Lung cancer lesion detection in histopathology images using graph‐based sparse PCA network

Ram, Sundaresh; Tang, Wenfei; Bell, Alexander J.; Spencer, Cara; Buschhaus, Alexander; Hatt, Charles R.; diMagliano, Marina Pasca; Rodríguez, Jeffrey J.; Galbán, Stefanie; Galbán, Craig J.

doi:10.1016/j.neo.2023.100911

Cited by 7 publications

(7 citation statements)

References 51 publications

(91 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“… Using the Enhanced Grasshopper Optimization Algorithm (EGOA), Pradhan and Sahu [ 7 ] achieved an accuracy of 98.50 % for lung cancer classification. Ram et al [ 8 ] employed a method called GS-PCANet and attained an accuracy of 90.80 % for lung cancer classification, with an AUC (a measure of accuracy) of 0.95. Reis and Turk [ 9 ] utilized DenseNet169 for colon cancer classification and achieved an accuracy of 95.0 %.…”

Section: Resultsmentioning

confidence: 99%

“…Ram et al [ 8 ] detailed the application of a machine learning technique named the graph-based sparse principal component analysis (GS-PCA) network for the auto-detection of malignant lesions in histological lung slides. Using methods like Support Vector Machine (SVM) classification, cascaded graph-based sparse , binary hashing, and block-wise histograms, the recommended method's detection accuracy clearly excels when compared to established techniques.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Colon and lung cancer classification from multi-modal images using resilient and efficient neural network architectures

Uddin,

Chen,

Akter

et al. 2024

Heliyon

View full text Add to dashboard Cite

Section: Resultsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Colon and lung cancer classification from multi-modal images using resilient and efficient neural network architectures

Uddin,

Chen,

Akter

et al. 2024

Heliyon

View full text Add to dashboard Cite

“…22, 23 We also investigated the selection bias for each input in the ML model and obtained the prediction accuracy for 10 different choices of training image patches, considering each input separately in the model. The prediction accuracy for each training run is fit to a Gaussian probability density function 24 . All processing and analyses were performed using in-house algorithms developed in MATLAB version 2020a (MathWorks, Natick, MA).…”

Section: Methodsmentioning

confidence: 99%

“…The prediction accuracy for each training run is fit to a Gaussian probability density function 24 .…”

Section: Predict Spirometric Declinementioning

confidence: 99%

“…Image sparsity has emerged as a significant property of images and sparsity-based regularization has been used for various image processing applications. [18][19][20]24,[27][28][29][30] Sparse image representations are at the heart of many modern approaches to medical image classification and include [31][32][33] . The sparse model assumes that each patch within an image can be accurately represented using a few elements of a basis set called a dictionary.…”

Section: Appendix 2: Dictionary Learning Algorithmmentioning

confidence: 99%

See 1 more Smart Citation

Quantitative CT of Normal Lung Parenchyma and Small Airways Disease Topologies are Associated With COPD Severity and Progression

Bell

Ram

Labaki³

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

Purpose: Small airways disease (SAD), a major cause of airflow obstruction in COPD patients, has been identified as a precursor to emphysema. Nevertheless, there is a lack of clinical techniques that can quantify the progression of SAD. We aim to determine if our Parametric Response Mapping (PRM) method to quantify SAD offers insight into lung progression from healthy to emphysema. Materials and Methods: PRM metrics quantifying normal lung (PRMNorm) and functional SAD (PRMfSAD) were generated from CT scans collected as part of the COPDGene study (n=8956). Volume density (V) and Euler-Poincaré Characteristic (χ) maps, measures of extent and coalescence of pocket formations, respectively, were determined for both PRMNorm and PRMfSAD. Association with COPD severity, emphysema, and spirometric measures were assessed via multivariable regression models. Results: Across all GOLD, we observed a strong linear correlation between χfSAD and χNorm (ρ =-0.745, p<0.001). Values of χfSAD and χNorm were found to flip signs in unison between GOLD 2 and 4, demonstrating an inversion in parenchymal topology. For subjects with COPD, multivariable analysis showed that both χfSAD (β of 0.106, p<0.001) and VfSAD (β of 0.065, p=0.004) were independently associated with FEV1% predicted. V and χ measures for PRMfSAD and PRMNorm were independently associated with the amount of emphysema. Conclusions: We demonstrated that χ of fSAD and Norm have independent value when associated with lung function and emphysema, even when considering the amount of each (i.e., VfSAD, VNorm). Our approach for quantifying pocket formations of PRMfSAD from normal lung parenchyma (PRMNorm) may show promise as a CT readout of emphysema onset.

show abstract

Transformer-based semantic segmentation and CNN network for detection of histopathological lung cancer

Talib,

Amin,

Sharif

et al. 2024

Biomedical Signal Processing and Control

View full text Add to dashboard Cite

Lung cancer lesion detection in histopathology images using graph‐based sparse PCA network

Cited by 7 publications

References 51 publications

Colon and lung cancer classification from multi-modal images using resilient and efficient neural network architectures

Colon and lung cancer classification from multi-modal images using resilient and efficient neural network architectures

Quantitative CT of Normal Lung Parenchyma and Small Airways Disease Topologies are Associated With COPD Severity and Progression

Transformer-based semantic segmentation and CNN network for detection of histopathological lung cancer

Contact Info

Product

Resources

About