Effective diagnosis of Alzheimer's disease by means of large margin-based methodology.

BACKGROUND Functional brain images such as Single-Photon Emission Computed Tomography (SPECT) and Positron Emission Tomography (PET) have been widely used to guide the clinicians in the Alzheimer's Disease (AD) diagnosis. However, the subjectivity involved in their evaluation has favoured the development of Computer Aided Diagnosis (CAD) Systems. METHODS It is proposed a novel combination of feature extraction techniques to improve the diagnosis of AD. Firstly, Regions of Interest (ROIs) are selected by means of a t-test carried out on 3D Normalised Mean Square Error (NMSE) features restricted to be located within a predefined brain activation mask. In order to address the small sample-size problem, the dimension of the feature space was further reduced by: Large Margin Nearest Neighbours using a rectangular matrix (LMNN-RECT), Principal Component Analysis (PCA) or Partial Least Squares (PLS) (the two latter also analysed with a LMNN transformation). Regarding the classifiers, kernel Support Vector Machines (SVMs) and LMNN using Euclidean, Mahalanobis and Energy-based metrics were compared. RESULTS Several experiments were conducted in order to evaluate the proposed LMNN-based feature extraction algorithms and its benefits as: i) linear transformation of the PLS or PCA reduced data, ii) feature reduction technique, and iii) classifier (with Euclidean, Mahalanobis or Energy-based methodology). The system was evaluated by means of k-fold cross-validation yielding accuracy, sensitivity and specificity values of 92.78%, 91.07% and 95.12% (for SPECT) and 90.67%, 88% and 93.33% (for PET), respectively, when a NMSE-PLS-LMNN feature extraction method was used in combination with a SVM classifier, thus outperforming recently reported baseline methods. CONCLUSIONS All the proposed methods turned out to be a valid solution for the presented problem. One of the advances is the robustness of the LMNN algorithm that not only provides higher separation rate between the classes but it also makes (in combination with NMSE and PLS) this rate variation more stable. In addition, their generalization ability is another advance since several experiments were performed on two image modalities (SPECT and PET).

Pattern of recurrence of early breast cancer is different according to intrinsic subtype and proliferation index.

INTRODUCTION Recurrence risk in breast cancer varies throughout the follow-up time. We examined if these changes are related to the level of expression of the proliferation pathway and intrinsic subtypes. METHODS Expression of estrogen and progesterone receptor, Ki-67, human epidermal growth factor receptor 2 (HER2), epidermal growth factor receptor (EGFR) and cytokeratin 5/6 (CK 5/6) was performed on tissue-microarrays constructed from a large and uniformly managed series of early breast cancer patients (N = 1,249). Subtype definitions by four biomarkers were as follows: luminal A (ER + and/or PR+, HER2-, Ki-67 <14), luminal B (ER + and/or PR+, HER2-, Ki-67 ≥14), HER2-enriched (any ER, any PR, HER2+, any Ki-67), triple-negative (ER-, PR-, HER2-, any Ki-67). Subtype definitions by six biomarkers were as follows: luminal A (ER + and/or PR+, HER2-, Ki-67 <14, any CK 5/6, any EGFR), luminal B (ER + and/or PR+, HER2-, Ki-67 ≥14, any CK 5/6, any EGFR), HER2-enriched (ER-, PR-, HER2+, any Ki-67, any CK 5/6, any EGFR), Luminal-HER2 (ER + and/or PR+, HER2+, any Ki-67, any CK 5/6, any EGFR), Basal-like (ER-, PR-, HER2-, any Ki-67, CK5/6+ and/or EGFR+), triple-negative nonbasal (ER-, PR-, HER2-, any Ki-67, CK 5/6-, EGFR-). Each four- or six-marker defined intrinsic subtype was divided in two groups, with Ki-67 <14% or with Ki-67 ≥14%. Recurrence hazard rate function was determined for each intrinsic subtype as a whole and according to Ki-67 value. RESULTS Luminal A displayed a slow risk increase, reaching its maximum after three years and then remained steady. Luminal B presented most of its relapses during the first five years. HER2-enriched tumors show a peak of recurrence nearly twenty months post-surgery, with a greater risk in Ki-67 ≥14%. However a second peak occurred at 72 months but the risk magnitude was greater in Ki-67 <14%. Triple negative tumors with low proliferation rate display a smooth risk curve, but with Ki-67 ≥14% show sharp peak at nearly 18 months. CONCLUSIONS Each intrinsic subtype has a particular pattern of relapses over time which change depending on the level of activation of the proliferation pathway assessed by Ki-67. These findings could have clinical implications both on adjuvant treatment trial design and on the recommendations concerning the surveillance of patients.