Limits...
Cellular phone enabled non-invasive tissue classifier.

Laufer S, Rubinsky B - PLoS ONE (2009)

Bottom Line: The results of the tissue analysis were returned to the remote data measurement site.When used for the detection of malignant tumors, classifiers can be designed to produce false positives in order to ensure that no tumors will be missed.This mode of operation has applications in remote non-invasive tissue diagnostics in situ in the body, in combination with medical imaging, as well as in remote diagnostics of biopsy samples in vitro.

View Article: PubMed Central - PubMed

Affiliation: Center for Bioengineering in the Service of Humanity and Society, School of Computer Science and Engineering, Hebrew University of Jerusalem, Jerusalem, Israel. Shlomi.laufer@mail.huji.ac.il

ABSTRACT
Cellular phone technology is emerging as an important tool in the effort to provide advanced medical care to the majority of the world population currently without access to such care. In this study, we show that non-invasive electrical measurements and the use of classifier software can be combined with cellular phone technology to produce inexpensive tissue characterization. This concept was demonstrated by the use of a Support Vector Machine (SVM) classifier to distinguish through the cellular phone between heart and kidney tissue via the non-invasive multi-frequency electrical measurements acquired around the tissues. After the measurements were performed at a remote site, the raw data were transmitted through the cellular phone to a central computational site and the classifier was applied to the raw data. The results of the tissue analysis were returned to the remote data measurement site. The classifiers correctly determined the tissue type with a specificity of over 90%. When used for the detection of malignant tumors, classifiers can be designed to produce false positives in order to ensure that no tumors will be missed. This mode of operation has applications in remote non-invasive tissue diagnostics in situ in the body, in combination with medical imaging, as well as in remote diagnostics of biopsy samples in vitro.

Show MeSH

Related in: MedlinePlus

Implementation of classifier using the cellular phone.a) Measured rat heart data sent; b) Heart result received; c) Measured rat kidney data sent; and d) Kidney result received.
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC2664897&req=5

pone-0005178-g004: Implementation of classifier using the cellular phone.a) Measured rat heart data sent; b) Heart result received; c) Measured rat kidney data sent; and d) Kidney result received.

Mentions: In order to test the cellular phone concept, the data were arranged in a text file containing about 1200 bytes and sent via e-mail from the cellular phone to the remote computer. Each File had seven lines, one for each different electrode configuration. Each line consisted of eleven complex numbers, one for each frequency. Once the file was received on the remote computer (AMD Athlon 64 X2 Dual Core Processor 5000+, 2.61 GHz, 2 GB RAM, Microsoft Windows XP), the SVM classifier program was applied to the data and a classifier score was calculated. Accordingly, the tissue was classified as either kidney or heart. The remote computer then sent an email reply to the DAD site with the word heart or the word kidney. The process is demonstrated in Figure 4.


Cellular phone enabled non-invasive tissue classifier.

Laufer S, Rubinsky B - PLoS ONE (2009)

Implementation of classifier using the cellular phone.a) Measured rat heart data sent; b) Heart result received; c) Measured rat kidney data sent; and d) Kidney result received.
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC2664897&req=5

pone-0005178-g004: Implementation of classifier using the cellular phone.a) Measured rat heart data sent; b) Heart result received; c) Measured rat kidney data sent; and d) Kidney result received.
Mentions: In order to test the cellular phone concept, the data were arranged in a text file containing about 1200 bytes and sent via e-mail from the cellular phone to the remote computer. Each File had seven lines, one for each different electrode configuration. Each line consisted of eleven complex numbers, one for each frequency. Once the file was received on the remote computer (AMD Athlon 64 X2 Dual Core Processor 5000+, 2.61 GHz, 2 GB RAM, Microsoft Windows XP), the SVM classifier program was applied to the data and a classifier score was calculated. Accordingly, the tissue was classified as either kidney or heart. The remote computer then sent an email reply to the DAD site with the word heart or the word kidney. The process is demonstrated in Figure 4.

Bottom Line: The results of the tissue analysis were returned to the remote data measurement site.When used for the detection of malignant tumors, classifiers can be designed to produce false positives in order to ensure that no tumors will be missed.This mode of operation has applications in remote non-invasive tissue diagnostics in situ in the body, in combination with medical imaging, as well as in remote diagnostics of biopsy samples in vitro.

View Article: PubMed Central - PubMed

Affiliation: Center for Bioengineering in the Service of Humanity and Society, School of Computer Science and Engineering, Hebrew University of Jerusalem, Jerusalem, Israel. Shlomi.laufer@mail.huji.ac.il

ABSTRACT
Cellular phone technology is emerging as an important tool in the effort to provide advanced medical care to the majority of the world population currently without access to such care. In this study, we show that non-invasive electrical measurements and the use of classifier software can be combined with cellular phone technology to produce inexpensive tissue characterization. This concept was demonstrated by the use of a Support Vector Machine (SVM) classifier to distinguish through the cellular phone between heart and kidney tissue via the non-invasive multi-frequency electrical measurements acquired around the tissues. After the measurements were performed at a remote site, the raw data were transmitted through the cellular phone to a central computational site and the classifier was applied to the raw data. The results of the tissue analysis were returned to the remote data measurement site. The classifiers correctly determined the tissue type with a specificity of over 90%. When used for the detection of malignant tumors, classifiers can be designed to produce false positives in order to ensure that no tumors will be missed. This mode of operation has applications in remote non-invasive tissue diagnostics in situ in the body, in combination with medical imaging, as well as in remote diagnostics of biopsy samples in vitro.

Show MeSH
Related in: MedlinePlus