COMMENT & RESPONSEIn Reply Thank you to Drs Huang and Liu for their thoughtful comments on our analysis of a randomized clinical trial evaluating the relationship between residual cancer burden (RCB) and event-free survival in the I-SPY2 trial. 1 They raise the point that local therapy received may influence recurrence events, as prior reports have shown an increased risk of locoregional recurrence in patients receiving neoadjuvant chemotherapy. 2 Additionally, they suggest incorporation of stromal features into prognostic tools, such as RCB.We have previously presented data from the I-SPY2 trial 3 showing that patients with improved response to chemotherapy do have significantly higher rates of breast conservation than those with higher RCB. However, type of local therapy (breast conservation, mastectomy, with or without postmastectomy radiotherapy) was not associated with local recurrence in the I-SPY2 trial. Rather, we found that RCB was the strongest predictor of local recurrence events regardless of type of local therapy received. While older studies may have shown slightly higher local recurrence events in neoadjuvant trials, this may be mitigated by improvements in systemic therapy and imaging. Indeed, serial imaging with breast magnetic resonance imaging is a key component of the I-SPY2 trial and may allow for more precise surgical planning and improved selection of breast conservation candidates compared with trials that relied primarily on physical examination and mammography for assessing local disease extent.Regardless, if downstaging a tumor with neoadjuvant therapy to allow for breast conservation does result in higher rates of local recurrence, we would expect that patients with lower RCB would have more recurrence events, biasing our results toward the null hypothesis. That we found the opposite strengthens the conclusion regarding the prognostic ability of RCB. 1 An updated analysis of local therapy and recurrence events with longer follow-up time in the I-SPY2 trial is currently ongoing.Regarding the incorporation of stromal features into response evaluation, we agree that additional information can contribute to prognostic tools. While RCB is already highly prognostic, we are actively evaluating tumor stromal features, including immune infiltrates and their spatial heterogeneity. 4 Additionally, radiomics features, such as functional volume change on dynamic contrast-enhanced magnetic resonance imaging, also provide response information that can be included. 5 By integrating these features, we hope to not only better prognosticate, but also better predict response so that we can treat patients with the right drugs at the right time.