Current Medical Imaging - Volume 20, Issue 1, 2024

Deep learning-based diagnosis systems are useful to identify abnormalities in medical images with the greatly increased workload of doctors. Specifically, the rate of new cases and deaths from malignancies is rising for liver diseases. Early detection of liver lesions plays an extremely important role in effective treatment and gives a higher chance of survival for patients. Therefore, automatic detection and classification of common liver lesions are essential for doctors. In fact, radiologists mainly rely on Hounsfield Units to locate liver lesions but previous studies often pay little attention to this factor.

In this paper, we propose an improved method for the automatic classification of common liver lesions based on deep learning techniques and the variation of Hounsfield Unit densities on CT images with and without contrast. Hounsfield Unit is used to locate liver lesions accurately and support data labeling for classification. We construct a multi-phase classification model developed on the deep neural networks of Faster R-CNN, R-FCN, SSD, and Mask R-CNN with the transfer learning approach.

The experiments are conducted on six scenarios with multi-phase CT images of common liver lesions. Experimental results show that the proposed method improves the detection and classification of liver lesions compared with recent methods because its accuracy achieves up to 97.4%.

The proposed models are very useful to assist doctors in the automatic segmentation and classification of liver lesions to solve the problem of depending on the clinician’s experience in the diagnosis and treatment of liver lesions.

Chronic kidney disease has become one of the world's major public health problems, immunoglobulin A (IgA) nephropathy is a common pathological type of CKD. Delaying the progression of IgA nephropathy has currently become the main clinical treatment strategy, precise evaluation of renal pathological injury during follow-up of patients with IgA nephropathy is important. Therefore, it is imperative to develop an accurate and non-invasive imaging technique for effective follow-up of renal pathological injury in patients with IgA nephropathy.

To investigate the clinical value of intravoxel incoherent motion diffusion-weighted imaging (IVIM-DWI) in assessing renal pathological injury in patients with immunoglobulin A (IgA) nephropathy compared with a mono-exponential model.

Altogether, 80 patients with IgA nephropathy were divided into the mild (41 cases) andmoderate–severe (m–s) renal injury groups (39 cases) according to pathology scores, and 20 healthy volunteers were recruited as controls. All participants underwent IVIM-DWI of the kidneys, and renal parenchymal apparent diffusion coefficient (ADC), pure molecular diffusion coefficient (D), pseudo-diffusion coefficient (D*), and perfusion fraction (f) values were measured. One-way analysis of variance, receiver operating characteristic (ROC) curve analysis, and Pearson correlation analysis were performed for all the DWI-derived parameters.

The DWI-derived parameters of the m–s renal injury group were significantly lower than those of the mild renal injury and control groups (P < 0.01). The ROC analysis revealed that f had the largest area under the ROC curve for differentiation between the m–s and mild renal injury groups and between the m–s renal injury and control groups. The f had the largest correlation coefficient with renal pathology scores (r=−0.81), followed by the D* (−0.69), ADC (−0.54), and D values (−0.53), respectively (all P<0.01).

IVIM-DWI demonstrated better diagnostic performance than the mono-exponential model in assessing renal pathological injury in patients with IgA nephropathy.

Many studies have reported Xp 11.2 translocation renal cancer in radioimaging,but there is little literature on the evaluation of Xp11.2 translocation renal cell carcinoma by ultrasound.

To investigate the ultrasonographic features and diagnostic value of renal cell carcinoma associated with Xp11.2 translocation/TFE3 gene fusion in children and adolescents.

The clinical and ultrasonographic data of 10 patients with renal cell carcinoma associated with Xp11.2 translocation/TFE3 gene fusion confirmed by pathology in our hospital were analyzed retrospectively. The age ranged from 3 to 18 years old, including 7 males and 3 females. The tumor location, size, boundary, echo, hemorrhage, cystic change, calcification, blood flow, lymph node status and metastasis were mainly observed, and the results were compared with the pathological results.

There were 10 masses in 10 cases of renal cell carcinoma associated with Xp11.2 translocation/TFE3 gene fusion, including 4 in the right kidney and 6 in the left kidney; the maximum diameter line is 5-23cm; 9 cases had clear mass boundary (90%); 9 masses (90%) showed mixed cystic and solid masses with high echo of solid components, and 1 mass (10%) showed huge multilocular cystic mass with multiple septations; necrosis and cystic changes were seen in all 10 masses (100%); calcification in 5 masses (50%); blood flow signals were seen in the solid components of the mass (100%).

Renal cell carcinoma associated with Xp11.2 translocation/TFE3 gene fusion in children and adolescents are mostly large cystic and solid mixed echo masses, with high echo of solid components, and often accompanied by cystic changes and calcification. Its ultrasonic manifestations have certain characteristics. Color Doppler ultrasound has a certain diagnostic value for this disease.

This article aims to explore the ADC value of ultrahigh b-value DWI and the diagnostic cutoff point in prostate cancer.

A total of 78 patients were included in this study. T2 weighted imaging (T2WI), conventional diffusion-weighted imaging (DWI) (1000 s/mm²), and DWI with ultrahigh b-values of 2000 s/mm² and 3000 s/mm² were performed in each patient. With reference biopsy as the gold standard, the apparent diffusion coefficient (ADC)s of each b-value DWI image were analyzed. According to different b-value receiver operating characteristic (ROC) curves, the ADC diagnostic cutoff point for prostate cancer was determined.

A total of 154 lesions were identified as prostate cancer. The ADC values for conventional DWI and ultrahigh b-value DWI with 2000 s/mm² and 3000 s/mm² were 1.097×10^-3 mm²/s (1.040-1.153), 0.809×10^-3 mm²/s (0.766-0.851) and 0.622×10^-3 mm²/s (0.591-0.652), respectively, in the peripheral zone and 1.085×10^-3 mm²/s (1.022-1.147), 0.815×10^-3 mm²/s (0.770-0.861) and 0.651×10^-3 mm²/s (0.617-0.685) in the transition zone. The area under the curve (AUC)s of the ADC values from ultrahigh b-value DWI (2000 s/mm² and 3000 s/mm²) were 0.824 and 0.852 in the peripheral zone and 0.905 for the ADC values from ultrahigh b-value DWI (3000 s/mm²) in the transition zone. In the peripheral zone, the ADC diagnostic cutoff values for prostate cancer were 0.75×10^-3 mm²/s and 0.685×10^-3 mm²/s in DWI at 2000 s/mm² and 3000 s/mm², respectively, and the diagnosis of transition zone cancer was 0.8×10^-3 mm²/s and 0.634×10^-3 mm²/s, respectively.

The ADC values from ultrahigh b-value DWI demonstrated better consistency and diagnostic efficacy in the diagnosis of prostate cancer.

Diagnosis of pulmonary thromboembolism (PTE) can be delayed if the signs and symptoms of patients are nonspecific.

To assess the clinical value of deep vein thrombosis (DVT) density on computed tomography (CT) venography for predicting PTE.

From 2016 to 2021, patients with DVT diagnosed on lower-extremity CT venography were included. Of these patients, those without PTE were classified into ‘DVT-only group’ and those with PTE were classified into the ‘DVT with PTE group’. The DVT Hounsfield unit (HU) density was measured by drawing free-hand region-of-interests within the thrombus at the most proximal filling defect level. The risk factors associated with PTE were identified by using multivariate logistic regression analysis. A receiver operating characteristic (ROC) analysis was used to evaluate the value of DVT density for predicting the risk of PTE.

This study included 177 patients with a mean age of 41.7 ± 10.3 years (DVT-only group: 105 patients; DVT with PTE group: 72 patients). DVT density was significantly higher in DVT with the PTE group than DVT-only group (66.8HU ± 8.7 vs. 57.9HU ± 11.1, p < 0.001). The ROC analysis revealed that the area under the curve (AUC), sensitivity, and specificity for predicting the risk of PTE were 0.737, 72.2%, and 66.7%, respectively, at a DVT density cutoff of 63.0 HU. On univariate and multivariate analysis, DVT density was the only significant risk factor associated with PTE.

Higher DVT density was a significant risk factor for PTE. In addition, DVT density could be a predictive factor for PTE.

DWI and ADC-mapping was performed to analyze hepatic metastasis of GIST, GEP-NET.

The objective of this study is to present hepatic metastasis of GIST and GEP-NET with Diffusion weighted MR imaging(DWI) and the Apparent diffusion coefficients (ADC) values of masses.

18 GIST patients and 8 GEP-NET patients were examined retrospectively. 11 males and 6 females were present in GIST group, 7 males to 5 females were involved in GEP-NET group. 18 primary GIST and 10 hepatic metastasis of GIST, 8 original GEP-NET and 19 hepatic metastasis of GEP-NET; total 55 GIST and GEP-NET masses were analysed by ADC mapping. MR images were acquired by 1,5 T MR units (32 mT/min gradient strength- Achieva; Philips Healthcare, Best, Netherlands and 32 channel GE Signa GE-Wisconsin-USA); by using a 4-8 channel standard phased-array torso XL coil, all images were evaluated by an Abdominal MRI experienced radiologist. DWI was performed in the transverse plane by using spin-echo-planar imaging sequence.

No statistical differences were observed between GIST and GEP-NET patients according to age and gender variations. No significant statistical differences were observed according to the diameters and ADC values of GIST and GEP-NET patients. A significant statistical difference was observed between GIST and GEP-NET groups in terms of size of liver metastasis which was significantly higher in GIST patients. All three groups (GIST_Hep. MET, GEP-NET_Liver_Met and normal) were statistically differed according to ADC values. With the ROC curve analysis: Hepatic metastasis of GIST(n=10) and normal liver (n:47) had cut-off value for ADC: 0.925 under AUC: 0.939 with regard to ADC values and regarded 89.4% Sensitivity, 100% Specificity, 100% PPV and 66.7% PPV. ROC curve of GEP NET_ Hepatic metastasis (n=19) group and normal liver (n:47) group presented cut-off value for ADC: 0.860 under AUC: 0.967 correlated to ADC values with 93.6% sensitivity, 89.5% specificity, 95.7% PPV and 85% PPV.

High cellular tumors resulted from liver metastasis of GIST and GEP-NET’s, and a positive correlation was observed between ADC values and cellularity/differentiation ratios of metastatic masses.

Clear cell hepatocellular carcinoma (HCC) is a rare subtype of HCC. Histologically, clear cell HCC is characterized by the cytoplasmic accumulation of glycogen with a clear cell appearance, constituting > 80% of tumor cells. Radiologically, clear cell HCC demonstrates early enhancement and washout similar to conventional HCC. Occasionally, enhancing capsule and intratumoral fat are accompanied by clear cell HCC.

A 57-year-old male presented to our hospital with right upper quadrant abdominal pain. Ultrasonography, computed tomography, and magnetic resonance imaging revealed a large mass with a well-defined margin in the right hemiliver. The patient underwent a right hemihepatectomy, and the final histopathology revealed clear cell-type HCC.

Distinguishing clear cell types from other types of HCC solely based on radiological findings is challenging. If hepatic tumors exhibit encapsulated margins, enhancing rims, intratumoral fat, and arterial phase hyperenhancement/washout pattern despite their large size, consideration of clear cell subtypes in the differential diagnosis list will aid patient management, implying better prognosis than not-otherwise-specified HCC.

The liver plays a significant role in the digestive system, and disease in the liver initiates various other problems. The liver is severely affected due to alcohol use, and it initiates various chronic diseases, including Alcohol-Related Liver Disease (ARLD). Alcoholic/Nonalcoholic Fatty Liver Disease (AFLD/NFLD) is a severe medical emergency, and early screening and treatment are necessary to cure the patient. The untreated AFLD/NFLD will cause various problems, including fatigue, weight loss, and discomfort in the abdomen.

This study aims to present a detailed review and investigation of schemes considered in medical clinics to identify the AFLD/NFLD coupled problems and present the merit of the Transient Elastography (TE) combined with Fibroscan® practice to identify liver abnormality.

This research aims to study the clinical significance and the accuracy of TE-supported liver illness screening.

This work aims to collect the recent research works and clinical reports published from 2011 to 2021 from the chosen databases and provide a detailed review using the clinical information discussed in the selected articles.

The essential statistical investigation of the collected data is executed with Review Manager (RevMan^®) software, and the significance of the TE is confirmed using the articles supporting a 2x2 contingency table, and each case is evaluated using a p-score and the Region of Convergence (RoC) curve for 95% confidence intervals.

Scimitar syndrome is a rare developmental anomaly with an incidence of 2/100.000 births. Major components of this disease are partial anomalous pulmonary venous drainage, pulmonary hypoplasia, systemic arterialization of the right basal lung, and dextroposition of the heart. Horseshoe lung and accessory hemidiaphragm are two rarer components of this disease.

In this paper, horseshoe lung and accessory diaphragm associated with Scimitar syndrome have been reported in two cases.

In conclusion, being aware of rare manifestations of rare diseases is important to fully describe the pathologic spectrum of the disease. This will assist in better management and decision-making process.

Absent contractility is a kind of esophageal motility disorder. Patients are often diagnosed by endoscopic ultrasonography or high-resolution manometry (HRM). In this paper, we report two cases of absent esophageal contractility first discovered by transabdominal ultrasonography.

The main symptom of both cases was acid reflux, and the main diagnosis was absent esophageal contractility. They were all discovered by transabdominal ultrasonography. After the treatment of surgery or drugs, the symptoms relieved during follow-up, without any other discomforts.

Transabdominal ultrasound as a primary screening method for absent contractility is proposed in this paper, which is non-invasive, real-time, and fast. Ultrasound is expected to improve the diagnostic efficiency of peristalsis and patient experience in combination with the above invasive examinations. The use of sonography in the diagnosis of absent contractility has seldom been studied. Further investigation is required to determine whether this technique may better assess absent contractility.

The usage of artificial intelligence in medical image analysis has significantly surpassed that of earlier related technologies. This paper aimed to investigate the diagnostic accuracy of Artificial Intelligence (AI) based-deep learning models for breast cancer detection.

We used the PICO (Patient/Population/Problem, Intervention, Comparison, Outcome) scheme to formulate the research question and construct our search terms. Studies were systematically examined from the available literature using the constructed search terms from PubMed, and ScienceDirect according to the PRISMA guidelines. The quality of the included studies was assessed using the QUADAS-2 checklist. The characteristics of each included study such as the study design, population, index test, and reference standard, were extracted. The sensitivity, specificity, and AUC for each study were also reported.

In this systematic review, 14 studies were analyzed. Eight studies showed that AI was more accurate than radiologists in evaluating mammographic images, while one comprehensive study found AI to be less precise. Studies that reported sensitivity and specificity without radiologist intervention showed performance scores ranging from 16.0% to 89.71%. With radiologist intervention, sensitivity was between 62% to 86%. Only three studies reported a specificity of 73.5% to 79%. The AUC of the studies was between 0.79 and 0.95. Thirteen studies were retrospective, and one was prospective.

There's inadequate evidence on the effectiveness of AI-based deep learning for breast cancer screening in clinical settings. More research is needed, including studies evaluating accuracy, RCTs, and large-scale cohort studies. This systematic review found that AI-based deep learning improves radiologists' accuracy, especially for novice radiologists. Younger, tech-savvy clinicians may be more accepting of AI. Although it can't replace radiologists, the encouraging results suggest it will play a significant role in identifying breast cancer in the future.

PROSPERO - CRD42021278655

Intervertebral disc degeneration (IVD) is now the most prevalent disease in the world; thus, precise intervertebral disc segmentation is essential for the assessment and diagnosis of spinal diseases. Multi-modal magnetic resonance (MR) imaging is more multi-dimensional and thorough than unimodal imaging. However, manual segmentation of multi-modal MRI not only imposes a huge burden on physicians but also has a high error rate.

In this study, we propose a new method that can efficiently and accurately segment intervertebral discs from multi-modal MR spine images, providing a reproducible usage scheme for the diagnosis of spinal disorders.

We suggest a network structure called MLP-Res-Unet that reduces the amount of computational load and the number of parameters while maintaining performance. Our contribution is two-fold. First, a medical image segmentation network that fuses residual blocks and a multilayer perceptron (MLP) is proposed. Secondly, we design a new deep supervised method and pass the features extracted from the encoder to the decoder through the residual path to achieve a new full-scale residual connection.

We evaluate the network on the MICCAI-2018 IVD dataset and obtain Dice similarity coefficient equal to 94.77 (%) and Jaccard coefficient equal to 84.74 (%), while we reduce the amount of parameters by a factor of 3.9 and computation by a factor of 2.4 compared to the IVD-Net.

Experiments show that MLP-Res-Unet improves segmentation performance and creates a simpler model structure while reducing the number of parameters and computation.