Scientific Reports · 2017 · 63 citations · 61 references
MedicineOncologyPathologyDiagnostic ImagingUltrasonographyUltrasoundBreast Cancer ImagingBreast UltrasonographyImaging BiomarkersTherapy Response AssessmentTumor CharacterizationUltrasound ImagingBiomarkersBreast Cancer Response AssessmentCell Death DetectionCell MorphologyChemotherapy Response AssessmentIntra-tumor HeterogeneityQuantitative UltrasoundUltrasound Imaging Biomarkers
Anti-cancer therapies including chemotherapy aim to induce tumour cell death. Cell death introduces alterations in cell morphology and tissue micro-structures that cause measurable changes in tissue echogenicity. This study investigated the effectiveness of quantitative ultrasound (QUS) parametric imaging to characterize intra-tumour heterogeneity and monitor the pathological response of breast cancer to chemotherapy in a large cohort of patients (n = 100). Results demonstrated that QUS imaging can non-invasively monitor pathological response and outcome of breast cancer patients to chemotherapy early following treatment initiation. Specifically, QUS biomarkers quantifying spatial heterogeneities in size, concentration and spacing of acoustic scatterers could predict treatment responses of patients with cross-validated accuracies of 82 ± 0.7%, 86 ± 0.7% and 85 ± 0.9% and areas under the receiver operating characteristic (ROC) curve of 0.75 ± 0.1, 0.80 ± 0.1 and 0.89 ± 0.1 at 1, 4 and 8 weeks after the start of treatment, respectively. The patients classified as responders and non-responders using QUS biomarkers demonstrated significantly different survivals, in good agreement with clinical and pathological endpoints. The results form a basis for using early predictive information on survival-linked patient response to facilitate adapting standard anti-cancer treatments on an individual patient basis.
61
Nearest neighbor pattern classification
Thomas M. Cover, Peter E. Hart · IEEE Transactions on Information Theory · 1967 · 16.1K citations
Haibo He, Edwardo A. Garcia · IEEE Transactions on Knowledge and Data Engineering · 2009 · 9.7K citations