Skip to main content
Log in

High-resolution computed tomography to differentiate chronic diffuse interstitial lung diseases with predominant ground-glass pattern using logical analysis of data

  • Chest
  • Published:
European Radiology Aims and scope Submit manuscript

Abstract

Objectives

We evaluated the performance of high-resolution computed tomography (HRCT) to differentiate chronic diffuse interstitial lung diseases (CDILD) with predominant ground-glass pattern by using logical analysis of data (LAD).

Methods

A total of 162 patients were classified into seven categories: sarcoidosis (n = 38), connective tissue disease (n = 32), hypersensitivity pneumonitis (n = 18), drug-induced lung disease (n = 15), alveolar proteinosis (n = 12), idiopathic non-specific interstitial pneumonia (n = 10) and miscellaneous (n = 37). First, 40 CT attributes were investigated by the LAD to build up patterns characterising a category. From the association of patterns, LAD determined models specific to each CDILD. Second, data were recomputed by adding eight clinical attributes to the analysis. The 20 × 5 cross-folding method was used for validation.

Results

Models could be individualised for sarcoidosis, hypersensitivity pneumonitis, connective tissue disease and alveolar proteinosis. An additional model was individualised for drug-induced lung disease by adding clinical data. No model was demonstrated for idiopathic non-specific interstitial pneumonia and the miscellaneous category. The results showed that HRCT had a good sensitivity (≥64%) and specificity (≥78%) and a high negative predictive value (≥93%) for diseases with a model. Higher sensitivity (≥78%) and specificity (≥89%) were achieved by adding clinical data.

Conclusion

The diagnostic performance of HRCT is high and can be increased by adding clinical data.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1

Similar content being viewed by others

Abbreviations

HRCT:

high-resolution computed tomography

CDILD:

chronic diffuse interstitial lung diseases

GGO:

ground-glass opacity

References

  1. Hansell DM, Bankier AA, MacMahon H, McLoud TC, Muller NL, Remy J (2008) Glossary of terms for thoracic imaging. Fleischner Society. Radiology 246:697–711

    Article  PubMed  Google Scholar 

  2. Hansell DM, Wells AU, Padley SP, Muller NL (1996) Hypersensitivity pneumonitis: correlation of individual CT patterns with functional abnormalities. Radiology 199:123–128

    CAS  PubMed  Google Scholar 

  3. Cleverley JR, Screaton NJ, Hiorns MP, Flint JD, Muller NL (2002) Drug-induced lung disease: high-resolution CT and histological findings. Clin Radiol 57:292–299

    Article  PubMed  Google Scholar 

  4. Godwin JD, Muller NL, Takasugi JE (1988) Pulmonary alveolar proteinosis: CT findings. Radiology 169:609–613

    CAS  PubMed  Google Scholar 

  5. Kim EA, Lee KS, Johkoh T et al (2002) Interstitial lung diseases associated with collagen vascular diseases: radiologic and histopathologic findings. Radiographics 22 Spec No:S151–165

  6. MacDonald SL, Rubens MB, Hansell DM et al (2001) Nonspecific interstitial pneumonia and usual interstitial pneumonia: comparative appearances at and diagnostic accuracy of thin-section CT. Radiology 221:600–605

    Article  CAS  PubMed  Google Scholar 

  7. Miller WT Jr, Shah RM (2005) Isolated diffuse ground-glass opacity in thoracic CT: causes and clinical presentations. AJR Am J Roentgenol 184:613–622

    PubMed  Google Scholar 

  8. Shah RM, Miller W Jr. (2003) Widespread ground-glass opacity of the lung in consecutive patients undergoing CT: Does lobular distribution assist diagnosis? AJR Am J Roentgenol 180:965–968

    PubMed  Google Scholar 

  9. Engeler CE, Tashjian JH, Trenkner SW, Walsh JW (1993) Ground-glass opacity of the lung parenchyma: a guide to analysis with high-resolution CT. AJR Am J Roentgenol 160:249–251

    CAS  PubMed  Google Scholar 

  10. Leung AN, Miller RR, Muller NL (1993) Parenchymal opacification in chronic infiltrative lung diseases: CT-pathologic correlation. Radiology 188:209–214

    CAS  PubMed  Google Scholar 

  11. Collins J, Stern EJ (1997) Ground-glass opacity at CT: the ABCs. AJR Am J Roentgenol 169:355–367

    CAS  PubMed  Google Scholar 

  12. Nowers K, Rasband JD, Berges G, Gosselin M (2002) Approach to ground-glass opacification of the lung. Semin Ultrasound CT MR 23:302–323

    Article  PubMed  Google Scholar 

  13. Sundaram B, Gross BH, Martinez FJ et al (2008) Accuracy of high-resolution CT in the diagnosis of diffuse lung disease: effect of predominance and distribution of findings. AJR Am J Roentgenol. 191:1032–1039

    Article  PubMed  Google Scholar 

  14. Crama Y, Hammer PL, Ibaraki T (1988) Cause-effect relationships and partially defined Boolean functions. 16:299–325

  15. Alexe G, Alexe S, Liotta LA, Petricoin E, Reiss M, Hammer PL (2004) Ovarian cancer detection by logical analysis of proteomic data. Proteomics 4:766–783

    Article  CAS  PubMed  Google Scholar 

  16. Lauer MS, Alexe S, Pothier Snader CE, Blackstone EH, Ishwaran H, Hammer PL (2002) Use of the logical analysis of data method for assessing long-term mortality risk after exercise electrocardiography. Circulation 106:685–690

    PubMed  Google Scholar 

  17. Alexe S, Blackstone E, Hammer PL, Ishwaran H, Lauer MS, Snader CEP (2003) Coronary risk prediction by logical analysis of data. Ann Oper Res 119:15–42

    Article  Google Scholar 

  18. Hammer PL, Bonates TO (2006) Logical analysis of data—An overview: from combinatorial optimization to medical applications. Ann Oper Res 148:203–225

    Article  Google Scholar 

  19. Grenier P, Chevret S, Beigelman C, Brauner MW, Chastang C, Valeyre D (1994) Chronic diffuse infiltrative lung disease: determination of the diagnostic value of clinical data, chest radiography, and CT and Bayesian analysis. Radiology 191:383–390

    CAS  PubMed  Google Scholar 

  20. Fukushima A, Ashizawa K, Yamaguchi T et al (2004) Application of an artificial neural network to high-resolution CT: usefulness in differential diagnosis of diffuse lung disease. AJR Am J Roentgenol 183:297–305

    PubMed  Google Scholar 

  21. Statement on sarcoidosis. Joint Statement of the American Thoracic Society (ATS), the European Respiratory Society (ERS) and the World Association of Sarcoidosis and Other Granulomatous Disorders (WASOG) (1999) Am J Respir Crit Care Med 160:736–755

    Google Scholar 

  22. Masi AT, Rodnan GP, Medsger TA, Altman RD et al (1980) Preliminary criteria for the classification of systemic sclerosis (scleroderma). Arthritis Rheum 23:581–590

    Article  Google Scholar 

  23. Dalakas MC, Hohlfeld R (2003) Polymyositis and dermatomyositis. Lancet 362:971–982

    Article  CAS  PubMed  Google Scholar 

  24. Lee DM, Weinblatt ME (2001) Rheumatoid arthritis. Lancet 358:903–911

    Article  CAS  PubMed  Google Scholar 

  25. Vitali C, Bombardieri S, Jonsson R et al (2002) Classification criteria for Sjogren’s syndrome: a revised version of the European criteria proposed by the American-European Consensus Group. Ann Rheum Dis 61:554–558

    Article  CAS  PubMed  Google Scholar 

  26. Sharp GC (1987) Diagnostic criteria for classification of MCTD. In: Kasukawa R, Sharp G (eds) Mixed connective tissue diseases and antinuclear antibodies. Elsevier, Amsterdam, pp 41–47

  27. Webb WR, Müller NL, Naidich DP (2009) Diffuse pulmonary neoplasms and pulmonary lymphoproliferative diseases (Table 7.3). In: High-resolution CT of the lung, 3rd edn. Lippincott Williams & Wilkins, Philadelphia, pp 241–272

  28. Abehsera M, Valeyre D, Grenier P, Jaillet H, Battesti JP, Brauner MW (2000) Sarcoidosis with pulmonary fibrosis: CT patterns and correlation with pulmonary function. AJR Am J Roentgenol 174:1751–1757

    CAS  PubMed  Google Scholar 

  29. Alexe S, Hammer PL (2005) Pattern-based discriminants in the logical analysis of data. In: Pardalos P (ed) Data mining in biomedicine, biocomputing, vol 3. Springer, Berlin

    Google Scholar 

  30. Silva CI, Müller NL, Lynch DA et al (2008) Chronic hypersensitivity pneumonitis: differentiation from idiopathic pulmonary fibrosis and nonspecific interstitial pneumonia by using thin-section CT. Radiology 246:288–297

    PubMed  Google Scholar 

  31. Johkoh T, Itoh H, Muller NL et al (1999) Crazy-paving appearance at thin-section CT: spectrum of disease and pathologic findings. Radiology 211:155–160

    CAS  PubMed  Google Scholar 

  32. Nishimura K, Itoh H, Kitaichi M, Nagai S, Izumi T (1993) Pulmonary sarcoidosis: correlation of CT and histopathologic findings. Radiology 189:105–109

    CAS  PubMed  Google Scholar 

  33. Naccache JM, Marchand-Adam S, Kambouchner M, Guillon F, Monnet I, Girard F, Brauner M, Valeyre D (2003) Ground-glass computed tomography pattern in chronic beryllium disease: pathologic substratum and evolution. J Comput Assist Tomogr 27:496–500

    Article  PubMed  Google Scholar 

  34. Lee KN, Levin DL, Webb WR, Chen D, Storto ML, Golden JA (1997) Pulmonary alveolar proteinosis: high-resolution CT, chest radiographic, and functional correlations. Chest 111:989–995

    Article  CAS  PubMed  Google Scholar 

  35. Kinder BW, Collard HR, Koth L, Daikh DI, Wolters PJ, Elicker B, Jones KD, King TE Jr (2007) Idiopathic nonspecific interstitial pneumonia: lung manifestation of undifferentiated connective tissue disease? Am J Respir Crit Care Med 176:691–697

    Article  PubMed  Google Scholar 

  36. Witten IH, Frank E (2005) Data mining: Practical machine learning tools and techniques, 2nd edn. Morgan Kaufmann, San Francisco

    Google Scholar 

Download references

Acknowledgements

This paper is dedicated to the memory of P. Hammer who is the founder of the LAD method and who initiated N.B. and L.P.K. in its use.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Pierre-Yves Brillet.

ELECTRONIC SUPPLEMENTARY MATERIAL

Below is the link to the electronic supplementary material.

Supplementary Table 1

(XLS 75.5 KB)

Appendix

Appendix

Comparison of results obtained with the LAD with those of other classical data-mining methods

Data mining [1] is the science of extracting useful information from large data sets or databases (here it refers to hidden patterns from a large set of attributes). Most data-mining methods are supervised methods, meaning that there is a particular prespecified target variable (here it refers to the seven defined categories of diseases).

Tables 15 and 16 present the results of preliminary experiments using classical learning methods (logistic regression, neural networks, support vector machine, decision tree, nearest neighbours) available in the Weka software package (http://www.cs.waikato.ac.nz/ml/weka/). Weka is open source software including a collection of machine learning algorithms for data-mining tasks.

Table 15 Mean performance of LAD based on HRCT compared with other classical data-mining methods
Table 16 Mean performance of LAD based on HRCT and clinical data compared with other classical data-mining methods [36]

The performance measure we used is the mean between specificity and sensitivity estimated on 20 × 5 cross-folding validations. This criterion is adapted to problems where the sizes of the classes are not balanced. We observe that the LAD is either as or more efficient than other data-mining methods. This can partly be explained by our research group’s more accurate knowledge of the LAD optimisation. It is noteworthy that for drug-induced lung disease when using HRCT data alone, the performances of methods other than the LAD were close to 50%. This is equivalent to a random generation and is in agreement with the fact that no model could be derived with the LAD. Similar results were obtained for non-specific interstitial pneumonia and the miscellaneous category (data not shown).

Rights and permissions

Reprints and permissions

About this article

Cite this article

Martin, S.G., Kronek, LP., Valeyre, D. et al. High-resolution computed tomography to differentiate chronic diffuse interstitial lung diseases with predominant ground-glass pattern using logical analysis of data. Eur Radiol 20, 1297–1310 (2010). https://doi.org/10.1007/s00330-009-1671-4

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00330-009-1671-4

Keywords

Navigation