Computer-aided detection of lung nodules via 3D fast radial transform, scale space representation, and Zernike MIP classification.

PURPOSE: The authors presented a novel system for automated nodule detection in lung CT exams. METHODS: The approach is based on (1) a lung tissue segmentation preprocessing step, composed of histogram thresholding, seeded region growing, and mathematical morphology; (2) a filtering step, whose aim is the preliminary detection of candidate nodules (via 3D fast radial filtering) and estimation of their geometrical features (via scale space analysis); and (3) a false positive reduction (FPR) step, comprising a heuristic FPR, which applies thresholds based on geometrical features, and a supervised FPR, which is based on support vector machines classification, which in turn, is enhanced by a feature extraction algorithm based on maximum intensity projection processing and Zernike moments. RESULTS: The system was validated on 154 chest axial CT exams provided by the lung image database consortium public database. The authors obtained correct detection of 71% of nodules marked by all radiologists, with a false positive rate of 6.5 false positives per patient (FP/patient). A higher specificity of 2.5 FP/patient was reached with a sensitivity of 60%. An independent test on the ANODE09 competition database obtained an overall score of 0.310. CONCLUSIONS: The system shows a novel approach to the problem of lung nodule detection in CT scans: It relies on filtering techniques, image transforms, and descriptors rather than region growing and nodule segmentation, and the results are comparable to those of other recent systems in literature and show little dependency on the different types of nodules, which is a good sign of robustness.

Computer-aided mass detection in mammography: false positive reduction via gray-scale invariant ranklet texture features.

In this work, gray-scale invariant ranklet texture features are proposed for false positive reduction (FPR) in computer-aided detection (CAD) of breast masses. Two main considerations are at the basis of this proposal. First, false positive (FP) marks surviving our previous CAD system seem to be characterized by specific texture properties that can be used to discriminate them from masses. Second, our previous CAD system achieves invariance to linear/nonlinear monotonic gray-scale transformations by encoding regions of interest into ranklet images through the ranklet transform, an image transformation similar to the wavelet transform, yet dealing with pixels' ranks rather than with their gray-scale values. Therefore, the new FPR approach proposed herein defines a set of texture features which are calculated directly from the ranklet images corresponding to the regions of interest surviving our previous CAD system, hence, ranklet texture features; then, a support vector machine (SVM) classifier is used for discrimination. As a result of this approach, texture-based information is used to discriminate FP marks surviving our previous CAD system; at the same time, invariance to linear/nonlinear monotonic gray-scale transformations of the new CAD system is guaranteed, as ranklet texture features are calculated from ranklet images that have this property themselves by construction. To emphasize the gray-scale invariance of both the previous and new CAD systems, training and testing are carried out without any in-between parameters' adjustment on mammograms having different gray-scale dynamics; in particular, training is carried out on analog digitized mammograms taken from a publicly available digital database, whereas testing is performed on full-field digital mammograms taken from an in-house database. Free-response receiver operating characteristic (FROC) curve analysis of the two CAD systems demonstrates that the new approach achieves a higher reduction of FP marks when compared to the previous one. Specifically, at 60%, 65%, and 70% per-mammogram sensitivity, the new CAD system achieves 0.50, 0.68, and 0.92 FP marks per mammogram, whereas at 70%, 75%, and 80% per-case sensitivity it achieves 0.37, 0.48, and 0.71 FP marks per mammogram, respectively. Conversely, at the same sensitivities, the previous CAD system reached 0.71, 0.87, and 1.15 FP marks per mammogram, and 0.57, 0.73, and 0.92 FPs per mammogram. Also, statistical significance of the difference between the two per-mammogram and per-case FROC curves is demonstrated by the p-value < 0.001 returned by jackknife FROC analysis performed on the two CAD systems.

Comprehensive characterization of IGHV3-21-expressing B-cell chronic lymphocytic leukemia: an Italian multicenter study.

IGHV3-21-using chronic lymphocytic leukemia (CLL) is a distinct entity with restricted immunoglobulin gene features and poor prognosis and is more frequently encountered in Northern than Southern Europe. To further investigate this subset and its geographic distribution in the context of a country (Italy) with both continental and Mediterranean areas, 37 IGHV3-21 CLLs were collected out of 1076 cases enrolled by different institutions from Northern or Central Southern Italy. Of the 37 cases, 18 were identified as homologous (hom)HCDR3-IGHV3-21 CLLs and were found almost exclusively (16 of 18) in Northern Italy; in contrast, 19 nonhomHCDR3-IGHV3-21 cases were evenly distributed throughout Italy. Clinically, poor survivals were documented for IGHV3-21 CLLs as well as for subgroups of mutated and homHCDR3-IGHV3-21 CLLs. Negative prognosticators CD38, ZAP-70, CD49d, and CD79b were expressed at higher levels in homHCDR3 than nonhomHCDR3-IGHV3-21 cases. Differential gene expression profiling (GEP) of 13 IGHV3-21 versus 52 non-IGHV3-21 CLLs identified, among 122 best-correlated genes, TGFB2 and VIPR1 as down- and up-regulated in IGHV3-21 CLL cases, respectively. Moreover, GEP of 7 homHCDR3 versus 6 nonhomHCDR3-IGHV3-21 CLLs yielded 20 differentially expressed genes, with WNT-16 being that expressed at the highest levels in homHCDR3-IGHV3-21 CLLs. Altogether, IGHV3-21 CLLs, including those with homHCDR3, had a peculiar global phenotype in part explaining their worse clinical outcome.

Comparison of different commercial FFDM units by means of physical characterization and contrast-detail analysis.

The purpose of this study was to perform a complete evaluation of three pieces of clinical digital mammography equipment. Image quality was assessed by performing physical characterization and contrast-detail (CD) analysis. We considered three different FFDM systems: a computed radiography unit (Fuji "FCR 5000 MA") and two flat-panel units, the indirect conversion a-Si based GE "Senographe 2000D" and the direct conversion a-Si based IMS "Giotto Image MD." The physical characterization was estimated by measuring the MTF, NNPS, and DQE of the detectors with no antiscatter grid and over the clinical range of exposures. The CD analysis was performed using a CDMAM 3.4 phantom and custom software designed for automatic computation of the contrast-detail curves. The physical characterization of the three digital systems confirms the excellent MTF properties of the direct conversion flat-panel detector (FPD). We performed a relative standard deviation (RSD) analysis, for investigating the different components of the noise presented by the three systems. It turned out that the two FPDs show a significant additive component, whereas for the CR system the statistical noise is dominant. The multiplicative factor is a minor constituent for all the systems. The two FPDs demonstrate better DQE, with respect to the CR system, for exposures higher than 70 microGy. The CD analysis indicated that the three systems are not statistically different for detail objects with a diameter greater than 0.3 mm. However, the IMS system showed a statistically significant different response for details smaller than 0.3 mm. In this case, the poor response of the a-Se detector could be attributed to its high-frequency noise characteristics, since its MTF, NEQ, and DQE are not inferior to those of the other systems. The CD results were independent of exposure level, within the investigated clinical range. We observed slight variations in the CD results, due to the changes in the visualization parameters (window/level and magnification factor). This suggests that radiologists would benefit from viewing images using varied window/level and magnification.

Surface-antigen expression profiling of B cell chronic lymphocytic leukemia: from the signature of specific disease subsets to the identification of markers with prognostic relevance.

Studies of gene expression profiling have been successfully used for the identification of molecules to be employed as potential prognosticators. In analogy with gene expression profiling, we have recently proposed a novel method to identify the immunophenotypic signature of B-cell chronic lymphocytic leukemia subsets with different prognosis, named surface-antigen expression profiling. According to this approach, surface marker expression data can be analysed by data mining tools identical to those employed in gene expression profiling studies, including unsupervised and supervised algorithms, with the aim of identifying the immunophenotypic signature of B-cell chronic lymphocytic leukemia subsets with different prognosis. Here we provide an overview of the overall strategy employed for the development of such an "outcome class-predictor" based on surface-antigen expression signatures. In addition, we will also discuss how to transfer the obtained information into the routine clinical practice by providing a flow-chart indicating how to select the most relevant antigens and build-up a prognostic scoring system by weighing each antigen according to its predictive power. Although referred to B-cell chronic lymphocytic leukemia, the methodology discussed here can be also useful in the study of diseases other than B-cell chronic lymphocytic leukemia, when the purpose is to identify novel prognostic determinants.

Surface-antigen expression profiling (SEP) in B-cell chronic lymphocytic leukemia (B-CLL): Identification of markers with prognostic relevance.

Studies of gene expression profiling (GEP) have been successfully used for the identification of molecules to be employed as potential prognosticators. With the aim of identifying the immunophenotypic profile of B-CLL subsets with different prognoses, we investigated by flow cytometry the expression of 36 surface antigens in 117 cases, 113 with survival data. In analogy with GEP, results were analyzed by applying unsupervised hierarchical algorithms (surface-antigen expression profiling, SEP). Distinct immunophenotypic groups (A, B1, B2 and C) were identified, group C (57/117) with longer survivals, as compared to groups A (23/117), B1 (16/117) and B2 (21/117). The immunophenotypic signatures of these groups were characterized by the coordinated and differential over-expression of: i) CD62L, CD54 and CD49c (group C); ii) CD38 and CD49d (group A); iii) none of the above markers (group B1 and B2). Other molecules were either not expressed, widely expressed by all samples, or were variably expressed within the observed B-CLL subgroups, although without a clearly distinguishable pattern. By employing an identical approach for investigating the reactivity of B-cell panel monoclonal antibodies (B-mAbs) in B-CLLs (29 cases) and in 19 B and non-B leukemia/lymphoma cell lines, we found mAbs (B012, B001, B006, B018, B019, B020, B017) mainly unreactive in all the samples, mAbs (B002, B010, B013, B014, B015) strongly reactive in B-CLLs and B-cell lines but not in non-B-cell lines, and mAbs recognizing antigens variably expressed in cell lines and B-CLLs. A hierarchical clustering focused on B-CLLs alone, combining reactivity values for B-mAbs with the expression of CD62L and CD38, these latter antigens identified as leader markers of B-CLL subsets with different prognosis, demonstrated a correlation between CD62L expression and the reactivity of B007, B003, B011 and B005 mAbs. These mAbs may represent potentially novel markers with prognostic relevance in B-CLLs.

Signature of B-CLL with different prognosis by Shrunken centroids of surface antigen expression profiling.

With the aim of identifying the immunophenotypic profile of B-cell chronic lymphocytic leukemia (B-CLL) subsets with different prognosis, we investigated by flow cytometry the expression of 36 surface antigens in 123 cases, all with survivals. By analyzing results with unsupervised (hierarchical and K-means clustering) algorithms, three distinct immunophenotypic groups (I, II, and III) were identified, group I (51/123) with longer survivals, as compared to the group II (36/123) and III (36/123). The immunophenotypic signatures of these groups, as determined by applying the nearest Shrunken centroids method as class predictor, were characterized by the coordinated and differential expression of 12 surface markers, that is, group I: above-average expression of CD62L, CD54, CD49c, and CD25, below-average expression of CD38; group II: above-average expression of CD38, CD49d, CD29, and CD49e; and group III: below-average expression of the above markers, overexpression of CD23, CD20, SmIg, and CD79b. As opposed to groups II-III, group I B-CLLs lacked expression of ZAP-70 and activation-induced cytidine deaminase in the majority of cases, while more frequently had mutated IgV(H) genes and IgV(H) mutations consistent with antigen-driven selection. Our findings contribute to improve the immunophenotypical identification of disease subsets with different prognosis and suggest a set of surface antigens to be employed as prognosticators in routine diagnostic/prognostic procedures.

A novel featureless approach to mass detection in digital mammograms based on support vector machines.

In this work, we present a novel approach to mass detection in digital mammograms. The great variability of the appearance of masses is the main obstacle to building a mass detection method. It is indeed demanding to characterize all the varieties of masses with a reduced set of features. Hence, in our approach we have chosen not to extract any feature, for the detection of the region of interest; in contrast, we exploit all the information available on the image. A multiresolution overcomplete wavelet representation is performed, in order to codify the image with redundancy of information. The vectors of the very-large space obtained are then provided to a first support vector machine (SVM) classifier. The detection task is considered here as a two-class pattern recognition problem: crops are classified as suspect or not, by using this SVM classifier. False candidates are eliminated with a second cascaded SVM. To further reduce the number of false positives, an ensemble of experts is applied: the final suspect regions are achieved by using a voting strategy. The sensitivity of the presented system is nearly 80% with a false-positive rate of 1.1 marks per image, estimated on images coming from the USF DDSM database.