Fig. 3.

Tumor volume plotted against patient’s age for the dominant PNs (A) and DNLs (B). Each tumor is shown as a series of line segments. (C, D) Tumor volume plotted against patient’s age for the dominant PNs (C) and DNLs (D), but here the PNST volume is transformed (cube root), quartiles and 10th/90th percentiles are indicated on both axes, and linear regression lines are plotted along with the data (dots). (E, F) Association between estimated tumor growth rate to patient’s age at initial MRI for dominant PNs (E) and DNLs (F). A moderate negative correlation was observed for PNs, whereas only a weak Fig. 3 Continued. association was noted for DNLs. (G, H) Association between estimated tumor growth rate to initial tumor volume for dominant PNs (3G) and DNLs (H). A moderate negative correlation was observed for PNs and DNLs. (E–H) The 42 dominant PNs and 24 DNLs which had not received any tumor-directed medical therapies prior to the “treatment-free” portion of follow-up for which growth rates were calculated (no prior treatment subset) are shown in red, and the remaining PNs and DNLs are shown in blue. (I) Growth rates of the non-dominant PN relative to the growth rate of the dominant PN in the 7 patients (8 non-dominant PNs) who had multiple PNs where tumor volumes were measured over the same follow-up period. No pattern could be noted in our small number of patients. (J) Growth rates of the DNLs relative to the growth rate of the patients’ dominant PN in 26 patients (42 DNLs) where volumes were measured over the same follow-up period. The diagonal lines on I and J indicate the points where the growth rates are equal.