A new computed tomography score-based staging for melioidosis pneumonia to predict progression

Abstract

Background: Melioidosis pneumonia, caused by the bacterium Burkholderia pseudomallei, is a serious infectious disease prevalent in tropical regions. Chest computed tomography (CT) has emerged as a valuable tool for assessing the severity and progression of lung involvement in melioidosis pneumonia. However, there persists a need for the quantitative assessment of CT characteristics and staging methodologies to precisely anticipate disease progression. This study aimed to quantitatively extract CT features and evaluate a CT score-based staging system in predicting the progression of melioidosis pneumonia.

Methods: This study included 97 patients with culture-confirmed melioidosis pneumonia who presented between January 2002 and December 2021. Lung segmentation and annotation of lesions (consolidation, nodules, and cavity) were used for feature extraction. The features, including the involved area, amount, and intensity, were extracted. The CT scores of the lesion features were defined by the feature importance weight and qualitative stage of melioidosis pneumonia. Gaussian process regression (GPR) was used to predict patients with severe or critical melioidosis pneumonia according to CT scores.

Results: The melioidosis pneumonia stages included acute stage (0-7 days), subacute stage (8-28 days), and chronic stage (>28 days). In the acute stage, the CT scores of all patients ranged from 2.5 to 6.5. In the subacute stage, the CT scores for the severe and mild patients were 3.0-7.0 and 2.0-5.0, respectively. In the chronic stage, the CT score of the mild patients fluctuated approximately between 2.5 and 3.5 in a linear distribution. Consolidation was the most common type of lung lesion in those with melioidosis pneumonia. Between stages I and II, the percentage of severe scans with nodules dropped from 72.22% to 47.62% (P<0.05), and the percentage of severe scans with cavities significantly increased from 16.67% to 57.14% (P<0.05). The GPR optimization function yielded area under the receiver operating characteristic curves of 0.71 for stage I, 0.92 for stage II, and 0.87 for all stages.

Conclusions: In patients with melioidosis pneumonia, it is reasonable to divide the period (the whole progression of melioidosis pneumonia) into three stages to determine the prognosis.

Keywords: Melioidosis pneumonia; chest computed tomography (chest CT); computed tomography score (CT score); prognostic prediction; staging.

Figure 1

Framework of the modeling. This study aimed to quantitatively extract computed tomography features and evaluate a computed tomography score—based staging system in predicting the progression of melioidosis pneumonia. (A) First, for easy feature extraction, lung segmentation and the annotation of lesions (consolidation: yellow, nodules: green, and cavities: orange) were used in preprocessing. (B) Second, the features, including the involved area, amount, and intensity, for each of the three lesions were extracted. (C) Third, the computed tomography scores of the lesion features were defined by the feature importance weight. We could qualitatively stage melioidosis pneumonia into three stages. (D) Finally, we employed Gaussian process regression to predict severe or critical melioidosis pneumonia using computed tomography scores. CT, computed tomography.

Figure 2

Flowchart showing the study population and exclusion criteria for the data.

Figure 3

Examples of the lung segmentation (SCOAT-Net) and lesion annotation. For each pair of images, (A) the original computed tomography image and (B) the visualization of the lung segmentation and lesion annotation are presented (blue: consolidation; red: nodules; yellow-green: cavities). SCOAT-Net, spatial- and channel-wise coarse-to-fine attention network.

Figure 4

The normalized importance of the twelve characteristic features.

Figure 5

CT scores of the patients with melioidosis pneumonia based on chest CT scans from the time of onset of the initial symptoms. (A) Graph showing the dynamic changes in the computed tomography score for each scan. The lines between the dots indicate that the scans belong to the same patient. (B) Graph showing the scans of the patients who would not develop severe disease (linear fitting: y=−0.0058x + 3.1881, where x is the time from the onset of the initial symptoms, and y is the computed tomography score; R²=0.0656; P<0.005). (C) Graph showing the scans of the patients who would develop severe disease. CT, computed tomography; d, days.

Figure 6

The L-BFGS Gaussian process for predicting a severe outcome in patients, with the receiver operating characteristic curve for the CT score in predicting severe disease in the mild test set and severe set. (A) Posterior mean and 95% (±1.96) (dashed lines) and 99.7% (±3) credible intervals (dotted lines) estimated from the training data (gray). The test set includes the data from mild patients (blue) and the data from severe patients (yellow) and connects multiple samples belonging to the same patient. (B) Scatter plot of the Z score vs. time showing that there are different boundaries in the different stages. The Z scores tended to be higher in the patients with severe pneumonia. (C) For stages I and II, the L-BFGS yielded AUCs of 0.71 and 0.92, respectively, while OLS yielded AUCs of 0.71 and 0.88, respectively. (D) In all the scans for all stages, L-BFGS and OLS versions yielded AUCs of 0.87 and 0.85, respectively. CT, computed tomography; d, days; L-BFGS, limited-memory Broyden-Fletcher-Goldfarb-Shanno algorithm; AUC, area under the curve; OLS, ordinary least squares.