Micro-computed tomography images of lung adenocarcinoma: detection of lepidic growth patterns.

Abstract

ABSTRACTMicro-computed tomography (µCT) provides extremely high-resolution images of samples and can be employed as a non-destructive inspection tool. Using µCT, we can obtain images comparable with microscopic images. In this work, we have attempted to take high-resolution images of the human lung using µCT. Compared to clinical high-resolution computed tomography (HRCT) images of living body (in-vivo imaging), we can obtain extremely high-resolution images by µCT of ex-vivo tissues (resected lungs) as three-dimensional data. The purpose of this study was to distinguish between areas of normal lung and lung cancer by µCT images in order to study the feasibility of cancer diagnosis using this novel radiological image modality. Ten resected human lungs containing primary cancer were fixed by Heitzman’s methods to obtain high-resolution µCT images. After fixation of the lung, images of the specimens were taken by µCT between January 2016 and November 2017. The imaging conditions were tube voltage: 90 kV and tube current: 110 µA. To compare details of images gained by conventional HRCT and µCT, we measured the thickness of the alveolar walls of the normal lung area and the cancer area of which alveoli might be replaced by tumor cells, and compared their appearance by means of histopathological images. All the nodules were diagnosed as adenocarcinoma. The median whole tumor size was 18 mm (9 mm–24 mm). Each specimen was clearly divided into areas of normal alveolar wall and of thickened alveolar wall on µCT ‘visually’. Median thickness of alveolar walls of the normal lung was 0.037 mm (0.034 mm–0.048 mm), and that of the cancer area was 0.084 mm (0.074 mm–0.094 mm); there was a statistically significant difference between both thicknesses by Student’s t-test (P < 0.01). The area of thickened alveolar walls on µCT corresponded well with the area of microscopically lepidic growth patterns of adenocarcinoma. We found that µCT images could be correctly divided by alveolar walls into normal lung area and lung cancer area. Further detailed investigations with regard to µCT are needed to make comparable histological diagnoses using µCT images with conventional microscopic methods of pathological diagnoses.