Breast Cancer Diagnosis Using Relational Discriminant Analysis of Malignancy-Associated Changes

Breast Cancer Diagnosis Using Relational Discriminant Analysis of Malignancy-Associated Changes

Dmitry Klyushin (Kyiv National Taras Shevchenko University, Ukraine), Natalia Boroday (National Academy of Sciences of Ukraine, Ukraine), Kateryna Golubeva (Kyiv National Taras Shevchenko University, Ukraine), Maryna Prysiazhna (Kyiv National Taras Shevchenko University, Ukraine) and Maksym Shlykov (Kyiv National Taras Shevchenko University, Ukraine)
Copyright: © 2017 |Pages: 21
DOI: 10.4018/978-1-5225-2229-4.ch002
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

The chapter is devoted to description of a novel method of breast cancer diagnostics based on the analysis of the distribution of the DNA concentration in interphase nuclei of epitheliocytes of buccal epithelium with the aid of novel algorithms of statistical machine learning, namely: novel proximity measure between multivariate samples, novel algorithm of construction of tolerance ellipsoids, novel statistical depth and novel method of multivariate ordering. In contrast to common methods of diagnostics used in oncology, this method is a non-invasive and offers a high rate of accuracy and sensitivity.
Chapter Preview
Top

Material And Method

For investigation the groups of the women suffering from the breast cancer (stages T2 and T3) and fibroadenomatosis from 25 to 53 years old (25 cases of breast cancer and 25 cases of firboadenomatosis) were taken. The scrapes were taken from the spinous layer of buccal mucous after gargling and taking down of the superficial cell layer. The smears were dried out under the room temperature and fixed during 30 min in the Nikiforov mixture. Then, the Feulgen reaction with the cold hydrolysis in 5 n. HCl during 15 min under the temperature t=21-22 °C was made. The optical density of the nuclei was registerated by cytospectrophotometer LOMO with the help of the scanning method for the wave length 575 nm and probe diameter 0.05 mkm. We investigated from 10 to 30 nuclei in each preparation. The DNA- fuchsine content in the nuclei of the epitheliocytes was defined as a product of the optical density on area. Thus, under investigation of the interphase nucleus a rectangular matrix , whose entries characterizes the DNA content in the corresponding grid cell was obtained (mostly, n and m were equal to 8 or 9).

On the basis of these cytophotospectrometric indices the following morpho- and densitometric features that characterize structural and textural peculiarities of chromatin were calculated.

  • 1.

    Area of nuclei: x1 is a number of the elements of the matrix R where rij ≥ 0.08.

  • 2.

    Area of condensed chromatin: x2 is a number of the elements of the R where rij ≥ 0.35.

  • 3.

    Area of decondensed chromatin: x3 is a number of the elements of the R where 0.08 ≤ rij < 0.35.

  • 4.

    Area of strongly decondenced chromatin: x4 is a number of the elements of the R where 0.08 ≤ rij < 0.15.

  • 5.

    Specific area of condensed chromatin:

    .

  • 6.

    Specific area of decondensed chromatin:

    .

  • 7.

    Integral density:

where the inequality rij ≥ 0.08 placing under the sum means that summation is taken over indices i and j for which rij ≥ 0.08.
  • 8.

    Mean density:

    ,

where p is a number of the elements rij < 0.08.
  • 9.

    Averaged sum of overfalls:

    ,

where q is a number of the elements such that

Complete Chapter List

Search this Book:
Reset