Vergleich von 2D- und 3D-Sequenzen für die MRCP

Using deep learning models in magnetic resonance cholangiopancreatography images to diagnose common bile duct stones

BACKGROUNDS AND AIMS

Magnetic resonance cholangiopancreatography (MRCP) plays a significant role in diagnosing common bile duct stones (CBDS). Currently, there are no studies to detect CBDS by using the deep learning (DL) model in MRCP. This study aimed to use the DL model You Only Look Once version 5 (YOLOv5) to diagnose CBDS in MRCP images and verify its validity compared to the accuracy of radiologists.

METHODS

By collecting the thick-slab MRCP images of patients diagnosed with CBDS, 4 submodels of YOLOv5 were used to train and validate the performance. Precision, recall rate, and mean average precision (mAP) were used to evaluate model performance. Analyze possible reasons that may affect detection accuracy by validating MRCP images in 63 CBDS patients and comparing them with radiologist detection accuracy. Calculate the correctness of YOLOv5 for detecting one CBDS and multiple CBDS separately.

RESULTS

The precision of YOLOv5l (0.970) was higher than that of YOLOv5x (0.909), YOLOv5m (0.874), and YOLOv5s (0.939). The mAP did not differ significantly between the 4 submodels, with the following results: YOLOv5l (0.942), YOLOv5x (0.947), YOLO5s (0.927), and YOLOv5m (0.946). However, in terms of training time, YOLOv5s was the fastest (4.8 h), detecting CBDS in only 7.2 milliseconds per image. In 63 patients the YOLOv5l model detected CBDS with an accuracy of 90.5% compared to 92.1% for radiologists, analyzing the difference between the positive group successfully identified and the unidentified negative group not. The incorporated variables include common bile duct diameter > 1 cm (p = .560), combined gallbladder stones (p = .706), maximum stone diameter (p = .057), combined cholangitis (p = .846), and combined pancreatitis (p = .656), and the number of CBDS (p = .415). When only one CBDS was present, the accuracy rate reached 94%. When multiple CBDSs were present, the recognition rate dropped to 70%.

CONCLUSION

YOLOv5l is the model with the best results and is almost as accurate as the radiologist's detection of CBDS and is also capable of detecting the number of CBDS. Although the accuracy of the test gradually decreases as the number of stones increases, it can still be useful for the clinician's initial diagnosis.

Efficacy of an artificial neural network algorithm based on thick-slab magnetic resonance cholangiopancreatography images for the automated diagnosis of common bile duct stones

BACKGROUND AND AIM

Magnetic resonance cholangiopancreatography (MRCP) can accurately diagnose common bile duct (CBD) stones but is laborious to interpret. We developed an artificial neural network (ANN) capable of automatically assisting physicians with the diagnosis of CBD stones. This study aimed to evaluate the ANN's diagnostic performance for detecting CBD stones in thick-slab MRCP images and identify clinical factors predictive of accurate diagnosis.

METHODS

The presence of CBD stones was confirmed via direct visualization through endoscopic retrograde cholangiopancreatography (ERCP). The absence of CBD stones was confirmed by either a negative endoscopic ultrasound accompanied by clinical improvements or negative findings on ERCP. Our base networks were constructed using state-of-the-art EfficientNet-B5 neural network models, which are widely used for image classification.

RESULTS

In total, 3156 images were collected from 789 patients. Of these, 2628 images from 657 patients were used for training. An additional 1924 images from 481 patients were prospectively collected for validation. Across the entire prospective validation cohort, the ANN achieved a sensitivity, specificity, positive predictive value, negative predictive value, and overall accuracy of 93.03%, 97.05%, 97.01%, 93.12%, and 95.01%, respectively. Similarly, a radiologist achieved a sensitivity, specificity, positive predictive value, negative predictive value, and overall accuracy 91.16%, 93.25%, 93.22%, 90.20%, and 91.68%, respectively. In multivariate analysis, only bile duct diameter > 10 mm (odds ratio = 2.45, 95% confidence interval [1.08-6.07], P = 0.040) was related to ANN diagnostic accuracy.

CONCLUSION

Our ANN algorithm automatically and quickly diagnoses CBD stones in thick-slab MRCP images, therein aiding physicians with optimizing clinical practice, such as whether to perform ERCP.

Comparison of diagnostic imaging modalities for the evaluation of pancreatic duct injury in children: a multi-institutional analysis from the Pancreatic Trauma Study Group

PURPOSE

Determining the integrity of the pancreatic duct is important in high-grade pancreatic trauma to guide decision making for operative vs non-operative management. Computed tomography (CT) is generally an inadequate study for this purpose, and magnetic resonance cholangiopancreatography (MRCP) is sometimes obtained to gain additional information regarding the duct. The purpose of this multi-institutional study was to directly compare the results from CT and MRCP for evaluating pancreatic duct disruption in children with these rare injuries.

METHODS

Retrospective study of data obtained from eleven pediatric trauma centers from 2010 to 2015. Children up to age 18 with suspected blunt pancreatic duct injury who had both CT and MRCP within 1 week of injury were included. Imaging findings of both studies were directly compared and analyzed using descriptive statistics, Chi square, Wilcoxon rank-sum, and McNemar's tests.

RESULTS

Data were collected for 21 patients (mean age 7.8 years). The duct was visualized more often on MRCP than CT (48 vs 5%, p < 0.05). Duct disruption was confirmed more often on MRCP than CT (24 vs 0%), suspected based on secondary findings equally (38 vs 38%), and more often indeterminate on CT (62 vs 38%). Overall, MRCP was not superior to CT for determining duct integrity (62 vs 38%, p = 0.28).

CONCLUSIONS

In children with blunt pancreatic injury, MRCP is more useful than CT for identifying the pancreatic duct but may not be superior for confirmation of duct integrity. Endoscopic retrograde cholangiogram (ERCP) may be necessary to confirm duct disruption when considering pancreatic resection.

LEVEL OF EVIDENCE

III.

[Tumor-like lesion of the pancreas in chronic pancreatitis : imaging characteristics of computed tomography]

OBJECTIVE

The aim of this study was to describe the imaging findings for inflammatory pseudotumors in patients with chronic pancreatitis as detected in computed tomography (CT).

MATERIALS AND METHODS

In this retrospective study 20 patients with chronic pancreatitis were included, who underwent an abdominal CT scan. In all patients the diagnosis was confirmed by surgery and histopathology. Imaging findings which have previously been described as typical for chronic pancreatitis and for inflammatory pseudotumors were assessed by two radiologists in consensus. Values of tissue density (HU) at CT were measured within the lesions and in the surrounding pancreatic tissue.

RESULTS

In 90% of patients with histologically proven chronic pancreatitis, CT showed corresponding indicative findings. In 10 patients the resected specimen revealed an inflammatory pseudotumor, which was located in all cases within the pancreatic head. Using CT these 10 patients presented with calcifications within the lesion in 50% of the cases, an irregular dilatation of the main pancreatic duct in 90%, a "double duct sign" in 70%, an interrupted main pancreatic duct in the area of the lesion in 50%, a "duct penetrating sign" in 30%, an infiltration of adjacent structures in 10% and pathologically enlarged lymph nodes in 100% of the cases. In the venous contrasting phase six tumors were hypodense and four isodense compared to the surrounding pancreatic tissue. In six patients biphasic CT was performed and the mean difference in attenuation between inflammatory pseudotumors and surrounding parenchyma was significantly higher in the pancreatic phase than in the venous phase.

DISCUSSION

Differentiation between inflammatory pseudotumors and adenocarcinoma remains difficult or even impossible. Typical signs indicative of an adenocarcinoma of the pancreatic head, such as dilatation of the common bile duct and/or the main pancreatic duct as well as enlarged lymph nodes, were also found in patients with inflammatory pseudotumors. Inflammatory pseudotumors showed low contrast between lesions and parenchyma in the venous phase and calcifications within the solid part of the tumor.

Technology Insight: advances in liver imaging

The role of diagnostic imaging in the assessment of liver disease continues to gain in importance. The classic techniques used for liver imaging are ultrasonography, CT and MRI. In the past decade, there have been significant advances in all three techniques. In this article, we discuss the advances in ultrasonography, CT and MRI that have improved assessment of focal and diffuse liver disease, including the development of hardware, software, processing algorithms and procedural innovations.

[Value of radiological methods in the diagnosis of biliary diseases]

The modern cross-sectional radiological methods of ultrasound, computed tomography, and magnetic resonance imaging provide manifold possibilities for the diagnosis of inflammatory and tumorous diseases of the gallbladder and the bile ducts as well as stone-related diseases. The invasive methods for direct imaging of the bile ducts, such as endoscopic retrograde cholangiopancreatography, are mainly used within therapeutic concepts. According to the literature, ultrasound and magnetic resonance imaging show a sensitivity of up to 100% in the diagnosis of intra- and extrahepatic cholestasis, but concerning the diagnosis of the cause of cholestasis these methods are limited. Therefore, additional MRI sequences or computed tomography are necessary. Computed tomography is particularly efficient for the diagnosis of the biliary system and adjacent anatomical and pathological structures within the pre- and postoperative period.