Valentina Giannini

Computer-aided diagnosis for dynamic contrast-enhanced breast MRI of mass-like lesions using a multiparametric model combining a selection of morphological, kinetic, and spatiotemporal features

PURPOSE Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) is a radiological tool for the detection and discrimination of breast lesions. The aim of this study is to evaluate a computer-aided diagnosis (CAD) system for discriminating malignant from benign breast lesions at DCE-MRI by the combined use of morphological, kinetic, and spatiotemporal lesion features. METHODS Fifty-four malignant and 19 benign breast lesions in 51 patients were retrospectively evaluated. Images were acquired at two centers at 1.5 T. Mass-like lesions were automatically segmented after image normalization and elastic coregistration of contrast-enhanced frames. For each lesion, a set of 28 3D features were extracted: ten morphological (related to shape, margins, and internal enhancement distribution); nine kinetic (computed from signal-to-time curves); and nine spatiotemporal (related to the variation of the signal between adjacent frames). A support vector machine (SVM) was trained with feature subsets selected by a genetic search. Best subsets were composed of the most frequent features selected by majority rule. The performance was measured by receiver operator characteristics analysis with a stratified tenfold cross-validation and bootstrap method for confidence intervals. RESULTS SVM training by the three separated classes of features resulted in an area under the curve (AUC) of 0.90 ± 0.04 (mean ± standard deviation), 0.87 ± 0.06, and 0.86 ± 0.06 for morphological, kinetic, and spatiotemporal feature, respectively. Combined training with all 28 features resulted in AUC of 0.96 ± 0.02 obtained with a selected feature subset composed by two morphological, one kinetic, and two spatiotemporal features. CONCLUSIONS Quantitative combination of morphological, kinetic, and spatiotemporal features is feasible and provides a higher discriminating power than using the three different classes of features separately.

Detection of prostate cancer index lesions with multiparametric magnetic resonance imaging (mp-MRI) using whole-mount histological sections as the reference standard

To evaluate the sensitivity of multiparametric magnetic resonance imaging (mp‐MRI) for detecting prostate cancer foci, including the largest (index) lesions.

Performance of a fully automatic lesion detection system for breast DCE-MRI.

To describe and test a new fully automatic lesion detection system for breast DCE‐MRI.

A fully automatic algorithm for segmentation of the breasts in DCE-MR images

Automatic segmentation of the breast and axillary region is an important preprocessing step for automatic lesion detection in breast MR and dynamic contrast-enhanced-MR studies. In this paper, we present a fully automatic procedure based on the detection of the upper border of the pectoral muscle. Compared with previous methods based on thresholding, this method is more robust to noise and field inhomogeneities. The method was quantitatively evaluated on 31 cases acquired from two centers by comparing the results with a manual segmentation. Results indicate good overall agreement within the reference segmentation (overlap=0.79±0.09, recall=0.95± 0.02, precision=0.82 ± 0.1).

A fully automatic computer aided diagnosis system for peripheral zone prostate cancer detection using multi-parametric magnetic resonance imaging

Multiparametric (mp)-Magnetic Resonance Imaging (MRI) is emerging as a powerful test to diagnose and stage prostate cancer (PCa). However, its interpretation is a time consuming and complex feat requiring dedicated radiologists. Computer-aided diagnosis (CAD) tools could allow better integration of data deriving from the different MRI sequences in order to obtain accurate, reproducible, non-operator dependent information useful to identify and stage PCa. In this paper, we present a fully automatic CAD system conceived as a 2-stage process. First, a malignancy probability map for all voxels within the prostate is created. Then, a candidate segmentation step is performed to highlight suspected areas, thus evaluating both the sensitivity and the number of false positive (FP) regions detected by the system. Training and testing of the CAD scheme is performed using whole-mount histological sections as the reference standard. On a cohort of 56 patients (i.e. 65 lesions) the area under the ROC curve obtained during the voxel-wise step was 0.91, while, in the second step, a per-patient sensitivity of 97% was reached, with a median number of FP equal to 3 in the whole prostate. The system here proposed could be potentially used as first or second reader to manage patients suspected to have PCa, thus reducing both the radiologists reporting time and the inter-reader variability. As an innovative setup, it could also be used to help the radiologist in setting the MRI-guided biopsy target.

A fully automatic multiscale 3-dimensional hessian-based algorithm for vessel detection in breast DCE-MRI

ObjectivesThe objectives of this study were to develop a fully automatic method for detecting blood vessels in dynamic contrast-enhanced magnetic resonance imaging of the breast on the basis of a multiscale 3-dimensional Hessian-based algorithm and to evaluate the improvement in reducing the number of vessel voxels incorrectly classified as parenchymal lesions by a computer-aided diagnosis (CAD) system. Materials and MethodsThe algorithm has been conceived to work on images obtained with different sequences, different acquisition parameters, such as the use of fat-saturation, and different contrast agents. The analysis was performed on 28 dynamic contrast-enhanced magnetic resonance imaging examinations, with 39 malignant (28 principal and 11 satellite) and 8 benign lesions, acquired at 2 centers using 2 different 1.5-T magnetic resonance scanners, radiofrequency coils, and contrast agents (14 studies from group A and 14 studies from group B). The method consists of 2 main steps: (a) the detection of linear structures on 3-dimensional images, with a multiscale analysis based on the second-order image derivatives and (b) the exclusion of non-vessel enhancements based on their morphological properties through the evaluation of the covariance matrix eigenvalues. To evaluate the algorithm performances, the identified vessels were converted into a 2-dimensional vasculature skeleton and then compared with manual tracking performed by an expert radiologist. When assessing the outcome of the algorithm performances in identifying vascular structures, the following terms must be considered: the correct-detection rate refers to pixels identified by both the algorithm and the radiologist, the missed-detection rate refers to pixels detected only by the radiologist, and the incorrect-detection rate refers to pixels detected only by the algorithm. The Wilcoxon rank sum test was used to assess differences between the performances of the 2 subgroups of images obtained from the different scanners. ResultsFor the testing set, which is composed of 28 patients from 2 different clinical centers, the median correct-detection rate was 89.1%, the median missed-detection rate was 10.9%, and the median incorrect-detection rate was 27.1%. The difference between group A and group B was not significant (P > 0.25). The exclusion of vascular voxels from the lesion detection map of a CAD system leads to a reduction of 68.4% (30.0%) (mean [SD]) of the total number of false-positives because of vessels, without a significant difference between the 2 subgroups (P = 0.50). ConclusionsThe system showed promising results in detecting most vessels identified by an expert radiologist on both fat-saturated and non–fat-saturated images obtained from different scanners with variable temporal and spatial resolutions and types of contrast agent. Moreover, the algorithm may reduce the labeling of vascular voxels as parenchymal lesions by a CAD system for breast magnetic resonance imaging, improving the CAD specificity and, consequently, further stimulating the use of CAD systems in clinical workflow.

MR-T2-weighted signal intensity: a new imaging biomarker of prostate cancer aggressiveness

Prostate cancer (PCa) is the most common solid neoplasm in males and a major cause of cancer-related death. Behaviour of PCa is dichotomous, as patients may either have an indolent clinical course or rapidly progress towards metastatic disease. Unfortunately, biopsy Gleason score (GS) may fail to predict cancer aggressiveness; tumour heterogeneity and inaccurate sampling during biopsy are major causes of underestimation. As a consequence, this frequently results in over-treatment, i.e. low-risk patients overcautiously undergo radical prostatectomy or radiotherapy, frequently with devastating side-effects. Some patients with PCa could be offered a more conservative approach if it were possible to predict patient risk confidently, especially in subjects lying in the grey zone of intermediate risk (i.e. GS = 7), which are in the majority. Recent studies have demonstrated that magnetic resonance (MR) imaging may help improve risk stratification in patients with PCa, providing imaging biomarkers of cancer aggressiveness. The aim of this study is to implement an automatic algorithm pipeline to discriminate different risks of progression from T2-weighted (T2-w) MR imaging. The obtained results confirm that T2-w signal intensity, together with other imaging biomarkers, may represent a new non-invasive approach to assess cancer aggressiveness, potentially helping to plan personalised treatments, and thus dramatically limiting over-diagnosis and over-treatment risks, and reducing the costs for the national healthcare system.

A new algorithm for automatic vascular mapping of DCE-MRI of the breast

BACKGROUND AND OBJECTIVE Vascularity evaluation on breast dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) has a potential diagnostic value, but it represents a time consuming procedure, affected by intra- and inter-observer variability. This study tests the application of a recently published method to reproducibly quantify breast vascularity, and evaluates if the vascular volume of cancer-bearing breast, calculated from automatic vascular maps (AVMs), may correlate with pathologic tumor response after neoadjuvant chemotherapy (NAC). METHODS Twenty-four patients with unilateral locally advanced breast cancer underwent DCE-MRI before and after NAC, 8 responders and 16 non-responders. A validated algorithm, based on multiscale 3D Hessian matrix analysis, provided AVMs and allowed the calculation of vessel volume before the initiation and after the last NAC cycle for each breast. For cancer bearing breast, the difference in vascular volume before and after NAC was compared in responders and non-responders using the Wilcoxon two-sample test. A radiologist evaluated the vascularity on the subtracted images (first enhanced minus unenhanced), before and after treatment, assigning a vascular score for each breast, according to the number of vessels with length ≥30mm and maximal transverse diameter ≥2mm. The same evaluation was repeated with the support of the simultaneous visualization of the AVMs. The two evaluations were compared in terms of mean number of vessels and mean vascular score per breast, in responders and non-responders, by use of Wilcoxon two sample test. For all the analysis, the statistical significance level was set at 0.05. RESULTS For breasts harboring the cancer, evidence of a difference in vascular volume before and after NAC for responders (median=1.71cc) and non-responders (median=0.41cc) was found (p=0.003). A significant difference was also found in the number of vessels (p=0.03) and vascular score (p=0.02) before or after NAC, according to the evaluation supported by the AVMs. CONCLUSIONS The encouraging, although preliminary, results of this study suggest the use of AVMs as new biomarker to evaluate the pathologic response after NAC, but also support their application in other breast DCE-MRI vessel analysis that are waiting for a reliable quantification method.

A fully automatic method to register the prostate gland on T2-weighted and EPI-DWI images

Prostate adenocarcinoma (PCa) is the most frequent noncutaneous cancer among men in developed countries. Magnetic Resonance (MR) has been used to detect PCa and several clinical trials report on the accuracy of the test. Multiparametric MR imaging (mpMRI) is defined as the integration of information from different morphological and functional datasets. mpMRI could be used to increase the performances of prostate MR, therefore allowing a more accurate assessment of the tumor gland extent, while reducing reporting time and interobserver variability. The first step to perform such a multiparametric analysis is to correct for voluntary and involuntary movements during the acquisitions, as well as for image distortion in the Diffusion Weighted (DWI) images. The aim of this work is to present a fully automatic registration algorithm between T2w and DWI images, able to realign the images and to correct the distortions in the DWI. Results showed a good overlap after registration and a strong decrease of mean surface distance in both the central gland and peripheral zone. These promising results suggest that the algorithm could be integrated in a CAD system which will combine the pharmacokinetic parameters derived from DCE-MRI, T2w MRI and DWI MR to generate one comprehensive value assessing the risk of malignancy. However to perform such a multiparametric analysis, it is necessary to correct for voluntary and involuntary (breathing, heart beating) movements during the DCE-MRI acquisition, and to realign also the DCE-MRI sequence to the T2w sequence.

A prostate CAD system based on multiparametric analysis of DCE T1-w, and DW automatically registered images

Prostate specific antigen (PSA)-based screening reduces the rate of death from prostate cancer (PCa) by 31%, but this benefit is associated with a high risk of overdiagnosis and overtreatment. As prostate transrectal ultrasound-guided biopsy, the standard procedure for prostate histological sampling, has a sensitivity of 77% with a considerable false-negative rate, more accurate methods need to be found to detect or rule out significant disease. Prostate magnetic resonance imaging has the potential to improve the specificity of PSA-based screening scenarios as a non-invasive detection tool, in particular exploiting the combination of anatomical and functional information in a multiparametric framework. The purpose of this study was to describe a computer aided diagnosis (CAD) method that automatically produces a malignancy likelihood map by combining information from dynamic contrast enhanced MR images and diffusion weighted images. The CAD system consists of multiple sequential stages, from a preliminary registration of images of different sequences, in order to correct for susceptibility deformation and/or movement artifacts, to a Bayesian classifier, which fused all the extracted features into a probability map. The promising results (AUROC=0.87) should be validated on a larger dataset, but they suggest that the discrimination on a voxel basis between benign and malignant tissues is feasible with good performances. This method can be of benefit to improve the diagnostic accuracy of the radiologist, reduce reader variability and speed up the reading time, automatically highlighting probably cancer suspicious regions.