Almost all patients received combination therapy, respectively. The in vivo response data reported by clinical oncologist were used to define the RECIST status of the patient. Any observable response (complete or partial remission) was considered in vivo sensitive, and all other responses were considered as in vivo resistant (stable disease, progressive disease).
Subsequently, a true-positive (TP) correlation corresponds to clinical response (complete or partial remission) while in vitro treatment induced at least a 15% drop in the viability index compared to controls (readout values below 85%). A true-negative (TN) correlation corresponds to in vitro viability index higher than 85% (readout over the 85% cut-off point) and the patient showing no response. False-positive (FP) correlations correspond to patients showing no response but the in vitro viability index was below 85% and, finally, false-negative (FN) correlation means a viability index over 85% although the patient had a clinical response.
Sensitivity (%) was calculated as 100× TP/(TP+FN), specificity as 100 x TN/(TN+FP) and diagnostic accuracy was calculated as (TP+TN)/N, while “N” being the sample size
Based on the above, data showed that Humeltis’ Predictive Chemosensitivity Assay is highly sensitive (100%), highly specific (86%), which altogether leads to highly accurate (89%) treatment predictions.