Janus Uhd Nybing

Janus Uhd Nybing

Auto-generated profile, based on publicly available information

Janus Uhd Nybing

Department of Radiology, Bispebjerg and Frederiksberg Hospital, Copenhagen, Denmark

Research Focus & Metrics

Academic Career Length

11Years (2016-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 May 2023 in Radiology
Verified
Journal Article
Closed access
LPLouis L. Plesner
Louis L. Plesner
FMFelix C. Müller
Felix C. Müller
JNJanus D. Nybing
Janus D. Nybing
LLLene C. Laustrup
Lene C. Laustrup
FRFinn Rasmussen
Finn Rasmussen
MAMichael Brun Andersen
Michael Brun Andersen
Background Automated interpretation of normal chest radiographs could alleviate the workload of radiologists. However, the performance of such an artificial intelligence (AI) tool compared with clinical radiology reports has not been established. Purpose To perform an external evaluation of a commercially available AI tool for (a) the number of chest radiographs autonomously reported, (b) the sensitivity for AI detection of abnormal chest radiographs, and (c) the performance of AI compared with that of the clinical radiology reports. Materials and Methods In this retrospective study, consecutive posteroanterior chest radiographs from adult patients in four hospitals in the capital region of Denmark were obtained in January 2020, including images from emergency department patients, in-hospital patients, and outpatients. Three thoracic radiologists labeled chest radiographs in a reference standard based on chest radiograph findings into the following categories: critical, other remarkable, unremarkable, or normal (no abnormalities). AI classified chest radiographs as high confidence normal (normal) or not high confidence normal (abnormal). Results A total of 1529 patients were included for analysis (median age, 69 years [IQR, 55-69 years]; 776 women), with 1100 (72%) classified by the reference standard as having abnormal radiographs, 617 (40%) as having critical abnormal radiographs, and 429 (28%) as having normal radiographs. For comparison, clinical radiology reports were classified based on the text and insufficient reports excluded (n = 22). The sensitivity of AI was 99.1% (95% CI: 98.3, 99.6; 1090 of 1100 patients) for abnormal radiographs and 99.8% (95% CI: 99.1, 99.9; 616 of 617 patients) for critical radiographs. Corresponding sensitivities for radiologist reports were 72.3% (95% CI: 69.5, 74.9; 779 of 1078 patients) and 93.5% (95% CI: 91.2, 95.3; 558 of 597 patients), respectively. Specificity of AI, and hence the potential autonomous reporting rate, was 28.0% of all normal posteroanterior chest radiographs (95% CI: 23.8, 32.5; 120 of 429 patients), or 7.8% (120 of 1529 patients) of all posteroanterior chest radiographs. Conclusion Of all normal posteroanterior chest radiographs, 28% were autonomously reported by AI with a sensitivity for any abnormalities higher than 99%. This corresponded to 7.8% of the entire posteroanterior chest radiograph production. © RSNA, 2023 Supplemental material is available for this article. See also the editorial by Park in this issue.

1 July 2024 in Radiology
Verified
Journal Article
Closed access
MBMathias W. Brejnebøl
Mathias W. Brejnebøl
ALAnders Lenskjold
Anders Lenskjold
KZKatharina Ziegeler
Katharina Ziegeler
HRHuib Ruitenbeek
Huib Ruitenbeek
FMFelix C. Müller
Felix C. Müller
MBMikael Boesen
Mikael Boesen
In this retrospective, three-center, multireader study using consecutive knee radiographs in patients with suspected osteoarthritis, concurrent artificial intelligence assistance improved osteoarthritis grading performance of junior readers and increased interobserver agreement across all readers.


5 November 2024 in Scientific Reports
Verified
Journal Article
Open access
PDF file
ALAnders Lenskjold
Anders Lenskjold
MBMathias W Brejnebøl
Mathias W Brejnebøl
MRMartin H Rose
Martin H Rose
HGHenrik Gudbergsen
Henrik Gudbergsen
ACAkshay Chaudhari
Akshay Chaudhari
MBMikael Boesen
Mikael Boesen
Humans have been shown to have biases when reading medical images, raising questions about whether humans are uniform in their disease gradings. Artificial intelligence (AI) tools trained on human-labeled data may have inherent human non-uniformity. In this study, we used a radiographic knee osteoarthritis external validation dataset of 50 patients and a six-year retrospective consecutive clinical cohort of 8,273 patients. An FDA-approved and CE-marked AI tool was tested for potential non-uniformity in Kellgren-Lawrence grades between the right and left sides of the images. We flipped the images horizontally so that a left knee looked like a right knee and vice versa. According to human review, the AI tool showed non-uniformity with 20–22% disagreements on the external validation dataset and 13.6% on the cohort. However, we found no evidence of a significant difference in the accuracy compared to senior radiologists on the external validation dataset, or age bias or sex bias on the cohort. AI non-uniformity can boost the evaluated performance against humans, but image areas with inferior performance should be investigated.

17 June 2016 in JMIR Research Protocols
Verified
Journal Article
Open access
MHMorten Sejer Hansen
Morten Sejer Hansen
MAMohammad Sohail Asghar
Mohammad Sohail Asghar
JWJørn Wetterslev
Jørn Wetterslev
CPChristian Pipper
Christian Pipper
JMJohan Mårtensson
Johan Mårtensson
JDJørgen Berg Dahl
Jørgen Berg Dahl
Experience and development of pain may be influenced by a number of physiological, psychological, and psychosocial factors. In a previous study we found differences in neuronal activation to noxious stimulation, and microstructural neuroanatomical differences, when comparing healthy volunteers with differences in size of the area of secondary hyperalgesia following a standardized burn injury. We aim to investigate the degree of association between the volume of pain-relevant structures in the brain and the size of the area of secondary hyperalgesia following brief thermal sensitization. The study consists of one experimental day, in which whole-brain magnetic resonance imaging (MRI) scans will be conducted including T1-weighed three-dimensional anatomy scan, diffusion tensor imaging, and resting state functional MRI. Before the experimental day, all included participants will undergo experimental pain testing in a parallel study (Clinicaltrials.gov Identifier: NCT02527395). Results from this experimental pain testing, as well as the size of the area of secondary hyperalgesia from the included participants, will be extracted from this parallel study. The association between the volume of pain-relevant structures in the brain and the area of secondary hyperalgesia will be investigated by linear regression of the estimated best linear unbiased predictors on the individual volumes of the pain relevant brain structures. We plan to investigate the association between experimental pain testing parameters and the volume, connectivity, and resting state activity of pain-relevant structures in the brain. These results may improve our knowledge of the mechanisms responsible for the development of acute and chronic pain. Danish Research Ethics Committee (identifier: H-15010473). Danish Data Protection Agency (identifier: RH-2015-149). Clinicaltrials.gov NCT02567318; http://clinicaltrials.gov/ct2/show/NCT02567318 (Archived by WebCite at http://www.webcitation.org/6i4OtP0Oi).

11 September 2025 in Osteoarthritis and Cartilage
Verified
Journal Article
Open access
CNCamilla T. Nielsen
Camilla T. Nielsen
MBMikael Boesen
Mikael Boesen
HGHenrik Gudbergsen
Henrik Gudbergsen
PHPhilip Hansen
Philip Hansen
FMFelix Müller
Felix Müller
TTTom D. Turmezei
Tom D. Turmezei
Bone remodelling plays a role in osteoarthritis. Shape analysis using Computed Tomography (CT) enables 3-D evaluation of periarticular bone by combining statistical shape modelling (SSM) and cortical bone mapping (CBM). In a cross-sectional analysis, we explored the relationship between 3-D bone shape, cortical thickness (CTh) and trabecular attenuation (TA) in individuals with knee osteoarthritis. The distal femur and proximal tibia were segmented from 133 CT-scans with knee osteoarthritis. Canonical objects were registered to each individual object for SSM. Generalized estimating equations looked for associations of shape modes with Kellgren-Lawrence-Grade (KLG). Statistical parametric mapping (SPM) then established any dependence of CTh and TA on bone shape. Shape mode 2 (SM2) was significantly associated with KLG (OR 1.43 per SD, 95%CI: 1.28-1.59, p<0.05). Marginal articular bone showed significant CTh dependence on SM2 (increase of up to 0.2 mm per SD, p<0.05). TA was significantly dependent on SM2 across nearly all the femoral articular surface (decrease of 40 HU per SD, p<0.05) and in the posterolateral tibial plateau (decrease of 20 HU per SD, p<0.05). We uncovered a 3-D shape mode recognisable as osteoarthritic that was associated with radiographic disease status in patients with pre- to moderate knee OA and overweight. Increase in this shape mode was associated with greater cortical thickening at the joint margins, likely related to osteophytosis, and reduced subchondral trabecular attenuation, the implications of which are not yet understood. Further work on combining 3-D shape and bone parameter distributions could improve identification of structural disease progression.