International Colloquium on Lung and Airway Fibrosis • September 20-24, 2014 • Mont-Tremblant, Quebec, Canada

Structural and Functional Quantitative Imaging Techniques are Complimentary in Retrospective Analysis of PRM-151 Data in Idiopathic Pulmonary Fibrosis (IPF)B van den Blink, MD, PhD1; J Burggraaf, MD, PhD2; LD Morrison, MD3; LC Ginns, MD4; MS Wijsenbeek, MD, PhD1; M Moerland, PhD2; MR Dillingh, MSc2; B Bartholmai, MD5; R Chamberlain, PhD6; W Vos, PhD7; L Nuyttens, MSc7; JP Hanrahan, MD8; EG Trehu, MD8

1 Erasmus MC, Rotterdam, The Netherlands; 2 CHDR, Leiden, The Netherlands; 3 Duke University Hospital, Durham, NC, USA; 4 Massachusetts General Hospital, Boston, MA, USA; 5 Mayo Clinic, Rochester, MN, USA; 6 Imbio, Minneapolis, MN, USA; 7 FluidDA, Kontich, Belgium and New Brunswick, NJ, USA; 8 Promedior, Lexington, MA, USA

Conclusions / Interpretation

• In this trial, imaging technologies that assess IPF-related structural and functional lung pathology demonstrate changes consistent with well-characterized pulmonary function outcomes.

• Mean increase in FVC % predicted in subjects treated with PRM-151 was accompanied by increases in normal lung texture and lobar volumes in some subjects.

• Decline in predicted FVC % was observed in all placebo subjects, and these changes were generally associated with decreases in quantitative volumes of normal lung texture and lobar volumes in CT scans at the second time point.

• These correlations are encouraging given the limitations of retrospective applications of these technologies, and support their prospective inclusion in future clinical trials utilizing standardized methods that support both platforms.

• The imaging technologies utilized in this trial may permit more rapid and efficient assessment of treatments for IPF and other interstitial lung diseases.

References: 1) Van Den Blink B et al. A Phase I Study Of PRM-151 In Patients With Idiopathic Pulmonary Fibrosis doi:10.1164/ajrccm-conference.2013.187.1_MeetingAbstracts.A5707. 2) Zavaletta VA. Acad Radiol. 2007 Jul; 14(7):772-87. 3) Raghunath S. PLoS One. 2014; 9(3):e93229. 4) Maldonado F. Eur Respir J. 2014 Jan; 43(1):204-12. 5) De Backer L. et al. Eur Respir J. 2012 Aug;40(2):298-305. 6) Vos W et al. Respiration. 2013;86(5):393-401. 7) De Backer W et al. Eur Respir J.DOI 10.1183/09031936.00011714

Mean Changes in Clinical and Imaging Outcomes Over 8 Weeks of PRM-151 Treatment

VariablePre-Rx Mean Post-Rx Mean Δ over Rx Period Δ PRM-151

vs. PBO Δ 95% CIPRM-151 PBO PRM-151 PBO PRM-151 PBO

FVC % 77.7 66.7 82.1 61.7 4.4 -5.0 9.4 [ 4.3 , 14.5 ]

DLCO % 48.4 38.2 46.2 34.5 -2.3 -3.7 1.4 [ -4.7 , 7.5 ]

6MWD (m) 467.2 457.7 468.8 447.2 1.6 -10.5 12.1 [ -42.2 ; 66.3 ]

Non-ILA-Involved Lung (%) 72.4 65.5 69.5 52.3 -2.9 -13.2 10.3 [ -6.2 ; 26.7 ]

ILA-Involved Lung (%) 27.0 33.8 29.7 46.8 2.7 13.0 -10.3 [ -27.5 ; 6.8 ]

iV Lobar (%p) 68.9 73.5 68.3 69.8 -0.6 -3.8 3.2 [ -8.5 ; 14.9 ]

Result Qualifications• No clear dose-response among the 3 PRM-151 doses was observed.

• PRM-151 treatment-related differences for clinical outcomes were consistent in both the Rotterdam and Duke IPF subject cohorts.

• Some HRCT scans were not suitable for retrospective analysis with the specific imaging software.

• Results suggest concordance between imaging and pulmonary function changes over the course of the trial with PRM-151 therapy

• An increase in mean change in FVC % over 8 weeks for PRM-151-treated subjects was observed

• Only small changes in 6MWD and DLCO % were observed

• Extent of ILA-involved lung and lobar lung volumes also showed trend for improvement with treatment

DLCO %

PRM-151(-2.25 ± 6.28)

Placebo(-3.67 ± 4.46)

-15

-10

-5

0

5

Δ D

LCO

(%)

Non-ILA-Involved Lung (%)

PRM-151(-2.92 ± 17.15)

Placebo(-13.17 ± 11.27)

Δ N

on-IL

A-In

volv

ed L

ung

(%)

-40

-30

-20

-10

0

10

20

iV Lobar (%p)

PRM-151(-0.57 ± 8.77)

Placebo(-3.76 ± 12.92)

-20

0

20

40

Δ iV

Lob

ar (%

p)

�

�

� �

�

�

�

�

�

�

�

�

�

�

�

�

Average/Subject

6MWD (m)

PRM-151(1.58 ± 51.36)

Placebo(-10.50 ± 50.73)

-100

-50

0

50

Δ 6

MW

D (m

)

FVC %

-10

-5

0

5

10

Δ F

VC (%

p)

PRM-151(4.42 ± 5.18)

Placebo(-5.00 ± 3.90)

ILA-Involved Lung (%)

PRM-151(2.67 ± 17.53)

Placebo(13.00 ± 12.62)

-20

-10

0

10

20

30

40

Δ IL

A-In

volv

ed L

ung

(%)

Dose Group Placebo 1 mg/kg 5 mg/kg 10 mg/kg

Subject 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Screening FVC % 49 91 79 63 67 51 88 76 62 101 84 78 71 79 77 48 82 86

Δ FVC % -5 -1 -6 -4 -2 -12 13 9 1 -3 12 -1 8 4 2 0 6 2

Imbio

Pre-RxRU

RM

RL LL

LM

LURU

RM

RL LL

LM

LU RU

RM

RL LL

LM

LU RU

RM

RLLL

LM

LU RU

RM

RLLL

LM

LURU

RMRL LL

LM

LURU

RM

RLLL

LM

LURU

RM

RL LL

LM

LURU

RM

RL LL

LM

LURU

RM

RLLL

LM

LURU

RM

RLLL

LM

LU RU

RM

RL LL

LM

LURU

RM

RL LL

LM

LU RU

RM

RL LL

LM

LU RU

RM

RL LL

LM

LURU

RMRL

LLLM

LURU

RM

RLLL

LM

LURU

RM

RL LL

LM

LU

Post-RxRU

RMRLLL

LM

LURU

RM

RL LL

LM

LU RU

RM

RL LL

LM

LU RU

RM

RLLL

LM

LU RU

RM

RLLL

LM

LU RU

RM

RL LLLM

LURU

RM

RLLL

LM

LURU

RM

RL LL

LM

LURU

RM

RL LL

LM

LURU

RM

RLLL

LM

LURU

RMRL

LLLM

LU RU

RM

RL LL

LM

LU RU

RM

RL LL

LM

LURU

RM

RL LL

LM

LURU

RM

RL LL

LM

LURU

RMRL

LLLM

LURU

RM

RLLL

LM

LURU

RM

RL LL

LM

LU

Δ Normal + Mild LAA -23 -3 -8 -30 -2 -13 5 -4 -9 0 -42 -24 24 16 5 -1 -1 -4

FluidDAΔ over Rx Image Not

AvailableImage Not Available

Mean Δ iV Lobar (%p) 10.0 0.4 — -24.2 2.0 -7.1 -0.5 -1.9 10.5 -0.6 -21.6 -6.2 9.0 3.7 -5.9 3.7 — 3.6

Individual Structural Lung (Imbio) and Lobar Volume (FluidDA) Changes Following 8 Weeks of Placebo or PRM-151 Treatment

Imbio FibrosisScale

NormalGround Glass

ReticularHoneycombing

Mild LAA*Moderate LAA*

Severe LAA* (*Low Areas of Attenuation)[White perimeter] Reduction from predicted lung volume

20151050-5-10-15-20-25-30-35

Worsening Improving

FluidDA Lobar VolumeChanges (%p)

Clinical TrialThis trial was a randomized, double blind, ascending-dose design that has been described previously.1 In brief, subjects with CT or biopsy-confirmed usual interstitial pneumonia consistent with IPF were randomly allocated to placebo or PRM-151 treatment at escalating doses of 1 mg/kg, 5 mg/kg, or 10 mg/kg, administered on Days 1, 3, 5, 8, and 15 by intravenous (IV) infusion over 30 minutes. Subjects were followed for 8 weeks after the initiation of therapy, with evaluations of tolerability and safety at multiple time points. Pulmonary function test (PFT) outcomes and 6MWD were assessed multiple times over the 8 weeks, including at Screening and 8 weeks after the initiation of treatment. HRCT scans were obtained at or prior to Screening and again 8 weeks after the initiation of treatment.

Study Design

Analytical MethodsDifferences in clinical/pulmonary function and CT imaging outcomes between Screening and 8 weeks after the initiation of therapy were assessed for each subject. Pre- to post-treatment outcomes were compared for placebo vs. PRM-151-treated subjects. Data from all PRM-151-treated subjects were combined, as there was no clear dose-response or by-dose difference for subjects receiving active therapy. Differences for PRM-151 vs. placebo treated subjects were calculated with 95% confidence intervals (CI) for observed differences.

Results

Twenty-one (21) subjects with IPF participated in this trial; HRCT imaging data were available on 18 subjects. Of these, 14 were males and 4 females; mean age was 64 years. Randomized dosing was: placebo (n=6) and PRM-151 1 mg/kg (n=4), 5 mg/kg (n=5), and 10 mg/kg (n=3). Two additional subjects did not have HRCT imaging data suitable for the FluidDA imaging software (one each in the 10 mg/kg and placebo groups), limiting FluidDA results to 16 subjects.

Methods

Background

Recent advances in pulmonary computerized tomography (CT) imaging technologies may help to identify effective therapies for interstitial lung disease (ILD). Two such techniques were evaluated retrospectively in a randomized, placebo-controlled, dose-escalation Phase 1B trial of PRM-151 (human recombinant Pentraxin 2) in idiopathic pulmonary fibrosis (IPF). This molecule selectively binds to human monocytes and preferentially directs their differentiation to regulatory macrophages, rather than pro-inflammatory or pro-fibrotic phenotypes. Clinical results previously reported from this trial1 suggested that 3 doses of PRM-151 (1 mg/kg, 5 mg/kg, 10 mg/kg) were well tolerated by subjects with IPF, and when administered 5 times over 15 days were associated with improvement in forced vital capacity (FVC) 8 weeks after the start of therapy in some subjects compared to placebo-treated subjects.

This investigation used high resolution CT (HRCT) pulmonary scans performed at Screening and Day 57 to explore PRM-151-related changes in quantification of image-based interstitial lung abnormalities (ILA) (Imbio, Minneapolis, Minnesota, USA) and lobar lung volumes (FluidDA Inc., Kontich, Belgium). These results were compared to changes in clinical outcomes including FVC, diffusion capacity (DLCO), and 6-minute walk distance (6MWD).

PRM-151 dosing: 1 mg/kg, 5 mg/kg, or 10 mg/kg or Placebo IV on Days 1, 3, 5, 8, and 15

Data used for Retrospective Quantitative Imaging Analysis

-4 -3 -2 -1 W1 W2 W3 W4 W5 W6 W7 W8 W9

Screening Observation

(S=Safety) SS S S S S SSS SS SSSSS

PFT PFT PFT PFT

6-min.walk

6-min.walk

6-min.walk

HRCT HRCT

Imbio CALIPER CT Parenchymal AnalysisImbio’s image analysis technology, CALIPER, was used for quantitative characterization of pulmonary parenchyma on CT scans obtained before and after treatment. CALIPER analysis requires volumetric scans with slice thickness ≤5 mm. The CALIPER algorithm classifies each voxel of lung parenchyma based on density characteristics and morphology: normal, ground glass opacity (GGO), reticular densities, honeycombing, or low attenuation areas (LAA; of mild, moderate, or severe subclasses). The volume of these parenchymal features can be represented in cubic centimeters or percent of total lung volume. It has been previously shown that:

• The quantitative features are comparable to but more reproducible than the semi-quantitative assessment by radiologists2

• These quantitative features correlate with physiologic features of disease severity such as DLCO, FEV1, and 6-minute walk tests3

• Overall extent of ILD features (GGO + reticular + honeycombing) and change in these features over time have independent predictive value in predicting mortality in ILD4

The current pilot study examines the extent of ILD features and non-involved parenchyma (normal + mild LAA) measured by inspiratory (TLC) chest CT scans performed at each of the two time points, and quantifies the magnitude and type of parenchymal changes during the study period.

FluidDA Functional Respiratory Imaging (FRI)In the FRI workflow, CT images are converted into 3D patient-specific quantifiable endpoints. FRI yields the following parameters:

Previously it was shown that:

• FRI is 3-8 times more sensitive than the classic PFT to evaluate treatment5,6

• FRI is optimal to understand the exact mode of action of a treatment in early clinical research7

• Changes in FRI parameters correlate with changes in lung function and changes in patient feeling6

For the current pilot study the main focus of the analysis was the expiratory iVlobe change after treatment.

• Internal airflow lobar distribution (IALD)

• Lung volume (iVlung)

• Lobe volume (iVlobe)

• Airway volume (iVaw)

• Airway resistance (iRaw)

• Lobar perfusion (BVD)

• Aerosol deposition