Automated volumetric radiomic analysis of breast cancer vascularization improves survival prediction in primary breast cancer.

To investigate whether automated volumetric radiomic analysis of breast cancer vascularization (VAV) can improve survival prediction in primary breast cancer. 314 consecutive patients with primary invasive breast cancer received standard clinical 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-center (outcome: disease specific survival/DSS vs. disease specific death/DSD). The Nottingham Prognostic Index (NPI) was used as the reference method with which to predict survival of breast cancer. Based on the MRI scans, VAV was accomplished by commercially available, FDA-cleared software. DSD served as endpoint. Integration of VAV into the NPI gave NPIVAV. Prediction of DSD by NPIVAV compared to standard NPI alone was investigated (Cox regression, likelihood-test, predictive accuracy: Harrell's C, Kaplan Meier statistics and corresponding hazard ratios/HR, confidence intervals/CI). DSD occurred in 35 and DSS in 279 patients. Prognostication of the survival outcome by NPI (Harrell's C = 75.3%) was enhanced by VAV (NPIVAV: Harrell's C = 81.0%). Most of all, the NPIVAV identified patients with unfavourable outcome more reliably than NPI alone (hazard ratio/HR = 4.5; confidence interval/CI = 2.14-9.58; P = 0.0001). Automated volumetric radiomic analysis of breast cancer vascularization improved survival prediction in primary breast cancer. Most of all, it optimized the identification of patients at higher risk of an unfavorable outcome. Future studies should integrate MRI as a "gate keeper" in the management of breast cancer patients. Such a "gate keeper" could assist in selecting patients benefitting from more advanced diagnostic procedures (genetic profiling etc.) in order to decide whether are a more aggressive therapy (chemotherapy) is warranted.

Association between survival in patients with primary invasive breast cancer and computer aided MRI.

PURPOSE: To identify the potential of semi-quantitative enhancement-analysis in breast MRI to predict disease-related death in primary breast cancer patients. MATERIALS AND METHODS: The present study was planned and conducted according to international recommendations. All patients referred for pretherapeutic staging of primary breast cancer during 24 consecutive months were included into the study collective. They were followed-up by our multidisciplinary breast center. For semi-quantitative MRI-analysis dedicated CAD-software (computer assisted diagnosis) was used. Association between enhancement parameters and disease-related survival was investigated using Cox proportional-hazards -regression (CR). RESULTS: A total of 115 patients were eligible for CR analysis. Median follow-up time was 52 months. In 15 patients, disease-related death occurred. CR analysis identified four enhancement parameters as independent and significant (P < 0.001) predictors of the endpoint. Coefficients were "Initial enhancement" (B = 0.0166), "Time to peak-enhancement" (B = 1.0573), "Tumor volume" (B = 0.0175), and proportion of "tumor volume" showing "slow initial enhancement" followed by a "persistent" curve-type (B = -0.0586). CONCLUSION: This study demonstrates the significant relationship between semi-quantitative enhancement analysis in breast MRI and disease-related death of breast cancer patients. As results were extracted from a routine staging examination, MRI noninvasively provides not only diagnostic information but also outcome data at one step. Future studies should address the impact of these findings on patient management and therapeutic approach.

Computer assisted analysis of MR-mammography reveals association between contrast enhancement and occurrence of distant metastasis.

Dynamic contrast enhanced magnetic resonance imaging (DCE-MRI) is able to detect breast cancer with high sensitivity. Furthermore, this method provides functional information on tissue composition and vascularization. This study aims to identify the potential of DCE-MRI to predict distant metastasis in breast cancer patients using computer assisted interpretation of dynamic enhancement data. For this purpose, 59 consecutive patients with newly diagnosed invasive breast cancer received pretherapeutic DCE-MRI at 1.5 Tesla according to international recommendations. In all patients, follow up interval and occurrence of distant metastasis was documented. For DCE-MRI analysis dedicated software was used (Brevis, Siemens Healthcare, Erlangen, Germany). It allows semiautomatic identification of the most suspect curve in a lesion analyzed. Enhancement parameters assessed were "Initial Enhancement", "Washout", "Peak-Enhancement", and "Time to Peak Enhancement". Cox proportional hazards regression (CPHR) was used to analyze the effect of these parameters on the probability of metachronous distant metastasis. Median follow up period was 52.0 months. 6 patients developed distant metastases between 11 and 35 months after breast cancer diagnosis. In CPHR, Washout could be identified as significant and independent predictor for occurrence of distant metastasis (P = 0.0134). Our initial data demonstrate an association between computer measured enhancement parameters in DCE-MRI and occurrence of distant metastasis by quantification of Washout.

Artificial Neural Networks for differential diagnosis of breast lesions in MR-Mammography: a systematic approach addressing the influence of network architecture on diagnostic performance using a large clinical database.

RATIONALE AND OBJECTIVES: Differential diagnosis of lesions in MR-Mammography (MRM) remains a complex task. The aim of this MRM study was to design and to test robustness of Artificial Neural Network architectures to predict malignancy using a large clinical database. MATERIALS AND METHODS: For this IRB-approved investigation standardized protocols and study design were applied (T1w-FLASH; 0.1 mmol/kgBW Gd-DTPA; T2w-TSE; histological verification after MRM). All lesions were evaluated by two experienced (>500 MRM) radiologists in consensus. In every lesion, 18 previously published descriptors were assessed and documented in the database. An Artificial Neural Network (ANN) was developed to process this database (The-MathWorks/Inc., feed-forward-architecture/resilient back-propagation-algorithm). All 18 descriptors were set as input variables, whereas histological results (malignant vs. benign) was defined as classification variable. Initially, the ANN was optimized in terms of "Training Epochs" (TE), "Hidden Layers" (HL), "Learning Rate" (LR) and "Neurons" (N). Robustness of the ANN was addressed by repeated evaluation cycles (n: 9) with receiver operating characteristics (ROC) analysis of the results applying 4-fold Cross Validation. The best network architecture was identified comparing the corresponding Area under the ROC curve (AUC). RESULTS: Histopathology revealed 436 benign and 648 malignant lesions. Enhancing the level of complexity could not increase diagnostic accuracy of the network (P: n.s.). The optimized ANN architecture (TE: 20, HL: 1, N: 5, LR: 1.2) was accurate (mean-AUC 0.888; P: <0.001) and robust (CI: 0.885-0.892; range: 0.880-0.898). CONCLUSION: The optimized neural network showed robust performance and high diagnostic accuracy for prediction of malignancy on unknown data.

Fusion of dynamic contrast-enhanced magnetic resonance mammography at 3.0T with X-ray mammograms: pilot study evaluation using dedicated semi-automatic registration software.

RATIONALE AND OBJECTIVES: To evaluate the semi-automatic image registration accuracy of X-ray-mammography (XR-M) with high-resolution high-field (3.0T) MR-mammography (MR-M) in an initial pilot study. MATERIAL AND METHODS: MR-M was acquired on a high-field clinical scanner at 3.0T (T1-weighted 3D VIBE ± Gd). XR-M was obtained with state-of-the-art full-field digital systems. Seven patients with clearly delineable mass lesions >10mm both in XR-M and MR-M were enrolled (exclusion criteria: previous breast surgery; surgical intervention between XR-M and MR-M). XR-M and MR-M were matched using a dedicated image-registration algorithm allowing semi-automatic non-linear deformation of MR-M based on finite-element modeling. To identify registration errors (RE) a virtual craniocaudal 2D mammogram was calculated by the software from MR-M (with and w/o Gadodiamide/Gd) and matched with corresponding XR-M. To quantify REs the geometric center of the lesions in the virtual vs. conventional mammogram were subtracted. The robustness of registration was quantified by registration of X-MRs to both MR-Ms with and w/o Gadodiamide. RESULTS: Image registration was performed successfully for all patients. Overall RE was 8.2mm (1 min after Gd; confidence interval/CI: 2.0-14.4mm, standard deviation/SD: 6.7 mm) vs. 8.9 mm (no Gd; CI: 4.0-13.9 mm, SD: 5.4mm). The mean difference between pre- vs. post-contrast was 0.7 mm (SD: 1.9 mm). CONCLUSION: Image registration of high-field 3.0T MR-mammography with X-ray-mammography is feasible. For this study applying a high-resolution protocol at 3.0T, the registration was robust and the overall registration error was sufficient for clinical application.

Application of MR mammography beyond local staging: is there a potential to accurately assess axillary lymph nodes? evaluation of an extended protocol in an initial prospective study.

OBJECTIVE: The purpose of our study was to clinically test an extended MR mammography (MRM) protocol for combined local staging (T-staging) and locoregional staging (N-staging) of breast cancer within one single examination using a dedicated whole-body scanner. SUBJECTS AND METHODS: Fifty-six consecutive primary breast cancer patients without prior treatment underwent MRM and surgicopathological N-staging. The MRM protocol (10 minutes; axial T1-weighted gradient-recalled echo; dynamic contrast-enhanced; T2-weighted; turbo spin-echo) was extended to evaluate axillary lymph nodes (90 seconds; coronal T2-weighted HASTE; T1-weighted volumetric breath-hold examination; field of view, both axillae, supraclavicular nodes, and cervical nodes). A dedicated whole-body scanner was used. First, two experienced radiologists independently rated the presence of lymph node metastasis (present or absent, weighted kappa). Second, predefined descriptors were applied by both readers to differentiate lymph node status. These were statistically analyzed using univariate chi-square statistics, sensitivity and specificity, positive likelihood ratio, diagnostic odds ratio (OR), and multivariate statistics (binary logistic-regression, receiver operating characteristics, and chi-squared automatic interaction detection [CHAID] tree). RESULTS: Most significant predictors (p < 0.001) of present metastasis were "irregular margin" (diagnostic OR, 14.0), "inhomogeneous cortex" (diagnostic OR, 8.4), "perifocal edema" (positive likelihood ratio, 100) and "asymmetry" (diagnostic OR, 19.5). CHAID tree identified "asymmetry" and "irregular margin" as significant predictors (adjusted-p < 0.05) for present metastasis (PPV: 100%), whereas absence of "asymmetry" and "homogeneous internal structure" were highly predictive of absent metastasis (negative predictive value, 94.3%). Combination of significant descriptors using binary logistic regression revealed an area under the receiver operating characteristic curve of 0.93 (p < 0.001). Interrater agreement was "almost-perfect" (κ = 0.95). CONCLUSION: Combined T-staging and locoregional staging (N-staging) was possible within one imaging session using the proposed protocol. Despite a minimal increase in examination time, high diagnostic accuracy and excellent interrater reliability were achieved.

Kinetic analysis of lesions without mass effect on breast MRI using manual and computer-assisted methods.

OBJECTIVE: To analyse the kinetic characteristics of lesions without mass effect in dynamic breast MRI using manual and computer assisted methods. METHODS: The enhancement pattern of 82 histopathologically verified lesions without mass effect (36 malignant, 46 benign) was evaluated on breast MRI using manual placement of a region of interest. Commercially available computer analysis software automatically assessed volume enhancement characteristics of a lesion voxelwise. Kinetic features evaluated included classification of the signal-intensity time curve as washout, plateau or persistent enhancement. RESULTS: Unlike manual ROI placement, computer-aided analysis demonstrated a significant difference in enhancement pattern between benign (washout: 32.6%, plateau: 32.6%, persistent: 34.8%) and malignant lesions without mass effect (77.1%, 8.6%, 14.3% respectively, P < 0.01, two-sided Chi-squared test) following initial rapid signal increase. Mean percentage of washout voxel volumes within a lesion was significantly higher in malignant lesions than in benign lesions (11.9% +/-12.7 (SD) vs. 6.9% +/-11.3 (SD), P < 0.01, Mann-Whitney U Test). Conversely, the mean percentage of persistent voxel volumes was significantly lower in malignant lesions than in benign lesions (60.1% +/-21.1 (SD) vs. 79% +/-23 (SD), P < 0.01, Mann-Whitney U Test). CONCLUSION: Computer-assisted enhancement pattern analysis might have diagnostic benefit in the evaluation of lesions without mass effect.