Feb 19, 2012

Cancer Breath Test with 83% accuracy Enters Clinical Trials and next generation system is 100 times more sensitive

Metabolomx, has developed technology enabling the identification of lung cancer from breath. Using the first generation of our breath analysis system, the Cleveland Clinic announced the results of a 237 subject trial at the American College of Chest Physicians conference in November 2010, reporting 81%, accuracy of lung cancer detection, comparable to CT scan, the present gold standard. This study, further reported that lung cancer subtype (small cell, adenocarcinoma, and squamous cell) is identified by the breath exam. The Cleveland Clinic has now commenced testing of the current generation Metabolomx sensor, over 100 times more sensitive than the version used in the study

Human breath contains hundreds of volatile organic compounds (VOCs) produced both endogenously and from external environmental sources. To detect the few metabolite biomarkers of disease in the background of hundreds of other VOCs a breath analysis system must be highly dimensional, able to distinguish the signature pattern of diverse VOC biomarkers in a diverse chemical background. The metabolite biomarkers are often important at very low part per billion (ppb) concentrations. Therefore it is critical the breath analysis system sensor be very sensitive, able to detect diverse VOC biomarkers often in low single digit ppb concentration. The Metabolomx sensor has the high dimensionality and sensitivity to capture the chemical signature pattern of the complex mixture of VOC’s present in breath.

Technology Review - Metabolomx, a startup in Mountain View, California, recently completed a clinical trial that shows that its breath test can spot lung cancer with 83 percent accuracy and can also distinguish between several different types of the disease, something that usually requires a biopsy. The accuracy of the test matches what's possible with low-dose computerized tomography imaging of the lungs.

Please continue reading at: