Surface-enhanced Raman spectroscopy + support vector machine: a new noninvasive method for prostate cancer screening?

Li, Shaoxin; Guo, Zhouyi; Liu, Zhiming

doi:10.1586/14737140.2015.992419

Cited by 12 publications

(10 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Among various machine learning methods, the support vector machines (SVMs) were chosen due to their high accuracy, ability to deal with high-dimensional and large datasets, and their flexibility in modeling diverse sources of data [22,23,28,29,30,31]. For example, in comparison with neural networks, SVMs do not require arbitrary coefficients or weights, are computationally faster, avoid the problem of local extrema (SVM optimization always ends up in a global extremum due to its convex nature), and are designed to generalize well to hitherto unseen objects.…”

Section: Discussionmentioning

confidence: 99%

Diagnostic Performance of a Support Vector Machine for Dermatofluoroscopic Melanoma Recognition: The Results of the Retrospective Clinical Study on 214 Pigmented Skin Lesions

Szyc¹,

Hillen

Scharlach³

et al. 2019

Diagnostics

View full text Add to dashboard Cite

The need for diagnosing malignant melanoma in its earliest stages results in an increasing number of unnecessary excisions. Objective criteria beyond the visual inspection are needed to distinguish between benign and malignant melanocytic tumors in vivo. Fluorescence spectra collected during the prospective, multicenter observational study (“FLIMMA”) were retrospectively analyzed by the newly developed machine learning algorithm. The formalin-fixed paraffin-embedded (FFPE) tissue samples of 214 pigmented skin lesions (PSLs) from 144 patients were examined by two independent pathologists in addition to the first diagnosis from the FLIMMA study, resulting in three histopathological results per sample. The support vector machine classifier was trained on 17,918 fluorescence spectra from 49 lesions labeled as malignant (1) and benign (0) by three histopathologists. A scoring system that scales linearly with the number of the “malignant spectra” was designed to classify the lesion as malignant melanoma (score > 28) or non-melanoma (score ≤ 28). Finally, the scoring algorithm was validated on 165 lesions to ensure model prediction power and to estimate the diagnostic accuracy of dermatofluoroscopy in melanoma detection. The scoring algorithm revealed a sensitivity of 91.7% and a specificity of 83.0% in diagnosing malignant melanoma. Using additionally the image segmentation for normalization of lesions’ region of interest, a further improvement of sensitivity of 95.8% was achieved, with a corresponding specificity of 80.9%.

show abstract

Section: Discussionmentioning

confidence: 99%

Diagnostic Performance of a Support Vector Machine for Dermatofluoroscopic Melanoma Recognition: The Results of the Retrospective Clinical Study on 214 Pigmented Skin Lesions

Szyc¹,

Hillen

Scharlach³

et al. 2019

Diagnostics

View full text Add to dashboard Cite

show abstract

“…Classification experiments were implemented using stratified 5-fold cross validation to preserve the same percentage of samples for each class to improve robustness. For experiments with a relatively small sample set, SVM is usually an efficient while reliable option, for it is designed to find the optimal decision boundary as represented by a hyperplane that maximizes the margin of separation between different classes [53] . In our binary classification, we used linear SVM to find the optimal linear decision boundary between the two classes.…”

Section: Methodsmentioning

confidence: 99%

“…Although spectral comparison and principal component analysis (PCA) have been employed in Raman spectral analysis, molecule identification is unreliable when the intra-class spectral variation is too high. [17][18][19][20][21] In recent years, machine learning has been frequently employed in Raman spectral analyses for disease diagnosis such as AD, cancer, infectious disease, etc.. [22][23][24][25] High accuracy in diagnosis is enabled by machine learning models including support vector machine (SVM), 26 random forest classifier 27 and neural networks. 28 Besides achieving outstanding performance in classification, machine learning can also interpret the correlation between Raman modes and diseases by providing spectral feature importance map.…”

Section: Introductionmentioning

confidence: 99%

“…However, so far machine learning interpretation lacks quantitative correlation to molecular composition in the Raman analysis for biomedical systems. [22][23][24][25] In this work, we employed machine learning classification and interpretation on Raman spectra of mouse brain slices and screened AD biomarkers. Our workflow is primarily composed of threestep procedures: first, we collected Raman spectra on mice brain slices with and without AD; then, we used machine learning to classify the collected Raman spectra on AD and non-AD brain slices; finally, we used linear SVM to interpret the spectral feature importance map which differentiates AD and non-AD spectra and discovered potential AD biomarkers (Figure 1).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Rapid Biomarker Screening of Alzheimer’s Disease by Interpretable Machine Learning and Graphene-Assisted Raman Spectroscopy

Wang

Ding

et al. 2021

Preprint

View full text Add to dashboard Cite

As the most common cause of dementia, Alzheimer's disease (AD) faces challenges in terms of understanding of pathogenesis, developing early diagnosis and developing effective treatment. Rapid and accurate identification of AD biomarkers in the brain will be critical to provide novel insights of AD. To this end, in the current work, we developed a system that can enable a rapid screening of AD biomarkers by employing Raman spectroscopy and machine learning analyses in AD transgenic animal brains. Specifically, we collected Raman spectra on slices of mouse brains with and without AD, and used machine learning to classify AD and non-AD spectra. By contacting monolayer graphene with the brain slices, we achieved significantly increased accuracy from 77% to 98% in machine learning classification. Further, we identified the Raman signature bands that are most important in classifying AD and non-AD samples. Based on these, we managed to identify AD-related biomolecules, which have been confirmed by biochemical studies. Our Raman-machine learning integrated method is promising to greatly accelerate the study of AD and can be potentially extended to human samples and various other diseases.

show abstract

“…The signal intensities of individual specific Raman peaks were compared using principal component analysis (PCA) and discriminated between the groups with 90.7% sensitivity and 100% specificity. 29 Further studies of cervical, 30 colon, 31 stomach, 32 parotid gland, 33 and prostate 34 cancer found that the individual types of cancer had characteristic Raman peaks. The unwanted and uncontrollable aggregation of the nanoparticles can be avoided by replacing the colloids of noble metals by electrodeposited SERS substrates.…”

Section: Introductionmentioning

confidence: 99%

Electrochemically deposited silver detection substrate for surface-enhanced Raman spectroscopy cancer diagnostics

2018

J. Biomed. Opt.

View full text Add to dashboard Cite

Surface-enhanced Raman spectroscopy (SERS) of blood plasma on an electrochemically prepared silver surface has been studied as a label-free, noninvasive diagnostic test for colorectal cancer. Indium tin oxide glass substrates were modified with 0.01 mol dm - 3 silver nitrate using the pulsed double-potentiostatic method. The prepared silver substrates were tested with Rhodamine 6G as a model analyte and the surface with the highest signal enhancement was selected. This silver dendritic surface was used as a diagnostic substrate for SERS measurements of human blood plasma. A group of oncological patients with declared colorectal carcinoma (n = 15) and the control group of healthy volunteers (n = 15) were compared. The biomolecular changes in chemical composition in the cancer samples were detected by statistical processing of the resulting SERS spectra. About 94% specificity and 100% sensitivity were achieved for the analysis by the ratio of the SERS peak intensity at 725 cm - 1 for adenine to the peak intensity at 638 cm - 1 for tyrosine and 100% specificity and sensitivity by using principal component analysis. This method of SERS diagnostics of colorectal cancer, which does not require the nanoparticle preparation, mixing, and incubation of plasma with a colloidal solution as in conventional tests, is a rapid, inexpensive method, which could be introduced as a primary diagnostic test.

show abstract

Surface-enhanced Raman spectroscopy + support vector machine: a new noninvasive method for prostate cancer screening?

Cited by 12 publications

References 18 publications

Diagnostic Performance of a Support Vector Machine for Dermatofluoroscopic Melanoma Recognition: The Results of the Retrospective Clinical Study on 214 Pigmented Skin Lesions

Diagnostic Performance of a Support Vector Machine for Dermatofluoroscopic Melanoma Recognition: The Results of the Retrospective Clinical Study on 214 Pigmented Skin Lesions

Rapid Biomarker Screening of Alzheimer’s Disease by Interpretable Machine Learning and Graphene-Assisted Raman Spectroscopy

Electrochemically deposited silver detection substrate for surface-enhanced Raman spectroscopy cancer diagnostics

Contact Info

Product

Resources

About