. 2021 Apr 13;11:8076. doi: 10.1038/s41598-021-87470-x

Table 3.

Hypothesis testing results using the chosen mean functional form for each PDX model and sp(pow) covariance structure for all PDX models.

PDX	Comparison		P values				Conclusion
PDX	Comparison		Coincident curves (2df)	Intercept (AUC)	Slope	Quad	Conclusion
PH80	Diluent	MK	0.0590	0.0491	0.1299		Marginal activity
		Chemo	< 0.0001	< 0.0001	< 0.0001		Shows activity
		Chemo + MK	< 0.0001	< 0.0001	< 0.0001		Shows activity
	Chemo	Chemo + MK	0.2715	0.9441	0.1038		Not better than chemo
PH87	Diluent	MK	0.0989	0.2042	0.0757		Marginal activity
		Chemo	< 0.0001	0.0164	< 0.0001		Shows activity
		Chemo + MK	< 0.0001	0.0747	< 0.0001		Shows activity
	Chemo	Chemo + MK	0.7616	0.5204	0.5437		Not better than chemo
PH77	Diluent	MK	< 0.0001	0.0023	< 0.0001		Shows activity
		Chemo	< 0.0001	< 0.0001	< 0.0001		Shows activity
		Chemo + MK	< 0.0001	< 0.0001	< 0.0001		Shows activity
	Chemo	Chemo + MK	0.7587	0.9648	0.4575		Not better than chemo
PH95	Diluent	MK	0.9972	0.9865	0.9420		No activity
		Chemo	0.0007	0.0139	0.0008		Shows activity
		Chemo + MK	0.0001	0.0068	0.0002		Shows activity
	Chemo	Chemo + MK	0.9377	0.7845	0.7914		Not better than chemo
PH39	Diluent	MK	< 0.0001	0.0005	0.0030	0.8080	Shows activity
		Chemo	< 0.0001	< 0.0001	0.0021	0.9792	Shows activity
		Chemo + MK	< 0.0001	< 0.0001	0.0006	0.3725	Shows activity
	Chemo	Chemo + MK	0.0290	0.0090	0.2309	0.0255	Improvement above chemo in mean, not growth rate

Comparisons of Chemo versus Chemo + MK assess the primary question of the study, i.e., whether addition of the PARPi MK to chemo improves the performance. Comparisons with diluent assess the secondary question of whether each arm shows activity. The test of coincident curves used 2 degrees of freedom for PH80, PH87, PH77, PH95, and 3 degrees of freedom for PH39.

PDX Patient derived xenograft, Sp(pow) Spatial(power), Chemo Chemotherapy, MK MK-4827, PARPi Polymerase inhibitor, AUC Area under the curve.