Trust and stakeholder perspectives on the implementation of AI tools in clinical radiology.

OBJECTIVES: To define requirements that condition trust in artificial intelligence (AI) as clinical decision support in radiology from the perspective of various stakeholders and to explore ways to fulfil these requirements. METHODS: Semi-structured interviews were conducted with twenty-five respondents-nineteen directly involved in the development, implementation, or use of AI applications in radiology and six working with AI in other areas of healthcare. We designed the questions to explore three themes: development and use of AI, professional decision-making, and management and organizational procedures connected to AI. The transcribed interviews were analysed in an iterative coding process from open coding to theoretically informed thematic coding. RESULTS: We identified four aspects of trust that relate to reliability, transparency, quality verification, and inter-organizational compatibility. These aspects fall under the categories of substantial and procedural requirements. CONCLUSIONS: Development of appropriate levels of trust in AI in healthcare is complex and encompasses multiple dimensions of requirements. Various stakeholders will have to be involved in developing AI solutions for healthcare and radiology to fulfil these requirements. CLINICAL RELEVANCE STATEMENT: For AI to achieve advances in radiology, it must be given the opportunity to support, rather than replace, human expertise. Support requires trust. Identification of aspects and conditions for trust allows developing AI implementation strategies that facilitate advancing the field. KEY POINTS: • Dimensions of procedural and substantial demands that need to be fulfilled to foster appropriate levels of trust in AI in healthcare are conditioned on aspects related to reliability, transparency, quality verification, and inter-organizational compatibility. •Creating the conditions for trust to emerge requires the involvement of various stakeholders, who will have to compensate the problem's inherent complexity by finding and promoting well-defined solutions.

Comparison of the patient-derived modified Japanese Orthopaedic Association scale and the European myelopathy score.

PURPOSE: To compare the patient-derived modified Japanese Orthopaedic Association (P-mJOA) scale with the European myelopathy score (EMS) for the assessment of patients with degenerative cervical myelopathy (DCM). METHODS: In this register-based cohort study with prospectively collected data, included patients were surgically treated for DCM and had reported both P-mJOA and EMS scores at baseline, 1-year follow-up, and/or 2-year follow-up to the Swedish Spine Register. P-mJOA and EMS scores were defined as severe (P-mJOA 0-11 and EMS 5-8), moderate (P-mJOA 12-14 and EMS 9-12), or mild (P-mJOA 15-18 and EMS 13-18). P-mJOA and EMS mean scores were compared, and agreement was evaluated with Spearman's rank correlation coefficient (ρ), the intraclass correlation coefficient (ICC), and kappa (κ) statistics. RESULTS: Included patients (n = 714, mean age 63.2 years, 42.2% female) completed 937 pairs of the P-mJOA and the EMS. The mean P-mJOA and EMS scores were 13.9 ± 3.0 and 14.5 ± 2.7, respectively (mean difference -0.61 [95% CI -0.72 to -0.51; p < 0.001]). Spearman's ρ was 0.84 (p < 0.001), and intra-rater agreement measured with ICC was 0.83 (p < 0.001). Agreement of severity level measured with unweighted and weighted κ was fair (κ = 0.22 [p < 0.001]; κ = 0.34 [p < 0.001], respectively). Severity levels were significantly higher using the P-mJOA (p < 0.001). CONCLUSION: The P-mJOA and the EMS had similar mean scores, and intra-rater agreement was high, whereas severity levels only demonstrated fair agreement. The EMS has a lower sensitivity for detecting severe myelopathy but shows an increasing agreement with the P-mJOA for milder disease severity. A larger interval to define severe myelopathy with the EMS is recommended.

MRI-based measurements of spondylolisthesis and kyphosis in degenerative cervical myelopathy.

BACKGROUND: To provide normative data and to determine accuracy and reliability of preoperative measurements of spondylolisthesis and kyphosis on supine static magnetic resonance imaging (MRI) of patients with degenerative cervical myelopathy. METHODS: T2-weighted midsagittal images of the cervical spine were in 100 cases reviewed twice by one junior observer, with an interval of 3 months, and once by a senior observer. The spondylolisthesis slip (SSlip, mm) and the modified K-line interval (mK-line INT, mm) were assessed for accuracy with the standard error of measurement (SEm) and the minimum detectable change (MDC). Intraobserver and interobserver reliability levels were determined using the intraclass correlation coefficient (ICC). RESULTS: The SEm was 0.5 mm (95% CI 0.4-0.6) for spondylolisthesis and 0.6 mm (95% CI 0.5-0.7) for kyphosis. The MDC, i.e., the smallest difference between two examinations that can be detected with statistical certainty, was 1.5 mm (95% CI 1.2-1.8) for spondylolisthesis and 1.6 mm (95% CI 1.3-1.8) for kyphosis. The highest reliability levels were seen between the second observation of the junior examiner and the senior observer (ICC = 0.80 [95% CI 0.70-0.87] and ICC = 0.96 [95% CI 0.94-0.98] for SSlip and mK-line INT, respectively). CONCLUSIONS: This study provides normative values of alignment measurements of spondylolisthesis and kyphosis in DCM patients. It further shows the importance of taking measurement errors into account when defining cut-off values for cervical deformity parameters and their potential clinical application in surgical decision-making.

Improvement rates, adverse events and predictors of clinical outcome following surgery for degenerative cervical myelopathy.

PURPOSE: To investigate improvement rates, adverse events and predictors of clinical outcome after laminectomy alone (LAM) or laminectomy with instrumented fusion (LAM + F) for degenerative cervical myelopathy (DCM). METHODS: This is a post hoc analysis of a previously published DCM cohort. Improvement rates for European myelopathy score (EMS) and Neck Disability Index (NDI) at 2- and 5-year follow-ups and adverse events are presented descriptively for available cases. Predictor endpoints were EMS and NDI scores at follow-ups, surgeon- and patient-reported complications, and reoperation-free interval. For predictors, univariate and multivariable models were fitted to imputed data. RESULTS: Mean age of patients (LAM n = 412; LAM + F n = 305) was 68 years, and 37.4% were women. LAM + F patients had more severe spondylolisthesis and less severe kyphosis at baseline, more surgeon-reported complications, more patient-reported complications, and more reoperations (p ≤ 0.05). After imputation, the overall EMS improvement rate was 43.8% at 2 years and 36.3% at 5 years. At follow-ups, worse EMS scores were independent predictors of worse EMS outcomes and older age and worse NDI scores were independent predictors of worse NDI outcomes. LAM + F was associated with more surgeon-reported complications (ratio 1.81; 95% CI 1.17-2.80; p = 0.008). More operated levels were associated with more patient-reported complications (ratio 1.12; 95% CI 1.02-1.22; p = 0.012) and a shorter reoperation-free interval (hazard ratio 1.30; 95% CI 1.08-1.58; p = 0.046). CONCLUSIONS: These findings suggest that surgical intervention at an earlier myelopathy stage might be beneficial and that less invasive procedures are preferable in this patient population.

Amino acid tracers in PET imaging of diffuse low-grade gliomas: a systematic review of preoperative applications.

Positron emission tomography (PET) imaging using amino acid tracers has in recent years become widely used in the diagnosis and prediction of disease course in diffuse low-grade gliomas (LGG). However, implications of preoperative PET for treatment and prognosis in this patient group have not been systematically studied. The aim of this systematic review was to evaluate the preoperative diagnostic and prognostic value of amino acid PET in suspected diffuse LGG. Medline, Cochrane Library, and Embase databases were systematically searched using keywords "PET," "low-grade glioma," and "amino acids tracers" with their respective synonyms. Out of 2137 eligible studies, 28 met the inclusion criteria. Increased amino acid uptake (lesion/brain) was consistently reported among included studies; in 25-92% of subsequently histopathology-verified LGG, in 83-100% of histopathology-verified HGG, and also in some non-neoplastic lesions. No consistent results were found in studies reporting hot spot areas on PET in MRI-suspected LGG. Thus, the diagnostic value of amino acid PET imaging in suspected LGG has proven difficult to interpret, showing clear overlap and inconsistencies among reported results. Similarly, the results regarding the prognostic value of PET in suspected LGG and the correlation between uptake ratios and the molecular tumor status of LGG were conflicting. This systematic review illustrates the difficulties with prognostic studies presenting data on group-level without adjustment for established clinical prognostic factors, leading to a loss of additional prognostic information. We conclude that the prognostic value of PET is limited to analysis of histological subgroups of LGG and is probably strongest when using kinetic analysis of dynamic FET uptake parameters.

Normal Brain and Brain Tumor ADC: Changes Resulting From Variation of Diffusion Time and/or Echo Time in Pulsed-Gradient Spin Echo Diffusion Imaging.

OBJECTIVES: Increasing gradient performance on modern magnetic resonance imaging scanners has profoundly reduced the attainable diffusion and echo times for clinically available pulsed-gradient spin echo (PGSE) sequences. This study investigated how this may impact the measured apparent diffusion coefficient (ADC), which is considered an important diagnostic marker for differentiation between normal and abnormal brain tissue and for therapeutic follow-up. MATERIALS AND METHODS: Diffusion time and echo time dependence of the ADC were evaluated on a high-performance 3 T magnetic resonance imaging scanner. Diffusion PGSE brain scans were performed in 10 healthy volunteers and in 10 brain tumor patients using diffusion times of 16, 40, and 70 ms, echo times of 60, 75, and 104 ms at 3 b-values (0, 100, and 1000 s/mm 2 ), and a maximum gradient amplitude of 68 mT/m. A low gradient performance system was also emulated by reducing the diffusion encoding gradient amplitude to 19 mT/m. In healthy subjects, the ADC was measured in 6 deep gray matter regions and in 6 white matter regions. In patients, the ADC was measured in the solid part of the tumor. RESULTS: With increasing diffusion time, a small but significant ADC increase of up to 2.5% was observed for 6 aggregate deep gray matter structures. With increasing echo time or reduced gradient performance, a small but significant ADC decrease of up to 2.6% was observed for 6 aggregate white matter structures. In tumors, diffusion time-related ADC changes were inconsistent without clear trend. For tumors with diffusivity above 1.0 µm 2 /ms, with prolonged echo time, there was a pronounced ADC increase of up to 12%. Meanwhile, for tumors with diffusivity at or below 1.0 µm 2 /ms, no change or a reduction was observed. Similar results were observed for gradient performance reduction, with an increase of up to 21%. The coefficient of variation determined in repeat experiments was 2.4%. CONCLUSIONS: For PGSE and the explored parameter range, normal tissue ADC changes seem negligible. Meanwhile, observed tumor ADC changes can be relevant if ADC is used as a quantitative biomarker and not merely assessed by visual inspection. This highlights the importance of reporting all pertinent timing parameters in ADC studies and of considering these effects when building scan protocols for use in multicenter investigations.

Circulating Brain Injury Biomarkers: A Novel Method for Quantification of the Impact on the Brain After Tumor Surgery.

BACKGROUND: Clinical methods to quantify brain injury related to neurosurgery are scarce. Circulating brain injury biomarkers have recently gained increased interest as new ultrasensitive measurement techniques have enabled quantification of brain injury through blood sampling. OBJECTIVE: To establish the time profile of the increase in the circulating brain injury biomarkers glial fibrillary acidic protein (GFAP), tau, and neurofilament light (NfL) after glioma surgery and to explore possible relationships between these biomarkers and outcome regarding volume of ischemic injury identified with postoperative MRI and new neurological deficits. METHODS: In this prospective study, 34 adult patients scheduled for glioma surgery were included. Plasma concentrations of brain injury biomarkers were measured the day before surgery, immediately after surgery, and on postoperative days 1, 3, 5, and 10. RESULTS: Circulating brain injury biomarkers displayed a postoperative increase in the levels of GFAP ( P < .001), tau ( P < .001), and NfL ( P < .001) on Day 1 and a later, even higher, peak of NFL at Day 10 ( P = .028). We found a correlation between the increased levels of GFAP, tau, and NfL on Day 1 after surgery and the volume of ischemic brain tissue on postoperative MRI. Patients with new neurological deficits after surgery had higher levels of GFAP and NfL on Day 1 compared with those without new neurological deficits. CONCLUSION: Measuring circulating brain injury biomarkers could be a useful method for quantification of the impact on the brain after tumor surgery or neurosurgery in general.

Candesartan improves survival following severe hypovolemia in pigs; a role for the angiotensin II type 2 receptor?

OBJECTIVE: To investigate the involvement of intestinal angiotensin II type 2 receptors in the outcome of acute severe hypovolemia as well as systemic and regional mesenteric hemodynamics and intestinal mucosal functions in anesthetized pigs. DESIGN AND SETTING: Prospective, interventional animal study in a university research laboratory. SUBJECTS: 53 landrace pigs, 28-35 kg. INTERVENTIONS: 30+30% or 20+20% hemorrhage of estimated total blood volume followed by retransfusion performed in untreated controls, in animals treated with the angiotensin II type 1 receptor blocker candesartan or with a combination of candesartan and the angiotensin II type 2 receptor blocker PD123319. MEASUREMENTS AND RESULTS: Following 30+30% hemorrhage the candesartan-treated animals attained a significantly higher survival rate than controls and animals treated with PD123319 in combination with candesartan. Less pronounced hemorrhage (20+20%) resulted in no mortality and functional variables were assessed. A significantly higher output of jejunal intraluminal nitric oxide occurred during hypovolemia in the candesartan treated group than in controls and animals that received PD123319 in combination with candesartan. Jejunal transmucosal potential difference was significantly better preserved after retransfusion in candesartan-treated animals than in controls. Expression of angiotensin II type 2 receptors in intestinal tissue was significantly higher in animals surviving the 30+30% hemorrhage than in nonsurvivors. CONCLUSIONS: Lethal circulatory failure is possibly influenced by use of angiotensin receptor ligands, and activation of intestinal angiotensin II type 2 receptors may play a significant role in improving the outcome of severe hypovolemia.

The angiotensin II receptor type 2 agonist CGP 42112A stimulates NO production in the porcine jejunal mucosa.

BACKGROUND: This study was conducted to elucidate if nitric oxide is released by the porcine jejunal mucosa upon selective stimulation of AT2 receptors and the possible involvement of iNOS, and to investigate the presence of jejunal AT1 and AT2 receptors. Young landrace pigs were anaesthetized with ketamine and alpha-chloralose. Jejunal luminal NO output was assessed by intraluminal tonometry and analysed by chemiluminescense. Western blot analysis quantified mucosal iNOS and detected AT1 and AT2 receptor protein expression. AT1 and AT2 receptor RNA expression was detected by rtPCR. RESULTS: Baseline luminal NO output correlated significantly to baseline mucosal iNOS-protein content. In animals treated with the AT2-receptor agonist CGP42112A (n = 11) luminal NO output increased significantly (at 0.1 micrograms kg(-1) min(-1) and 1.0 micrograms kg(-1) min(-1)), but not in animals simultaneously treated with the AT2-receptor antagonist PD123319 (bolus 0.3 mgkg-1, infusion 0.03 mg kg(-1) h(-1)) (n = 7). No differences in iNOS protein expression were found between groups or before/after the administration of drugs. Western blot and rtPCR recognised expression of the AT1 and AT2 receptors in jejunal tissue. CONCLUSION: The results suggest that activation of AT2 receptors increases jejunal luminal NO output. This response was not due to an increase in the expression of the iNOS protein in the mucosa.