Page 303 - IJB-10-3
P. 303

International Journal of Bioprinting                              Design and optimization of 3DP bioscaffolds





                           First optimization step                        Second optimization step
              (a)                                               (f)

















               (b)                    (c)                      (g)                   (h)










               (d)                    (e)                      (i)                   (j)












            Figure 12. Optimization process of the channeled scaffold. (a) The structures used for model simulation in the first optimization step. The channel
            volume fraction increased by 2.5% in each step. The dotted line indicates the specific surface area ratio between channeled and solid structures. Panels b,
            c, d, and e illustrate the normalized average oxygen concentration, hypoxic region volume fraction, normalized average cell density, and cell count after
            five iterations of calculation, respectively. The red dashed rectangles indicate the optimal channel volume fraction where the cell count is maximum. (f)
            Structures adopted for model simulations in the second optimization step. The dotted line expresses the specific surface area ratio. Panels g, h, i, and j
            depict the normalized average oxygen concentration, hypoxic region volume fraction, normalized average cell density, and cell count for structures with
            various channels, respectively.


            and maximum cell growth rate influence the oxygen   carried out (Figure 12). The optimal channel size, number
            concentration and cell growth. A purely trial-and-error   and distribution ensuring maximum cell growth after a
            approach is thus problematic for obtaining the optimal   7-day culturing were obtained using a two-step method
            values of these parameters simultaneously. The modeling   assisted with the model. Figure 12a shows the process of
            system can help find the desired parameter values prior to   finding the threshold of the void volume fraction with the
            conducting actual experiments, or narrow the range of the   maximum cell number while the oxygen concentration
            values for testing.                                and cell density are high. Commencing from a solid
                                                               cylindrical scaffold, the channel volume fraction increased
            4.3. Structural optimization                       up to 12.5% with an increment of 2.5%. The average oxygen
            To further demonstrate the capability of the modeling   concentration and cell density ascended gradually within
            system in optimizing the design of a cell-laden scaffold, the   such a volume fraction range (Figure 12b and d). However,
            optimization for a channeled scaffold, as an example, was   the cell count reached the peak when the channel volume


            Volume 10 Issue 3 (2024)                       295                                doi: 10.36922/ijb.1838
   298   299   300   301   302   303   304   305   306   307   308