A community challenge to predict clinical outcomes after immune checkpoint blockade in non-small cell lung cancer

Table 1 Challenge questions and metrics used for performance evaluation [23]

Sub-challenge	Challenge Question	BM	Primary Metric^a for Performance Evaluation
PFS (Sub-challenge 1)	PFS/OS sub-challenges: Predict response to nivolumab, in terms of PFS/OS, via an immune checkpoint-specific model using clinical, demographic, and gene expression data	PFS/OS Harrel’s C-index	DSS BM of PFS/OS between the nivolumab and chemotherapy arms
OS (Sub-challenge 2)		PFS/OS Harrel’s C-index
BOR (Sub-challenge 3)	BOR sub-challenge: Predict which patients will not respond and have a BOR of PD	ROC-AUC	DSS BM of BOR between the nivolumab and chemotherapy arms

AUC area under the curve, BOR best overall response, C-index concordance index, DSS BM difference in squared scaled basal metrics, OS overall survival, PD progressive disease, PFS progression-free survival, ROC receiver operating characteristic
^aThe computing of the primary metric from the BM is shown in Additional file 1: Fig. S1

ISSN: 1479-5876