The virtual reference radiologist: comprehensive AI assistance for clinical image reading and interpretation

OBJECTIVES

Large language models (LLMs) have shown potential in radiology, but their ability to aid radiologists in interpreting imaging studies remains unexplored. We investigated the effects of a state-of-the-art LLM (GPT-4) on the radiologists' diagnostic workflow.

MATERIALS AND METHODS

In this retrospective study, six radiologists of different experience levels read 40 selected radiographic [n = 10], CT [n = 10], MRI [n = 10], and angiographic [n = 10] studies unassisted (session one) and assisted by GPT-4 (session two). Each imaging study was presented with demographic data, the chief complaint, and associated symptoms, and diagnoses were registered using an online survey tool. The impact of Artificial Intelligence (AI) on diagnostic accuracy, confidence, user experience, input prompts, and generated responses was assessed. False information was registered. Linear mixed-effect models were used to quantify the factors (fixed: experience, modality, AI assistance; random: radiologist) influencing diagnostic accuracy and confidence.

RESULTS

When assessing if the correct diagnosis was among the top-3 differential diagnoses, diagnostic accuracy improved slightly from 181/240 (75.4%, unassisted) to 188/240 (78.3%, AI-assisted). Similar improvements were found when only the top differential diagnosis was considered. AI assistance was used in 77.5% of the readings. Three hundred nine prompts were generated, primarily involving differential diagnoses (59.1%) and imaging features of specific conditions (27.5%). Diagnostic confidence was significantly higher when readings were AI-assisted (p > 0.001). Twenty-three responses (7.4%) were classified as hallucinations, while two (0.6%) were misinterpretations.

CONCLUSION

Integrating GPT-4 in the diagnostic process improved diagnostic accuracy slightly and diagnostic confidence significantly. Potentially harmful hallucinations and misinterpretations call for caution and highlight the need for further safeguarding measures.

CLINICAL RELEVANCE STATEMENT

Using GPT-4 as a virtual assistant when reading images made six radiologists of different experience levels feel more confident and provide more accurate diagnoses; yet, GPT-4 gave factually incorrect and potentially harmful information in 7.4% of its responses.

[Current MR imaging of cartilage in the context of knee osteoarthritis (part 2) : Cartilage pathologies and their assessment]

High-quality magnetic resonance (MR) imaging is essential for the precise assessment of the knee joint and plays a key role in the diagnostics, treatment and prognosis. Intact cartilage tissue is characterized by a smooth surface, uniform tissue thickness and an organized zonal structure, which are manifested as depth-dependent signal intensity variations. Cartilage pathologies are identifiable through alterations in signal intensity and morphology and should be communicated based on a precise terminology. Cartilage pathologies can show hyperintense and hypointense signal alterations. Cartilage defects are assessed based on their depth and should be described in terms of their location and extent. The following symptom constellations are of overarching clinical relevance in image reading and interpretation: symptom constellations associated with rapidly progressive forms of joint degeneration and unfavorable prognosis, accompanying symptom constellations mostly in connection with destabilizing meniscal lesions and subchondral insufficiency fractures (accelerated osteoarthritis) as well as symptoms beyond the "typical" degeneration, especially when a discrepancy is observed between (minor) structural changes and (major) synovitis and effusion (inflammatory arthropathy).

[Current MR imaging of cartilage in the context of knee osteoarthritis (part 1) : Principles and sequences]

Magnetic resonance imaging (MRI) is the clinical method of choice for cartilage imaging in the context of degenerative and nondegenerative joint diseases. The MRI-based definitions of osteoarthritis rely on the detection of osteophytes, cartilage pathologies, bone marrow edema and meniscal lesions but currently a scientific consensus is lacking. In the clinical routine proton density-weighted, fat-suppressed 2D turbo spin echo sequences with echo times of 30-40 ms are predominantly used, which are sufficiently sensitive and specific for the assessment of cartilage. The additionally acquired T1-weighted sequences are primarily used for evaluating other intra-articular and periarticular structures. Diagnostically relevant artifacts include magic angle and chemical shift artifacts, which can lead to artificial signal enhancement in cartilage or incorrect representations of the subchondral lamina and its thickness. Although scientifically validated, high-resolution 3D gradient echo sequences (for cartilage segmentation) and compositional MR sequences (for quantification of physical tissue parameters) are currently reserved for scientific research questions. The future integration of artificial intelligence techniques in areas such as image reconstruction (to reduce scan times while maintaining image quality), image analysis (for automated identification of cartilage defects), and image postprocessing (for automated segmentation of cartilage in terms of volume and thickness) will significantly improve the diagnostic workflow and advance the field further.

Pulmonary CT perfusion robustly measures cardiac output in the context of multilevel pulmonary occlusion: a porcine study

BACKGROUND

To validate pulmonary computed tomography (CT) perfusion in a porcine model by invasive monitoring of cardiac output (CO) using thermodilution method.

METHODS

Animals were studied at a single center, using a Swan-Ganz catheter for invasive CO monitoring as a reference. Fifteen pigs were included. Contrast-enhanced CT perfusion of the descending aorta and right and left pulmonary artery was performed. For variation purposes, a balloon catheter was inserted to block the contralateral pulmonary vascular bed; additionally, two increased CO settings were created by intravenous administration of catecholamines. Finally, stepwise capillary occlusion was performed by intrapulmonary arterial injection of 75-μm microspheres in four stages. A semiautomatic selection of AFs and a recirculation-aware tracer-kinetics model to extract the first-pass of AFs, estimating blood flow with the Stewart-Hamilton method, was implemented. Linear mixed models (LMM) were developed to calibrate blood flow calculations accounting with individual- and cohort-level effects.

RESULTS

Nine of 15 pigs had complete datasets. Strong correlations were observed between calibrated pulmonary (0.73, 95% confidence interval [CI] 0.6-0.82) and aortic blood flow measurements (0.82, 95% CI, 0.73-0.88) and the reference as well as agreements (± 2.24 L/min and ± 1.86 L/min, respectively) comparable to the state of the art, on a relatively wide range of right ventricle-CO measurements.

CONCLUSIONS

CT perfusion validly measures CO using LMMs at both individual and cohort levels, as demonstrated by referencing the invasive CO.

RELEVANCE STATEMENT

Possible clinical applications of CT perfusion for measuring CO could be in acute pulmonary thromboembolism or to assess right ventricular function to show impairment or mismatch to the left ventricle.

KEY POINTS

• CT perfusion measures flow in vessels. • CT perfusion measures cumulative cardiac output in the aorta and pulmonary vessels. • CT perfusion validly measures CO using LMMs at both individual and cohort levels, as demonstrated by using the invasive CO as a reference standard.

Outcome of transarterial radioembolization in patients with hepatocellular carcinoma as a first-line interventional therapy and after a previous transarterial chemoembolization

Purpose

Due to a lack of data, there is an ongoing debate regarding the optimal frontline interventional therapy for unresectable hepatocellular carcinoma (HCC). The aim of the study is to compare the results of transarterial radioembolization (TARE) as the first-line therapy and as a subsequent therapy following prior transarterial chemoembolization (TACE) in these patients.

Methods

A total of 83 patients were evaluated, with 38 patients having undergone at least one TACE session prior to TARE [27 male; mean age 67.2 years; 68.4% stage Barcelona clinic liver cancer (BCLC) B, 31.6% BCLC C]; 45 patients underwent primary TARE (33 male; mean age 69.9 years; 40% BCLC B, 58% BCLC C). Clinical [age, gender, BCLC stage, activity in gigabecquerel (GBq), Child-Pugh status, portal vein thrombosis, tumor volume] and procedural [overall survival (OS), local tumor control (LTC), and progression-free survival (PFS)] data were compared. A regression analysis was performed to evaluate OS, LTC, and PFS.

Results

No differences were found in OS (95% CI: 1.12, P = 0.289), LTC (95% CI: 0.003, P = 0.95), and PFS (95% CI: 0.4, P = 0.525). The regression analysis revealed a relationship between Child-Pugh score (P = 0.005), size of HCC lesions (>10 cm) (P = 0.022), and OS; neither prior TACE (Child-Pugh B patients; 95% CI: 0.120, P = 0.729) nor number of lesions (>10; 95% CI: 2.930, P = 0.087) correlated with OS.

Conclusion

Prior TACE does not affect the outcome of TARE in unresectable HCC.

Vitamin K1 and progression of cardiovascular calcifications in hemodialysis patients: The VitaVasK randomized controlled trial

Background

Cardiovascular calcifications are prevented by matrix Gla protein (MGP), a vitamin K-dependent protein. Hemodialysis patients exhibit marked vitamin K deficiency. The randomized, prospective, open-label, multicenter VitaVasK trial analyzed whether vitamin K1 supplementation reduces progression of coronary artery calcifications (CAC) and thoracic aortic calcifications (TAC).

Methods

Patients with preexisting CAC were randomized to continue on standard care or to additionally receive 5 mg vitamin K1 orally thrice weekly. Hierarchically ordered primary end points were progression of TAC and CAC in computed tomography scans at 18 months. Linear mixed effects models with repeated measures at baseline, 12 and 18 months assessed treatment effects after adjusting for study site.

Results

Of 60 randomized patients, 20 dropped out for reasons unrelated to vitamin K1, resulting in 23 control and 17 vitamin K1 patients. The trial was stopped early due to slow recruitment. At 18 months, the average TAC progression was 56% lower in the vitamin K1 compared to the control group (p = 0.039). CAC significantly progressed within the control group, but not within the vitamin K1 group. Average progression at 18 months was 68% lower in the vitamin K1 compared to the control group (p = 0.072). Vitamin K1 reduced plasma levels of pro-calcific uncarboxylated-MGP by 69% at 18 months. No treatment-related adverse events were noted.

Conclusion

Vitamin K1 intervention is a potent, safe and cost-effective approach to correct vitamin K deficiency and to potentially reduce cardiovascular calcification in this high-risk population.

Image-derived mean velocity measurement for prediction of coronary flow reserve in a canonical stenosis phantom using magnetic particle imaging

INTRODUCTION

Aim of this study is to evaluate whether magnetic particle imaging (MPI) is capable of measuring velocities occurring in the coronary arteries and to compute coronary flow reserve (CFR) in a canonical phantom as a preliminary study.

METHODS

For basic velocity measurements, a circulation phantom was designed containing replaceable glass tubes with three varying inner diameters, matching coronary-vessel diameters. Standardised boluses of superparamagnetic-iron-oxide-nanoparticles were injected and visualised by MPI. Two image-based techniques were competitively applied to calibrate the respective glass tube and to compute the mean velocity: full-duration-at-half-maximum (FDHM) and tracer dilution (TD) method. For CFR-calculation, four necessary settings of the circulation model of a virtual vessel with an inner diameter of 4 mm were generated using differently sized glass tubes and a stenosis model. The respective velocities in stenotic glass tubes were computed without recalibration.

RESULTS

On velocity level, comparison showed a good agreement (rFDHM = 0.869, rTD = 0.796) between techniques, preferably better for 4 mm and 6 mm inner diameter glass tubes. On CFR level MPI-derived CFR-prediction performed considerably inferior with a relative error of 20-44%.

CONCLUSIONS

MPI has the ability to reliably measure coronary blood velocities at rest as well as under hyperaemia and therefore may be suitable for CFR calculation. Calibration-associated accuracy of CFR-measurements has to be improved substantially in further studies.

Colloidal transport and agglomeration in column studies for advanced run-off filtration facilities--particle size and time resolved monitoring of effluents with flow-field-flow-fractionation

The efficiency of road run-off filtration facilities based on ion-exchange materials is reduced by pollutants which are transported bound to particles. To quantify the factors governing particle transport phenomena, a simplified model consisting of quartz sand-filled columns representing the filter/soil was set up. Suspensions of artificial clays, cold water-extracted natural clays, and real run-off were used as model effluents. Five experiments were performed: breakthrough of a natural soil suspension, remobilization of a natural soil suspension after ionic strength-drop, the same two experiments with a suspension of the artificial clay mineral Laponite, and the remobilization of run-off accumulated on a column at high ionic strength with an ionic strength down-gradient. Short-interval effluent fractions were analysed by flow-field-flow-fractionation (F4) to obtain the size distributions of the colloids present. The size distributions of subsequent fractions were then plotted in a staggered arrangement to give three-dimensional graphs that are time- and particle size-resolved. With this method the subsequent release of different agglomerate sizes formed on the column could be shown for the artificial clay mineral, questioning its use as a model colloid. The combined particle size- and time-resolved plots proved to be a powerful tool for monitoring colloidal solids in column effluents.

Comparison of growth and primary shunt product formation by Claviceps purpurea cultured on succinic acid and glucose as carbon sources

Growth on a medium containing succinic acid as the sole carbon source produced 1 g (dry weight) of mycelium per liter of medium by 50 days of incubation, whereas 25 g of mycelium was produced in 10 days when glucose was also present in the medium. Primary shunt metabolism took place during growth on succinic acid in spite of the extremely slow growth. Mycelia grown on succinic acid contained a higher percentage of residual mycelium and phosphate, but a lower percentage of mannitol, carbohydrate, lipid, and water-soluble nitrogen, than mycelia grown on a mixture of glucose and succinic acid. Thus, although primary shunt metabolism is favored by rapid growth on a rich, balanced sugar medium, it can also take place during extremely restricted growth in a medium containing succinic acid as the sole carbon source.