Table A3.

A phenomenological Power model gives the best fit for the subset of the data. B16 tumor dynamics in two settings (at $T={10}^6\mathrm{cell}/\mathrm{mL}$ and $E={10}^6\mathrm{cell}/\mathrm{mL}$ from Dataset 3 and $T={10}^5\mathrm{cell}/\mathrm{mL}$ and $E={10}^6\mathrm{cell}/\mathrm{mL}$ from Dataset 4) is not monotonic (Figure A1). We fit 4 alternative models (Equations (3)–(6)) to the subset of the data that excludes these two settings for Datasets 1–4 (top) or Datasets 1–5 (bottom). Other details are similar to those given in Table 1.

Datasets 1–4  (Subset)  $\mathit{n}=371$
Models	$\alpha$	r	k	h	n	$g_0$	$g_1$	$g_2$	SSR	AIC	$\Delta$	w
MA	2.88	0.72	3.84 $\times {10}^{-7}$						88	526	99.7	0
Sat	2.74	0.72	4.8	2.49 $\times {10}^6$					71	451	24.7	${10}^{-6}$
Power	2.67	0.74	0.004		0.423				66.7	426.3	0	0.93
SiGMA	2.68		3.17 $\times {10}^{-7}$			6.84 $\times {10}^{-8}$	0.72	7930	67.3	431.4	5.1	0.072
Datasets 1–5  (subset)   $\mathit{n}=\mathbf{378}$
Models	$\alpha$	r	k	h	n	$g_0$	$g_1$	$g_2$	SSR	AIC	$\Delta$	w
MA	2.85	0.32	1.87 $\times {10}^{-7}$						724	1327	889	0
Sat	2.86	0.72	9.36	1.39 $\times {10}^7$					82	503	65	0
Power	2.62	0.72	0.01		0.37				69	438	0	1
SiGMA	2.94		1.76 $\times {10}^{-7}$			1.18 $\times {10}^{-7}$	0.84	252	544	1222	784	0