In many areas of research, the p-value of .05 is customarily treated as a "border-line acceptable" error level.
A Receiver Operating Characteristic (ROC) analysis of the sensitivity and specificity was shown to have an associated "p value" of less than .001, resulting in a statistically significant finding of CyPath(R)'s ability to discriminate between high-risk and cancer groups.
The Rainmaker, interesting post regarding p-values and have you ever seen many pre-clinical posted p-values?
.001 is certainly quite impressive, but what would you think of a p-valoue of .0048 in a pre-clinical setting for combination immunotherapy treatment ?
Pre-clinical is seldom the place where any biotech dares to find the resulting p-value, or I should say dares to publish it.... because in pre-clinical its seldom successful. Phase III I see some biotechs increasing and increasing their n... in order to just reach a good p-value that the FDA will like. Combination therapy / immunotherapy will be the biggest portion of all biotechs pipelines in percentage of revenue in the near term and here we have a global biomarker in flipped PS, that seems to be one hell of a biotech value.
Specifically, following once weekly treatments of ch1N11 plus anti-PD-1, tumor growth was inhibited by 78.7% (p= 0.0048 on day 23) compared anti-PD-1 alone. In addition, 50% of the tumors treated with the combination either regressed or did not progress compared to 0% for anti-PD-1 alone. Also, the once weekly combination treatment with ch1N11 and anti-PD-1 led to a 78% and 81% increase in intratumoral CD4+ and CD8+ T cells, two key indicators that show that tumor fighting immune cells are present in the local tumor environment, compared to the single agent anti-PD-1.