X-ray scatter signatures for normal and neoplastic breast tissues

Kidane, G.; Speller, Robert; Royle, G.; Hanby, Andrew M.

doi:10.1088/0031-9155/44/7/316

Cited by 172 publications

(215 citation statements)

References 9 publications

Supporting

Mentioning

160

Contrasting

Unclassified

Order By: Relevance

“…More recently this technique has been used as a new imaging method 1 -3 and a new tissue-characterization tool in biomedical sciences. 4 Several measurements of scattering profiles from excised breast samples have been reported using conventional x-ray diffractometer, 5 -7 synchrotron radiation 8 -10 and a broadbeam x-ray source at a fixed scattering angle 11,12 or with a stationary position-sensitive detector. 13 In these works, it had been pointed out that the scattering properties of breast tissues could be used for breast tissue characterization, on the basis of the fact that the number of coherently scattered photons is a function of the scattering angle and depends on the molecular environment within the biological tissues, meaning that each scattering profile is a 'signature' of a specific molecular or supramolecular structure.…”

Section: Introductionmentioning

confidence: 99%

X-ray scattering profiles of some normal and malignant human breast tissues

et al. 2006

View full text Add to dashboard Cite

The angular distribution of photons scattered by breast tissue samples (20 normal and 13 malignant) was measured in the interval of momentum transfer 0.02Å −1 Ä x Ä 0.62Å −1 , using a powder diffractometer apparatus operating in reflection mode and equipped with a monochromator in the scattering beam. The x-ray tube had a Cu anode and added Ni filtration. Experimental data was corrected for all background sources, polarization, self-attenuation and geometric effect in order to provide the linear differential scattering coefficient m s (scattering profile) of each sample. Discriminant analysis was applied to these profiles in order to quantify the differences existing between normal and malignant tissues. In this analysis, each momentum transfer measured was considered as an independent variable, and the scores of each sample on a discriminant function were obtained. The classification method of decision tree induction was applied to these scores, which allowed tissue characterization. The classification results were validated and compared with histological sample classifications obtained by microscopic analyses. The results have shown that discriminant analysis was able to successfully identify differences between the scattering profiles from normal and malignant tissues. Also, classification results allowed the correct identification of 97% of the analyzed samples. These results suggest that the scattering profile may be an effective marker to be used in the diagnosis of breast cancer.

show abstract

Section: Introductionmentioning

confidence: 99%

X-ray scattering profiles of some normal and malignant human breast tissues

et al. 2006

View full text Add to dashboard Cite

show abstract

“…The predominance of the two above mentioned momentum transfer components in describing variations between different tissue types has been observed by several authors (Kidane et al, 1999;Ryan and Farquharson, 2007;Oliveira et al, 2008 to include those who have addressed different type of tumours). A quantitative comparison is difficult due to the likely differences in the experimental set-ups used (for instance, different angular resolutions are likely to affect the peak width and height), to the exiguity and sparseness of data available for non-adipose tissue types and to lack of information on some experimental set-ups.…”

Section: Discussionmentioning

confidence: 85%

“…Although several authors (Kidane et al, 1999;Castro et al, 2004;Geraki et al 2004;LeClair et al, 2006;Ryan and Farquharson, 2007;Oliveira et al, 2008) have proved that a different diffraction signal is achieved from normal and neoplastic breast tissue in the momentum transfer range 0.5 -4 nm -1 , there are few data about identification of different lesion types. In this study, the diffraction patterns of different lesions have been investigated using EDXRDCT.…”

Section: Discussionmentioning

confidence: 99%

“…Several studies, based both on the SAXS (Lewis et al, 2000;Fernandez et al, 2002;Round et al, 2005;Suhonen et al, 2005;Sidhu et al, 2009) and on the XRD modality (Kidane et al, 1999;Poletti et al, 2002;LeClair et al, 2006;Ryan and Farquharson, 2007;Oliveira et al, 2008), have shown the feasibility of breast tissue characterization. Of particular interest is the possibility of discriminating healthy and neoplastic tissue.…”

Section: Introductionmentioning

confidence: 99%

“…This was done using conventional diffractometers (Poletti et al, 2002;Oliveira et al, 2008), monochromatic synchrotron radiation (Castro et al 2004), or energy-dispersive systems (Kidane et al, 1999;Geraki et al, 2004, LeClair et al, 2006Ryan and Farquharson, 2007).…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Characterization of breast tissue using energy-dispersive X-ray diffraction computed tomography

Pani

Cook

Horrocks

et al. 2010

Applied Radiation and Isotopes

View full text Add to dashboard Cite

Application of machine learning classifiers to X‐ray diffraction imaging with medically relevant phantoms

2021

View full text Add to dashboard Cite

Recent studies have demonstrated the ability to rapidly produce large field of view X-ray diffraction (XRD) images, which provide rich new data relevant to the understanding and analysis of disease. However, work has only just begun on developing algorithms that maximize the performance toward decision-making and diagnostic tasks. In this study, we present the implementation of and comparison between rules-based and machine learning (ML) classifiers on XRD images of medically relevant phantoms to explore the potential for increased classification performance. Methods: Medically relevant phantoms were utilized to provide wellcharacterized ground-truths for comparing classifier performance. Water and polylactic acid (PLA) plastic were used as surrogates for cancerous and healthy tissue, respectively, and phantoms were created with varying levels of spatial complexity and biologically relevant features for quantitative testing of classifier performance. Our previously developed X-ray scanner was used to acquire co-registered X-ray transmission and diffraction images of the phantoms. For classification algorithms, we explored and compared two rules-based classifiers (cross-correlation, or matched-filter, and linear least-squares unmixing) and two ML classifiers (support vector machines and shallow neural networks). Reference XRD spectra (measured by a commercial diffractometer) were provided to the rules-based algorithms, while 60% of the measured XRD pixels were used for training of the ML algorithms. The area under the receiver operating characteristic curve (AUC) was used as a comparative metric between the classification algorithms, along with the accuracy performance at the midpoint threshold for each classifier. Results:The AUC values for material classification were 0.994 (crosscorrelation [CC]), 0.994 (least-squares [LS]), 0.995 (support vector machine [SVM]), and 0.999 (shallow neural network [SNN]). Setting the classification threshold to the midpoint for each classifier resulted in accuracy values of CC = 96.48%, LS = 96.48%, SVM = 97.36%, and SNN = 98.94%. If only considering pixels ±3 mm from water-PLA boundaries (where partial volume effects could occur due to imaging resolution limits), the classification accuracies were CC = 89.32%, LS = 89.32%, SVM = 92.03%, and SNN = 96.79%, demonstrating an even larger improvement produced by the machine-learned algorithms in spatial regions critical for imaging tasks. Classification by transmission data alone produced an AUC of 0.773 and accuracy of 85.45%, well below the performance levels of any of the classifiers applied to XRD image data. Conclusions: We demonstrated that ML-based classifiers outperformed rulesbased approaches in terms of overall classification accuracy and improved the spatially resolved classification performance on XRD images of medical phantoms. In particular, the ML algorithms demonstrated considerably improved 532

show abstract

X-ray scatter signatures for normal and neoplastic breast tissues

Cited by 172 publications

References 9 publications

X-ray scattering profiles of some normal and malignant human breast tissues

X-ray scattering profiles of some normal and malignant human breast tissues

Characterization of breast tissue using energy-dispersive X-ray diffraction computed tomography

Application of machine learning classifiers to X‐ray diffraction imaging with medically relevant phantoms

Contact Info

Product

Resources

About