Fourier Transformation-Based Analysis of X-Ray Diffraction Pattern of Keratin for Cancer Detection
Investigates how Fourier transformations enhance the interpretation of XRD data from biological tissues.
With the growing number of cancer cases and deaths around the world, fast, non-invasive, and inexpensive screening is paramount. We examine the feasibility of such cancer detection using the X-ray scattering properties of nails in the canine model. A total of 945 samples taken from 266 dogs were measured, with 84 animals diagnosed with cancer. To analyze the obtained X-ray diffraction patterns of keratin, we propose a method based on the two-dimensional Fourier transformation of the images. We compare 745 combinations of data preprocessing steps and machine learning classifiers and determine the corresponding performance metrics. Excellent classification results are demonstrated, with sensitivity or specificity achieving 100% and the best value for balanced accuracy being 87.5%. We believe that our approach can be extended to human samples to develop a non-invasive, convenient, and cheap method for early cancer detection.
Overview
The study focuses on developing a new method for early cancer detection in dogs using X-ray diffraction (XRD) patterns from their claws. This research is significant because:
Dogs develop similar types of cancer as humans, making them excellent models for comparative oncology
An estimated 6 million pet dogs are diagnosed with cancer in the US
The findings could potentially benefit both canine and human cancer detection
Methodology:
Collected samples from 266 dogs (104 with cancer, 162 without)
Total dataset included 945 XRD patterns from thin slices of dogs' claws
Developed a new analytical approach using 2D Fourier transformation of XRD images
Compared various machine learning algorithms and data preprocessing techniques
Key Findings:
2D Fourier transformation methods performed significantly better than 1D approaches
The best results came from:
Random Forest model using imaginary parts of low-pass-filtered 2D Fourier coefficients
Light Gradient-Boosting Machine model using amplitudes of low-pass-filtered 2D Fourier coefficients
For some combinations, either sensitivity or specificity reached 1.0
The balanced accuracy achieved up to 0.875
Averaging results from multiple samples from the same patient improved accuracy
Important Technical Insights:
Low-Pass Filtering improved diagnostics by reducing unimportant features
Removing the primary beam area from XRD patterns enhanced results
Principal Component Analysis was most effective with 50-100 components
Most classifiers (except Logistic Regression) performed similarly well
This research demonstrates promising results for non-invasive cancer detection in dogs using XRD patterns from claws, with potential implications for future human cancer diagnostics.