Purpose: To assess the accuracy of answers provided by ChatGPT-3 when prompted with questions from the daily routine of radiologists and to evaluate the text response when ChatGPT-3 was prompted to ...provide references for a given answer. Methods: ChatGPT-3 (San Francisco, OpenAI) is an artificial intelligence chatbot based on a large language model (LLM) that has been designed to generate human-like text. A total of 88 questions were submitted to ChatGPT-3 using textual prompt. These 88 questions were equally dispersed across 8 subspecialty areas of radiology. The responses provided by ChatGPT-3 were assessed for correctness by cross-checking them with peer-reviewed, PubMed-listed references. In addition, the references provided by ChatGPT-3 were evaluated for authenticity. Results: A total of 59 of 88 responses (67%) to radiological questions were correct, while 29 responses (33%) had errors. Out of 343 references provided, only 124 references (36.2%) were available through internet search, while 219 references (63.8%) appeared to be generated by ChatGPT-3. When examining the 124 identified references, only 47 references (37.9%) were considered to provide enough background to correctly answer 24 questions (37.5%). Conclusion: In this pilot study, ChatGPT-3 provided correct responses to questions from the daily clinical routine of radiologists in only about two thirds, while the remainder of responses contained errors. The majority of provided references were not found and only a minority of the provided references contained the correct information to answer the question. Caution is advised when using ChatGPT-3 to retrieve radiological information.
Graphical Abstract
Objective
To establish a fully automated, robust imaging marker for cerebral small vessel disease (SVD) and related cognitive impairment that is easy to implement, reflects disease burden, and is ...strongly associated with processing speed, the predominantly affected cognitive domain in SVD.
Methods
We developed a novel magnetic resonance imaging marker based on diffusion tensor imaging, skeletonization of white matter tracts, and histogram analysis. The marker (peak width of skeletonized mean diffusivity PSMD) was assessed along with conventional SVD imaging markers. We first evaluated associations with processing speed in patients with genetically defined SVD (n = 113). Next, we validated our findings in independent samples of inherited SVD (n = 57), sporadic SVD (n = 444), and memory clinic patients with SVD (n = 105). The new marker was further applied to healthy controls (n = 241) and to patients with Alzheimer's disease (n = 153). We further conducted a longitudinal analysis and interscanner reproducibility study.
Results
PSMD was associated with processing speed in all study samples with SVD (p‐values between 2.8 × 10−3 and 1.8 × 10−10). PSMD explained most of the variance in processing speed (R2 ranging from 8.8% to 46%) and consistently outperformed conventional imaging markers (white matter hyperintensity volume, lacune volume, and brain volume) in multiple regression analyses. Increases in PSMD were linked to vascular but not to neurodegenerative disease. In longitudinal analysis, PSMD captured SVD progression better than other imaging markers.
Interpretation
PSMD is a new, fully automated, and robust imaging marker for SVD. PSMD can easily be applied to large samples and may be of great utility for both research studies and clinical use. Ann Neurol 2016;80:581–592.
Aim
To evaluate clinical and imaging features in patients with acute necrotizing encephalopathy of childhood (ANEC) to identify predictors of RANBP2 mutations, influenza association, and long‐term ...outcomes.
Method
A retrospective chart review in patients with ANEC (2012–2020) seen at a tertiary pediatric center was performed. Children were included if they had acute inflammatory lesions in the basal ganglia and pons. Variables included presenting features, imaging characteristics, RANBP2 gene testing, nasopharyngeal swab findings, therapies, and long‐term outcomes.
Results
Twenty patients were included (average age at presentation 3y 6mo, interquartile range 3y 7mo, SD 2y 8mo; 14 females, six males). Three of the 20 experienced recurrences; one of the 20 died. Ten patients were influenza positive. Seven patients were RANBP2 mutation positive. A higher likelihood of hemorrhage was observed in patients who were influenza positive compared to influenza negative (p=0.048). Patients with influenza had a higher degree of thalamic hemorrhage (2, p=0.035) and greater extent of diffusion restriction (3, p=0.035) in semiquantitive analysis.
Interpretation
Children with ANEC who are positive for influenza are more likely to have hemorrhage and greater thalamic swelling. RANBP2 status was predictive of relapse but not predictive of overall outcome.
Purpose
Artificial intelligence (AI) is playing an ever-increasing role in Neuroradiology.
Methods
When designing AI-based research in neuroradiology and appreciating the literature, it is important ...to understand the fundamental principles of AI. Training, validation, and test datasets must be defined and set apart as priorities. External validation and testing datasets are preferable, when feasible. The specific type of learning process (supervised vs. unsupervised) and the machine learning model also require definition. Deep learning (DL) is an AI-based approach that is modelled on the structure of neurons of the brain; convolutional neural networks (CNN) are a commonly used example in neuroradiology.
Results
Radiomics is a frequently used approach in which a multitude of imaging features are extracted from a region of interest and subsequently reduced and selected to convey diagnostic or prognostic information. Deep radiomics uses CNNs to directly extract features and obviate the need for predefined features.
Conclusion
Common limitations and pitfalls in AI-based research in neuroradiology are limited sample sizes (“small-n-large-p problem”), selection bias, as well as overfitting and underfitting.
Newborns with congenital heart disease undergoing cardiac surgery are at risk of neurodevelopmental impairment with limited understanding of the impact of intra‐operative cardiopulmonary bypass ...(CPB), deep hypothermia and selective cerebral perfusion on the brain. We hypothesized that a novel ultrasound technique, ultrafast power Doppler (UPD), can assess variations of cerebral blood volume (CBV) in neonates undergoing cardiac surgery requiring CPB. UPD was performed before, during and after surgery in newborns with hypoplastic left heart syndrome undergoing a Norwood operation. We found that global CBV was not significantly different between patients and controls (P = 0.98) and between pre‐ and post‐surgery (P = 0.62). UPD was able to monitor changes in CBV throughout surgery, revealing regional differences in CBV during hypothermia during which CBV correlated with CPB flow rate (R2 = 0.52, P = 0.021). Brain injury on post‐operative magnetic resonance imaging was observed in patients with higher maximum variation in CBV. Our findings suggest that UPD can quantify global and regional brain perfusion variation during neonatal cardiac surgery with this first intra‐operative application demonstrating an association between CBV and CPB flow rate, suggesting loss of autoregulation. Therefore, the measurement of CBV by UPD could enable optimization of cerebral perfusion during cardiac surgery in neonates.
Key points
The impact of cardiopulmonary bypass (CPB) on the neonatal brain undergoing cardiac surgery is poorly understood.
Ultrafast power Doppler (UPD) quantifies cerebral blood volume (CBV), a surrogate of brain perfusion.
CBV varies throughout CPB surgery and is associated with variation of the bypass pump flow rate during deep hypothermia.
Association between CBV and bypass pump flow rate suggests loss of cerebrovascular autoregulatory processes.
Quantitative monitoring of cerebral perfusion by UPD could provide a direct parameter to optimize CPB flow rate.
figure legend Cerebral blood volume (CBV) during neonatal cardiopulmonary bypass (CPB) assessed by ultrafast power Doppler (UPD). Using freehand transfontanellar UPD, CBV is shown to vary throughout and immediately following neonatal CPB surgery and is associated with the variation of the CPB flow rate during induced deep hypothermia. This association suggests loss of cerebrovascular autoregulatory mechanisms, which could be linked with acquired brain injury.
We investigated a patient who developed multiple sclerosis (MS) during treatment with the CTLA4‐blocking antibody ipilimumab for metastatic melanoma. Initially he showed subclinical magnetic ...resonance imaging (MRI) changes (radiologically isolated syndrome). Two courses of ipilimumab were each followed by a clinical episode of MS, 1 of which was accompanied by a massive increase of MRI activity. Brain biopsy confirmed active, T‐cell type MS. Quantitative next generation sequencing of T‐cell receptor genes revealed distinct oligoclonal CD4+ and CD8+ T‐cell repertoires in the primary melanoma and cerebrospinal fluid. Our results pinpoint the coinhibitory molecule CTLA4 as an immunological checkpoint and therapeutic target in MS. Ann Neurol 2016;80:294–300
Objective/Hypothesis
Sensorineural hearing loss (SNHL) is a common sequela of congenital cytomegalovirus (cCMV), potentially exacerbating neurocognitive delay. The objectives of this study were to ...assess: (1) age at which SNHL in children with cCMV; (2) stimulability of the auditory system in children with cCMV following cochlear implantation (CI); and (3) whether features of magnetic resonance imaging (MRI) potentially are predictive of hearing outcomes.
Methods
In this retrospective study of a prospectively acquired cohort, 123 children with cCMV who were referred for hearing loss at a single tertiary referral hospital over 20 years were compared with an unmatched comparative group of 90 children with GJB2‐related deafness. Outcome measures were results of newborn hearing screening (NHS), behavioral audiograms, and, in a subgroup of cochlear implant (CI) users, responses from the auditory nerve and brainstem evoked by CI at initial activation, as well as lesional volume of FLAIR‐hyperintense signal alterations on MRI.
Results
All but 3 of 123 children with cCMV had confirmed and persistent SNHL. At birth, 113 children with cCMV underwent NHS, 31 (27%) passed in both ears and 23 (20%) passed in one ear (no NHS data in 10 children). At the first audiologic assessment, 32 of 123 (26%) had normal hearing bilaterally; 35 of 123 (28%) had unilateral SNHL; and 57 of 123 (46%) had bilateral SNHL. More than half (67 of 123, 54%) experienced hearing deterioration in at least one ear. Survival analyses suggested that 60% of children who developed SNHL did so by 2.5 years and 80% by 5 years. In the children who passed NHS in one or both ears, 50% developed hearing loss by 3.5 years in the ear, which passed unilaterally (n = 23 ears), and 50% by 5 years in bilateral passes (n = 62 ears). Hearing loss was significant enough in all but one child with isolated high‐frequency loss for rehabilitation to be indicated. Hearing thresholds in individual ears were in the CI range in 83% (102 of 123), although duration of deafness was sufficient to preclude implantation at our center in 13 children with unilateral SNHL. Hearing aids were indicated in 16% (20 of 123). Responses from the auditory nerve and brainstem to initial CI stimulation were similar in children with cCMV‐related SNHL compared with GJB2‐related SNHL. Characteristic white matter changes on MRI were seen in all children with cCMV‐related SNHL (n = 91), but the lesion volume in each cortical hemisphere did not predict degree of SNHL.
Conclusions
cCMV‐related SNHL is often not detected by NHS but occurs with high prevalence in early childhood. Electrophysiological measures suggest equivalent stimulability of the auditory nerve and brainstem with CI in children with cCMV and GJB2‐related SNHL. Hyperintense white matter lesions on FLAIR MRI are consistently present in children with cCMV‐related SNHL but cannot be used to predict its time course or degree. Combined, the data show early and rapid deterioration of hearing in children with cCMV‐related SNHL with potential for good CI outcomes if SNHL is identified and managed without delay. Findings support universal newborn screening for cCMV followed by careful audiological monitoring.
Level of Evidence
3 Laryngoscope, 132:S1–S24, 2022
PDF
