adam feuerstein: NEOP -- a view from the short side... Here's what one short seller is saying about the Lymphoseek data...
Concordance with Vital Blue dye for both the -05 and -09 studies remained 100% and should not be surprising. Key is the per patient data as that’s all that’s really clinically relevant.
As previously released, the per node assessment showed that in 133 patients in the -09 study Lymphoseek failed to detect 0% of LN’s whereas Vital blue dye missed 25% or 33 patients. Key question was are these simply nodes down stream from the sentinel node al;ready detected by blue dye….if so then ID’ing added nodes is meaningless.
In fact in the per patient assessment the company had to POOL both the 0-5 and 09 studies in order to demonstrate any findings and did not break out the individual studies alone. Further more while the Lymphoseek Failed detection rate remained at 0% the vital bue dye only missed 7.3% of 4 patients…..way down from the 33 in the 09 study alone on a per node basis…..this basically confirms that the vast majority of cases that Lymphoseek picked up were simply added nodes beyond the sentinel node…..which is clinically irrelevant. Plus to hit a p value of <0.044…..barely stat sig…..they had to pool the studies…..which you can’t do from the FDA’s point of view.
Intuitively, you probably think that P=0.0001 is more statistically significant than P=0.04. Using strict definitions, this is not correct. Once you have set a threshold P value for statistical significance, every result is either statistically significant or is not statistically significant. Some statisticians feel very strongly about this.
You must pool data from all pivotals trials. Companies are not allowed to selectively choose data or they would eliminate anything negative. all data must be included in the Nda and Fda looks at at it all.