FIGURE 2 from A Pilot Study on Patient-specific Computational Forecasting of Prostate Cancer Growth during Active Surveillance Using an Imaging-informed Biomechanistic Model

Mapping ADC to N and GS. A, Nonlinear least-squares fit of a hyperbolic tangent model (blue solid line) of the mean values of ADC in the tumor across GS groups in previous studies (red dots), which were normalized with respect to the mean ADC in healthy tissue (ADCh) reported for each of them (41–45). We leverage a continuous, extended GS scale in which 0 < GS < 2 indicates healthy and pretumoral tissue and GS > 2 corresponds to prostate cancer (i.e., overlapping the standard discrete values of the GS used to assess histopathologic samples of prostate cancer; see refs. 2, 4). Further details on the hyperbolic tangent model and the fitting method are reported in Supplementary Data S1. B, Mapping of the hyperbolic tangent model obtained for the normalized ADC values (ADC/ADCh, A) to normalized tumor cell density values (i.e., , B). We use the linear mapping in Eq. (1), which introduces a negative proportionality between ADC and tumor cell density, as in other mpMRI-informed biomechanistic models of solid tumor growth (18, 21, 28, 29).