Nicolas J. Merlet

Multiple-Instance Learning Improves CAD Detection of Masses in Digital Mammography

We propose a novel multiple-instance learning(MIL) algorithm for designing classifiers for use in computer aided detection(CAD). The proposed algorithm has 3 advantages over classical methods. First, unlike traditional learning algorithms that minimize the candidate level misclassification error, the proposed algorithm directly optimizes the patient-wise sensitivity. Second, this algorithm automatically selects a small subset of statistically useful features. Third, this algorithm is very fast, utilizes all of the available training data (without the need for cross-validation etc.), and requires no human hand tuning or intervention. Experimentally the algorithm is more accurate than state of the art support vector machine (SVM) classifier, and substantially reduces the number of features that have to be computed.

Feasibility study of breast tomosynthesis CAD system

The purpose of this study was to investigate feasibility of computer-aided detection of masses and calcification clusters in breast tomosynthesis images and obtain reliable estimates of sensitivity and false positive rate on an independent test set. Automatic mass and calcification detection algorithms developed for film and digital mammography images were applied without any adaptation or retraining to tomosynthesis projection images. Test set contained 36 patients including 16 patients with 20 known malignant lesions, 4 of which were missed by the radiologists in conventional mammography images and found only in retrospect in tomosynthesis. Median filter was applied to tomosynthesis projection images. Detection algorithm yielded 80% sensitivity and 5.3 false positives per breast for calcification and mass detection algorithms combined. Out of 4 masses missed by radiologists in conventional mammography images, 2 were found by the mass detection algorithm in tomosynthesis images.

Addressing image variability while learning classifiers for detecting clusters of micro-calcifications

Computer aided detection systems for mammography typically use standard classification algorithms from machine learning for detecting lesions. However, these general purpose learning algorithms make implicit assumptions that are commonly violated in CAD problems. We propose a new ensemble algorithm that explicitly accounts for the small fraction of outlier images which tend to produce a large number of false positives. A bootstrapping procedure is used to ensure that the candidates from these outlier images do not skew the statistical properties of the training samples. Experimental studies on the detection of clusters of micro-calcifications indicate that the proposed method significantly outperforms a state-of-the-art general purpose method for designing classifiers (SVM), in terms of FROC curves on a hold out test set.

Optimizing the CAD Process for Detecting Mammographic Lesions by a New Generation Algorithm Using Linear Classifiers and a Gradient Based Approach

This study evaluates the performance of a new generation algorithm designed to both increase detection sensitivity of cancers and to markedly reduce the false mark rate. In the advanced algorithm, several improvements were implemented. The algorithm for the initial detection of potential mass candidates was upgraded to ignore dense areas that do not represent masses. For the initial detection of potential clusters candidates, the advanced algorithm considers interdependence between various stages of the parametric clusterization process and implements automatic performance optimization. Moreover, the advanced algorithm includes a one-step global classification model, which assigns a score to each candidate lesion, instead of sequential multi-step filtration at various steps of the algorithm. Both the advanced and the previous algorithm were run on 83 malignant cases, with proven pathology, and on 523 normal screening cases that were consecutively culled from 4 clinical sites. The overall sensitivity of the advanced algorithm was 86%, compared to a sensitivity of 84% for the previous one. The false mark (FM) rate per case, decreased from 3.20 for the previous algorithm, to 1.39 for the advanced one. The advanced algorithm reduced both mass FMs and cluster FMs. In conclusion, the new algorithm outperforms the old one with a slight increase in sensitivity and with a substantial reduction in false mark rate for both masses and clusters.