ABSTRACT
OBJECTIVES: To perform a meta-analysis of the diagnostic performance of learning (ML) algorithms (conventional and deep learning algorithms) for the classification of malignant versus benign focal liver lesions (FLLs) on US and CEUS. METHODS: Available databases were searched for relevant published studies through September 2022. Studies met eligibility criteria if they evaluate the diagnostic performance of ML for the classification of malignant and benign focal liver lesions on US and CEUS. The pooled per-lesion sensitivities and specificities for each modality with 95% confidence intervals were calculated. RESULTS: A total of 8 studies on US, 11 on CEUS, and 1 study evaluating both methods met the inclusion criteria with a total of 34,245 FLLs evaluated. The pooled sensitivity and specificity of ML for the malignancy classification of FLLs were 81.7% (95% CI, 77.2-85.4%) and 84.8% (95% CI, 76.0-90.8%) for US, compared to 87.1% (95% CI, 81.8-91.0%) and 87.0% (95% CI, 83.1-90.1%) for CEUS. In the subgroup analysis of studies that evaluated deep learning algorithms, the sensitivity and specificity of CEUS (n = 4) increased to 92.4% (95% CI, 88.5-95.0%) and 88.2% (95% CI, 81.1-92.9%). CONCLUSIONS: The diagnostic performance of ML algorithms for the malignant classification of FLLs was high for both US and CEUS with overall similar sensitivity and specificity. The similar performance of US may be related to the higher prevalence of DL models in that group.
Subject(s)
Liver Neoplasms , Humans , Liver Neoplasms/pathology , Contrast Media , Ultrasonography/methods , Tomography, X-Ray Computed , Sensitivity and Specificity , Machine Learning , Liver/diagnostic imagingABSTRACT
In this review, the authors address the analysis of different types of malignant focal liver lesions (FLLs) using contrast-enhanced ultrasonography (CEUS). The specific enhancing patterns of hepatocellular carcinoma, cholangiocarcinoma, and metastases are discussed and exemplified with images. In addition, the use of CEUS in malignant portal vein thrombosis is discussed. The advantages and limitations of CEUS for the analysis of malignant FLLs are also discussed.
ABSTRACT
OBJECTIVES: To evaluate the feasibility and diagnostic value of using a 2D spin-echo MR elastography (SE-MRE) sequence at 3.0 Tesla for solid focal liver lesions (FLL) characterization. METHODS: This prospective study included 55 patients with solid FLL (size > 20 mm), who underwent liver SE-MRE at 3 Tesla between 2016 and 2019. Stiffness measurements were performed by two independent readers blinded to the complete MRI exam or patient information. Histological confirmation or typical behavior on the complete MRI exam evaluated in consensus by expert abdominal radiologists was used as reference standard. FLLs were grouped and compared (malignant vs. benign) using the Mann-Whitney and Kruskal-Wallis tests. MRE diagnostic performance was assessed, and stiffness cutoffs were obtained by analysis of ROC curves from accuracy maximization. A linear regression plot was used to evaluate inter-rater agreement for FLLs stiffness measurements. p values < 0.05 were considered statistically significant. RESULTS: The final study group comprised 57 FLLs (34 malignant, 23 benign). Stiffness measurements were technically successful in 91.23% of lesions. To both readers, the median stiffness of the lesions categorized as benign was 4.5 ± 1.5 kPa and in the malignant group 6.8 ± 1.7 and 7.5 ± 1.5 kPa depending on the reader. A cutoff of 5.8 kPa distinguished malignant and benign lesions with 88% specificity and 75-85% accuracy depending on the reader. The inter-rater agreement was 0.90 ± 0.04 with a correlation coefficient of 0.94. CONCLUSION: 2D-SE-MRE at 3.0 T provides high specificity and PPV to differentiate benign from malignant liver lesions. Trial registration 18FFUA-A02.
Subject(s)
Elasticity Imaging Techniques , Liver Neoplasms , Humans , Liver/diagnostic imaging , Liver Neoplasms/diagnostic imaging , Magnetic Resonance Imaging , Prospective Studies , Reproducibility of ResultsABSTRACT
Las lesiones focales del hígado comprenden un grupo heterogéneo de entidades como tumores benignos y malignos, sólidos o quistes, vasculares, masas inflamatorias y metástasis, entre otras. Reportar lesiones focales hepáticas en niños, características clínicas, estudios por imágenes y conducta terapéutica. Estudio descriptivo, retrospectivo, basado en la revisión de 16 historias clínicas de niños con lesiones focales hepáticas en el periodo comprendido entre enero 2007 a enero 2012. Se excluyeron los abscesos hepáticos. El 68,75%11 eran del sexo masculino y 31,25%5 del sexo femenino; con edad promedio de 9,5 años (rango: 10 días - 14 años). Las lesiones focales benignas fueron: hemangioma 4/16 (25%), hematoma hepático 4/16 (25%), quiste hepático 3/16 (18,75%), hiperplasia nodular focal 1/16 (6,25%) adenoma 1/16 (6,25%) y las malignas: hepatocarcinoma 2/16 (12,50%) y hepatoblastoma 1/16 (6,25%). La mayoría de los pacientes (56,25%) se encontraban asintomáticos, 3/16 (18,75%) presentaron dolor abdominal, 3/16 (18,75%) masa palpable y 2/16 (12,5%) ictericia. La lesión fue identificada por ultrasonido abdominal en todos los pacientes y sólo el 56,25% (9/16) ameritaron tomografía con contraste. Al 43,75% (7/16) se les realizó marcadores tumorales y de estos el 42,85% (3/7) reportaron neoplasias. El funcionalismo hepático reportó normal en el 75% de los casos. El ultrasonido abdominal es el método de elección para el diagnóstico y seguimiento de las lesiones focales hepáticas, mientras el origen de las mismas, guía la conducta terapéutica
Focal liver lesions comprise a heterogeneous group of entities such as benign and malignant, solid or cysts, vascular, inflammatory masses and metastasis among others. To report focal liver lesions in children, clinical features, imaging and therapeutic behavior. A descriptive, retrospective study based on a review of 16 case histories of children with focal liver lesions in the period from January 2007 to January 2012. Liver abscesses were excluded. 68.75%11 were male and 31.25%5 females, mean age 9.5 years (range: 10 days - 14 years). Focal lesions were benign: Hemangioma 4/16 (25%), liver hematoma 4/16 (25%), hepatic cyst 3/16 (18.75%), nodular focal hyperplasia 1/16 (6.25%) and adenomas 1/16 (6.25%) and malignant: Hepatocellular carcinoma 2/16 (12.50%) and hepatoblastoma 1/16 (6.25%). Most patients (56.25%) were asymptomatic, 3/16 (18.75%) had abdominal pain, 3/16 (18.75%) and palpable mass 2/16 (12.5%) jaundice. The lesion was identified by abdominal ultrasound in all patients and only 56.25% (9/16) needed tomography with contrast. To 43.75% (7/16) underwent tumor markers and of these 42.85% (3/7) reported neoplasm. Normal hepatic function reported in 75% of cases. An abdominal ultrasound is the method of choice for the diagnosis and management of focal liver lesions, while the origin of them, guide therapeutic conduct