Loading...
1 results
Search Results
Now showing 1 - 1 of 1
- Lung cancer screening and the impact of lung parenchyma on pulmonary nodule volumetryPublication . Pereira, Diana Graça Ramos Penha; Irion, Klaus Loureiro; Barata, Luís Manuel TabordaLung Cancer Screening (LCS) reduces lung cancer mortality, allowing for the early identification of pulmonary nodules using low-dose computed tomography (LDCT). Appropriate measurement and management of pulmonary nodules are essential in LCS programs. Guidelines recommend nodule volumetry analysis as the preferred method for nodule size measurement. However, artificial intelligence (AI) tools for volumetric analysis still present limiting factors, some of the most common being the nodule's location, size, shape and density. The study investigated whether changes in the attenuation of the lung parenchyma adjacent to a nodule affect the performance of nodule segmentation using computed tomography (CT) studies and volumetric tools. Two radiologists retrospectively applied two commercially available volumetric tools to assess lung nodules with diameters of 5– 8 mm detected by low-dose chest CT during a lung cancer screening program. The radiologists recorded the following parameters: • segmentation success and adequacy; • volume, longest and shortest diameters of the nodule; • mean attenuation value of the adjacent lung parenchyma; • presence of interstitial changes, emphysema, pleural plaques, or atelectasis. Predictors of appropriate volumetric segmentation of the nodule were assessed by regression analysis. Intraclass correlation coefficient assessed the interobserver agreement, and software-dependent appropriateness of nodule segmentation. In total, data on 1265 nodules (mean patient age, 68.3 ± 5.1 years; 70.2% male) were included in the study. The regression model revealed that the attenuation of the adjacent lung parenchyma effect in the size of the nodules was highly significant (odds ratio 0.987, p < 0.001). Interobserver and inter software agreement on appropriate segmentation were good. It was also observed that a software package performed better and that the measurements differed consistently between software packages. The conclusion from the main study showed that the likelihood of good segmentation of lung nodules with diameters of 5–8 mm declines with increasing attenuation of the adjacent parenchyma. These results support the hypothesis that lung nodule volume may not be assessed without critical analysis of the adjacent parenchyma. Numerous incidental findings were detected during the review of the cases for this study. Lack of uniformity in the approach to the incidental findings observed during the study justified the need for a separate publication proposing a more uniform approach to incidental findings on LCS studies, providing recommendations for the reporting and management, including the description and analysis of changes that increase the attenuation of the lung parenchyma. In my studies during this research, I observed that lung cancer with cystic characteristics could be misdiagnosed or overlooked and that they are not well detected or segmented by the current AI systems. I, therefore, felt appropriate to prepare an additional publication to alert the medical community regarding types of lung cancer associated with cystic airspaces. In summary, the central hypothesis that the volumetry of nodules detected on LCS can be affected by the attenuation values of the lung parenchyma surrounding the nodule was confirmed. These results have been published to alert that critical analysis of the parenchyma surrounding the nodule is needed before blindly accepting the numeric values from nodule volumetry to define stability or growth. Observation of inconsistencies in the approach to incidental findings and cystic lung cancer types led to the decision to publish the two additional papers related to this research.