Intention to treat.
The Study: Harvard researchers attempted to further tease out the confounded relationship between immune checkpoint inhibition (ICI) and corticosteroid use in another (ok, yes) retrospective study. It included 650 patients receiving ICI for non-small cell lung cancer (NSCLC) divided into 3 groups by prednisone-equivalent dose within 24 hours of ICI initiation:  0 to <10 mg,  ≥10 mg for cancer-related palliation (read: anorexia, dyspnea, pain, brain mets), or  ≥10 mg for other indications. Quick rad onc translation: 10 mg pred = 1.5 mg dex. Not surprisingly, a superficial glance at patients in group  on <10 mg pred demonstrated significantly better objective responses (11 → 20%) and median progression-free (PFS 2 → 3 months) and overall (OS 5 → 11 months) survival than groups [2-3] on ≥10 mg pred. But the real outlier, in a bad way, emerged as group  on cancer-related pred who suffered significantly shorter survival outcomes than the other groups: median PFS was  3,  1, and  5 months and median OS  11,  2, and  11 months. In fact, there was no significant difference in survival outcomes between groups  and  at all, suggesting the reason for steroids rather than the dose of steroids is the real tell. A strength of this study worth noting is that tumor mutation burden and PD-L1 expression are actually balanced between the groups receiving ≥ 10 mg prednisone.
TBL: The relationship between steroid dose, ICI efficacy, and oncologic outcomes may simply be that patients with worse disease require more steroids. | Ricciuti, J Clin Oncol 2019