 Home > Active Ingredient News > Antitumor Therapy > Photoacoustics: Combining photoacoustic spectroscopy and machine learning is beneficial in diagnosing prostate cancer

Photoacoustics: Combining photoacoustic spectroscopy and machine learning is beneficial in diagnosing prostate cancer

 Last Update: 2021-06-28
 Source: Internet
 Author: User

Tags

cpr machine

oxidation reduction potential

Search more information of high quality chemicals, good prices and reliable suppliers, visit www.echemi.com

Prostate cancer is a disease with high incidence and low cure rate after metastasis
.

According to cancer statistics in 2019 , prostate cancer is the second most common cancer in men worldwide ( after lung cancer and bronchial cancer )

Prostate cancer is a disease with high incidence and low cure rate after metastasis

Photoacoustic Physicochemical Analysis (PAPCA) has shown great potential in the diagnosis of prostate cancer

Photoacoustic spectroscopy technology can generate abundant chemical and physical information of biological tissues

Recently, Photoacoustics magazine published an article describing the application of machine learning methods (including unsupervised hierarchical clustering and supervised classification) to diagnose prostate cancer by photoacoustic spectrum analysis

The research combines PAPCA with machine learning to analyze the characteristic changes of the photoacoustic physicochemical spectrum corresponding to the evolution of prostate cancer
.

It aims to clarify the feasibility of prostate cancer diagnosis based on photoacoustic physicochemical spectroscopy

The research combines PAPCA with machine learning to analyze the characteristic changes of the photoacoustic physicochemical spectrum corresponding to the evolution of prostate cancer

101 specimens were obtained from 22 cases of radical prostatectomy .
Then, based on unsupervised clustering, the correlation analysis of the power spectra of normal and tumor prostate tissues at different wavelengths is carried out .
First, it aims to explore the changes in the microstructure distribution of different biological macromolecules during the evolution of prostate cancer .
Second, clarify the feasibility of realizing prostate cancer diagnosis based on photoacoustic physicochemical spectroscopy .
Subsequently, based on the supervised linear discriminant analysis (LDA) and quadratic discriminant analysis (QDA) algorithms, the dimensionality reduction and feature parameter extraction of the photoacoustic physicochemical spectrum were used for the diagnosis of prostate cancer .

Machine learning algorithm block diagram

The research focused on the content and distribution of hemoglobin, collagen, and lipids, because these molecules change during the development of prostate cancer
.

The photoacoustic physicochemical spectra of typical normal and tumor tissues are shown in the figure below
.

The spectrum shows the difference between normal tissue and tumor tissue

The photoacoustic physicochemical spectra of typical normal and tumor tissues are shown in the figure below

The figure above shows the comparison of photoacoustic physicochemical spectra of cancerous samples and normal samples
.

(a) Photoacoustic physicochemical spectra of typical normal (top) and cancer (bottom)

The figure above shows the comparison of photoacoustic physicochemical spectra of cancerous samples and normal samples

Effectiveness of cluster analysis
.
(a) Typical cluster tree of UPGMA results
.
(b) Schematic diagram of light absorption spectra and wavelength numbers of biological macromolecules
.
(c) Statistical clustering results of normal and cancerous samples
.

Using PCCs and UPGMA to group 77 wavelengths into 6 wavelength groups ( Figure 4a)
.
Compare the clustering results with the light absorption spectra of biomolecules ( Figure 4b)
.
Although the clustering tree 77 end nodes are not to 0-76 arrangement order, but FIG.
4a wavelength grouping and figure 4b highly consistent in different biological macromolecules light absorbing group
.
In addition, due to the difference in molecular bonds and vibration modes, different biological macromolecules have unique light absorption spectra
.
UPGMA helps to analyze the difference in photoacoustic spectra and determine the dominant detection wavelength range of photoacoustic signals generated by different biological macromolecules
.
Since the main biomolecules in normal tissues and cancer tissues did not change, there was no significant difference in the clustering results between normal tissues and cancer tissues
.

Using PCCs and UPGMA to group 77 wavelengths into 6 wavelength groups ( Figure 4a)
.
Compare the clustering results with the light absorption spectra of biomolecules ( Figure 4b)
.
Although the clustering tree 77 end nodes are not to 0-76 arrangement order, but FIG.
4a wavelength grouping and figure 4b highly consistent in different biological macromolecules light absorbing group
.
In addition, due to the difference in molecular bonds and vibration modes, different biological macromolecules have unique light absorption spectra
.
UPGMA UPGMA helps to analyze the difference in photoacoustic spectra and determine the dominant detection wavelength range of photoacoustic signals generated by different biological macromolecules
.
Since the main biomolecules in normal tissues and cancer tissues did not change, there was no significant difference in the clustering results between normal tissues and cancer tissues
.

Compare the correlation between normal and cancerous samples
.
(a) The clustering tree of normal samples
.
(b) Heat map of normal samples
.
(c) Cluster tree of cancerous samples
.
(d) Heat map of cancerous samples
.
(e) The correlation weight (WR) network diagram of the normal population
.
(f) WR network diagram of tumor group
.
Note the increase in the number of links (green arrows) and tags (purple arrows), as well as larger nodes (orange arrows)
.

Comparison of traditional signal processing-based methods and machine learning-based methods

The ultrasound power spectra of these chemical components have a higher correlation in cancerous tissues than in normal tissues, indicating that the microstructure distribution of cancerous tissues is more consistent
.
The accuracy of different classification methods in tumor tissue diagnosis reached 82% ( better than clinical standard methods )
.
Therefore, this technology shows great potential for early diagnosis of aggressive prostate cancer without pain
.

The study used the photoacoustic physicochemical spectra of normal and cancerous prostate tissue to establish a prostate cancer diagnostic model
.
Based on the UPGMA cluster analysis method, the correlation difference between different macromolecules in normal and cancerous tissues is analyzed, and the characteristic light absorption bands V1-V6 of different biological macromolecules are distinguished
.
It was found that the power spectrum correlation of hemoglobin, collagen, and lipids in prostate cancer tissue was significantly higher than that of normal tissues, which reflected that the power spectrum correlation of hemoglobin, collagen, and lipids in prostate cancer tissues based on normal and tumor samples was significantly higher.
For normal tissues, this reflects the increased similarity of the microstructure of the distribution of these biomolecules based on the V1-V6 V1-V6 visualization results of
normal and tumor samples .
As a result of visualization, the microstructure similarity of the distribution of these biomolecules increases
.
This may be because the complex tumor environment increases the complexity of the photoacoustic spectrum and makes the distribution characteristics of a single substance obscure, thereby increasing the correlation of the photoacoustic spectrum.
The complex tumor environment increases the complexity of the photoacoustic spectrum and makes the photoacoustic spectrum more complicated.
The distribution characteristics of a single substance become blurred, thereby increasing the correlation of the photoacoustic spectrum
.
These results indicate that photoacoustic physicochemical spectroscopy can effectively reflect the changes of biological macromolecules during tumor development and can be used for the diagnosis of prostate cancer
.
The photoacoustic physicochemical spectrum can effectively reflect the changes of biological macromolecules in the process of tumor development, and can be used for the diagnosis of prostate cancer
.

The results of this study proved the feasibility and effectiveness of combining machine learning and photoacoustic physicochemical spectroscopy technology in exploring the changes in the microstructure and chemical composition of prostate cancer, thereby promoting the diagnosis of prostate cancer
.
The results of this study proved the feasibility and effectiveness of combining machine learning and photoacoustic physicochemical spectroscopy technology in exploring the changes in the microstructure and chemical composition of prostate cancer, thereby promoting the diagnosis of prostate cancer
.

Original source

Yingna Chen, Qian Cheng,Prostate cancer identification via photoacoustic spectroscopy and machine learning,Photoacoustics,Volume 23,2021,100280,ISSN 2213-5979,https://doi.
org/10.
1016/j.
pacs.
2021.
100280.
Yingna Chen, Qian Cheng,Prostate cancer identification via photoacoustic spectroscopy and machine learning,Photoacoustics,Volume 23,2021,100280,ISSN 2213-5979,https://doi.
org/10.
1016/j.
pacs.
2021.
100280.
Leave a message here

This article is an English version of an article which is originally in the Chinese language on echemi.com and is provided for information purposes only. This website makes no representation or warranty of any kind, either expressed or implied, as to the accuracy, completeness ownership or reliability of the article or any translations thereof. If you have any concerns or complaints relating to the article, please send an email, providing a detailed description of the concern or complaint, to service@echemi.com. A staff member will contact you within 5 working days. Once verified, infringing content will be removed immediately.