“…Such a practice invalidates the usual interpretation attached to a P-value. [18] Univariate analysis was performed using Kaplan-Meier method [19] for survival and log-rank test [20] for comparisons. Variables examined as dichotomous were: Age (<55 years vs. ≥55 years), gender (male vs. female), smoking (non-smoker vs. smoker), stage (limited vs. extensive), KPS (≥70 vs. <70), symptom burden (≤5 vs. >5), SVC obstruction (no vs. yes), brain metastases (no vs. yes), bone metastases (no vs. yes), liver metastases (no vs. yes), adrenal metastases (no vs. yes) and all the lab parameters (<median vs. ≥median).…”