Study ID | Analysis form | ROI for BPE segmentation | Calculation formula for BPE levels | Calculation formula for change in BPE | DCE phases for BPE analysis | DCE acquisition time points or temporal resolution | MRI follow-up time points | No. of readers | Blindness to clinical data | Major findings |
---|---|---|---|---|---|---|---|---|---|---|
You et al. 2017 [14] | Fully automated | All the fifibroglandular tissue | BPE = (the enhanced fibroglandular tissue volume / total fibroglandular tissue volume) × 100% | ΔBPE1/2/3 = (BPE2nd/4th/6th follow-up MRI – BPEbaseline MRI) / BPEbaseline MRI *100% | The subtraction image of pre- and post-contrast MRI scans | At 90, 180, 270, and 360 s after injection | Before and after the 2nd, 4th, 6th, and 8th cycle of NAC | NR | NR | Reduction of BPE at the early stage of NAC was positively associated with pCR, especially in HR-negative status |
Rella et al. 2020 [15] | Semi-automated | All the fifibroglandular tissue | BPE = [(SI_post–SI_pre) / SI_pre] × 100% | The enhancement rate identified in the MRI after chemotherapy minus the enhancement rate identifified at the baseline MRI (total BPE change) | The pre- and first post-contrast acquisitions | NR | Before and after NAC (n = 101); Before NAC, after the 4th cycle of NAC and after NAC (n = 127) | 2 | NR | In the subgroup of patients with stages 3 and 4 breast cancers and who were diagnosed with a HER2-negative tumor phenotype, a significant association was found between early BPE change and pCR (P = 0.020) |
Onishi et al. 2021 [19] | Fully automated | The central 50% of axial sections in the breast | BPE = [(SI_early post – SI_pre) / SI_pre] × 100% | Evaluated as suppressed if ΔBPE was less than 0 | The pre- and early post-contrast acquisitions | 80–100 s per dynamic acquisition | Before treatment (T0), early treatment (3 weeks after treatment initiation, T1), interregimen (T2), and before surgery (T3) | 1 | NR | The association between BPE nonsuppression and lower pCR rate was detected at T2 and T3 in the HR-positive cohort |
Chen et al. 2015 [17] | Computer-based segmentation algorithm | All the fifibroglandular tissue | BPEi = ((SI_posti–SI_pre)/ SI_pre) × 100%; BPE = ∑BPEi / 12 | NR | 4 pre- and 12 post-contrast acquisitions | 42 s per dynamic acquisition | Before NAC; After 1st cycle or 2ed cycle of AC; After 4th cycle of AC or 2ed cycle of AC + 3 weekly second-line taxane-based regimen | NR | NR | Compared to baseline, BPE at F/U-1 was significantly decreased in the pCR group but not in the non-pCR group |
Arasu et al. 2020 [23] | Semi-automated | All the fifibroglandular tissue | BPE = (SI_early post – SI_pre) / SI_pre | %ΔBPE0_1 = (BPE_1 – BPE_0)/BPE_0 | Pre-contrast (time 0) and the first post-contrast acquisition (time 1) | Continued for at least 8 min after injection | Before NAC (T0), after 3 weeks of NAC, or early treatment (T1), after 12 weeks of therapy, or inter-regimen (T2), and after NAC and prior to surgery, (T3) | NR | NR | Among women with HER2-negative cancer, BPE alone demonstrated association with pCR in women with HR-positive HER2-negative breast cancer |
Moliere et al. 2019 [26] | Fully automated | All the fifibroglandular tissue | BPE20% = VBPE / VFGT × 100% | ΔBPE20% = (BPE20%post—BPE20%pre) / BPE20%pre | 3 subtraction images of pre- and the three post-contrast sequences, respectively | 90 s per dynamic acquisition; The late acquisition was centered at 6 min | Before and after NAC | 2 | Blinded to pathology data | There was no signifcant diference between complete responders and non-complete responders in term of pre- and post-therapeutic BPE |