A nonrigid registration of MR breast images using complex-valued wavelet transform.

In this paper, a fast, slice-by-slice, nonrigid registration algorithm of dynamic magnetic resonance breast images is presented. The method is based on a multiresolution motion estimation of the breast using complex discrete wavelet transform (CDWT): the pyramid of oriented complex subimages is used to implement a hierarchical phase-matching-based motion estimation algorithm. The resulting motion estimate is nonrigid and pixel-independent. To assess the method performance, we computed the correlation coefficient and the normalized mutual information between pre- and postcontrast images with and without realignment. The indices increased after using our approach and the improvement was superior to rigid or affine registration. A set of clinical scores was also evaluated. The clinical validation demonstrated an increased readability in the subtraction images. In particular, CDWT registration allowed a best definition of breast and lesion borders and greater detail detectability.

BMRI in early detection of breast cancer in patients with increased genetic risk: our preliminary results.

Because of the higher risk of developing breast cancer and the early onset of the disease in women proved or suspected to be carriers of a breast cancer susceptibility gene, a dedicated screening should be offered as a less invasive approach with respect to the otherwise suggested prophylactic mastectomy. This should be optimized in order to overcome the limitations of conventional breast imaging with the application of new technologies such as Breast Magnetic Resonance Imaging (BMRI). A diagnostic protocol for routine control in patients with high risk for developing breast cancer has been prepared. Within a 7 months period, 23 patients suspected or proved to carry a breast cancer susceptibility gene underwent BMRI. Four breast cancers were identified with BMRI. In these cases mammography was negative because of the density of the parenchyma or for its fibroglandular pattern. US was negative in two cases, not specific for malignancy in one case and considered as only possibly malignant but with biopsy recommendation on the basis of MR findings in the last one. Clinic analysis was positive for mass in two cases. The accuracy of BMRI is known to be higher than that of conventional imaging in the study of breast parenchyma. High spatial resolution and no breast density influence can give more detailed information about smaller lesions and the right extent of the disease.

The Italian multi-centre project on evaluation of MRI and other imaging modalities in early detection of breast cancer in subjects at high genetic risk.

This report presents the preliminary results of the first phase (21 months) of a multi-centre, non-randomised, prospective study, aimed at evaluating the effectiveness of contrast-enhanced magnetic resonance imaging (MRI), X-ray mammography (XM) and ultrasound (US) in early diagnosis of breast cancer (BC) in subjects at high genetic risk. This Italian national trial (coordinated by the Istituto Superiore di Sanità, Rome) so far recruited 105 women (mean age 46.0 years; median age 51.0; age range 25-77 years), who were either proven BRCA1 or BRCA2 mutation carriers or had a 1 in 2 probability of being carriers (40/105 with a previous personal history of BC). Eight cases of breast carcinomas were detected in the trial (mean age 55.3 years, median age 52.5; age range 35-70 years; five with previous personal history of BC). All trial-detected BC cases (8/8) were identified by MRI, while XM and US correctly classified only one. MRI had one false positive case, XM and US none. Seven "MRI-only" detected cancers (4 invasive, 3 in situ) occurred in both pre- (n = 2) and post-menopausal (n = 5) women. With respect to the current XM screening programmes addressed to women in the age range 50-69 years, the global incidence of BC in the trial (7.6%) was over ten-fold higher. The cost per "MRI-only" detected cancer in this particular category of subjects at high genetic risk was substantially lower than that of an XM-detected cancer in the general women population. These preliminary results confirmed that MRI is a very useful tool to screen subjects at high genetic risk for breast carcinoma, not only in pre-, but also in post-menopausal age, with a low probability of false positive cases.

Regularization networks: fast weight calculation via Kalman filtering.

Regularization networks are nonparametric estimators obtained from the application of Tychonov regularization or Bayes estimation to the hypersurface reconstruction problem. Their main drawback back is that the computation of the weights scales as O(n(3)) where n is the number of data. In this paper, we show that for a class of monodimensional problems, the complexity can be reduced to O(n) by a suitable algorithm based on spectral factorization and Kalman filtering. Moreover, the procedure applies also to smoothing splines.

Locally advanced breast cancer treated with primary chemotherapy: comparison between magnetic resonance imaging and pathologic evaluation of residual disease.

AIMS AND BACKGROUND: We evaluated the response of locally advanced breast cancer to induction chemotherapy using MRI techniques. The size and vitality of any residual pathologic tissue was quantified by means of morphologic and dynamic analysis. A curve derived from the dynamic parameters shows the uptake intensity with respect to the time elapsed since administration, which is related to vascularization and therefore indirectly reflects the angiogenesis of malignant tissue. METHODS AND STUDY DESIGN: A group of 30 patients were examined with MRI for staging purposes before undergoing treatment and subsequently to assess the response to treatment. Alterations in size and dynamic parameters were closely monitored. RESULTS: The overall accuracy was 90%, the sensitivity 96%, the specificity 75%, the positive predictive value 92.5% and the negative predictive value 66%. Interestingly, analysis of the dynamic curves made it possible to obtain additional information regarding the angiogenetic activity of the residual tumor. CONCLUSIONS: Evaluation of the response to treatment by means of conventional imaging and clinical examination can be particularly difficult because of the fibrosis induced by cytotoxic drugs or the small volume of residual disease. The additional information supplied by MRI could therefore allow a more conservative surgical approach in selected cases of optimal response to treatment, as well as a much more accurate follow-up. Furthermore, the variation in dynamic parameters according to the vitality of residual disease could in the future become a useful tool for monitoring the effectiveness of anti-angiogenetic drugs.

Lobular breast cancer: how useful is breast magnetic resonance imaging?

RATIONALE AND OBJECTIVES: To review magnetic resonance imaging (MRI) findings in lobular breast carcinoma, the in situ or infiltrating subtype, with special attention to the dynamic curves with the aim to evaluate possible differences with ductal carcinoma. METHODS: In 2 years, 27 patients with lobular and one with tubular carcinoma underwent MRI at the Istituto Nazionale Tumori of Milan. RESULTS: All lobular carcinomas demonstrated early or late enhancement (100% sensitivity), without significant differences in morphology compared with ductal carcinoma, but frequently with a different shape of the dynamic curves. CONCLUSIONS: Due to its infiltrative growth associated to only limited connective tissue reaction, lobular carcinoma often encounters difficulties in mammographic diagnosis. In contrast, MRI can be very helpful in evaluating the true extension of the disease, especially when breast conservation is considered. Due to a more consistent fibrotic stroma, these lesions sometimes show a delayed enhancement, which suggests that more than one set of subtracted images should be evaluated during MRI analysis.

[Magnetic resonance in primary bone tumors: a review of 10 years of activities]. / La Risonanza Magnetica nei tumori primitivi dell'osso: revisione di dieci anni di attività.

All the MR exams of primary bone tumors performed during ten years were reviewed by three different radiologists. In all, 484 exams in 220 patients were considered--namely, 160 exams (33.1%) for staging purposes, 219 (45.2%) during therapy and 105 (21.7%) performed more than 8 months after the last treatment. Its well-known accuracy in the assessment of intra/extraosseous spread confirms the major role of MRI in the staging of primary bone tumors. During treatment, the overall accuracy of this method decreased to 88.8% because of the presence of therapy-induced tissue changes. MRI was 95.2% reliable in the detection of persistent disease or relapse in the exams performed long after therapy. Conventional radiology is still the method of choice in the study of primary bone tumors at presentation as it detects the lesion, differentiates malignant tumors and usually suggests the possible histotype. Nevertheless, MRI seems to be needed to depict actual tumor extent and to find the correct therapeutic approach. In the follow-up, MRI is the best single method to assess the response to therapy and to detect tumor persistence.

Breast magnetic resonance imaging: a computer-based analysis of enhancement curves.

The aim of this work is to report our experience in contrast-enhanced magnetic resonance (MR) automatic analysis. This technique is usually used in our practice for the diagnosis of breast cancer. Study execution and semiautomatic data analysis require at least 1 hour of the radiologist. So, we decided to develop a tool that aids the radiologist in locating the regions of neoplastic angiogenesis and quickly plotting the enhancement curves. This is not a computer-aided diagnostic tool; of course, the curves are interpreting directly by the radiologist. We used a programmable graphical and calculus environment (Matlab, The Mathworks, Natick, MA) to develop a tool that interfaces directly with our picture archiving and communication system (PACS) via Digital Imaging and Communications in Medicine (DICOM), receives images, performs images subtraction, and shows some possible regions of interest (ROIs), each with an enhancement curve. The radiologist can also select other ROIs. The software shows a good performance (sensitivity, 100%). A low specificity (25%) is acceptable because the software allows the radiologist to save some time during each study.

[Locally advanced breast tumors. Role of magnetic resonance in the assessment of response to preoperative therapy and of neoplastic residue before the operation]. / Tumori della mammella localmente avanzati. Ruolo della Risonanza Magnetica nella valutazione della risposta alla terapia prechirurgica e del residuo neoplastico prima dell'intervento.

INTRODUCTION AND PURPOSE: Induction chemotherapy is the preoperative treatment for locally advanced breast carcinoma. The patients affected with this kind of tumor were previously considered inoperable. The sequential use of different cytotoxic drugs reduces the tumor mass effectively, thus allowing resection and improving patients prognosis. Tumor debulking is at times so significant that conservative treatment can even be considered. A reliable assessment of the response to drug therapy by conventional diagnostic procedures is usually hindered by chemotherapy-induced fibrosis. Magnetic resonance imaging (MRI) is a better tool for distinguishing fibrosis from still vascularized pathologic tissue and thus permits more accurate evaluation of tumor response to chemotherapy, namely tumor debulking and residual viability. MATERIAL AND METHODS: We selected 27 patients with breast cancer and submitted them to MRI both before and after chemotherapy. All examinations were performed with a high field system using 3D Flash sequences with optimized spatial and temporal resolution. RESULTS AND DISCUSSION: The morphologic and dynamic parameters of MRI were in agreement with pathologic findings. In case of persistent disease after chemotherapy, MRI demonstrated increased contrast agent uptake at restaging, with dynamic curves indicating early and intense uptake. In case of marked post-chemotherapy changes, the dynamic curves had a shorter and less steep trend. Finally, when no or very little (few microns) tumor tissue was left, MRI showed no uptake. CONCLUSIONS: Our initial experience indicates MRI as a valid too for monitoring chemotherapy response in breast cancer patients.