Primary objective is to predict early for progression in both IPF and non-IPF ILD
population using an artificial intelligence (AI)/Machine Learning (ML) algorithm of STP
score. The primary interest is to validate STP score in identifying a cohort early for
the candidate of anti-fibrotic treatment. The study plans to collect clinical information
such as pulmonary function tests (PFT), symptom scores, 6-minute walk tests (6MWT), and
radiologic information from HRCT. This study does not intervene with patient's standard
medical care.
This proposal is a prospective study that will enroll patients from the UCLA ILD Center.
STP scores of subjects' baseline HRCT images will be grouped to one of 2 arms based on
the baseline HRCT.
A subject's allocation will be determined by the baseline HRCT scan. STP score will be
derived from the baseline HRCT to compare the early prediction of progression in ILD, STP
of 30% threshold is expected to be close to the mean of overall population. In addition,
a multi-scale guided attention (MSGA) is an imaging marker from deep learning model with
two attention models to classify an IPF-likeliness using HRCT.
In IPF, progression-free survival (PFS) is defined by the reduction of 10% or more by FVC
in volume or 15% or more by DLCO (DLCO) or death from any cause, whichever came first.
In non-IPF ILD, PFS is defined by two worsening outcomes out of three elements of PFT
worsening, radiological worsening or symptom or disease-related death alone.
Worsening in PFT is defined by 5% or more absolute decreases in the percent
predicted FVC or 10% or more absolute decrease in the percent predicted DLCO.
Radiological evidence of disease progression is defined by visual worsening (one or
more of the following) from a radiological report or quantitative lung fibrosis
(QLF) changes >=2% in whole lung
Symptomatic worsening can be measure by the modified Medical Research Council (mMRC)
Dyspnea scale or King's Brief Interstitial Lung Disease (K-BILD).
Secondary outcomes of this study are:
To compare overall survival between the two arms of STP
To compare the changes in 6-minute walk tests between the two arms of STP
To compare PFS between two groups of MSGA marker positive and negative
To compare overall survival between two groups of MSGA marker positive and negative
With a chronic ILD or IPF, lung function may be stable for a few years or continue to
deteriorate slowly or rapidly develop more scar tissues over time. While it is known that
age, biological sex, and lung function are factors that can impact risk of worsening lung
function, there is a great need for better methods to predict which patients are at risk
of worsening lung function. Having better methods to predict disease progression could
allow more timely treatment with anti-fibrotic treatment to prevent the disease
progression.
In both IPF and non-IPF ILD, HRCT scan is required for diagnosis. Imaging patterns
derived from HRCT, called STP is designed to predict the areas in lung that may be likely
to progress in the next 6 to 12 months. High STP scores are associated with poor
prognosis and worsening the pulmonary function. The goal of this study is to test whether
an AI-algorithm, the STP score from a single CT study, can predict disease progression in
subjects with IPF and non IPF-ILD in a prospective study. This AI-algorithm was developed
under NIH-sponsored study.
The purpose of prospective observational cohort study from UCLA is to test for the early
sign of progressive fibrosis using baseline HRCT. This study, Imaging Signature of
Progressive Pulmonary Fibrosis (IS-PPF) Research is a prospective study that will collect
information regarding HRCT images, pulmonary function test, 6-minute walk, symptomatic
score, and patients' clinical information to set up AI-driven imaging signature for
evaluating the STP in predicting progression in IPF and non-IPF ILD.
This is an observational study; only minimally invasive procedures will be performed with
study subjects (blood draws and nasal swabs). These biological samples will support
future research studies. The study subject's will participation in the study for up to 3
years, the length of participation may vary. All subjects will continue to receive their
usual care and treatment.
In summary, this research will create an opportunity to test and validate the imaging
score and early prediction for IPF and non-IPF ILD that can impact current and future
care practices.