Towards practical digital twins to optimize TNBC treatment stratification

About This Grant

PROJECT ABSTRACT Triple-negative breast cancer (TNBC) has the poorest prognosis among breast cancer subtypes due to its inherently aggressive clinical behavior and the absence of well-defined molecular targets. Achievement of pCR to neoadjuvant therapy has been widely recognized as a surrogate marker for improved long-term outcomes, including reduced risk of recurrence and enhanced overall survival. However, current neoadjuvant therapy stratification for TNBC is far from satisfying. With the conventional neoadjuvant chemotherapy (NAC), over half of TNBC patients do not achieve pCR and face poor prognoses. The new neoadjuvant chemoimmunotherapy (NACI) combined pembrolizumab with the NAC and has improved pCR and survival rates (by less than 10%), but also introduces substantial toxicity risks without reliable predictive methods to determine individual patient benefit. Given TNBC’s aggressive nature and lack of effective targeted therapies in practice, early identification of responders versus non-responders to a specific therapeutic regimen is of great important for assisting clinicians to tailor therapy accordingly. However, there is currently no method to foretell how much an individual patient may benefit from the addition of immunotherapy, nor to practically guide optimization of therapy on a patient-specific basis. Digital twin techniques have gained emerging attention in this context, which offer great promise for precision management of TNBC by integrating real-time clinical and multi-modal data into virtual patient models to support decision-making. However, their clinical adoption faces substantial challenges, including developing accurate predictive models capable of comparing multiple treatment options, ensuring continuous updates, and seamless integration into clinical workflows. To overcome these barriers, our study will establish cancer digital twins (CDT) to deliver reliable, real-time predictive metrics for personalized TNBC treatment. In particular, to support pre-treatment stratification (Aim 1), we will leverage the clinically available multi-modal data and state-of-the-art AI models to improve the accuracy of response prediction and benefit quantification, so identifying the optimal therapy option (i.e., NAC vs. NACI) for individual patient. To support on- treatment adaption (Aim 2), we will develop a mechanism-informed data assimilation framework to continuously update response prediction given individual monitoring data, so guiding a timely, interpretable adjustment of therapeutic regimens. To address implementation (Aim 3), we will assess the impact of real-world data heterogeneity on model uncertainty and bias, establish a CDT prototype to deliver model outputs within clinical workflows, and collect stakeholders feedback on its feasibility. We hypothesize that the developed CDT can not only accurately predict patient-specific response, but also provide a deployable workflow that assists clinical decision-making. Along with training to gain comprehensive understanding on real-world data heterogeneities, deployment considerations, and modern clinical trials for computational devices, this work will lay the groundwork for clinical validation and practical translation of personalized TNBC treatment optimization.