Accurate medical image segmentation is essential for diagnosis, surgical planning and many other applications. Convolutional Neural Networks (CNNs) have become the state-of-the-art automatic ...segmentation methods. However, fully automatic results may still need to be refined to become accurate and robust enough for clinical use. We propose a deep learning-based interactive segmentation method to improve the results obtained by an automatic CNN and to reduce user interactions during refinement for higher accuracy. We use one CNN to obtain an initial automatic segmentation, on which user interactions are added to indicate mis-segmentations. Another CNN takes as input the user interactions with the initial segmentation and gives a refined result. We propose to combine user interactions with CNNs through geodesic distance transforms, and propose a resolution-preserving network that gives a better dense prediction. In addition, we integrate user interactions as hard constraints into a back-propagatable Conditional Random Field. We validated the proposed framework in the context of 2D placenta segmentation from fetal MRI and 3D brain tumor segmentation from FLAIR images. Experimental results show our method achieves a large improvement from automatic CNNs, and obtains comparable and even higher accuracy with fewer user interventions and less time compared with traditional interactive methods.
High-resolution volume reconstruction from multiple motion-corrupted stacks of 2D slices plays an increasing role for fetal brain Magnetic Resonance Imaging (MRI) studies. Currently existing ...reconstruction methods are time-consuming and often require user interactions to localize and extract the brain from several stacks of 2D slices. We propose a fully automatic framework for fetal brain reconstruction that consists of four stages: 1) fetal brain localization based on a coarse segmentation by a Convolutional Neural Network (CNN), 2) fine segmentation by another CNN trained with a multi-scale loss function, 3) novel, single-parameter outlier-robust super-resolution reconstruction, and 4) fast and automatic high-resolution visualization in standard anatomical space suitable for pathological brains. We validated our framework with images from fetuses with normal brains and with variable degrees of ventriculomegaly associated with open spina bifida, a congenital malformation affecting also the brain. Experiments show that each step of our proposed pipeline outperforms state-of-the-art methods in both segmentation and reconstruction comparisons including expert-reader quality assessments. The reconstruction results of our proposed method compare favorably with those obtained by manual, labor-intensive brain segmentation, which unlocks the potential use of automatic fetal brain reconstruction studies in clinical practice.
Convolutional neural networks (CNNs) have achieved state-of-the-art performance for automatic medical image segmentation. However, they have not demonstrated sufficiently accurate and robust results ...for clinical use. In addition, they are limited by the lack of image-specific adaptation and the lack of generalizability to previously unseen object classes (a.k.a. zero-shot learning). To address these problems, we propose a novel deep learning-based interactive segmentation framework by incorporating CNNs into a bounding box and scribble-based segmentation pipeline. We propose image-specific fine tuning to make a CNN model adaptive to a specific test image, which can be either unsupervised (without additional user interactions) or supervised (with additional scribbles). We also propose a weighted loss function considering network and interaction-based uncertainty for the fine tuning. We applied this framework to two applications: 2-D segmentation of multiple organs from fetal magnetic resonance (MR) slices, where only two types of these organs were annotated for training and 3-D segmentation of brain tumor core (excluding edema) and whole brain tumor (including edema) from different MR sequences, where only the tumor core in one MR sequence was annotated for training. Experimental results show that: 1) our model is more robust to segment previously unseen objects than state-of-the-art CNNs; 2) image-specific fine tuning with the proposed weighted loss function significantly improves segmentation accuracy; and 3) our method leads to accurate results with fewer user interactions and less user time than traditional interactive segmentation methods.
Aims
The development of paediatric specific devices appears to lag behind advancements in our specialty. Children could therefore be limited in the number of procedures available to them unless we ...continue to use and modify adult devices ‘off-label’. This study quantifies the proportion of IR devices in which paediatric use is indicated by the manufacturer.
Materials & Methods
Cross-sectional analysis of device instructions for use (IFU), assessing inclusion of children was performed. Vascular access, biopsy, drainage, and enteral feeding devices, from 28 companies who sponsored BSIR, CIRSE and SIR (2019–2020) as determined by the meeting websites, were included. Devices for which the IFU was not available were excluded.
Results
190 (106 vascular access, 40 biopsy, 39 drainage and 5 feeding) devices with IFU’s from 18 medical device manufacturers were assessed. 49/190 (26%) IFU’s referenced children. 6/190 (3%) explicitly stated the device could be used in children and 1/190 (0.5%) explicitly stated the device was not for use in children. 55/190 (29%) implied they could be used in children through caution notes. The most common caution was a reference to the size of the device that could be accommodated in a child (26/190, 14%).
Conclusions
This data identifies an unmet need for paediatric IR devices and could be used to support the future development of devices intended for the children we treat. There is potentially a larger proportion (29%) of devices suitable for paediatric use, but which lack manufacturer explicit support.
Level of Evidence
Level 2c, Cross-Sectional study.
Background
Traditionally, testicular biopsy is performed using an open surgical approach. Ultrasound-guided percutaneous biopsy is a less invasive alternative and can be performed in children.
...Objective
The aim of this study is to report our technique and to assess the diagnostic accuracy and safety of ultrasound-guided percutaneous biopsy of testicular masses in children.
Materials and methods
This is a 16-year retrospective review of ultrasound-guided percutaneous testicular biopsies at a single pediatric hospital.
Results
We performed nine ultrasound-guided testicular biopsies in 9 patients (median age: 3 years, range: 4 months–11 years; median weight: 20.9 kg, range: 8.4–35 kg; median volume of testicular lesion biopsied: 4.4 mL, range: 1.2–17 mL). A percutaneous co-axial technique was used for 5/9 biopsies with absorbable gelatin sponge tract embolization performed in 4 of those biopsies. A non-co-axial technique was used in 4/9 biopsies. A median of three cores, range 2–6, were obtained. The diagnostic yield was 89% with one biopsy yielding material suggestive of, but insufficient for, a definitive diagnosis. The most common histological diagnosis was leukemic infiltration, occurring in 6/9 biopsies. Of the remaining three biopsies, one biopsy was suggestive of, but not confirmatory for, juvenile granulosa cell tumor and two biopsies confirmed normal testicular tissue; the long-term follow-up of which demonstrated normal growth and no lasting damage. There was one (clinically insignificant) complication out of nine biopsies (11%, 95% confidence interval 0–44%): a mild, self-resolving scrotal hematoma.
Conclusion
Ultrasound-guided testicular biopsy can be performed safely in children as an alternative to open surgical biopsy, with a high diagnostic yield and low complication rate.
Graphical abstract
Background
Gastrojejunal tubes are important feeding devices for children with gastro-esophageal reflux, allowing medication and feeding into the small bowel, and allowing gastric venting to prevent ...reflux. As with many medical devices, there are multiple manufacturers and designs, including balloon-retained tubes and disc-retained tubes.
Objective
This study evaluated the cost difference between these two types of gastrojejunal tube.
Materials and methods
We conducted a 3.5-year retrospective cost evaluation for all pediatric patients undergoing an insertion or change of gastrojejunal tube using a bottom-up micro-costing analysis. We calculated days between encounters and a subsequent cost per day for each patient.
Results
A total of 187 children and adolescents were included, with an average age of 9.2 years. They underwent a total of 1,240 encounters, an average of 6.6 encounters per patient during the study period. A total of 82% of these encounters were related to balloon-retained tubes and 18% to disc-retained tubes. The most common reason for an encounter was a routine change (57%), with mechanical complications accounting for 31%. Disc-retained tubes had a longer period between encounters (117.5 days) than balloon-retained tubes (95 days;
P
=0.038). However, disc-retained tubes cost 6.9 British pound sterling (GBP) per day, which was significantly higher than balloon-retained tubes at 5.2 GBP per day (
P
<0.0001).
Conclusion
Despite being more expensive to purchase, balloon-retained tubes were noted to be the least costly device in a cost-per-day analysis.
Purpose
Magnetic resonance (MR) cholangiopancreatography (MRCP) is an established specialist method for imaging the upper abdomen and biliary/pancreatic ducts. Due to limitations of either MR image ...contrast or low through‐plane resolution, patients may require further evaluation with contrast‐enhanced computed tomography (CT) images. However, CT fails to offer the high tissue‐ductal‐vessel contrast‐to‐noise ratio available on T2‐weighted MR imaging.
Methods
MR super‐resolution reconstruction (SRR) frameworks have the potential to provide high‐resolution visualizations from multiple low through‐plane resolution single‐shot T2‐weighted (SST2W) images as currently used during MRCP studies. Here, we (i) optimize the source image acquisition protocols by establishing the ideal number and orientation of SST2W series for MRCP SRR generation, (ii) optimize post‐processing protocols for two motion correction candidate frameworks for MRCP SRR, and (iii) perform an extensive validation of the overall potential of upper abdominal SRR, using four expert readers with subspeciality interest in hepato‐pancreatico‐biliary imaging.
Results
Obtained SRRs show demonstrable advantages over traditional SST2W MRCP data in terms of anatomical clarity and subjective radiologists’ preference scores for a range of anatomical regions that are especially critical for the management of cancer patients.
Conclusions
Our results underline the potential of using SRR alongside traditional MRCP data for improved clinical diagnosis.
Background
Endovenous laser ablation is well recognized as the first-line treatment for superficial venous reflux with varicose veins in adults. It is not widely reported and is not an established ...practice in pediatric patients.
Objective
To illustrate a variety of pediatric venous conditions in which endovenous laser ablation can be utilized and to demonstrate its feasibility and safety in children.
Materials and methods
We conducted a retrospective review of endovenous laser ablation procedures performed between January 2007 and July 2014 at two large pediatric institutions.
Results
We included 35 patients (17 males) who underwent endovenous laser ablation to 43 veins. Median age at first treatment was 14 years (range: 3–18 years). Median weight was 56 kg (range: 19–97 kg). Underlying diagnoses were common venous malformation (15), Klippel–Trenaunay syndrome (8), superficial venous reflux with varicose veins (5), verrucous hemangioma-related phlebectasia (4), venous varix (2) and arteriovenous fistula (1). The most common aim of treatment was to facilitate sclerotherapy. Thirty-four patients had treatment in the lower limbs and one patient in an upper limb. Ten of the veins treated with endovenous laser ablation had an additional procedure performed to close the vein. Complications attributable to endovenous laser ablation occurred in two patients (6%). One patient experienced post-procedural pain and one patient developed a temporary sensory nerve injury. Median clinical follow-up was 13 months (range: 28 days–5.7 years). The aim of the treatment was achieved in 29 of the 35 (83%) patients.
Conclusion
Endovenous laser ablation is technically feasible and safe in children. It can be used in the management of a range of pediatric venous diseases with good outcomes.
Angiomatoid fibrous histiocytoma (AFH) is a soft tissue neoplasm of intermediate biological potential. Typically a slow-growing tumor, it can recur locally. Rarely, it manifests as a soft tissue ...sarcoma capable of metastasis. When metastases are nonamenable to local therapy, it is believed uniformly fatal. We present 3 patients with metastatic AFH who demonstrated a sustained response to chemotherapy; including one who achieved complete remission with cryoablation. These cases reinforce the potential value of chemotherapy in some patients with unresectable metastatic AFH and provide the first case in the literature of cryoablation in AFH.