Keywords

CAD, Density index, Histogram, Frequency, Entropy

Introduction

There has been a rapid progress in the modality of CT since it was developed by Mr. Hounsfield. The earlier CT had been upgraded to the helical CT and for the time being the MDCT is widely used in the radiology practice. The up-to-dated MDCT is supported by the detectors of 64 rows or more [3-5].

The slice thickness of CT has been reached to 0.1 mm, based on the noble high resolution it reflects the indirect the pathological findings and enables working on the algorithm and image density index anticipating the pathological diagnosis. Some researchers reported the characteristics of image density in healthy lungs using the various indices including the mean value, standard deviation, maximum in the region of interesting [6-8].

Recently, the free service of MDCT is provided for diagnosis of patients at some health facilities. Taking the advantage of the high resolution, we have selected the new quantitative index of image density other than HU and obtained the basic data for CAD and prediction of the pathological findings in lung diseases [9,10].

Materials and Methods

We analyzed the CT findings of 40 males and 40 females without respiratory symptoms and abnormal X-ray findings.

The findings of target group was disaggregated by sex and analyzed.

And we selected the image density index and compare the index results of the bilateral lungs at the level of vertebrae thoracales from No. 1 to 8.

Each image density indices were evaluated when setting range of HU from 0 to 5000 instead of -1000 to 4000.

Result

1. The indices were selected to quantitatively evaluate the image density.

Density grade is Density ratio of interior half and exterior half of healthy lung

Average value of energy

M_f = S_f/L_y

S_f : Integration of quarter division of spectrum

L_y: Area of quarter division of spectrum

Average range ratio of high frequency and low frequency

C_f = H_f/L_f

C_f is the average range ratio of high frequency and low frequency of the regions in healthy lungs.

H_f: Average range of high frequency

L_f: Average range of low frequency

Histogram of parenchyma of healthy lungs

R_h = $H\max -H\min$

H_max: Maximum density of histogram

H_min: Minimum density of histogram

Average histogram value of healthy lung parenchyma

M_h = T_h/R_h

T_h: Total value of histogram

R_h: Range of histogram

Entropy

Entropy = ${\displaystyle {\sum }_{It=0}^{L-1}p\left(Zi\right)}\text{ log2P}\left(Zi\right)$

2. The value of image density indices of bilateral lungs at each level of vertebrae thoracales are shown in the tables below.

As mentioned above, the quantitative indices have been newly identified to indicate the characteristics of image density of healthy lungs. Subsequently the diagnostic algorithm was drafted to compare the image density of ill lung and healthy lung at slice levels.

According to the results of study on image density of healthy lungs at the level of vertebrae thoracales from No. 1 to 8, the significant deviation was observed in the histogram at the level of vertebrae thoracales from No. 4 to 8, though not observed in the density grade, ratio of high frequency and low frequency, average frequency, histogram and entropy.

We could note that the comparison of image density of healthy and ill lungs at the appropriate slice level would be one of key approaches to anticipate the pathological diagnosis regarding the values of the indices vary from disease to disease.

Conclusion

The quantitative indices have been newly identified to indicate the characteristics of image density of healthy lungs. The significant deviation was observed in the histogram at the level of vertebrae thoracales from No. 4 to 8, though not observed in the density grade, ratio of high frequency and low frequency, average frequency, histogram and entropy.

References

1. CR Haider, BJ Bartholmai, DR Holmes, JJ Camp, RA Robb (2005) Quantitative characterization of lung disease. Comput Med Imaging Graph 29: 555-563.
2. Isaure de Lavernhe, Alain Le Blanche, Dégrugilliers L, Carette MF, Bayat S (2015) CT density distribution analysis in patients with cystic fibrosis: Correlation with pulmonary function and radiologic scores. Acad Radiol 22: 179-185.
3. Ehiichi Kohda, Naoyuki Shigematsu (1989) Measurement of lung density by computed tomography: Implication for radiotherapy. Keio J Med 38: 454-463.
4. Shiloah Elizabeth Darmanayagam, Khanna Nehemiah Harichandran, Sunil Retmin Raj Cyril, Kannan Arputharaj (2013) A novel supervised approach for segmentation of lung parenchyma from chest CT for computer-aided diagnosis. J Digit Imaging 26: 496-509.
5. Tomohiro Hirose, Norihisa Nitta, Junji Shiraishi, Nagatani Y, Takahashi M, et al. (2008) Evaluation of computer-aided diagnosis (CAD) software for the detection of lung nodules on multidetector row computed tomography (MDCT). JAFROC study for the improvement in radiologists' diagnostic accuracy. Acad Radiol 15: 1505-1512.
6. Xu DM, van Klaveren RJ, de Bock GH, Leusveld A, Zhao Y, et al. (2008) Limited value of shape, margin and CT density in the discrimination between benign and malignant screen detected solid pulmonary nodules of the NELSON trial. Eur J Radiol 68: 347-352.
7. Howard Lee, Yi-Ping Phoebe Chen (2015) Image based computer aided diagnosis system for cancer detection. Expert Systems with Applications 42: 5356-5365.
8. Xiangrong Zhou, Tatsuro Hayashi, Takeshi Hara, Hiroshi Fujita, Ryujiro Yokoyama, et al. (2006) Automatic segmentation and recognition of anatomical lung structures from high-resolution chest CT images. Computerized Medical Imaging and Graphics 30: 299-313.
9. Pia Opulencia, David S Channin, Daniela S Raicu, Jacob D Furst (2011) Mapping LIDC, RadLex™, and lung nodule image features. J Digit Imaging 24: 256-270.
10. Shingo Iwano, Tatsuya Nakamura, Yuko Kamioka, Takeo Ishigaki (2005) Computer-aided diagnosis: A shape classiffication of pulmonary nodules imaged by high-resolution CT. Comput Med Imaging Graph 29: 565-570.

---