Here the explanation why the results from 31/07 we
Post# of 30028
Posted On: 08/03/2014 7:21:13 AM
Here the explanation why the results from 31/07 were so outstanding (from yahoo board):
fibonacci1170
.
Whoever he is, he's posted some challenging questions for GC on seeking alpha. And an SA commenter asked him if lympro sensitivity/specificity numbers would improve using multivariate analysis. Fibonacci's response:
" Absolutely! ... When multiple disease markers are measured on an individual, the information obtained from these markers can be combined to diagnose the disease. The idea is to choose linear combinations of these markers in order to optimize diagnostic accuracy. Specifically, the accuracy index to be maximized is the area under the Receiver Operating Characteristic ROC curve. Such multivariate analysis is routinely done in many fields, astronomy, medicine (cancer diagnosis) etc.
Kirby's slide-12 with CD19 data was a jaw dropping moment. Why? Because that data comes from a mere single (I repeat, single) disease marker, ie. CD19. And yet it had a p-value of 0.0104 (meaning, very highly significant finding that cannot be attributed to mere chance). On Slide-13, look at the ROC plot for CD19. The Area Under Curve (AUC) is freaking 80% (with just a single disease marker, CD19) and the p-value is a staggering p=0.003.
AMBS is going to combine at least 6 biomarkers by forming their linear combinations and maximize the area under the ROC curve. From what I know, LymPro is on track to cross 93% spec/sens numbers.
An average investor on a stock market can not be expected to know medicine/biostatistics in great detail. Kirby's slides were ok if presented to medical audience, but not ok when presenting to investor audience. That is where AMBS got disconnected from its investors and the market punished it.
Please share this info with your social contacts
fibonacci1170
.
Whoever he is, he's posted some challenging questions for GC on seeking alpha. And an SA commenter asked him if lympro sensitivity/specificity numbers would improve using multivariate analysis. Fibonacci's response:
" Absolutely! ... When multiple disease markers are measured on an individual, the information obtained from these markers can be combined to diagnose the disease. The idea is to choose linear combinations of these markers in order to optimize diagnostic accuracy. Specifically, the accuracy index to be maximized is the area under the Receiver Operating Characteristic ROC curve. Such multivariate analysis is routinely done in many fields, astronomy, medicine (cancer diagnosis) etc.
Kirby's slide-12 with CD19 data was a jaw dropping moment. Why? Because that data comes from a mere single (I repeat, single) disease marker, ie. CD19. And yet it had a p-value of 0.0104 (meaning, very highly significant finding that cannot be attributed to mere chance). On Slide-13, look at the ROC plot for CD19. The Area Under Curve (AUC) is freaking 80% (with just a single disease marker, CD19) and the p-value is a staggering p=0.003.
AMBS is going to combine at least 6 biomarkers by forming their linear combinations and maximize the area under the ROC curve. From what I know, LymPro is on track to cross 93% spec/sens numbers.
An average investor on a stock market can not be expected to know medicine/biostatistics in great detail. Kirby's slides were ok if presented to medical audience, but not ok when presenting to investor audience. That is where AMBS got disconnected from its investors and the market punished it.
Please share this info with your social contacts
(0)
(0)Amarantus Bioscience Holdings (AMBS) Stock Research Links
Scroll down for more posts ▼