MRI-informed mechanistic model to guide patient-specific optimization triple-negative breast cancer response to neoadjuvant chemotherapy
Please login to view abstract download link
Neoadjuvant systemic therapy has been the standard-of-care of locally advanced triple-negative breast cancer (TNBC). However, only 50% of TNBC patients achieve a pathological complete response (pCR) to conventional neoadjuvant chemotherapy (NAC). Aside from the need to develop new therapies with higher efficacy and lower toxicity, a critical barrier is the lack of rigorous ways to personally tailor therapeutic regimens. In this study, we aim to improve TNBC response to NAC by establishing an image-informed math model to identify optimal therapeutic schedules for individual patients. This study employed a cohort of 105 TNBC patients in ARTEMIS trial, who received four cycles of Adriamycin/Cytoxan (A/C) every 2-3 weeks followed by 12 cycles of weekly Taxol (T). Longitudinal multiparametric MRI were acquired to evaluate tumor morphology, perfusion, and cellularity. A biology-based mathematical model was calibrated to the MRI data collected before, during, and after A/C, creating personalized digital twins. The digital twins predicted pCR status after NAC with an AUC of 0.82. Each patient’s digital twin was used to predict individual pCR status under 128 clinically feasible NAC schedules. This optimization significantly improved pCR rates over the observed rates in the cohort by 20.95%-24.76%. The digital twins were validated by virtually recapitulating A/C-T regimens investigated in previous clinical trials (INT C9741, ECOG 1199, SWOG S0221). The predictions indicated that bi-weekly A/C-T led to a pCR rate (73.33%) significantly higher than tri-weekly A/C-T (49.52%), weekly and bi-weekly T led to higher pCR rates (55.24%, 60.00%) than tri-weekly T (49.52%), and combinations of weekly or bi-weekly A/C and T showed comparable pCR rates (79.05%, 72.38%, 73.33%, 69.52%). These findings align with previous trials. Overall, our digital twin approach offers a practical method for patient-specific NAC tailoring, with significant implications for adaptive clinical trials.