Highly accurate diagnosis of lung adenocarcinoma and squamous cell carcinoma tissues by deep learning

Qi, Yafeng; Yang, Lin; Liu, Bangxu; Liu, Li; Liu, Yuhong; Zheng, Qingfeng; Liu, Dameng; Luo, Jianbin

doi:10.1016/j.saa.2021.120400

Cited by 24 publications

(14 citation statements)

References 53 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Thus, the deep learning methods in the field of computer vision and natural language processing can be used in Raman data analysis. Just like in our previous work, we have demonstrated the application of speech command recognition by deep learning in lung tissue diagnosis with excellent accuracy. , In this work, we expanded on previous ideas and came up with a new concept called 2D Raman figures, which could be converted by the recurrence plot, the Gramian angular field, and the Markov transition field. The deep learning methods combined with 2D Raman figures obtained higher performances than the PCA-LDA method and the RS-CNN method, showing their huge potential in tumor diagnosis.…”

Section: Discussionmentioning

confidence: 82%

“…Therefore, we should find a way to transform the typical Raman spectrum into a 2D figure. According to our previous work, , we have converted it into a 2D Raman spectrogram by spectral short-time Fourier transform. Then, these 2D Raman spectrograms were fed into deep learning models for training and testing, which yielded a higher accurate diagnosis of lung tissues compared with the PCA-LDA method.…”

Section: Discussionmentioning

confidence: 99%

“…However, SRS instruments are very expensive and complex, which is not conducive to rapid and accurate clinical detection. According to our previous works, , we have proposed an unusual method wherein we considered the Raman spectral signal as a sequence and then converted it into a two-dimensional (2D) Raman spectrogram by spectral short-time Fourier transform (SSTFT). This novel method combined with deep learning yielded excellently accurate diagnosis of lung tissues.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

High-Precision Intelligent Cancer Diagnosis Method: 2D Raman Figures Combined with Deep Learning

Zhang

Yang

et al. 2022

Anal. Chem.

Self Cite

View full text Add to dashboard Cite

Raman spectroscopy, as a label-free detection technology, has been widely used in tumor diagnosis. However, most tumor diagnosis procedures utilize multivariate statistical analysis methods for classification, which poses a major bottleneck toward achieving high accuracy. Here, we propose a concept called the two-dimensional (2D) Raman figure combined with convolutional neural network (CNN) to improve the accuracy. Twodimensional Raman figures can be obtained from four transformation methods: spectral recurrence plot (SRP), spectral Gramian angular field (SGAF), spectral short-time Fourier transform (SSTFT), and spectral Markov transition field (SMTF). Two-dimensional CNN models all yield more than 95% accuracy, which is higher than the PCA-LDA method and the Raman-spectrum-CNN method, indicating that 2D Raman figure inputs combined with CNN may be one reason for gaining excellent performances. Among 2D-CNN models, the main difference is the conversion, where SRP is based on the structure of wavenumber series with the best performances (98.9% accuracy, 99.5% sensitivity, 98.3% specificity), followed by SGAF on the wavenumber series, SSTFT on wavenumber and intensity information, and SMTF on wavenumber position information. The inclusion of external information in the conversion may be another reason for improvement in the accuracy. The excellent capability shows huge potential for tumor diagnosis via 2D Raman figures and may be applied in other spectroscopy analytical fields.

show abstract

Section: Discussionmentioning

confidence: 82%

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

High-Precision Intelligent Cancer Diagnosis Method: 2D Raman Figures Combined with Deep Learning

Zhang

Yang

et al. 2022

Anal. Chem.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Qi et al adopted a novel approach to classify Raman spectra of lung tissue as adenocarcinoma, squamous cell carcinoma or normal in a pairwise fashion [ 29 ]. They transformed the data into 2D spectrograms in a similar process used to classify audio data.…”

Section: Resultsmentioning

confidence: 99%

“…Of those studies that split the data at the level of spectra, the best accuracies of 90%, 96.6%, 97.7%, 93.8% and 94.8% were achieved [ 19 , 25 , 26 , 29 , 35 ]. Of those studies splitting data at the level of subject or sample, the accuracies were: 84.4%, 99%, 81.8%, 98.3%, 83%, 87%, 92.9%, 81.3% and 62.5% [ 20 , 22 , 24 , 30 , 31 , 32 , 34 , 35 , 40 ].…”

Section: Discussionmentioning

confidence: 99%

Machine Learning of Raman Spectroscopy Data for Classifying Cancers: A Review of the Recent Literature

et al. 2022

View full text Add to dashboard Cite

Raman Spectroscopy has long been anticipated to augment clinical decision making, such as classifying oncological samples. Unfortunately, the complexity of Raman data has thus far inhibited their routine use in clinical settings. Traditional machine learning models have been used to help exploit this information, but recent advances in deep learning have the potential to improve the field. However, there are a number of potential pitfalls with both traditional and deep learning models. We conduct a literature review to ascertain the recent machine learning methods used to classify cancers using Raman spectral data. We find that while deep learning models are popular, and ostensibly outperform traditional learning models, there are many methodological considerations which may be leading to an over-estimation of performance; primarily, small sample sizes which compound sub-optimal choices regarding sampling and validation strategies. Amongst several recommendations is a call to collate large benchmark Raman datasets, similar to those that have helped transform digital pathology, which researchers can use to develop and refine deep learning models.

show abstract

ResNet models for rapid identification of species and geographical origin of wild boletes from Yunnan, and MaxEnt model for delineation of potential distribution

Chen

Liu

et al. 2022

Journal of Chemometrics

View full text Add to dashboard Cite

Yunnan is known for its rich biodiversity and is known as the Wild Mushroom Kingdom. Boletes are a world-renowned wild edible mushroom, with unique sensory characteristics, nutritional value and medicinal value extraordinary.However, the species and geographical origin of boletes influence their price and quality. In this study, a method was developed to identify species and geographical origin simultaneously. Therefore, Fourier transform near-infrared (FT-NIR) data sets of boletes were collected and converted to two-dimensional correlation spectroscopy (2D-COS). On this basis, the species and geographic origins of boletes were identified using Residual neural network (ResNet) image analysis model. The results showed that FT-NIR could identify boletes species and geographical origins, 7000-4000 cm À1 band was more suitable for species identification, 7000-5300 cm À1 band was more suitable for geographical origins identification. In addition, the environmental factors with high contribution to the distribution of boletes were screened based on the maximum entropy (MaxEnt) model. This allows characterization of the potential geographic distribution of boletes. The results showed that precipitation factors played a vital role in its distribution and might even be responsible for the difference in chemical composition.

show abstract

Highly accurate diagnosis of lung adenocarcinoma and squamous cell carcinoma tissues by deep learning

Cited by 24 publications

References 53 publications

High-Precision Intelligent Cancer Diagnosis Method: 2D Raman Figures Combined with Deep Learning

High-Precision Intelligent Cancer Diagnosis Method: 2D Raman Figures Combined with Deep Learning

Machine Learning of Raman Spectroscopy Data for Classifying Cancers: A Review of the Recent Literature

ResNet models for rapid identification of species and geographical origin of wild boletes from Yunnan, and MaxEnt model for delineation of potential distribution

Contact Info

Product

Resources

About