Texture Analysis of Lung ImageS for Medical diagnosis AssistaNce

Interstitial lung diseases (ILDs) are a relatively heterogeneous group of around 150 different illnesses with often very unspecific symptoms. The most complete imaging method for the characterization of ILDs is the highresolution computed tomography (HRCT) of the chest, but a correct interpretation of these images is difficult even for specialists as many diseases are rare and thus little experience exists. A computerized diagnostic aid tool based on HRCT images with associated medical data to retrieve similar cases of ILDs from a dedicated database can bring quick and precious information for example for emergency radiologists. The experience from a pilot project [1] highlighted the need for detailed database containing high-quality annotations in addition to clinical data [2].

The selection of the most relevant clinical parameters was done in collaboration with lung specialists from current literature [3,4], along with knowledge bases of computer-based diagnostic decision support systems [5]. In order to perform high-quality annotations of the interstitial lung tissue in the HRCT images, a specific annotation software and its own file format was implemented for DICOM images.

Several visual features are used to characterize the texture of the lung tissue such as grey-level histograms, Gabor filters responses and features derived from cooccurence matrices. Finally, a multimodal classification with Support Vector Machines (SVM) organizes the retrieval of similar cases from a database constituted of 150 cases of typical ILDs. First results show an important contribution for diagnosing ILDs in emergency radiology.

Keywords: quantitative image analysis, database construction, content-based image retrieval, feature extraction, texture analysis, chest high-resolution CT, similar case retrieval.

References

[1] H. Müller, S. Marquis, G. Cohen, and A. Geissbuhler, “Lung CT analysis and retrieval as a diagnostic aid”, in Medical Informatics Europe (MIE 2005), pp. 453–458, (Geneva, Switzerland), August 2005.
[2] A. Depeursinge, H. Müller, A. Hidki, P.-A. Poletti, T. Rochat, and A. Geissbuhler, “Building a library of annotated pulmonary CT cases for diagnostic aid”, in Swiss conference on medical informatics (SSIM 2006), (Basel, Switzerland), April 2006.
[3] T. King, “Approach to the adult with interstitial lung diseases”, UpToDate December, 2005.
[4] A. Hidki, H. Müller, A. Depeursinge, P.-A. Poletti, and A. Geissbuhler, “Putting the image into perspective: The need for domain knowledge when performing image-based diagnostic aid”, in Swiss conference on medical informatics (SSIM 2006), (Basel, Switzerland), April 2006.
[5] C. P. Friedman, A. S. Elstein, F. M. Wolf, G. C. Murphy, T. M. Franz, P. S. Heckerling, P. L. Fine, T. M. Miller, and V. Abraham, “Enhancement of clinician’s diagnostic reasoning by computer-based consultation,” Journal of the American Medical Association 282, pp. 1851–1856, November 1999.

Members

Adrien Depeursinge, PhD Student
Henning Müller, PhD
Asmâa Hidki, MD
Pierre-Alexandre Poletti, MD
Alexandra Platon, MD

Grants

This work is supported by the Swiss National Science Foundation (FNS) with grant 200020-118638/1, the equalization fund of University and Hospitals of Geneva (grant 05-9-II) and a grant from the Swiss Confederation for the work of Asmâa Hidki.