6000 Shoreline Court, Suite 300 T 844.464.LUNG (844.464.5864)South San Francisco, CA 94080 USA percepta@veracyte.com

©2016 Veracyte, Inc. All rights reserved. The Veracyte name and logo are registered trademarks of Veracyte, Inc. All other logos and names are marks of their respective owners in the United States and other countries. PC070.1.1605

A multicenter, prospective, blinded comparison of local histopathology and central pathology panel diagnoses in ILDImtiaz, U., Colby, T., Myers, J., Cleveland, K., Edejer, S., Bianchi, P., Choi, Y., Pankratz, D., Huang, J., Kennedy, G.C.

Presented at the

2016 ATS International Conference

May 13-18, 2016San Francisco, California

The ATS/ERS 20132 pathology criteria were used, with prospectively documented modifications to ensure consistency across reviewers and sample types.

A total of 96 samples from 56 patients with diagnostic local and central pathology interpretations were included in the analysis.

RESULTS• For the subset of samples with a local diagnosis of UIP (40 samples), 100%

of the cases were classified as UIP by central review. • In 21 samples (34%) with centrally reviewed diagnosis of UIP, local

pathology reported a non-UIP diagnosis (19) or a non-diagnostic result (2).

• Focusing on 96 samples from 56 patients where both local and central provided a diagnosis and defining the central diagnosis as “truth”, local pathology has a specificity of 100% and a sensitivity of 68% for UIP with an overall concordance of 80% (κ=0.62).

CONCLUSIONOur study suggests that there may be lower sensitivity for identifying ‘true positive’ UIP histopathology patterns by local pathologists when compared to a highly structured central pathology conferral process. Given that a UIP pattern is a pre-requisite for a clinical diagnosis of IPF, this discordance may limit patient access to potentially effective therapy. Furthermore, these findings suggest that a diagnostic tool with high sensitivity for the UIP pattern would be useful to physicians without access to central pathology review.

BACKGROUNDDiagnosis of interstitial lung disease (ILD) ideally requires multidisciplinary evaluation of clinical history, radiographic images, and often will be accompanied by histopathologic data from surgical lung biopsies. When radiology is ambiguous, histopathology may provide essential information to diagnose ILDs. A Usual Interstitial Pneumonia1 (UIP) pattern histologically or radiologically is required for a diagnosis of IPF diagnosis, which is associated with a much poorer prognosis compared to other ILDs.

METHODSIn a prospective, IRB-approved, sample-collection clinical protocol we collected a data set of 127 histopathology biopsy samples from 70 patients with ILD (BRAVE Trial) who underwent surgical lung biopsy, cryobiopsy, or transbronchial biopsy. Local histopathology diagnoses was annotated from the clinical records.

A central panel of 3 pathologists, blinded to the local diagnosis and the clinical findings, reviewed all the samples. Local diagnoses were compared to the consensus (2 of 2 or 2 of 3) of central review. 28 non-diagnostic samples by central review and 1 sample where cancer was co-diagnosed with an ILD were excluded from the analysis.

© 2016 Veracyte, Inc. All rights reserved. PC064.1.1605 The Veracyte name and logo are trademarks of Veracyte.

FIGURE 3.

Central Pathology Diagnostic Process

Central Pathology Diagnostic Process

Pth.1

Pth.2

Pth.3

Pth.1

Pth.2

Pth.3

Pth.1

Pth.2

Pth.3

SAMPLE A

Slide A

Blinded Review Blinded Review

SAMPLE B

Slide B

Blinded Review

PATIENT

Slide BSlide A

Pathologist 1Blinded Review

Pathologist 2Blinded Review

Pathologist 3Blinded Review

Pth.1

Pth.2

Pth.3

2/3 Agreement Disagreement

ConferralUn-blinded Review

Sample DiscardedFinal Central Pathology Diagnosis(converted to classi�cation label)

2/3 Agreement Disagreement

2/2 Agreement Disagreement

FIGURE 1.

BRAVE Studies Site MapUNITED STATES OF AMERICA EUROPE

Potential site for 2015

Sites currently enrolling patients

Potential site for 2015

Sites currently enrolling patients

FIGURE 5.

KappaCohen’s kappa coefficient is calculated for comparison among groups

• a statistic which measures inter-rater agreement for qualitive (categorical) items.

- More robust measures than simple percent agreement calculations.

Noagreement

0 0.2 0.4 0.6 0.8 1

Slightagreement

Fairagreement

Moderateagreement

Substantialagreement

Almost perfect agreement

0.62

A multicenter, prospective, blinded comparison of local histopathology and central pathology panel diagnoses in ILD Imtiaz, U., Colby, T., Myers, J., Cleveland, K., Edejer, S., Bianchi, P., Choi, Y., Pankratz, D., Huang, J., Kennedy, G.C.

Veracyte, Inc., South San Francisco, CA

Disclosure of Commercial Support and Relevant Financial Interests: Veracyte, Inc. provided all commercial support for the research presented in this study. Urooj Imtiaz, Katie Cleveland, Shella Edejer, Pauline Bianchi, Yoonha Choi, Dan Pankratz, Jing Huang and Giulia Kennedy are employees and shareholders of Veracyte, Inc.

REFERENCES1. American Thoracic Society; European Respiratory Society. Idiopathic pulmonary fibrosis: diagnosis and treatment: international consensus statement. Am J Respir

Crit Care Med 2000; 161:646–664.2. American Thoracic Society, European Respiratory Society. American Thoracic Society/European Respiratory Society International Multidisciplinary Consensus

Classification of the Idiopathic Interstitial Pneumonias. Am J Respir Crit Care Med 2002;165:277–304.

Central

UIP Not UIP

LocalUIP 40 0

Not UIP 21 35

FIGURE 4. Concordance Between Local and Central Pathologists

Non-UIP UIP

Acute Lung Injury HP Classic UIP

Bronchiolitis Lung Cancer Difficult UIP

Cellular NSIP NSIP Favor UIP

Chronic Bronchiolitis OP UIP

Conferral-disagree Pneumocystis Pneumonia

Emphysema RB

Favor HP Sarcoidosis

Favor NSIP SRIF

FIGURE 2.

Central Pathology Categorical Subtypes