Chest radiography remains widely used for diagnosing and monitoring thoracic disease, but its two-dimensional format can obscure lung parenchyma where structures overlap. Dense areas such as the mediastinum and diaphragm are recognised as difficult regions for interpretation, and missed lung cancers have occurred in these zones. A prospective analysis published in Radiology: Cardiothoracic Imaging used upright multidetector row CT to estimate how much lung and lobar volume is superimposed on these structures during chest radiograph projection. The assessment included 132 asymptomatic Asian participants who underwent both upright and supine CT within 2 hours. Results indicate that upright positioning reduces the proportion of superimposed lung volume, even though total lung volume is greater when standing.
Measuring Overlap Under Upright Conditions
The approach compared thoracic anatomy in standing and supine positions during deep inspiration breath hold. Upright imaging used a prototype 320-detector-row CT scanner, while supine imaging used a conventional 320-detector-row CT scanner. Imaging settings were matched for each participant, allowing comparison between positions under similar technical conditions. Pulmonary function testing took place on the same day. Lung and lobar volumes were measured with a commercial workstation that automatically segmented lungs and lobes, with manual adjustment when the automated result was inadequate.
CT volumetric data were then converted into ray-sum projection images, which were used to approximate the frontal projection of a chest radiograph. Mediastinal and diaphragmatic contours defined the regions where lung tissue overlapped dense thoracic structures in the posteroanterior direction. These contours included major mediastinal interfaces, cardiac borders and the diaphragmatic margins. Repeated assessments in a subset of participants showed high measurement reproducibility. This method allowed the proportion of lung and lobar volume hidden behind the mediastinum and diaphragm to be quantified in both postures.
Must Read: CT-Based Model Aids Lung Lesion Diagnosis
Upright Position Reduces Most Superimposition
Total lung volume was higher in the upright position than in the supine position, yet the proportion of lung volume superimposed on the mediastinum and diaphragm was lower. The mean superimposed percentage was 23.9% when upright and 25.7% when supine. This difference was statistically significant. Lobar measurements showed that overlap was not distributed evenly across the lungs. In the upright position, the left lower lobe had the largest superimposed proportion, while the right middle lobe had the smallest.
Superimposition was lower upright than supine for both lungs and for most lobes. The exception was the left lower lobe, where the upright and supine values were similar. These results provide a more detailed view of radiographic blind spots than whole-lung assessment alone. They also show that upright imaging reduces overlap across much of the lung but does not remove the challenge entirely. The left lower lobe remains a prominent area of superimposition in both positions.
Patient Factors and Interpretation Limits
Participant characteristics were linked with variation in superimposed lung volume. Higher body mass index and older age were independently associated with greater superimposition in both upright and supine positions. Male sex was also independently associated with greater superimposition in the upright position after adjustment for pulmonary function measures. These associations indicate that the extent of overlap on frontal chest radiographs can vary according to patient characteristics. The results may support radiological education by providing quantitative reference points for regions where lung tissue is more likely to be hidden behind dense structures. They may also support optimisation of artificial intelligence-based detection systems, especially because reduced sensitivity can occur in superimposed regions.
Several constraints limit generalisability. The participants were asymptomatic, otherwise healthy and recruited at a single institution. Most had body mass index values within the normal range. The cohort also consisted of Asian participants, so results may differ in populations with other body habitus profiles, smoking histories or chronic lung disease. CT-derived projections approximate chest radiograph geometry, but do not fully reproduce conventional radiographs.
Upright multidetector row CT provides a quantitative way to assess lung regions that overlap the mediastinum and diaphragm on chest radiograph projections. The upright position shows less proportional superimposition than the supine position, despite larger total lung volume when standing. Lobar measurements identify the left lower lobe as the most consistently overlapped region. Age, body mass index and sex contribute to variation in upright superimposition. The results add anatomical context for chest radiograph interpretation, radiological training and future development of artificial intelligence-based detection systems.
Source: Radiology: Cardiothoracic Imaging
Image Credit: iStock
References:
Yamada Y, Yamada M, Chubachi S et al. (2026) Estimation of Lung Volume Superimposed on the Mediastinum and Diaphragm on Chest Radiographs Using Upright Multidetector Row CT. Radiology: Cardiothoracic Imaging; 8:3.
