Background: Pathological complete response (pCR) is a well validated endpoint for early-stage breast cancer (EBC) patients treated with neoadjuvant chemotherapy (NAT), with a strong correlation with long term outcomes such as event free survival (EFS) and overall survival (OS). However, given the growing armamentarium of NAT regimens, pre-treatment predictors of longterm outcome are urgently needed in order to identify patients for escalation or de-escalation of therapy. SimBioSys TumorScope (TS) is a commercially available biophysical simulation platform that utilizes baseline MRI, receptor status, and planned treatment regimen to simulate response to NACT over time. TS has demonstrated accurate prediction of pCR (accuracy >90%) in prior studies. Here, we evaluate TS predicted pCR as a prognostic biomarker which can be assessed prior to treatment initiation.
Methods: An independent blinded retrospective study was performed by University of Chicago for patients who received NACT for EBC from Jan 2010 – March 2020. Patients must have had a pretreatment breast MRI. TS classified patients as likely pCR/low risk of recurrence when predicted residual tumor volume was < 0.01 cm3 or there was a 99.9% or greater reduction in tumor volume. The log-rank test was used to assess the prognostic value of pCR and TS predictions.
Results: In total, 144 tumors from 141 pts were analyzed. Average age was 52 yrs; 65% had stage II and 19% had stage III disease; 41% had TNBC, 34% were HER2+, and 25% were HR-/HER2+. TS classified 54 tumors as having low risk of recurrence/likely pCR and 90 as having high risk of recurrence/likely residual disease. With a median follow-up of 4.7 years, 4-yr EFS was 100% in low-risk patients, and 80% for high-risk patients (hazard ratio [HR] 6.71, p = 0.002). This compared favorably to pCR as a prognostic indicator (HR for EFS 4.07, p = 0.02). Out of 94 patients who had residual disease after NAT, TS identified 10 patients as low-risk and 84 patients as high-risk – no events were seen in the low-risk group (HR not evaluable, p = 0.10).
Conclusions: TS accurately predicts EFS of EBC patients treated with NAT from pre-treatment MRI and clinicopathologic data, comparable to the predictive accuracy of pCR. There was also a trend towards improved EFS among patients with residual disease predicted by TS to be low-risk. TS may have utility to guide the escalation/de-escalation of treatment regimens, and further evaluation in the neoadjuvant and adjuvant setting is ongoing.