Research Article: Predicting all-cause and lung cancer mortality using emphysema score progression rate between baseline and follow-up chest CT images: A comparison of risk model performances

Date Published: February 21, 2019

Publisher: Public Library of Science

Author(s): Anton Schreuder, Colin Jacobs, Leticia Gallardo-Estrella, Mathias Prokop, Cornelia M. Schaefer-Prokop, Bram van Ginneken, Ulas Bagci.


Normalized emphysema score is a protocol-robust CT biomarker of mortality. We aimed to improve mortality prediction by including the emphysema score progression rate–its change over time–into the models.

CT scans from 6000 National Lung Screening Trial CT arm participants were included. Of these, 1810 died (445 lung cancer-specific). The remaining 4190 survivors were sampled with replacement up to 24432 to approximate the full cohort. Three overlapping subcohorts were formed which required participants to have images from specific screening rounds. Emphysema scores were obtained after resampling, normalization, and bullae cluster analysis of the original images. Base models contained solely the latest emphysema score. Progression models included emphysema score progression rate. Models were adjusted by including baseline age, sex, BMI, smoking status, smoking intensity, smoking duration, and previous COPD diagnosis. Cox proportional hazard models predicting all-cause and lung cancer mortality were compared by calculating the area under the curve per year follow-up.

In the subcohort of participants with baseline and first annual follow-up scans, the analysis was performed on 4940 participants (23227 after resampling). Area under the curve for all-cause mortality predictions of the base and progression models 6 years after baseline were 0.564 (0.564 to 0.565) and 0.569 (0.568 to 0.569) when unadjusted, and 0.704 (0.703 to 0.704) to 0.705 (0.704 to 0.705) when adjusted. The respective performances predicting lung cancer mortality were 0.638 (0.637 to 0.639) and 0.643 (0.642 to 0.644) when unadjusted, and 0.724 (0.723 to 0.725) and 0.725 (0.725 to 0.726) when adjusted.

Including emphysema score progression rate into risk models shows no clinically relevant improvement in mortality risk prediction. This is because scan normalization does not adjust for an overall change in lung density. Adjusting for changes in smoking behavior is likely required to make this a clinically useful measure of emphysema progression.

Partial Text

Visual assessment of emphysema extent in CT images of the lungs is a time-consuming process that requires training and is prone to subjectivity [1]. To solve these issues, a quantitative measure of emphysema was introduced [2]. Emphysema score (ES) is defined as the percentage voxels below -950 HU in the lungs [3, 4]. However, its value as a clinical risk predictor is limited due to its dependence on different imaging protocols and reconstruction algorithms [5, 6]; ES from different centers cannot be compared. Recently, normalized ES was introduced to overcome this issue by applying a normalization algorithm before calculating the ES [7]. It was subsequently shown using images from the National Lung Screening Trial (NLST) CT cohort that mortality risk groups based on normalized ES could better distinguish high- from low-risk individuals compared to those based on the original ES [8, 9]. From this point on in this paper, “normalized ES” is referred to as “ES” unless stated otherwise.

To our best knowledge, no other study has yet attempted to measure pulmonary emphysema progression in CT scan using quantitative methods for potential use as a measure of disease progression over time. Our study demonstrated that when a prior chest CT scan is available, calculating the EPR in addition to the latest ES showed a very minor improvement to the all-cause and lung cancer mortality prediction in a lung cancer screening population. On the other hand, when baseline patient information was available and included in the model–more specifically age, sex, weight, height, smoking history, and disease history–the value of EPR is negated completely. These results are also reflected in Table 7, which shows the predicted clinical outcomes when the population is separated into high versus low risk groups.

It is expected that the prognosis accuracy of a disease should improve considerably when the rate of disease progression is known, which can only be recorded when measurements are made at multiple time points. This was not the case in our study, where we have shown that simply calculating the change in normalized emphysema score between two CT images does not clinically improve the prediction of all-cause and lung cancer mortality. We surmised that recent smoking behavior has a crucial impact on the emphysema score: Though the normalization method remains useful for comparing CT images created using different protocols and algorithms, it does not adjust for changes in the overall lung density which, for each individual, is prone to change over time. It is expected that a future study utilizing follow-up clinical information or adjusting for the mean lung density is necessary before emphysema score progression rate may be clinically useful as a quantitative CT biomarker.




Leave a Reply

Your email address will not be published.