Abstract:
With the rapidly changing landscape of early-stage breast cancer, there is a significant need for upfront biomarkers to enable physicians to “right-size” therapy for patients. It is well known that tumor heterogeneity drives the variability of response to neoadjuvant therapy (NAT) in breast cancer. We developed TumorScope (TS), a 4D tumor modeling platform that captures the chemical and biophysical interactions within the tumor and the interactions with the surrounding tissue to predict response to a physician’s choice of therapy during treatment planning. TS combines models for tumor morphology, metabolism, drug delivery and behavior, and vascularity. The integration of different drug models within the platform allows visualization of response to multiple NATs for a specific patient tumor.
Here we present a preliminary meta-analysis of validation across 6 single site studies with prospectively defined end points. We used pre-treatment standard of care (SOC) diagnostic data (demographics, drug regimen, imaging (DCE MRI), and pathology) as TS inputs. Using only this data, TS predicted the weekly volumetric response throughout the course of treatment, pathological complete response(pCR) and likelihood of recurrence to the physician’s choice of therapy.
Ground Truth of pCR was available for 350 patients where TS predictions were assessed to have an accuracy of 92% with sensitivity and specificity of 89.9% and 92.5%, respectively. The performance of TS was robust across all breast cancer subtypes; accuracy, sensitivity and specificity were 92%, 91%, 92% in TNBC; 92%, 91%, 92% in HER2+; and 92%, 83%, 93% in HR+/HER2- patients. TS simulated volumetric response had a median absolute volumetric error of 3.4% compared to the radiographic volume extracted from pre-surgery DCE MRIs.
In summary, TS can accurately predict tumor response to all standard of care therapies currently approved by NCCN guidelines. With upfront radiographic response (volume) and pathologic (pCR, residual cancer burden) predictions TS allows better assessment of the risk and individualized benefit of certain regimens to individualize de-escalation, escalation or expedite enrollment in clinical trials.
View the full publication here.
