Subscribe to RSS
DOI: 10.1055/s-0039-1687955
Automated radiomic MRI phenotyping improves survival prediction in primary breast-cancer
Publication History
Publication Date:
28 May 2019 (online)
Background:
To investigate whether automated radiomic MRI-phenotyping (ARM) could improve survival prediction in primary breast-cancer.
Methods:
314 consecutive patients with primary invasive breast-cancer received standard staging breast-MRI before the initiation of treatment according to international recommendations. Diagnostic work-up, treatment, and follow-up was done at one tertiary care, academic breast-centre (disease-specific survival/DSS = 279; disease-specific death/DSD = 35; mean survival: 84.5 months). The Nottingham Prognostic Index (NPI) was used as the reference method with which to predict survival of breast-cancer. ARM was accomplished by commercially available, FDA-cleared software.
DSD served as endpoint. Integration of ARM into the NPI gave NPI+ (Cox regression). Prediction of DSD by NPI vs. NPI+ was investigated (Kaplan Meier statistics, HR: Hazard-ratio; alpha = 0.5, Logrank test; predictive accuracy: Harrell's C).
Results:
Prognostication of the endpoint by NPI (Harrell's C = 75.3%) was significantly enhanced by ARM (NPI+: Harrell's C = 81.0%, P = 0.03). Most of all, NPI+ more reliably identified patients with unfavourable outcome compared to NPI alone (HR = 3.14; P = 0.0001).
Conclusions:
Automated radiomic MRI-phenotyping (ARM) improved survival prediction in primary breast-cancer. Most of all ARM contributed to the identification of patients at higher risk of an unfavourable outcome.
Future studies should investigate ARM as a potential “gate keeper” in the management of breast-cancer patients. Such a “gate keeper” could assist in selecting patient benefitting from more advanced diagnostic procedures (genetic profiling etc.) in order to decide whether are a more aggressive therapy (chemotherapy) is warranted.