TY - JOUR
T1 - Data and implication based comparison of two chronic myeloid leukemia models
AU - Everett, R. A.
AU - Zhao, Y.
AU - Flores, K. B.
AU - Kuang, Y.
PY - 2013/10
Y1 - 2013/10
N2 - Chronic myeloid leukemia, a disorder of hematopoietic stem cells, is currently treated using targeted molecular therapy with imatinib. We compare two models that describe the treatment of CML, a multi-scale model (Model 1) and a simple cell competition model (Model 2). Both models describe the competition of leukemic and normal cells, however Model 1 also describes the dynamics of BCR-ABL, the oncogene targeted by imatinib, at the sub-cellular level. Using clinical data, we analyze the differences in estimated parameters between the models and the capacity for each model to predict drug resistance. We found that while both models fit the data well, Model 1 is more biologically relevant. The estimated parameter ranges for Model 2 are unrealistic, whereas the parameter ranges for Model 1 are close to values found in literature. We also found that Model 1 predicts long-term drug resistance from patient data, which is exhibited by both an increase in the proportion of leukemic cells as well as an increase in BCR-ABL/ABL%. Model 2, however, is not able to predict resistance and accurately model the clinical data. These results suggest that including sub-cellular mechanisms in a mathematical model of CML can increase the accuracy of parameter estimation and may help to predict long-term drug resistance.
AB - Chronic myeloid leukemia, a disorder of hematopoietic stem cells, is currently treated using targeted molecular therapy with imatinib. We compare two models that describe the treatment of CML, a multi-scale model (Model 1) and a simple cell competition model (Model 2). Both models describe the competition of leukemic and normal cells, however Model 1 also describes the dynamics of BCR-ABL, the oncogene targeted by imatinib, at the sub-cellular level. Using clinical data, we analyze the differences in estimated parameters between the models and the capacity for each model to predict drug resistance. We found that while both models fit the data well, Model 1 is more biologically relevant. The estimated parameter ranges for Model 2 are unrealistic, whereas the parameter ranges for Model 1 are close to values found in literature. We also found that Model 1 predicts long-term drug resistance from patient data, which is exhibited by both an increase in the proportion of leukemic cells as well as an increase in BCR-ABL/ABL%. Model 2, however, is not able to predict resistance and accurately model the clinical data. These results suggest that including sub-cellular mechanisms in a mathematical model of CML can increase the accuracy of parameter estimation and may help to predict long-term drug resistance.
KW - Cancer modeling
KW - Chronic myeloid leukemia
KW - Clinical data
KW - Drug resistance
UR - http://www.scopus.com/inward/record.url?scp=84884494686&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84884494686&partnerID=8YFLogxK
U2 - 10.3934/mbe.2013.10.1501
DO - 10.3934/mbe.2013.10.1501
M3 - Article
C2 - 24245631
AN - SCOPUS:84884494686
SN - 1547-1063
VL - 10
SP - 1501
EP - 1518
JO - Mathematical Biosciences and Engineering
JF - Mathematical Biosciences and Engineering
IS - 5-6
ER -