Exploring photoacoustic spectroscopy-based machine learning together with metabolomics to assess breast tumor progression in a xenograft model ex vivo

Abstract

In the current study, a breast tumor xenograft was established in athymic nude mice by subcutaneous injection of the MCF-7 cell line and assessed the tumor progression by photoacoustic spectroscopy combined with machine learning tools. The advancement of breast tumors in nude mice was validated by tumor volume kinetics and histopathology and corresponding image analysis by TissueQuant software compared to controls. The ex vivo tumors in progressive conditions belonging to time points, day 5th, 10th, 15th & 20th, were excited with 281 nm pulsed laser light and recorded the corresponding photoacoustic spectra in time domain. The spectra were then pre-processed, augmented for a 10-fold increase in the data strength, and subjected to wavelet packet transformation for feature extraction and selection using MATLAB software. In the present study, the top 10 features from all the time point groups under study were selected based on their prediction ranking values using the mRMR algorithm. The chosen features of all the time-point groups were then subjected to multi-class Support Vector Machine (SVM) algorithms for learning and classifying into respective time point groups under study. The analysis demonstrated accuracy values of 95.2%, 99.5%, and 80.3% with SVM- Radial Basis Function (SVM-RBF), SVM-Polynomial & SVM-Linear, respectively. The serum metabolomic levels during tumor progression complemented photoacoustic patterns of tumor progression, depicting breast cancer pathophysiology.

Fluorescence and Photoacoustic Spectroscopy-Based Assessment of Mitochondrial Dysfunction in Oral Cancer Together with Machine Learning: A Pilot Study

“…have evaluated the breast tumor progression in xenograft models at 10, 15, and 20 days by targeting the tumor proteins at 281 nm excitation and recording the corresponding PA spectra. The study results clearly showed the photoacoustic spectroscopy’s ability to discriminate tumor stages based on tryptophan PA patterns. , …”

“…34 Additionally, no studies have reported protein conformational changes through photoacoustic (PA) spectroscopy studies. 35,36 In most analytical methods, including photoacoustic spectroscopy, the spectral patterns are influenced by the background noises that arise from the instrumentation and the surroundings, affecting the overall signal quality. Further, when the signal is weak, the background noise dominates over the actual signal.…”

“…Usually, feature relevance is Previous studies have reported the application of mRMR algorithms in various fields, such as detection of breast tumor progression, protein domain detection, epileptic seizures prediction, etc. 36,42,43 Ding and Peng have applied mRMR feature selection algorithms to select genes from the microarray data to classify different cancer phenotypes, demonstrating better classification. 44 In another study using nearest neighbor and mRMR algorithms, the authors have attempted to predict the protein tyrosine sulfation.…”

Non-Invasive Monitoring of Human Health by Photoacoustic Spectroscopy

“…Several methods for monitoring tumor development by processing PA signals have also been reported recently. Rodrigues et al monitored the development of breast tumors in nude mice by utilizing PAS combined with support vector machine analysis [ 72 ]. Priya et al reported a method using wavelet principal component analysis-based logistic regression analysis to assess the tumor growth in mice [ 73 ].…”

Combining acoustic bioprinting with AI-assisted Raman spectroscopy for high-throughput identification of bacteria in blood

“…to remove any spectra with weak signal that likely were collected on the substrate without the presence of cells. We then transformed our data by taking log 10 (y) [85,86] and smoothed the spectra using wavelet denoising [87,88]. To perform our smoothing, we used the denoise wavelet function from the scikitimage Python library: denoise wavelet(y, method='BayesShrink', mode='soft', wavelet levels=1, wavelet='coif3', rescale sigma='True').…”