Method for Cancer Diagnosis using DNA Ploidy Analysis with a Combination of Fast Fourier Transform and Domain Method

DOI Web Site Open Access

Bibliographic Information

Other Title
  • FFT法と領域法を組み合わせたDNA ploidy解析によるがん診断法の研究
  • FFTホウ ト リョウイキホウ オ クミアワセタ DNA ploidy カイセキ ニ ヨル ガン シンダンホウ ノ ケンキュウ

Search this article

Abstract

We have previously reported the initial results of applying the fast Fourier transform (FFT) method to DNA ploidy analysis using a flow cytometer, aiming to develop a device that supports the diagnosis of solid tumors. In the present study, the diagnostic accuracy was improved by further increasing the number of cases and introducing additional analyses using screening, the domain method, and a tree-based model. In screening, tissues displaying clear aneuploidy on the histograms obtained (primary histograms)(the address of the maximum peak on the primary histogram deviates greatly from 200) were judged to be cancer. For the remaining tissues, the primary histograms were normalized (by setting the address of the maximum peak to 200, and the height of the maximum peak to 1) and analyzed using the FFT method. In addition, normalized histograms were separated into domains based on the cell cycle, and the integrated values of these domains were determined. Tissues collected from 46 patients who underwent resection of colorectal cancer at the Cancer Institute Hospital of JFCR were analyzed. Each resected tissue was divided into two pieces;one piece was subjected to hematoxylin and eosin staining for pathological diagnosis, and the remaining piece was analyzed for DNA ploidy by preparing a cell suspension and staining the nuclei using an automatic cell isolation and staining system and freeze-dry reagent. All 17 cases identified as cancer during screening were subsequently diagnosed as cancer pathologically. By the FFT method, the sensitivity and specificity using the Slope value (Maxdif in previous studies) were higher than 90%. Using the integrated values of the S and G2/M domains in the histograms, the sensitivity and specificity were both approximately 85%. Sensitivity and specificity of ≥95% were obtained using a combination of the FFT and domain methods. Furthermore, when all the parameters were used to construct a tree-based model, the cases were accurately classified into cancer, suspected cancer, and normal. These results indicate the potential clinical utility of this method.

Journal

Related Projects

See more

Keywords

Details 詳細情報について

Report a problem

Back to top