Aims and objectives
Computer based CAD-systems are well established in the analysis of contrast uptake of breast lesions in MRI.
peak uptake,
time to peak and distribution of vascularisation pattern of all voxel of the entire tumor volume are obtainable but not well analyzed regarding their diagnostic potential so far.
Methods and materials
100 patients (34malignant,
66 benign) with a suspicious breast lesion only visible in breast MRI (3T Philips Ingenia,
CAD-analysis by Aegis Breast 3.2.0) and histologically verified by MR-guided biopsy (9G vacuum assisted biopsy) were analyzed regarding peak uptake,
time to peak,
uptake within the first minute,
distribution of vascularisation pattern wash out related to entire segmented tumor volume (composition analysis) using several statistical tests (logistic regression,
ROC,
Chi-Square).
Cut off values were calculated.
Results
AUC of 1min uptake: 0.729,p<0.000; AUC of composition analysis 0.800,
p<0.001,
related cut off: 145% (Sens:76.5%; Spec: 65.2%).
Peak uptake is significantly higher at malignancies (asympt.
signif .000,
AUC 0.716).
Resulting Cut off values: 181% Uptake (Sen:0.71;Spec:0.67).
Time to peak was without statistical discriminatory power (p=0.066).
Most valuable parameter was composition analysis (logistic regression B=0.032;standard error 0.008;Wald14,69,p<0.001,
resulting cut off: 23.5% fo the entire tumor volume with wash out results in 73.5%sens and 71.2% spec.
1min uptake: B=0.005,standard error0.003;Wald3.138;p=0.076; Peak uptake: B=0.004; standard error:0.003; Wald:...
Conclusion
CAD-based analysis especially of the degree of wash out among the entire tumor voxels but also time to peak and peak uptake are useful and reliable parameters for increased differentiation of only MR-visible subtle breast lesion and might reduce number of unnecessary biopsies especially when using the recommended cut off values.
References
just a very short list,
if you are interested in further data and literature,
feel free to get in contact with me:
[email protected]
Malich A et al: Potential MRI Interpretation Model:Differentiation of Benign from Malignant Breast Masses.AJR 2005; 185:964–970 DOI: 0361–803X/05/1854–964
Han BK,
et al: Outcome of MRI-guided breast biopsy.
AJR Am J Roentgenol. 2008 Dec;191(6):1798-804.
doi: 10.2214/AJR.07.2827.
Kim SH,
et al: Dynamic contrast-enhanced MRI prefusion parameters as imaging biomarkers of angiogenesis.
PLoS One. 2016 Dec 30;11(12):e0168632.
doi: 10.1371/journal.pone.0168632.
eCollection 2016.
Strigel RM,
et al.:...