Search

Background Pectoral muscle removal is a fundamental preliminary step in computer-aided diagnosis systems for full-field digital mammography (FFDM). Currently, two open-source publicly available packages (LIBRA and OpenBreast) provide algorithms for pectoral muscle removal within Matlab environment. Purpose To compare performance of the two packages on a single database of FFDM images. Methods Only mediolateral oblique (MLO) FFDM was considered because of large presence of pectoral muscle on this type of projection. For obtaining ground truth, pectoral muscle has been manually segmented by two radiologists in consensus. Both LIBRA’s and OpenBreast’s removal performance with respect to ground truth were compared using Dice similarity coefficient and Cohen-kappa reliability coefficient; Wilcoxon signed-rank test has been used for assessing differences in performances; Kruskal–Wallis test has been used to verify possible dependence of the performance from the breast density or image laterality. Results FFDMs from 168 consecutive women at our institution have been included in the study. Both LIBRA’s Dice-index and Cohen-kappa were significantly higher than OpenBreast (Wilcoxon signed-rank test P P > 0.05). Conclusion: Libra has a better performance than OpenBreast in pectoral muscle delineation so that, although our study has not a direct clinical application, these results are useful in the choice of packages for the development of complex systems for computer-aided breast evaluation.

Background Radiology is an essential tool in the management of a patient. The aim of this manuscript was to build structured report (SR) Mammography based in Breast Cancer. Methods A working team of 16 experts (group A) was composed to create a SR for Mammography Breast Cancer. A further working group of 4 experts (group B), blinded to the activities of the group A, was composed to assess the quality and clinical usefulness of the SR final draft. Modified Delphi process was used to assess level of agreement for all report sections. Cronbach’s alpha (Cα) correlation coefficient was used to assess internal consistency and to measure quality analysis according to the average inter-item correlation. Results The final SR version was built by including n = 2 items in Personal Data, n = 4 items in Setting, n = 2 items in Comparison with previous breast examination, n = 19 items in Anamnesis and clinical context; n = 10 items in Technique; n = 1 item in Radiation dose; n = 5 items Parenchymal pattern; n = 28 items in Description of the finding; n = 12 items in Diagnostic categories and Report and n = 1 item in Conclusions. The overall mean score of the experts and the sum of score for structured report were 4.9 and 807 in the second round. The Cronbach’s alpha (Cα) correlation coefficient was 0.82 in the second round. About the quality evaluation, the overall mean score of the experts was 3.3. The Cronbach’s alpha (Cα) correlation coefficient was 0.90. Conclusions Structured reporting improves the quality, clarity and reproducibility of reports across departments, cities, countries and internationally and will assist patient management and improve breast health care and facilitate research.

To compare preoperative contrast-enhanced spectral mammography (CEM) versus digital mammography plus digital breast tomosynthesis (DM + DBT) in detecting breast cancer (BC) and assessing its size. We retrospectively included 78 patients with histological diagnosis of BC who underwent preoperative DM, DBT, and CEM over one year. Four readers, blinded to pathology and clinical information, independently evaluated DM + DBT versus CEM to detect BC and measure its size. Readers' experience ranged 3–10 years. We calculated the per-lesion cancer detection rate (CDR) and the complement of positive predictive value (1-PPV) of both methods, stratifying analysis on the total of lesions, index lesions, and additional lesions. The agreement in assessing cancer size versus pathology was assessed with Bland–Altman analysis. 100 invasive BCs (78 index lesions and 22 additional lesions) were analyzed. Compared to DM + DBT, CEM showed higher overall CDR in less experienced readers (range 0.85–0.90 vs. 0.95–0.96), and higher CDR for additional lesions, regardless of the reader (range 0.54–0.68 vs. 0.77–0.86). CEM increased the detection of additional disease in dense breasts in all readers and non-dense breasts in less experienced readers only. The 1-PPV of CEM (range 0.10–0.18) was comparable to that of DM + DBT (range 0.09–0.19). At Bland–Altman analysis, DM + DBT and CEM showed comparable mean differences and limits of agreement in respect of pathologic cancer size. Preoperative CEM improved the detection of additional cancer lesions compared to DM + DBT, particularly in dense breasts. CEM and DM + DBT achieved comparable performance in cancer size assessment.

Automated breast ultrasound (ABUS) is a recently introduced ultrasonography technique, developed with the purpose to standardize breast ultrasonography and overcome some limitations of handheld ultrasound (HHUS), such as operator dependence and the considerable amount of medical time necessary to perform and interpret HHUS. This new ultrasonography technique separates the moment of image acquisition (that may be performed also by a technician) from that of its interpretation, increasing reproducibility, reducing operator-dependence and physician time. Moreover, multiplanar reconstructions, especially the coronal view, introduce new diagnostic information. ABUS, with those advantages, has the potential to be used as an adjunctive tool to screening mammography, especially in the dense breast, where mammography has a relatively low sensitivity. Women's awareness of risks related to breast density is a hot topic, especially in the USA where legislative breast density notification laws increase the demand for supplemental ultrasound screening. Therefore, ABUS might have the potential to respond to this need. The purpose of this article is to present a summary of current state-of-the-art of ABUS technology and applications, with an emphasis on breast cancer screening. This article discusses also how to overcome some ABUS limitations, in order to be familiar with the new technique.

Women who were previously treated for breast cancer (BC) are an important particular subgroup of women at intermediate BC risk. Their breast follow-up should be planned taking in consideration a 1.0-1.5 % annual rate of loco-regional recurrences and new ipsilateral or contralateral BCs during 15-20 years, and be based on a regional/district invitation system. This activity should be carried out by a Department of Radiology integrating screening and diagnostics in the context of a Breast Unit. We recommend the adoption of protocols dedicated to women previously treated for BC, with a clear definition of responsibilities, methods for invitation, site(s) of visits, methods for clinical and radiological evaluation, follow-up duration, role and function of family doctors and specialists. These women will be invited to get a mammogram in dedicated sessions starting from the year after the end of treatment. The planned follow-up duration will be at least 10 years and will be defined on the basis of patient's age and preferences, taking into consideration organizational matters. Special agreements can be defined in the case of women who have their follow-up planned at other qualified centers. Dedicated screening sessions should include: evaluation of familial/personal history (if previously not done) for identifying high-risk conditions which could indicate a different screening strategy; immediate evaluation of mammograms by one or, when possible, two breast radiologists with possible addition of supplemental mammographic views, digital breast tomosynthesis, clinical breast examination, breast ultrasound; and prompt planning of possible further workup. Results of these screening sessions should be set apart from those of general female population screening and presented in dedicated reports. The following research issues are suggested: further risk stratification and effectiveness of follow-up protocols differentiated also for BC pathologic subtype and molecular classification, and evaluation of different models of survivorship care, also in terms of cost-effectiveness.

The purpose of this study is to describe the semiological features of rare breast neoplasms at magnetic resonance mammography (MRM).Out of 468 MRMs of patients with a histological diagnosis of breast malignancy, 27 (5.7%) cases of rare breast neoplasms, confirmed by definitive histological analysis on surgical specimens, were selected: four (0.9%) intracystic papillary neoplasms, four (0.9%) intraductal papillary neoplasms, five (1.0%) invasive papillary neoplasms, two (0.4%) medullary carcinomas, seven (1.5%) mucinous carcinomas, three (0.6%) tubular carcinomas, one (0.2%) tubulo-lobular carcinoma and one (0.2%) desmoid tumour. Two radiologists evaluated the MRM images according to Fischer criteria and indicated a level of diagnostic suspicion. In particular, MRM lesion morphology and enhancement characteristics were analysed.No semiologic features of malignancy or peculiar appearances indicating rare breast neoplasm were identified. On the contrary, MRM appearance was nonspecific and often suggestive of probably benign (Breast Imaging Reporting and Data System-BI-RADS 3) (40%) or benign lesions (BI-RADS 2) (7.5%), or lesions were undetectable at MRM (BI-RADS 1) (7.5%).Frequently, rare breast neoplasms show low suspicious morphologic and kinetic patterns at MRM, and they are often classified as indeterminate lesions. This is probably due to their high grade of differentiation and their histological features.

The introduction of systems for automated reading in mammography has been proposed to improve the sensitivity [computer-aided detection (CADe) systems] and, more recently, the specificity [computer-aided diagnosis (CADi) systems] of the test. Only CADe systems have been approved by the U.S. Food and Drug Administration (FDA) and are used in current practice. These systems are still under discussion. Several studies have demonstrated that they are beneficial to inexperienced readers and that, through comparison with the computer, radiologists are led to improve their performance. However, there is still considerable variation among different studies in the level of benefit deriving from CAD. Therefore the role of these systems in clinical practice is still debated, and their real contribution to the overall management of the diagnostic process is not yet clear.

The purpose of this study was to investigate supraspinatus tendon sonographic morphology in a population of young overhead athletes in correlation with main pathologic models of secondary shoulder impingement syndrome. Between April and May 2004, 20 subjects (ten professional basketball players and ten non-athlete controls of the same age, weight and height ranges) underwent bilateral, standardised, sonographic sholulder examination to evaluate supraspinatus echotexture, supraspinatus and subacromial bursa thickness, subacromial space width (cutoff of 7 mm) and dynamic anterior impingement beneath the acromial margin. All subjects were right handed. No symptomatic controls were found. Four players showed signs and symptoms of right atraumatic shoulder instability (two cases) or supraspinatus impingement with anterior pain (one right and one left shoulder). Subacromial space width was reduced in five right-dominant players’ shoulders (three symptomatic and two asymptomatic) versus zero controls (p

This study was undertaken to evaluate whether magnetic resonance (MR) imaging is able to rule out malignancy in the case of BI-RADS 3 microcalcifications, providing a sufficient negative predictive value (NPV) for early work-up, and to reduce unnecessary stereotactically guided vacuum-assisted biopsy (SVAB) procedures. We prospectively enrolled consecutive women with BI-RADS 3 microcalcifications, who subsequently underwent MR imaging and SVAB. The MR studies were reviewed according to the MR-BI-RADS classification system; lesions assessed as MR-BI-RADS 1 and 2 were considered negative for malignancy, categories MR-BI-RADS 3, 4 and 5 indicated malignant lesions. The presence of additional findings was recorded. Histologic analysis and follow-up were the reference standard. MR sensitivity, specificity, positive predictive value (PPV) and NPV were calculated. The final population consisted of 71 lesions. Histologic analysis showed malignancy in six cases (malignancy rate 8 %). At MR analysis, 60 (85 %) lesions were considered negative for malignancy and 11 (15 %) malignant. Additional MR imaging findings were identified in 19 (27 %) patients, all corresponding to nonmalignant lesions. MR sensitivity was 33 %, specificity 86 %, PPV 18 % and NPV 93 %. Because of its relatively low NPV, MR imaging is not able to safely exclude malignancy in the case of BI-RADS 3 microcalcifications. The relatively high malignancy rate found in this study might support SVAB in the case of BI-RADS 3 microcalcifications.

The authors compared the accuracy of diffusion-weighted imaging (DWI) visual analysis (VA) vs. apparent diffusion coefficient quantification (ADC-Q) in assessing malignancy of solid focal liver lesions (FLLs).Using a 1.5-T system, two radiologists retrospectively assessed as benign or malignant 50 solid FLLs: (a) by VA of signal intensity on DWI images at b=800 s/mm(2) and ADC map; (b) by quantifying lesion ADC. Reference standard included histology or follow-up confirmation of diagnosis by a consensus panel. Receiver operating characteristic (ROC) curve analysis was performed.because of 20 false-negative hepatocellular carcinomas, VA showed lower accuracy than ADC-Q (52.0% VS. 68.0%). however, stratified accuracy for metastases was higher with VA (75.0 VS. 66%). ADC and signal features of malignant and benign FLLs were found to largely overlap.VA performed worse than ADC-Q for hepatocellular carcinoma and better for metastases. Overall, the accuracy of both methods was limited because of the overlap in visual appearance and ADC values between solid benign and malignant FLLs.

This study was performed to assess the role of magnetic resonance cholangiography (MRC) in the clinical decision-making process of referring physicians when managing liver-transplanted patients.Over a 6-month period, 21 liver-transplanted patients with a suspected biliary complication were referred for MRC. Referring physicians were asked to prospectively state, before and after MRC, the leading diagnosis; the level of confidence (on a 0-100% scale); the most appropriate diagnostic/therapeutic plan. Data analysis assessed was the diagnostic yield of MRC; the proportion of change in the leading diagnosis; the therapeutic efficacy (i.e. proportion of change in the initial diagnostic/therapeutic plan); the diagnostic thinking efficacy (i.e., gain in diagnostic confidence). Statistical significance was assessed with the Mann-Whitney U test. MRC accuracy was also calculated.Data analysis showed a diagnostic yield of 85.7%; a proportion of change in leading diagnosis of 19.0%; a therapeutic efficacy of 42.8%; a diagnostic thinking efficacy for concordant and discordant leading diagnoses of 18.8% and 78.7%, respectively (p0.01). MRC accuracy was 92.3%.MRC significantly increased the diagnostic confidence, irrespective of the concordance between pre- and posttest diagnoses. Moreover, MRC determined a change in patient management in a significant proportion of cases, leading to clinical benefits.

The authors evaluated the diagnostic yield of chest multidetector computed tomography (MDCT) in acute pulmonary embolism (PE) and the proportion of other clinically relevant findings in a large cohort of consecutive inpatients and patients referred from the emergency department (outpatients).A total of 327 radiological reports of chest MDCT scans performed for suspected acute PE in 327 patients (158 men, 169 women; mean age 69 years, standard deviation 17.33 years; 233 inpatients, 94 outpatients) were retrospectively evaluated and classified into four categories: 1, positive for PE; 2, negative for PE but positive for other findings requiring specific and immediate intervention; 3, completely negative or positive for findings with a potential for significant morbidity requiring specific action on follow-up; 4, indeterminate. The distribution of findings by categories among the entire population and inpatients and outpatients separately was calculated (chi-square test, α=0.05).In the entire population, the diagnostic yield (i.e. proportion of cases classified as category 1) was 20.2% (66/327). Proportions of cases classified as categories 2, 3 and 4 were 27.5% (90/327), 44.3% (145/327) and 7.9% (26/327), respectively. No statistically significant difference was found between inpatients and outpatients (p=0.193).In patients with suspected acute PE, chest MDCT provides evidence of conditions requiring immediate and specific intervention (i.e. categories 1 and 2) in nearly 50% of cases, without differences between inpatients and outpatients.

The aim of this study was to evaluate the role of ultrasound (US)-guided core biopsy in the diagnosis of ductal carcinoma in situ (DCIS) and to correlate the histological results on percutaneous biopsy and surgical excision.Out of 2,423 consecutive core biopsies performed under US guidance, we evaluated 65 lesions with a histological diagnosis of DCIS. All patients underwent mammography, high-frequency broadband US and percutaneous breast biopsy with a 14-gauge needle and a mean number of five samples (range 4-7 passes). Surgical excision was performed in all cases, and the histological results on the surgical specimen were correlated with those on core biopsy samples. The sonographic features of DCIS lesions were described, comparing pure DCIS (those confirmed by definitive histology) and DCIS with invasive component at surgical excision.Twenty-seven out of 65 DCIS at core biopsy were found to have an invasive or microinvasive component at surgical excision, leading to rate of histological underestimation of core biopsy of 41.5%. The most frequent sonographic appearances were: (a) mass without microcalcifications (47.4% of pure DCIS, 63% of DCIS with invasive component); (b) mass with microcalcifications (23.7% of pure DCIS, 22% of DCIS with invasive component); (c) isolated microcalcifications (10.5% of pure DCIS); (d) ductal abnormalities (18.4% of pure DCIS, 15% of DCIS with invasive component).Due to the high underestimation rate of core biopsy, caution is mandatory in the case of DCIS diagnosis on core biopsy. Although some histological features (such as stromal fibrosis, periductal inflammatory infiltrate, high nuclear grade) can suggest the presence of an invasive component, the sonographic appearance of DCIS cannot be used to predict the cases that are underestimated on US-guided core biopsy. Nevertheless, a sonographically detectable solid component, either inside dilatated ducts or associated with microcalcifications, and a size greater than 20 mm are frequently associated with the presence of an invasive component.

This study was designed to establish whether the measurement of apparent diffusion coefficients (ADCs) is clinically accurate in diagnosing liver fibrosis in a selected series of cirrhotic patients.Twenty-eight cirrhotic patients (mean age 58.1 years) with histologically proven liver fibrosis and 29 healthy controls (mean age 43.8 years) underwent liver diffusion-weighted magnetic resonance (MR) using a 1.5-Tesla (T) magnet equipped with a phased-array coil. Diffusion studies with parallel imaging [generalized autocalibrating partially parallel acquisitions (GRAPPA)] were performed within a single breath-hold using a single-shot spin-echo echo-planar sequence (TE 74 ms) using four b values: b=0, 150, 250 and 400 s/mm(2). A unidirectional diffusion gradient was applied. ADCs were measured on ADC maps.Mean ADC was significantly lower in cirrhotic livers than in controls (1.11+/-0.16 vs. 1.54+/-0.12.10(-3)mm(2)/s) (p0.0001). Receiver operating characteristic (ROC) analysis showed an area under the curve (AUC) of 0.96 [confidence interval (CI) 95%:(0.87; 0.94)], demonstrating higher sensitivity and specificity (92.9% and 100%, respectively) for an ADC cutoff of 1.31.10(-3)mm(2)/s. Positive predictive value (PPV), negative predictive value (NPV) and overall accuracy were 100%, 99.9% and 96.4%, respectively.Diffusion-weighted MRI is an accurate tool in evaluating advanced liver fibrosis if an optimised single-shot spin-echo echo-planar sequence with maximum intermediate b value is used. The ADC threshold for liver fibrosis was 1.31.10(-3)mm(2)/s.

The purpose of this study was to assess the accuracy and clinical usefulness of stereotactic vacuum-assisted biopsy (VAB) for diagnosing suspicious, nonpalpable, only mammographically detectable breast lesions. We retrospectively evaluated the results of percutaneous stereotactic VAB with 11-gauge needles performed over a period of 34 months on 228 nonpalpable suspicious breast lesions detectable on mammography only [Breast Imaging Reporting and Data System (BI-RADS) 3: 25.9%; BI-RADS 4: 67.1%; BI-RADS 5: 7%]. The imaginghistological concordance was ascertained for each lesion. In cases of discordance, repeat biopsy or surgical excision were recommended; in cases of benign lesions, we urged a follow-up of at least 6 months and for borderline and malignant lesions a surgical excision. We also evaluated concordance between VAB results and subsequent examinations (surgical excision or followup). VAB demonstrated 123 (54%) benign lesions (with six cases of imaging–histological discordance), 26 (11.4%) borderline lesions and 79 (34.6%) malignant lesions. We obtained a suitable post-VAB mammographic or histological evaluation for 78 benign lesions, 17 borderline lesions and 76 malignant lesions, with one (1.3%) false negative (FN) case, two (11.8%) underestimations of borderline lesions, 14 (18.4%) underestimations of malignant lesions and no (0%) false positive cases. We did not observe any postbiopsy complications or scars. Percutaneous histological VAB with an 11-gauge needle proved to be, as reported in previous studies, a reliable method for diagnosing nonpalpable, mammographically detectable only breast lesions, with an underestimation rate lower than core biopsy and a FN rate similar to that of surgical biopsy, without any significant complications.

To retrospectively correlate high-risk proliferative breast lesions (radial scar, atypical lobular hyperplasia, lobular carcinoma in situ and papillary lesions) diagnosed on core biopsy with the definitive histopathological diagnosis obtained after surgical excision or with the follow-up, in order to assess the role of core biopsy in such lesions. To discuss the management of the patient after a core biopsy diagnosis of high-risk proliferative breast lesion.We evaluated 74 out of 1776 core biopsies consecutively performed on 67 patients. The histopathologic findings were as follows: 11 radial scars (RS), 3 atypical lobular hyperplasias (ALH), 3 lobular carcinomas in situ (LCIS), 57 benign papillary lesions. All patients underwent bilateral mammography, whole-breast ultrasound with a linear-array broadband transducer, and core biopsy with a 14 Gauge needle and a mean number of samples of 5 (range 4-7). Sixty-two of 67 patients, for a total of 69/74 lesions, underwent surgical biopsy despite benign histopathologic findings, mostly because of highly suspicious imaging for malignancy (BIRADS 4-5), whereas 5 patients refused surgery and have been followed up for a least 18 months and are still being followed up (2 with RS, 1 with ADH and 2 with papillary lesions).Among the core biopsied lesions with a diagnosis of RS (n = 11) pathology revealed one ductal carcinoma in situ (DCIS) (this case was characterized by granular microcalcifications on mammography and by a mass with irregular margins on ultrasound). Also in the group of ADH (n = 3) pathology revealed one DCIS (lesion not visible on mammography but depicted as a suspicious mass on US). In the group of LCIS (n = 3) pathologists found an invasive lobular carcinoma (ILC). Among the benign papillary lesions (n = 57) histopathologic analysis of the surgical specimen revealed 7 malignant lesions (4 papillary carcinomas and 3 DCIS), whose mammographic and ultrasound findings were indistinguishable from benign lesions. Altogether there were 10 false negative results (underestimation) out of 74 core biopsies with a diagnosis of high-risk proliferative breast lesions.The high rate of histological underestimation after core biopsy (10/74) (13.5%) demands a very careful management of patents with a core biopsy diagnosis of high-risk proliferative breast lesions, especially in the case of RS, lobular neoplasia and papillary lesions. However, the high imaging suspicion for malignancy prompts surgery. It is possible to assume that, when there is a low imaging suspicion for malignancy, when enough tissue has been sampled for pathology and no atypia is found within the lesions, surgery is not mandatory but a very careful follow-up is recommended. We must underline that there is no agreement regarding the quantity of tissue to sample. Vacuum-assisted biopsy may lead to better results, although there is as yet no proof that it can actually replace surgery in this group of lesions, since it seems only to reduce but not abolish the histological underestimation.

To validate a protocol for creating virtual models to be used in the construction of solid prototypes useful for the planning-simulation of maxillo-facial surgery, in particular for very complex anatomic and pathologic problems. To optimize communications between the radiology, engineering and surgical laboratories.We studied 16 patients with different clinical problems of the maxillo-facial district. Exams were performed with multidetector computed tomography (MDCT) and single slice computed tomography (SDCT) with axial scans and collimation of 0.5-2 mm, and reconstruction interval of 1 mm. Subsequently we performed 2D multiplanar reconstructions and 3D volume-rendering reconstructions. We exported the DICOM images to the engineering laboratory, to recognize and isolate the bony structures by software. With these data the solid prototypes were generated using stereolitography. To date, surgery has been preformed on 12 patients after simulation of the procedure on the stereolithographyc model.The solid prototypes constructed in the difficult cases were sufficiently detailed despite problems related to the artefacts generated by dental fillings an d prostheses. In the remaining cases the MPR/3D images were sufficiently detailed for surgical planning. The surgical results were excellent in all patients who underwent surgery, and the surgeons were satisfied with the improvement in quality and the reduction in time required for the procedure.MDCT enables rapid prototyping using solid replication, which was very helpful in maxillo-facial surgery, despite problems related to artifacts due to dental fillings and prosthesis within the acquisition field; solutions for this problem are work in progress. The protocol used for communication between the different laboratories was valid and reproducible.

Purpose To assess how frequently foci are identified on MRI in high-risk patients, and their association with malignancy, breast density, and background parenchymal enhancement (BPE). Materials and methods In this multicentric study, two readers, in consensus, retrospectively reviewed screening breast MRI of 245 high-risk women, performed between 2009 and 2014. Eligible patients had at least two consecutive screening MRI, and a follow-up of at least 1 year after a lesion was first detected; histology was available for all suspicious findings. Breast density, BPE (both using BI-RADS lexicon), presence, and changes at follow-up for foci were evaluated. Clinical history of the patients was reviewed. Chi-square test was used to define significant correlations. Results 166 women (mean age 43 years), who underwent a median of 4 MRI (range 2–6) during the study period, were included. 68 foci were found in 58 women [34.9 %, 95 % confidence interval (CI) 28.1–42.5 %]. Foci were more frequent in dense breasts (P = 0.079) and with moderate or marked BPE (P