Introduction: When considering the use of NACT for breast cancer, healthcare providers and patients must weigh the risks associated with different potential standard of care (SoC) treatment regimens against the likelihood of achieving a desired outcome. The decision of treatment option is currently made using standard NCCN guidelines matched with diagnostic data, population statistics derived from clinical trials of the regimens and a clinician’s expert judgement. Given the potential variability in decision making, there is a significant need to standardize how data is used in this process and the ability to systematically individualize the treatment decision for a patient. This individualization presents the opportunity to lower toxicity and costs where the response would be identical. This was a major topic at the 2019 American Society of Clinical Oncology (ASCO) Annual meeting with multiple oncologists and researchers seeking methods to deescalate care.
Method: Among the most common choices, anthracycline-based regimens are associated with only marginally better clinical outcomes than comparable taxane-based regimens (2.5% difference in 4-year disease free survival [Blum et al. 2017]), but they carry greater risks of cardiotoxicity and secondary leukemia. They are both preferred regimens per NCCN guidelines in HER2-negative breast cancer, which leaves the decision of when to use which regimen largely up the healthcare provider. With the rise of targeted but expensive therapies like trastuzumab and pertuzumab, and immunotherapies like pembrolizumab in clinical trials, the risk-benefit calculus is becoming even more complex. Moreover, the goals of NACT often vary from patient to patient. While complete pathological response is always the ideal, NACT is routinely used to downstage a tumor in order to enable breast conserving surgery or to simplify the tumor’s resection.
With the above in mind, we selected and reviewed tumor response of 50 randomly selected patients from the ISPY 1 dataset (measured via initial, inter-regimen and pre-surgery MRIs) and utilized the SimBioSys TumorScope in silico simulations to predict response to alternate therapies.
Results: Using the base data set and tumor response resulting from the AC (anthracycline + cyclophosphamide) regimen compared to resulting response from the T (Paclitaxel) arm of the regimen. 22 out of the 50 patients were shown to have a superior response to the T. For this analysis, we ignored the synergistic effect of the use of AC prior to the T. Since both Docetaxel and Paclitaxel are taxanes, the analysis provided ideal candidates for the in-silico simulation of these patients using the TC regimen (known to be less toxic). The analysis showed a majority number of patients that had similar (~5% variation) responses to TC as they did with AC. This was clearly depicted by Study ID 1217 where both ACT and TC would lead to pCR. More promisingly, there was a subset of patients that would achieve pCR with the TC regimen vs the ACT such as Study ID 1127.
Conclusion: While the results above are preliminary and will be further expanded via a future prospective study, there is clearly an opportunity for us to address the variation that exists in ACT vs TC selection. Whether it’s treatment in the East vs West Coast or a community vs an academic medical center, the variation in opinion is undeniable. There is a need to standardize yet individualize how this decision is made today – an in-silico determination might be solution.