-
Medical Image Analysis Jul 2024Supervised machine learning-based medical image computing applications necessitate expert label curation, while unlabelled image data might be relatively abundant....
Supervised machine learning-based medical image computing applications necessitate expert label curation, while unlabelled image data might be relatively abundant. Active learning methods aim to prioritise a subset of available image data for expert annotation, for label-efficient model training. We develop a controller neural network that measures priority of images in a sequence of batches, as in batch-mode active learning, for multi-class segmentation tasks. The controller is optimised by rewarding positive task-specific performance gain, within a Markov decision process (MDP) environment that also optimises the task predictor. In this work, the task predictor is a segmentation network. A meta-reinforcement learning algorithm is proposed with multiple MDPs, such that the pre-trained controller can be adapted to a new MDP that contains data from different institutes and/or requires segmentation of different organs or structures within the abdomen. We present experimental results using multiple CT datasets from more than one thousand patients, with segmentation tasks of nine different abdominal organs, to demonstrate the efficacy of the learnt prioritisation controller function and its cross-institute and cross-organ adaptability. We show that the proposed adaptable prioritisation metric yields converging segmentation accuracy for a new kidney segmentation task, unseen in training, using between approximately 40% to 60% of labels otherwise required with other heuristic or random prioritisation metrics. For clinical datasets of limited size, the proposed adaptable prioritisation offers a performance improvement of 22.6% and 10.2% in Dice score, for tasks of kidney and liver vessel segmentation, respectively, compared to random prioritisation and alternative active sampling strategies.
Topics: Humans; Algorithms; Tomography, X-Ray Computed; Neural Networks, Computer; Machine Learning; Markov Chains; Supervised Machine Learning; Radiography, Abdominal
PubMed: 38640779
DOI: 10.1016/j.media.2024.103181 -
Journal of the American Veterinary... Dec 2011To evaluate the sensitivity and specificity of abdominal computed radiography (CR) for the diagnosis of enterolithiasis in horses and to examine how these parameters are...
OBJECTIVE
To evaluate the sensitivity and specificity of abdominal computed radiography (CR) for the diagnosis of enterolithiasis in horses and to examine how these parameters are affected by the number and anatomic location of enteroliths and by gas distension of the gastrointestinal tract.
DESIGN
Retrospective case series.
ANIMALS
Horses ? 1 year old that underwent abdominal CR and subsequent exploratory laparotomy or postmortem examination.
PROCEDURES
3 reviewers blinded to signalment, history, clinical signs, and diagnoses separately evaluated abdominal computed radiographs of horses included in the study. Each set of radiographs was evaluated for the presence or absence of enteroliths, the amount of gas distention, and the image quality. Signalment, definitive diagnosis on the basis of findings on exploratory laparotomy or postmortem examination, and the number and location of enteroliths were obtained from medical records.
RESULTS
Of the 142 cases reviewed, 58.4% (83/142) had confirmed enterolithiasis. For the 3 reviewers, overall sensitivity was 85% and specificity was 93%. Sensitivity was lower for small colon enteroliths than for large colon enteroliths (50% and 94.5%, respectively) and was significantly affected by gas distention. Sensitivity was not significantly affected by the number of enteroliths.
CONCLUSIONS AND CLINICAL RELEVANCE
Computed radiography provided high sensitivity and high specificity for the diagnosis of enterolithiasis in horses. Caution should be exercised when the radiographic results are negative, as the sensitivity for small colon enterolithiasis was relatively low and gas distension negatively affected detection of enteroliths. Abdominal CR is indicated as a diagnostic test in horses examined for colic in geographic regions in which enterolithiasis is endemic.
Topics: Animals; Calculi; Diagnosis, Differential; Female; Horse Diseases; Horses; Intestinal Diseases; Male; Radiography, Abdominal; Retrospective Studies; Sensitivity and Specificity; Tomography, X-Ray Computed
PubMed: 22087725
DOI: 10.2460/javma.239.11.1483 -
Internal Medicine (Tokyo, Japan) 2014
Topics: Abdomen, Acute; Acute Kidney Injury; Diagnosis, Differential; Humans; Kidney; Male; Middle Aged; Radiography, Abdominal; Shock, Septic; Tomography, X-Ray Computed
PubMed: 25366022
DOI: 10.2169/internalmedicine.53.3008 -
IEEE Transactions on Medical Imaging Aug 2018Automatic segmentation of abdominal anatomy on computed tomography (CT) images can support diagnosis, treatment planning, and treatment delivery workflows. Segmentation...
Automatic segmentation of abdominal anatomy on computed tomography (CT) images can support diagnosis, treatment planning, and treatment delivery workflows. Segmentation methods using statistical models and multi-atlas label fusion (MALF) require inter-subject image registrations, which are challenging for abdominal images, but alternative methods without registration have not yet achieved higher accuracy for most abdominal organs. We present a registration-free deep-learning-based segmentation algorithm for eight organs that are relevant for navigation in endoscopic pancreatic and biliary procedures, including the pancreas, the gastrointestinal tract (esophagus, stomach, and duodenum) and surrounding organs (liver, spleen, left kidney, and gallbladder). We directly compared the segmentation accuracy of the proposed method to the existing deep learning and MALF methods in a cross-validation on a multi-centre data set with 90 subjects. The proposed method yielded significantly higher Dice scores for all organs and lower mean absolute distances for most organs, including Dice scores of 0.78 versus 0.71, 0.74, and 0.74 for the pancreas, 0.90 versus 0.85, 0.87, and 0.83 for the stomach, and 0.76 versus 0.68, 0.69, and 0.66 for the esophagus. We conclude that the deep-learning-based segmentation represents a registration-free method for multi-organ abdominal CT segmentation whose accuracy can surpass current methods, potentially supporting image-guided navigation in gastrointestinal endoscopy procedures.
Topics: Algorithms; Digestive System; Humans; Kidney; Radiographic Image Interpretation, Computer-Assisted; Radiography, Abdominal; Spleen; Tomography, X-Ray Computed
PubMed: 29994628
DOI: 10.1109/TMI.2018.2806309 -
Tanisal Ve Girisimsel Radyoloji : Tibbi... Mar 2004A wide spectrum of congenital anomalies may cause obstruction in the upper and lower gastrointestinal tract. Neonates with complete upper intestinal obstruction do not... (Review)
Review
A wide spectrum of congenital anomalies may cause obstruction in the upper and lower gastrointestinal tract. Neonates with complete upper intestinal obstruction do not usually require further radiological evaluation after radiography. Barium studies are sometimes needed. Barium studies and other comprehensive methods such as ultrasonography, computed tomography and magnetic resonance imaging are usually complementary procedures which are not usually helpful and may even delay surgery, resulting in some complications and death. The decision to perform a given imaging examination should be considered carefully to avoid unnecessary radiation exposure to the patient. The diagnosis of low intestinal obstruction is usually apparent at abdominal radiography because of the presence of many dilated loops. The differentiation between ileal and colonic obstruction can be made with a contrast enema study. Dilute ionic, water-soluble contrast agents and non-balloon tip catheter of appropriate size is preferred for neonatal contrast enemas. Barium sulphate suspensions typically should not be used because of their potential to exacerbate the impaction of meconium plugs in meconium ileus, whereas water-soluble enemas can be therapeutic.
Topics: Barium Sulfate; Child; Humans; Intestinal Obstruction; Magnetic Resonance Imaging; Radiography, Abdominal; Tomography, X-Ray Computed; Ultrasonography
PubMed: 15054709
DOI: No ID Found -
Journal of Medical Radiation Sciences Mar 2022The incidence of obesity has been steadily rising over the last few decades and is having a significant impact upon the health system. In radiography, a particular...
INTRODUCTION
The incidence of obesity has been steadily rising over the last few decades and is having a significant impact upon the health system. In radiography, a particular challenge of imaging obese patients is implementing the as low as reasonably achievable (ALARA) principle when determining radiation dose, and technical and patient-care adaptations. This study aimed to better understand the decision-making strategies of experienced radiographers in determining imaging and exposure factor selection in the context of imaging obese patients.
METHODS
This study employed a 'think-aloud,' methodology, and eight experienced diagnostic radiographers working in clinical education were recruited to perform routine AP abdominal X-ray projections on an anthropomorphic phantom. They were asked to simultaneously verbalise emerging thoughts as they considered positioning, exposure selection and image evaluation. This process was repeated with three different phantom sizes, each representing an increased BMI from 'healthy,' to, 'morbidly obese.' Audio recordings were transcribed and interpreted via Bowman's (1997) theory of radiographic judgement and decision-making.
RESULTS
Analysis of interview transcripts identified 12 key concepts considered by experienced radiographers. Differences in radiographic concepts were considered when imaging phantoms of different sizes was demonstrated. A shift from segmental (e.g. positioning) to more environmental factors (e.g. patient comfort) and an increase in the number of verbal considerations with increasing phantom size were identified. The shift in focus of decision-making stages identified the greater need to consider contextual factors such as patient comfort and repeatability when imaging obese patients.
CONCLUSION
Experienced radiographers find imaging obese patients challenging and alter their perception of image quality to accommodate for patient presentation. The findings will help inform future research, practice guidelines and learning resources to provide optimal imaging and care for obese patients, especially for student education.
Topics: Humans; Obesity, Morbid; Phantoms, Imaging; Radiation Dosage; Radiography, Abdominal
PubMed: 34496140
DOI: 10.1002/jmrs.543 -
Journal of Radiation Research Sep 2013The capacity of 4DCT to quantify organ motion is beyond conventional 3DCT capability. Local control could be improved. However we are unaware of any reports of organ...
The capacity of 4DCT to quantify organ motion is beyond conventional 3DCT capability. Local control could be improved. However we are unaware of any reports of organ dose measurements for helical 4DCT imaging. We therefore quantified the radiation doses for helical 4DCT imaging. Organ and tissue dose was measured for thoracic and abdominal 4DCT in helical mode using an adult anthropomorphic phantom. Radiation doses were measured with thermoluminescence dosimeter chips inserted at various anatomical sites on the phantom. For the helical thoracic 4DCT, organ doses were 57.2 mGy for the lung, 76.7 mGy for the thyroids, 48.1 mGy for the breasts, and 10.86 mGy for the colon. The effective doses for male and female phantoms were very similar, with a mean value of 33.1 mSv. For abdominal 4DCT imaging, organ doses were 14.4 mGy for the lung, 0.78 mGy for the thyroids, 9.83 mGy for breasts, and 58.2 mGy for the colon (all obtained by using ICRP 103). We quantified the radiation exposure for thoracic and abdominal helical 4DCT. The doses for helical 4DCT were approximately 1.5 times higher than those for cine 4DCT, however the stepwise image artifact was reduced. 4DCT imaging should be performed with care in order to minimize radiation exposure, but the advantages of 4DCT imaging mandates its incorporation into routine treatment protocols.
Topics: Abdominal Neoplasms; Humans; Imaging, Three-Dimensional; Phantoms, Imaging; Radiation Dosage; Radiography, Abdominal; Radiography, Thoracic; Radiotherapy, Image-Guided; Relative Biological Effectiveness; Reproducibility of Results; Respiratory-Gated Imaging Techniques; Sensitivity and Specificity; Thoracic Neoplasms; Tomography, X-Ray Computed; Viscera
PubMed: 23603303
DOI: 10.1093/jrr/rrt024 -
Canadian Association of Radiologists... Feb 2014
Review
Topics: Abdomen; Biopsy; Contrast Media; Humans; Image Enhancement; Magnetic Resonance Imaging; Multiple Myeloma; Radiography, Abdominal; Tomography, X-Ray Computed; Ultrasonography
PubMed: 22402107
DOI: 10.1016/j.carj.2011.12.010 -
BMC Pediatrics Mar 2023To examine the use of abdominal ultrasound (AUS) as a diagnostic adjunct in the diagnosis of necrotizing enterocolitis (NEC) in cases where abdominal radiography (AXR)...
BACKGROUND
To examine the use of abdominal ultrasound (AUS) as a diagnostic adjunct in the diagnosis of necrotizing enterocolitis (NEC) in cases where abdominal radiography (AXR) is equivocal in order to reduce unnecessary antibiotic use in neonates.
METHODS
Retrospective study (2017-2019) of infants undergoing NEC evaluation with equivocal AXR findings (n = 54). Paired AXR and AUS were reviewed with respect to presence or absence of pneumatosis. Concordance of AUS findings with decision to treat for NEC was evaluated.
RESULTS
Among 54 infants where AXR was equivocal, AUS demonstrated presence of pneumatosis in 22 patients (41%), absence of pneumatosis in 31 patients (57%), and was equivocal in 1 patient. All patients with pneumatosis on AUS were treated for NEC. Of 31 patients without pneumatosis on AUS, 25 patients (78%) were not treated for NEC. Patients without pneumatosis on AUS received a significantly shorter mean duration of antibiotics compared to those with pneumatosis (3.3 days (+/- 4.8 days) vs 12.4 days (+/- 4.7 days)); p < 0.001). Of those patients not treated, none required treatment within 1 week following negative AUS.
CONCLUSION
AUS is a valuable tool for evaluating the presence or absence of pneumatosis in the setting of equivocal AXR. Absence of pneumatosis on AUS informs clinical decision making and reduces unnecessary treatment and antibiotic usage.
Topics: Infant; Infant, Newborn; Humans; Enterocolitis, Necrotizing; Retrospective Studies; Radiography, Abdominal; Infant, Newborn, Diseases; Ultrasonography
PubMed: 36959616
DOI: 10.1186/s12887-023-03932-3 -
Abdominal Radiology (New York) Nov 2022
Topics: Education, Medical, Graduate; Humans; Radiography, Abdominal; Radiology; Surveys and Questionnaires
PubMed: 36216991
DOI: 10.1007/s00261-022-03696-x