Computers can calculate faster than humans as well as beat the best players at chess and Jeopardy, but can they describe a picture and explain its meaning? With the advent of the computed tomography (CT) scanner more than four decades ago, we now routinely look inside the living body for pathology. Emphysema was initially a disease that was only diagnosed at autopsy. With the use of CT scanning, we gained the technology to diagnose emphysema in the living being. It became apparent on CT that areas of low attenuation identified parenchymal destruction . The next step was to determine the cutoff value that defined emphysema on CT and correlated with the lung pathology , and then to calculate the extent and location of the disease. It is customary for radiologists to qualitatively score emphysema on a multipoint ordinal scale (ie, 0–5). Advances in imaging and software analysis have incorporated algorithms that can count the number of pixels below a specific density cutoff. Newer software programs can identify areas of low attenuation clusters to better define the extent and distribution of the disease.
However, just because we can measure something more accurately does not mean we have improved the system. As Malcolm Gladwell points out in his book Blink , experts can look at an object and immediately recognize when there is something amiss . There are obvious and nuanced changes in an image or an object that cannot be calculated but are quickly “seen” by an expert. Park et al showed that visual CT scoring by radiologists was better correlated with pulmonary function tests than an automated system . Can we teach computers to see these changes?
In the current issue of Academic Radiology , Gietema et al examine which factors radiologists take into account when estimating emphysema severity, and they assessed quantitative CT measurements of low attenuation areas .
This work is part of the ongoing study The Evaluation of chronic obstructive pulmonary disease (COPD) Longitudinally to Identify Predictive Surrogate End-points (ECLIPSE). ECLIPSE is a 3-year longitudinal study . One of the goals is to identify parameters that predict disease progression specific to defined COPD subtypes .
The current work shows that quantitative analysis of how these low attenuation areas are clustered has the highest odds of predicting an increase in the visual score assigned by the radiologist. Furthermore, the type of emphysema (centrilobular or noncentrilobular), the distribution (upper, lower, or diffuse), and the radiologist’s judgment as to whether the decreased attenuation was caused by “emphysematous destruction,” or what was considered to be “small airways disease,” also predicted the severity score assigned by the radiologist.
Comparing subjective and objective measures of lung disease is not new . That the software is good at matching the emphysema scores of the radiologists or vice versa may not be as important as the fact that we are advancing from simply quantifying the extent of the disease to developing a better understanding of what radiologists look for when they diagnose emphysema (eg, clusters, the type, and the distribution of the emphysema to assign a severity score). Others have applied texture-based algorithms to CT images to improve the ability to detect and quantify emphysema . The next step will be to use all the information to determine which are best related to the disease process and to patient outcomes. Clinically, these patients present with airway obstruction, increased lung volumes, and decreased diffusing capacities. Which CT patterns will best determine the clinical prognosis ?
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
References
1. Gould G.A., MacNee W., McLean A., et. al.: CT measurements of lung density in life can quantitate distal airspace enlargement–an essential defining feature of human emphysema. Am Rev Respir Dis 1988; 137: pp. 380-392.
2. Hayhurst M.D., MacNee W., Flenley D.C., et. al.: Diagnosis of pulmonary emphysema by computerised tomography. Lancet 1984; 2: pp. 320-322.
3. Muller N.L., Staples C.A., Miller R.R., et. al.: “Density mask”. An objective method to quantitate emphysema using computed tomography. Chest 1988; 94: pp. 782-787.
4. Coxson H.O., Rogers R.M., Whittall K.P., et. al.: A quantification of the lung surface area in emphysema using computed tomography. Am J Respir Crit Care Med 1999; 159: pp. 851-856.
5. Gevenois P.A., de Maertelaer V., De Vuyst P., et. al.: Comparison of computed density and macroscopic morphometry in pulmonary emphysema. Am J Respir Crit Care Med 1995; 152: pp. 653-657.
6. Gevenois P.A., Zanen J., de Maertelaer V., et. al.: Macroscopic assessment of pulmonary emphysema by image analysis. J Clin Pathol 1995; 48: pp. 318-322.
7. Madani A., De Maertelaer V., Zanen J., et. al.: Pulmonary emphysema: radiation dose and section thickness at multidetector CT quantification–comparison with macroscopic and microscopic morphometry. Radiology 2007; 243: pp. 250-257.
8. Gladwell M. Blink. New York: Little, Brown and Company, 2005.
9. Park K.J., Bergin C.J., Clausen J.L.: Quantitation of emphysema with three-dimensional CT densitometry: comparison with two-dimensional analysis, visual emphysema scores, and pulmonary function test results. Radiology 1999; 211: pp. 541-547.
10. Gietema H.A., Muller N., Nasute Fauerbach P.V., et. al.: Quantifying the extent of emphysema: Factors associated with radiologists’ estimations and quantitative indices of emphysema severity using the ECLIPSE Cohort. Acad Radiol 2011; 18: pp. 661-671.
11. Vestbo J., Anderson W., Coxson H.O., et. al.: Evaluation of COPD Longitudinally to Identify Predictive Surrogate End-points (ECLIPSE). Eur Respir J 2008; 31: pp. 869-873.
12. Bankier A.A., De Maertelaer V., Keyzer C., et. al.: Pulmonary emphysema: subjective visual grading versus objective quantification with macroscopic morphometry and thin-section CT densitometry. Radiology 1999; 211: pp. 851-858.
13. Gurney J.W., Jones K.K., Robbins R.A., et. al.: Regional distribution of emphysema: correlation of high-resolution CT with pulmonary function tests in unselected smokers. Radiology 1992; 183: pp. 457-463.
14. Uppaluri R., Mitsa T., Sonka M., et. al.: Quantification of pulmonary emphysema from lung computed tomography images. Am J Respir Crit Care Med 1997; 156: pp. 248-254.
15. Xu Y., Sonka M., McLennan G., et. al.: MDCT-based 3-D texture classification of emphysema and early smoking related lung pathologies. IEEE Trans Med Imaging 2006; 25: pp. 464-475.
16. Han M.K., Bartholmai B., Liu L.X., et. al.: Clinical significance of radiologic characterizations in COPD. Copd 2009; 6: pp. 459-467.
17. Gietema H.A., Zanen P., Schilham A., et. al.: Distribution of emphysema in heavy smokers: impact on pulmonary function. Respir Med 2010; 104: pp. 76-82.
18. Haruna A., Muro S., Nakano Y., et. al.: CT scan findings of emphysema predict mortality in COPD. Chest 2010; 138: pp. 635-640.
19. Mair G., Miller J.J., McAllister D., et. al.: Computed tomographic emphysema distribution: relationship to clinical features in a cohort of smokers. Eur Respir J 2009; 33: pp. 536-542.
20. Agusti A., Calverley P.M., Celli B., et. al.: Characterisation of COPD heterogeneity in the ECLIPSE cohort. Respir Res 2010; 11: pp. 122.
21. de Torres J.P., Bastarrika G., Zagaceta J., et. al.: Emphysema presence, severity, and distribution has little impact on the clinical presentation of a cohort of patients with mild to moderate COPD. Chest 2011; 139: pp. 36-42.
22. Regan E.A., Hokanson J.E., Murphy J.R., et. al.: Genetic epidemiology of COPD (COPDGene) study design. Copd 2010; 7: pp. 32-43.
23. Fishman A., Martinez F., Naunheim K., et. al.: A randomized trial comparing lung-volume-reduction surgery with medical therapy for severe emphysema. N Engl J Med 2003; 348: pp. 2059-2073.