Table 2.
Statistical analysis of cancer cell movements in vivo
| Cell type | MSD (α) | MLE for μ | Akaike weights | ||
|---|---|---|---|---|---|
| Truncated power law | TP | P | E | ||
| B16-F0 | |||||
| Zone 1 | 1.03 ± 0.04 | 3.58 ± 0.24 | 0.77 | 0.22 | < 0.01 |
| Zone 2 | 1.38 ± 0.04 | 3.57 ± 0.21 | 0.57 | 0.42 | < 0.01 |
| Zone 3 | 0.99 ± 0.04 | 3.85 ± 0.35 | 0.59 | 0.40 | < 0.01 |
| Zone 4 | – | – | – | – | – |
| B16-F10 | |||||
| Zone 1 | 1.07 ± 0.07 | 4.97 ± 0.55 | 0.58 | 0.41 | < 0.01 |
| Zone 2 | 1.63 ± 0.06 | 2.34 ± 0.21 | 0.90 | 0.09 | < 0.01 |
| Zone 3 | 1.47 ± 0.05 | 2.66 ± 0.26 | 0.93 | 0.06 | < 0.01 |
| Zone 4 | 1.87 ± 0.05 | 2.03 ± 0.22 | 0.99 | < 0.01 | < 0.01 |
TP truncated power law, P power law, E exponential distribution
Table summarizes exponent α values and the maximum likelihood estimates, MLEs70, of the truncated power law exponents μ for cancer cell movements in vivo for the data described in Fig. 3 and 4 (± values give the 95% confidence intervals). Lower cutoff value was a = 5 min for all in vivo data sets. Away from the tumor, in zones 2–4, Akaike weights heavily favor truncated power law with 2 < μ < 3 (shown in bold) indicating Lévy walks for B16-F10. For B16F0 for all zones, distributions correspond to truncated power law with μ > 3 indicating diffusive walks. See Supplementary Movies 7 and 8. Statistics were based on the following numbers of cells and time points − B16-F0, zone #1 (n = 59 cells, m = 2,925 data points); zone #2 (n = 57, m = 2,806); zone #3 (n = 28, m = 1,383). B16-F10, zone #1 (n = 30 cells, m = 990 data points); zone #2 (n = 30, m = 990); zone #3 (n = 29, m = 957); zone #4 (n = 23, m = 759)