Two fully automated data-driven 3D whole-breast segmentation strategies in MRI for MR-based breast density using image registration and U-Net with a focus on reproducibility.

Presence of higher breast density (BD) and persistence over time are risk factors for breast cancer. A quantitatively accurate and highly reproducible BD measure that relies on precise and reproducible whole-breast segmentation is desirable. In this study, we aimed to develop a highly reproducible and accurate whole-breast segmentation algorithm for the generation of reproducible BD measures. Three datasets of volunteers from two clinical trials were included. Breast MR images were acquired on 3 T Siemens Biograph mMR, Prisma, and Skyra using 3D Cartesian six-echo GRE sequences with a fat-water separation technique. Two whole-breast segmentation strategies, utilizing image registration and 3D U-Net, were developed. Manual segmentation was performed. A task-based analysis was performed: a previously developed MR-based BD measure, MagDensity, was calculated and assessed using automated and manual segmentation. The mean squared error (MSE) and intraclass correlation coefficient (ICC) between MagDensity were evaluated using the manual segmentation as a reference. The test-retest reproducibility of MagDensity derived from different breast segmentation methods was assessed using the difference between the test and retest measures (Δ2-1), MSE, and ICC. The results showed that MagDensity derived by the registration and deep learning segmentation methods exhibited high concordance with manual segmentation, with ICCs of 0.986 (95%CI: 0.974-0.993) and 0.983 (95%CI: 0.961-0.992), respectively. For test-retest analysis, MagDensity derived using the registration algorithm achieved the smallest MSE of 0.370 and highest ICC of 0.993 (95%CI: 0.982-0.997) when compared to other segmentation methods. In conclusion, the proposed registration and deep learning whole-breast segmentation methods are accurate and reliable for estimating BD. Both methods outperformed a previously developed algorithm and manual segmentation in the test-retest assessment, with the registration exhibiting superior performance for highly reproducible BD measurements.

Comparison of Gadobenate-Enhanced MRI and Gadoxetate-Enhanced MRI for Hepatocellular Carcinoma Detection Using LI-RADS Version 2018: A Prospective Intraindividual Randomized Study.

BACKGROUND. Gadobenate and gadoxetate show different degrees of intracellular accumulation within hepatocytes, potentially impacting these agents' relative performance for hepatocellular carcinoma (HCC) diagnosis. OBJECTIVE. The purpose of this article was to perform an intraindividual comparison of gadobenate-enhanced MRI and gadoxetate-enhanced MRI for detection of HCC and to assess the impact of inclusion of hepatobiliary phase images on HCC detection for both agents. METHODS. This prospective study enrolled 126 patients (112 men, 14 women; mean age, 52.3 years) at high risk for HCC who consented to undergo two 3-T liver MRI examinations (one using gadobenate [0.05 mmol/kg], one using gadoxetate [0.025 mmol/kg]) separated by 7-14 days. The order of the two contrast agents was randomized. All examinations included postcontrast dynamic and hepatobiliary phase images (120 minutes for gadobenate, 20 minutes for gadoxetate). Three radiologists independently reviewed the gadobenate and gadoxetate examinations in separate sessions and recorded the location of detected observations. Observations were classified using LI-RADS version 2018 and using a LI-RADS modification whereby hepatobiliary phase hypointensity may upgrade observations from category LR-4 to LR-5. Observations classified as LR-5 were considered positive interpretations for HCC. Diagnostic performance for histologically confirmed HCC (n = 96) was assessed. RESULTS. Across readers, sensitivity for HCC for gadobenate versus gadoxetate was 74.0-80.2% versus 54.2-67.7% using dynamic images alone and 82.1-87.4% versus 66.3-81.1% using dynamic and hepatobiliary phase images. For HCCs measuring 1.0-2.0 cm, sensitivity for gadobenate versus gadoxetate was 61.9% (all readers) versus 38.1-57.1% using dynamic images alone and 76.2-85.7% versus 52.4-61.9% using dynamic and hepatobiliary phase images. PPV for HCC ranged from 88.6% to 97.4% across readers, agents, and image sets. CONCLUSION. Sensitivity for HCC was higher for gadobenate than for gadoxetate, whether using dynamic images alone or dynamic and hepatobiliary phase images; the improved sensitivity using gadobenate was more pronounced for small HCCs. Whereas hepatobiliary phase images improved sensitivity for both agents, sensitivity of gadobenate using dynamic images alone compared favorably with that of gadoxetate using dynamic and hepatobiliary phase images. CLINICAL IMPACT. The findings support gadobenate as a preferred agent over gadoxetate when performing liver MRI in patients at high risk for HCC.

Editorial for "Utility of Diffusion-Weighted Imaging for Guiding Clinical Management of Patients With Kidney Transplant: A Prospective Study".

LEVEL OF EVIDENCE: 5 TECHNICAL EFFICACY STAGE: 5 J. Magn. Reson. Imaging 2020;52:575-576.

Liver-Directed Therapy for Hepatocellular Carcinoma: An Overview of Techniques, Outcomes, and Posttreatment Imaging Findings.

OBJECTIVE: The purposes of this article are to describe the indications, techniques, and results of arterially directed therapies and ablation and to review the imaging assessment of response and complications. CONCLUSION: Most patients with hepatocellular carcinoma are not eligible for surgery, and systemic treatments are suboptimal. Therefore, locoregional therapy plays a large role in this disease. Locoregional therapies include arterially directed therapies, ablation, and radiation therapy.

A Review and Update of Treatment Options and Controversies in the Management of Hepatocellular Carcinoma.

OBJECTIVE: To review the current management, outline recent advances and address controversies in the management of hepatocellular carcinoma (HCC). SUMMARY OF BACKGROUND DATA: The treatment of HCC is multidisciplinary involving hepatologists, surgeons, medical oncologists, radiation oncologists, radiologists, interventional radiologists, and other disciplines. Each of these disciplines brings its unique perspective and differing opinions that add to controversies in the management of HCC. METHODS: A focused literature review was performed to identify recent studies on the management of HCC and thereby summarize relevant information on the various therapeutic modalities and controversies involved in the treatment of HCC. RESULTS: The main treatment algorithms continue to rely on hepatic resection or transplantation with controversies involving patients harboring early stage disease and borderline hepatic function. The other treatment strategies include locoregional therapies, radiation, and systemic therapy used alone or in combination with other treatment modalities. Recent advances in locoregional therapies, radiation, and systemic therapies have provided better therapeutic options with curative intent potential for some locoregional therapies. Further refinements in combination therapies such as algorithms consisting of locoregional therapies and systemic or radiation therapies are likely to add additional options and improve survival. CONCLUSIONS: The management of HCC has witnessed significant strides with advances in existing options and introduction of several new treatment modalities of various combinations. Further refinements in these treatment options combined with enrollment in clinical trials are essential to improve the management and outcomes of patients with HCC.

Magnetic resonance imaging of acute appendicitis in pregnancy: a 5-year multiinstitutional study.

OBJECTIVE: The purpose of this study was to determine the diagnostic performance of magnetic resonance imaging (MRI) in the diagnosis of acute appendicitis during pregnancy in a multiinstitutional study. STUDY DESIGN: In this multicenter retrospective study, the cases of pregnant women who underwent MRI evaluation of abdominal or pelvic pain and who had clinical suspicion of acute appendicitis between June 1, 2009, and July 31, 2014, were reviewed. All MRI examinations with positive findings for acute appendicitis were confirmed with surgical pathologic information. Sensitivity, specificity, negative predictive values, and positive predictive values were calculated. Receiver operating characteristic curves were generated, and area under the curve analysis was performed for each participating institution. RESULTS: Of the cases that were evaluated, 9.3% (66/709) had MRI findings of acute appendicitis. Sensitivity, specificity, accuracy, positive predictive value, and negative predictive values were 96.8%, 99.2%, 99.0%, 92.4%, and 99.7%, respectively. There was no statistically significant difference between centers that were included in the study (pair-wise probability values ranged from 0.12-0.99). CONCLUSION: MRI is useful and reproducible in the diagnosis of suspected acute appendicitis during pregnancy.

Graduating 4th year radiology residents' perception of optimal imaging modalities for neoplasm and trauma: a pilot study from four U.S. universities / Percepção de médicos residentes em radiologia de 4º ano sobre as melhores modalidades de imagem na investigação de neoplasias e trauma: um estudo piloto de quatro universidades americanas

OBJECTIVE: Our purpose was to assess 4th year radiology residents' perception of the optimal imaging modality to investigate neoplasm and trauma. MATERIALS AND METHODS: Twenty-seven 4th year radiology residents from four residency programs were surveyed. They were asked about the best imaging modality to evaluate the brain and spine, lungs, abdomen, and the musculoskeletal system. Imaging modalities available were MRI, CT, ultrasound, PET, and X-ray. All findings were compared to the ACR appropriateness criteria. RESULTS: MRI was chosen as the best imaging modality to evaluate brain, spine, abdominal, and musculoskeletal neoplasm in 96.3 percent, 100 percent, 70.4 percent, and 63 percent of residents, respectively. CT was chosen by 88.9 percent to evaluate neoplasm of the lung. Optimal imaging modality to evaluate trauma was CT for brain injuries (100 percent), spine (92.6 percent), lung (96.3 percent), abdomen (92.6 percent), and major musculoskeletal trauma (74.1 percent); MRI was chosen for sports injury (96.3 percent). There was agreement with ACR appropriateness criteria. CONCLUSION: Residents' perception of the best imaging modalities for neoplasm and trauma concurred with the appropriateness criteria by the ACR.

OBJETIVO: Avaliar a percepção de médicos residentes em radiologia de 4º ano sobre as melhores modalidades de imagem na investigação de neoplasias e trauma. MATERIAIS E MÉTODOS: Vinte e sete médicos residentes de 4º ano de quatro programas de residência em radiologia americanos participaram do estudo. Aos participantes foi perguntado sobre a melhor modalidade de imagem para se avaliar o cérebro e a coluna vertebral, pulmões, abdome e o sistema musculoesquelético. As modalidades de imagem disponíveis foram: RM, TC, ultrassonografia, PET e radiografia simples. Todos os achados foram comparados com os Critérios de Adequação de Exames de Imagem e Radioterapia do ACR. RESULTADOS: A RM foi escolhida como melhor modalidade de imagem para se avaliar neoplasias encefálicas, espinhais, abdominais e musculoesqueléticas por 96,3 por cento, 100 por cento, 70,4 por cento e 63 por cento dos residentes, respectivamente. A TC foi escolhida por 88,9 por cento dos residentes para avaliar neoplasias pulmonares. A modalidade de imagem ótima para se avaliar trauma foi a TC para lesões encefálicas (100 por cento), espinhais (92,6 por cento), pulmonares (96,3 por cento), abdominais (92,6 por cento) e grandes lesões traumáticas musculoesqueléticas (74,1 por cento); a RM foi escolhida para lesões esportivas (96,3 por cento). Observou-se concordância com os critérios de adequação do ACR. CONCLUSÃO: Houve concordância entre a percepção dos residentes sobre as melhores modalidades de imagem para avaliação de neoplasias e trauma e os critérios de adequação do ACR.

Cirrhosis and lesion characterization at MR imaging.

Magnetic resonance (MR) imaging has emerged as an important imaging modality for the assessment of cirrhosis and its complications. Faster sequences now allow high-quality liver imaging with high intrinsic soft-tissue contrast. Automated contrast detection methods in combination with faster sequences allow reproducible capture of the arterial phase, which is essential for the detection and characterization of hepatocellular carcinoma. The lack of ionizing radiation permits routine use of gadolinium-enhanced three-dimensional (3D) fat-suppressed multiphasic imaging with high temporal and spatial resolution. In addition, MR imaging allows simultaneous evaluation of the background liver parenchyma and the liver lesions with the combined use of sequences that include T2-weighted sequences, T1-weighted sequences (including chemical shift imaging), echoplanar diffusion-weighted sequences, dynamic gadolinium-enhanced 3D multiphasic imaging, and liver-specific delayed phase sequences (if contrast agents with hepatobiliary excretion are used). The combination of findings from different sequences often helps pinpoint the nature of the liver abnormalities.

Liver imaging at 3.0 T: diffusion-induced black-blood echo-planar imaging with large anatomic volumetric coverage as an alternative for specific absorption rate-intensive echo-train spin-echo sequences: feasibility study.

Institutional Review Board approval and signed informed consent were obtained by all participants for an ongoing sequence optimization project at 3.0 T. The purpose of this study was to evaluate breath-hold diffusion-induced black-blood echo-planar imaging (BBEPI) as a potential alternative for specific absorption rate (SAR)-intensive spin-echo sequences, in particular, the fast spin-echo (FSE) sequences, at 3.0 T. Fourteen healthy volunteers (seven men, seven women; mean age +/- standard deviation, 32.7 years +/- 6.8) were imaged for this purpose. Liver coverage (20 cm, z-axis) was always performed in one 25-second breath hold. Imaging parameters were varied interactively with regard to echo time, diffusion b value, and voxel size. Images were evaluated and compared with fat-suppressed T2-weighted FSE images for image quality, liver delineation, geometric distortions, fat suppression, suppression of the blood signal, contrast-to-noise ratio (CNR), and signal-to-noise ratio (SNR). An optimized short- (25 msec) and long-echo (80 msec) BBEPI provided full anatomic, single breath-hold liver coverage (100 and 50 sections, respectively), with resulting voxel sizes of 3.3 x 2.7 x 2.0 mm and 3.3 x 2.7 x 4.0 mm, respectively. Repetition time was 6300 msec, matrix size was 160 x 192, and an acceleration factor of 2.00 was used. b Values of more than 20 sec/mm(2) showed better suppression of the blood signal but b values of 10 sec/mm(2) provided improved volume coverage and signal consistency. Compared with fat-suppressed T2-weighted FSE, the optimized BBEPI sequence provided (a) comparable image quality and liver delineation, (b) acceptable geometric distortions, (c) improved suppression of fat and blood signals, and (d) high CNR and SNR. BBEPI is feasible for fast, low-SAR, thin-section morphologic imaging of the entire liver in a single breath hold at 3.0 T.

Magnetic resonance imaging of liver lesions: exceptions and atypical lesions.

On state-of-the-art magnetic resonance imaging, most lesions can be detected and characterized with confidence according to well-known criteria. However, atypical characteristics in some common lesions and the incidental encounter with rare lesions may pose diagnostic difficulties. In this article, six challenging hepatic lesions will be discussed and evaluated on the most important magnetic resonance imaging sequences, with histological correlation when available. In addition, the background information concerning these lesions will be described based on the most recent available literature. By reading this article, the reader will be able to (1) categorize the lesion in solid and fluid-containing lesions, based on the T2 signal intensity; and (2) define the benign or malignant nature of the lesion, in relation to the signal intensity and dynamic enhancement pattern, despite the presence of atypical characteristics of some lesions.

Primary hepatocellular lesions: imaging findings on state-of-the-art magnetic resonance imaging, with pathologic correlation.

Magnetic resonance imaging is routinely used for the workup of patients with focal or diffuse liver disease, including primary hepatocellular lesions, storage diseases, metastatic liver disease, and diseases of the hepatobiliary tree. The most important magnetic resonance imaging sequences used for diagnostic imaging of the liver consist of T1-weighted sequences, T2-weighted sequences, and at least the arterial and delayed phases of dynamic gadolinium-enhanced imaging. This article provides an overview of magnetic resonance imaging of primary hepatocellular lesions and will describe the following: (1) the classification and etiology of primary hepatocellular lesions, including focal nodular hyperplasia, hepatocellular adenoma, and hepatocellular carcinoma; (2) the stepwise carcinogenesis of hepatocellular carcinoma in cirrhosis on magnetic resonance imaging; and (3) the typical imaging findings of primary hepatocellular lesions on magnetic resonance imaging, with differential diagnoses.

Hepatocellular carcinoma complicating biliary atresia after Kasai portoenterostomy.

Kasai portoenterostomy (PE) increases the survival for children with biliary atresia (BA) and consequently postpones subsequential liver transplantation. All long-term survivors, however, develop complications of biliary cirrhosis. We report a case of hepatocellular carcinoma (HCC) in a 19-year-old male patient with BA and Kasai PE. The preoperative abdominal ultrasound and magnetic resonance imaging showed a large hepatic mass (diameter 10 cm). The serum alpha-fetoprotein level was within normal range. Pathologic findings of the mass, after orthotopic liver transplantation, demonstrated a well-differentiated HCC (T1N0M0). HCC is a rare complication of BA, but will intensively impair the survival. Therefore, clinicians should be alert to the development of HCC in this very young patient group. Repeated sequential magnetic resonance imaging of the native liver in patients with Kasai PE is necessary to monitor possible malignant transformation of liver nodules that may potentially develop as a result of chronic cholestatic liver disease.

Extending slice coverage for breathhold fat-suppressed T2-weighted fast spin-echo of the liver at 3.0T: application of variable-rate selective-excitation (VERSE) RF pulses.

PURPOSE: To determine the benefits of variable-rate selective-excitation (VERSE) radio frequency (RF) pulses for increased slice coverage in breathhold (BH) fat-suppressed T2-weighted fast spin-echo (FS-T2W-FSE) liver imaging at 3.0T. MATERIALS AND METHODS: A total of 12 healthy volunteers were imaged on 3.0T, using FS-T2W-FSE. Slice coverage and specific absorption rate (SAR) levels were monitored for VERSE-RF and standard-RF (sRF), respectively. BH time was 25 seconds; slice thickness 3.5 mm. Maximum coverage was recorded for interactive variation of repetition time (TR), bandwidth (BW), and echo-train length (ETL). Image quality was assessed qualitatively and quantitatively. RESULTS: Total slice coverage was always higher using VERSE-RF, but varied depending on the selected parameters. For BW > or = 62.5 kHz, slice coverage using VERSE increased between 38% (TR = 8200 msec) and 58% (TR = 2500 msec). Maximum coverage was obtained for TR = 5000 msec, ETL = 12, and BW > or = 62.5 kHz; with a mean of 31.8 slices for VERSE-RF and 22.5 slices for sRF, respectively (P < 0.005, 41% increased coverage). SAR was lower for VERSE-RF using BW < 41.67 kHz (P < 0.05), and equal to sRF for higher BW. Image quality was best for TR < or = 5000 msec (P < 0.05). FS was more homogeneous for lower ETL (P < 0.05). Blood suppression was best for TR < or = 5000 msec (P < 0.05). CONCLUSION: VERSE-RF pulses can be applied for thin-slice BH FS-T2W liver imaging at 3.0T, with significantly improved slice coverage.

MR imaging of hepatocellular carcinoma: relationship between lesion size and imaging findings, including signal intensity and dynamic enhancement patterns.

PURPOSE: To assess the relationship between lesion size and MR imaging findings of pathologically-proven hepatocellular carcinoma (HCC). MATERIALS AND METHODS: In a retrospective, single-center study, 37 consecutive patients were identified between 1999 and 2005 that underwent preoperative MRI and surgical resection of HCC. A total of 47 lesions (mean size = 6.85 cm, range = 1-25 cm) were assessed for signal intensity (SI), enhancement patterns, and secondary morphologic features. Interobserver rating, percentage enhancement, and contrast-to-noise-ratio (CNR) were determined. Lesions were assessed for combinations of typical MRI features. Regression analysis was used to assess relations between MRI findings and tumor size. RESULTS: On fat-suppressed T2-weighted (T2w) fast-spin-echo, smaller lesions had lower SI compared to larger lesions (P < 0.05). In the arterial phase, smaller lesions showed significantly higher percentage enhancement compared to larger lesions (P < 0.05). In the delayed phase, smaller lesions showed less pronounced washout (P < 0.05). Heterogeneity of the lesions, including fatty infiltration, internal nodules, or mosaic pattern, was observed significantly more frequently in larger lesions (P < 0.001). The classic combination of high T2w signal, strong arterial enhancement, and delayed phase washout was present in 23 of 44 lesions (52%). CONCLUSION: Smaller HCC often showed lower SI on T2w, more intense arterial enhancement, and less pronounced delayed washout compared to larger HCC.

Stepwise carcinogenesis of hepatocellular carcinoma in the cirrhotic liver: demonstration on serial MR imaging.

PURPOSE: To demonstrate imaging findings of stepwise carcinogenesis of hepatocellular carcinoma (HCC) in cirrhosis at serial state-of-the-art MR imaging exams. MATERIALS AND METHODS: In a retrospective search of the hospital archives, three patients were identified in which developing HCC was observed in serial MR examinations, with histopathology or alpha-fetoprotein (AFP) correlation. Image findings were assessed for signal intensity of the lesions at multiple sequences, including dynamic gadolinium-enhanced imaging. RESULTS: Initial findings in patient A showed a small nodule with fatty infiltration that developed in HCC in follow-up MRI, comprised of low-grade dysplastic nodule (DN; DN I), high-grade DN (DN II), and eventually classic HCC. In patient B, increased signal intensity on T2-weighted images in a single DN among numerous regenerative nodules was the only initial sign. Follow up MRI showed further increase in signal intensity and increased neovascularity, which suggested focal HCC in a DN II. Patient C demonstrated gradually increasing neovascularity as only initial sign, with development of classic HCC over time. CONCLUSION: MR imaging provides insight in various pathways of stepwise carcinogenesis of developing HCC in cirrhosis. This may further explain the genetic heterogeneity, and may facilitate early detection and better selection of patients for follow-up.

Stereotactic body radiation therapy for primary and metastatic liver tumors: A single institution phase i-ii study.

The feasibility, toxicity and tumor response of stereotactic body radiation therapy (SBRT) for treatment of primary and metastastic liver tumors was investigated. From October 2002 until June 2006, 25 patients not suitable for other local treatments were entered in the study. In total 45 lesions were treated, 34 metastases and 11 hepatocellular carcinoma (HCC). Median follow-up was 12.9 months (range 0.5-31). Median lesion size was 3.2 cm (range 0.5-7.2) and median volume 22.2 cm3 (range 1.1-322). Patients with metastases, HCC without cirrhosis, and HCC < 4 cm with cirrhosis were mostly treated with 3 x 12.5 Gy. Patients with HCC > or =4 cm and cirrhosis received 5 x 5 Gy or 3 x 10 Gy. The prescription isodose was 65%. Acute toxicity was scored following the Common Toxicity Criteria and late toxicity with the SOMA/LENT classification. Local failures were observed in two HCC and two metastases. Local control rates at 1 and 2 years for the whole group were 94% and 82%. Acute toxicity grade > or =3 was seen in four patients; one HCC patient with Child B developed a liver failure together with an infection and died (grade 5), two metastases patients presented elevation of gamma glutamyl transferase (grade 3) and another asthenia (grade 3). Late toxicity was observed in one metastases patient who developed a portal hypertension syndrome with melena (grade 3). SBRT was feasible, with acceptable toxicity and encouraging local control. Optimal dose-fractionation schemes for HCC with cirrhosis have to be found. Extreme caution should be used for patients with Child B because of a high toxicity risk.

Focal nodular hyperplasia: lesion characteristics on state-of-the-art MRI including dynamic gadolinium-enhanced and superparamagnetic iron-oxide-uptake sequences in a prospective study.

PURPOSE: To image a cohort of patients with pathology-proven focal nodular hyperplasia (FNH) to assess which characteristics of state-of-the-art magnetic resonance imaging (MRI) of the liver are the most useful for improving the detection and characterization of FNH. MATERIALS AND METHODS: In 14 patients, pathology-proven FNH (N=33) were prospectively examined using gadolinium (Gd) and superparamagnetic iron-oxide (SPIO) contrast media. All lesions were evaluated for signal intensity (SI), fatty infiltration, central scar, mode of enhancement with Gd, and uptake of SPIO. The percentage of dynamic contrast enhancement in the arterial, portal, and delayed phases was assessed. The contrast-to-noise ratio (CNR) before and after administration of SPIO contrast was calculated. RESULTS: The SI of the lesions was low to isointense on T1-weighted (T1W) images, and intermediate to isointense on T2W images. Fatty infiltration of the lesions was present in 6%. The percentages of enhancement in the liver and lesion were 110%, 115%, and 95%, and 151%, 182%, and 160%, respectively (P<0.0001). All lesions showed uptake of SPIO with improved conspicuity of the central scar and septa. The CNR values precontrast and post-Gd/SPIO were significantly different for T1 in- and opposed-phase and black-blood echo-planar imaging (BBEPI). CONCLUSION: Combining dynamic Gd-enhanced imaging with T1W and T2W sequences after administration of SPIO facilitates comprehensive evaluation of FNH.

Diagnosis and treatment of benign focal liver lesions.

BACKGROUND: With the routine use of improved imaging modalities, more benign liver lesions are detected nowadays. An accurate characterization of these incidental lesions may be a challenge, and frequently a biopsy or even unnecessary surgery is being performed. However, these interventions are not always to the benefit of the patient. METHODS: A Medline search of studies relevant to imaging diagnosis and management of the most common, benign, solid and non-solid liver lesions was undertaken. References from identified articles were handsearched for further relevant articles. The authors' own experiences with benign liver lesions were also taken into account. RESULTS: Although atypical imaging features are the exception rather than the rule, it is sometimes difficult to differentiate between benign and malignant lesions, and knowledge of their imaging features is essential if unnecessary work-up is to be avoided. The use of tissue-specific contrast media, which has clearly improved the accuracy of highly advanced radiological techniques, may be helpful during differential diagnosis. Once having established an accurate diagnosis, surgery is rarely indicated for a benign liver lesion because of its asymptomatic nature. CONCLUSION: Knowledge of imaging features and a clear management strategy during diagnostic work-up, emphasizing the indications for surgery, will minimize the number of patients who have to undergo biopsy or unnecessary surgery.