 HRCT SLBx clinical data

Histopathologic Diagnosis in Diffuse Lung Disease An Ailing Gold Standard The recent view that a final diagnosis should be made by consensus between histopathologist, radiologist, and clinician is a radical departure from the diagnostic thinking of the late twentieth century. Athol Wells. Am J Respir Crit Care Med 2004; 170: 828 8 Goals to Improve Diagnosis and Reduce Diagnostic Error Facilitate more effective teamwork in the diagnostic process among health care professionals, patients, and their families Enhance health care professional education and training in the diagnostic process Ensure that health information technologies support patients and health care professionals in the diagnostic process Develop and deploy approaches to identify, learn from, and reduce diagnostic errors and near misses in clinical practice Establish a work system and culture that supports the diagnostic process and improvements in diagnostic performance Develop a reporting environment and medical liability system that facilitates improved diagnosis through learning from diagnostic errors and near misses Design a payment and care delivery environment that supports the diagnostic process Provide dedicated funding for research on the diagnostic process and diagnostic errors Diagnostic Process

Intra-observer Kappa Bell Sympoisum: Idiopathic Interstitial Pneumonias - Myers AJRCCM 2004; 170: 904-10 1 0.8 consensus almost perfect substantial 0.6 HRCT + history discussion moderate 0.4 0.2 0 Clin A Clin B Clin C Rad A Rad B SLBx results Step 2 Step 3 Step 4 Step 5 fair slight... Lancet Respir Med 2016 N = 70; 22 (31%) SLBx independent review by clinicians, radiologists, pathologists 7 multidisciplinary team meetings (MDTM) 4/7 non-significant increases toward greater prognostic separation for MDTM same analysis not significant for 5 pathologists small subgroup size (22) low IPF prevalence (15/154 diagnoses) Diagnostic Process

Diagnostic Process analytical models (slow) non-analytical models (fast) heuristics cognitive strategies/ mental shortcuts that are automatically & unconsciously employed Diagnostic Process analytical models (slow) non-analytical models (fast) representativeness bias tendency to make decisions based on a typical case overconfidence bias tendency to believe that we know more than we do search satisficing (premature closure) tendency to accept the first answer that comes along that explains the facts at hand, without considering whether there might be a different or better solution multidisciplinary discussion analytical models (slow) non-analytical models (fast) heuristics cognitive strategies/ mental shortcuts that are automatically & unconsciously employed

Test Case #1 Diagnostic Process Clinical correlation is recommended Diagnostic Process

Test Case #1 History 77-year-old man chest pain; no cough or shortness of breath bibasilar crackles on exam abnormal CT scan discovered in course of evaluation no relevant environmental exposures HRCT NSIP-like pattern representativeness bias? search satisficing (premature closure)? overconfidence bias? Diagnostic Process

Test Case #2 77-year-old man Test Case #2 History chest pain; no cough or shortness of breath bibasilar crackles on exam abnormal CT scan discovered in course of evaluation no relevant environmental exposures HRCT Definite UIP pattern with lower lobe predominant reticulations, traction bronchiectasis, and bibasilar honeycombing representativeness bias? search satisficing (premature closure)? overconfidence bias?

suspected UIP/IPF identifiable causes? DPLD of known cause or no association HRCT (eg, drugs, CTD, ±/not UIP occupational/environmental) Other clinically/radiologically surgical Not UIP distinct DPLDs Other histologically distinct lung biopsy ± UIP/NSIP DPLDs (eg, LCH, LAM) non-classifiable (eg, LCH, NSIP) fibrosis Granulomatous UIP DPLD multidisciplinary Granulomatous DPLD (eg, HP, sarcoidosis) review (eg, HP, sarcoidosis) UIP yes IPF IPF/Not IPF Not IPF! A subset (majority?) of patients with SLBx diagnoses of UIP/IPF will have an NSIPpattern on HRCT. A subset of patients (minority) with UIP/IPF will have SLBx diagnoses of NSIP ( pattern ) as a consequence of sampling bias. The sampling bias inherent in lung biopsies for diffuse lung disease is proportionally worse in patients with smaller closed (i.e., transbronchial forceps & cryo) biopsies. The smaller the biopsy the greater the risk!

A subset (majority?) of patients with SLBx diagnoses of UIP/IPF will have an NSIPpattern on HRCT. A subset of patients (minority) with UIP/IPF will have SLBx diagnoses of NSIP ( pattern ) as a consequence of sampling bias. Therefore, a subset of patients with UIP/IPF will be diagnosed as idiopathic NSIP even after multidisciplinary review. Lung, [site], wedge biopsy: UIP... sufficient to establish the diagnosis of, although not the clinical context for, UIP. Lung, [site], wedge biopsy: NSIP... necessary but insufficient to establish the diagnosis of NSIP.