Letters To The Editor Reconstruction Algorithms and Their Influence in Emphysema CT Measurements From: Bruno Hochhegger, MD1,2 Klaus L. Irion, MD, PhD3 Edson Marchiori, MD, PhD2,4 Jose´ S. Moreira, MD, PhD1 1 Pavilha˜o Pereira Filho, Postgraduation Program in Pulmonary Sciences, Federal University of Rio Grande do Sul, Porto Alegre, Brazil; 2 Rio de Janeiro Federal University, Rio de Janeiro, Brazil; 3 The Cardiothoracic Centre–Liverpool National Health Service Trust and The Royal Liverpool University Hospital, Liverpool, United Kingdom; 4 Fluminense Federal University, Rio de Janeiro, Brazil. To The Editor: We congratulate Koyama et al (1) on their recent report. They address important topics relating to the imaging of anatomic changes associated with emphysema. However, we would like to address some limitations regarding the technical settings used in their study. The choice of raw data reconstruction algorithm (RA), or kernel, influences the quantification of emphysema (2–4). Edge-enhancing algorithms modify the Hounsfield unit (HU) values of voxels in regions in which structures with significantly different attenuation coefficients interface. The HU values will be modified in regions in which blood vessels and air, chest wall and lungs, or areas of emphysematous change and preserved lungs adjoin. High-resolution algorithms, including bone, sharp, and FC50 algorithms, were developed to ease the visual identification of the margins or edges of structures by changing the original HUs of the interface zones and attributing new values that were similar to the ones of the adjacent structure with closer HU values. This image-processing technique is very helpful for the visual analysis of the qualitative assessment of emphysema progression. However, computers do not require artificial processing of the original raw data, because they can measure the attenua-
Acad Radiol 2010; 17:674 ªAUR, 2010 doi:10.1016/j.acra.2009.12.010
674
tion value of each individual pixel from the raw data. Once the high kernel or edge-enhancing filter is applied, the computer will measure the new artificial value in each interface zone. When computed tomographic data are reconstructed with different algorithms, the mean HU value of a region is expected to stay the same, but the distribution of attenuation values and, subsequently, the measures derived from this distribution will change (3). This explains why the total lung volume is stable, while emphysema indexes and percentiles are variable between measures. Studies have shown that different RAs can greatly influence the extent of emphysema measured using a threshold cutoff value (2–4). Despite this, Koyama et al (1) used high-frequency algorithms in the automatic quantification of emphysema. Gevenois et al (5) used a high-resolution RA and reported that the best threshold correlation with histopathology was 950 HU (5). Following this methodologic description, the influence of RAs was examined by numerous investigators (2–4). In a more recent paper, Gevenois et al (6) reported the correlations between standard RA multidetector computed tomography and pathology. These investigators determined that thresholds of 950 and 970HU were best correlated with histopathology. This important technical parameter should be critically evaluated, because it may dramatically influence the quantification of emphysema.
REFERENCES 1. Koyama H, Ohno Y, Yamazaki Y, et al. Quantitative and qualitative assessments of lung destruction and pulmonary functional loss from reduceddose thin-section CT in pulmonary emphysema patients. Acad Radiol 2010; 17:163–168. 2. Kemerink GJ, Kruize HH, Lamers RJ, et al. Density resolution in quantitative computed tomography of foam and lung. Med Phys 1996; 23:1697–1708. 3. Boedeker KL, McNitt-Gray MF, Rogers SR, et al. Emphysema: effect of reconstruction algorithm on CT imaging measures. Radiology 2004; 232: 295–301. 4. Ley-Zaporozhan J, Ley S, Weinheimer O, et al. Quantitative analysis of emphysema in 3D using MDCT: influence of different reconstruction algorithms. Eur J Radiol 2008; 65:228–234. 5. Gevenois PA, De Vuyst P, de Maertelaer V, et al. Comparison of computed density and microscopic morphometry in pulmonary emphysema. Am J Respir Crit Care Med 1996; 154:187–192. 6. Madani A, Zanen J, de Maertelaer V, et al. Pulmonary emphysema: objective quantification at multi-detector row CT—comparison with macroscopic and microscopic morphometry. Radiology 2006; 238:1036–1043.