Keeping Up With ChatGPT: Evaluating Its Recognition and Interpretation of Nuclear Medicine Images.

PURPOSE: The latest iteration of GPT4 (generative pretrained transformer) is a large multimodal model that can integrate both text and image input, but its performance with medical images has not been systematically evaluated. We studied whether ChatGPT with GPT-4V(ision) can recognize images from common nuclear medicine examinations and interpret them. PATIENTS AND METHODS: Fifteen representative images (scintigraphy, 11; PET, 4) were submitted to ChatGPT with GPT-4V(ision), both in its Default and "Advanced Data Analysis (beta)" version. ChatGPT was asked to name the type of examination and tracer, explain the findings and whether there are abnormalities. ChatGPT should also mark anatomical structures or pathological findings. The appropriateness of the responses was rated by 3 nuclear medicine physicians. RESULTS: The Default version identified the examination and the tracer correctly in the majority of the 15 cases (60% or 53%) and gave an "appropriate" description of the findings or abnormalities in 47% or 33% of cases, respectively. The Default version cannot manipulate images. "Advanced Data Analysis (beta)" failed in all tasks in >90% of cases. A "major" or "incompatible" inconsistency between 3 trials of the same prompt was observed in 73% (Default version) or 87% of cases ("Advanced Data Analysis (beta)" version). CONCLUSIONS: Although GPT-4V(ision) demonstrates preliminary capabilities in analyzing nuclear medicine images, it exhibits significant limitations, particularly in its reliability (ie, correctness, predictability, and consistency).

Performance of Deauville Criteria in [18F]FDG-PET/CT Diagnostics of Giant Cell Arteritis.

In this retrospective study, PET/CT data from 59 patients with suspected giant cell arteritis (GCA) were reviewed using the Deauville criteria to determine an optimal cut-off between PET positivity and negativity. Seventeen standardised vascular regions were analysed per patient by three investigators blinded to clinical information. Statistical analysis included ROC curves with areas under the curve (AUC), Cohen's and Fleiss' kappa (κ) to calculate sensitivity, specificity, accuracy, and agreement. According to final clinician's diagnosis and the revised 2017 ACR criteria GCA was confirmed in 29 of 59 (49.2 %) patients. With a diagnostic cut-off ≥ 4 (highest tracer uptake of a vessel wall exceeds liver uptake) for PET positivity, all investigators achieved high accuracy (range, 89.8-93.2%) and AUC (range, 0.94-0.97). Sensitivity and specificity ranged from 89.7-96.6% and 83.3-96.7%, respectively. Agreement between the three investigators suggested 'almost perfect agreement' (Fleiss' κ = 0.84) A Deauville score of ≥4 as threshold for PET positivity yielded excellent results with high accuracy and almost perfect inter-rater agreement, suggesting a standardized, reproducible, and reliable score in diagnosing GCA. However, the small sample size and reference standard could lead to biases. Therefore, verification in a multicentre study with a larger patient cohort and prospective setting is needed.

Comparative effectiveness and safety of anticoagulants for the treatment of heparin-induced thrombocytopenia.

BACKGROUND: The effectiveness and safety of non-heparin anticoagulants for the treatment of heparin-induced thrombocytopenia (HIT) are not fully established, and the optimal treatment strategy is unknown. In a systematic review and meta-analysis, we aimed to determine precise rates of platelet recovery, new or progressive thromboembolism (TE), major bleeding, and death for all non-heparin anticoagulants and to study potential sources of variability. METHODS: Following a detailed protocol (PROSPERO: CRD42020219027), EMBASE and Medline were searched for all studies reporting clinical outcomes of patients treated with non-heparin anticoagulants (argatroban, danaparoid, fondaparinux, direct oral anticoagulants [DOAC], bivalirudin, and other hirudins) for acute HIT. Proportions of patients with the outcomes of interest were pooled using a random-effects model for each drug. The influence of the patient population, the diagnostic test used, the study design, and the type of article was assessed. RESULTS: Out of 3194 articles screened, 92 studies with 119 treatment groups describing 4698 patients were included. The pooled rates of platelet recovery ranged from 74% (bivalirudin) to 99% (fondaparinux), TE from 1% (fondaparinux) to 7% (danaparoid), major bleeding from 1% (DOAC) to 14% (bivalirudin), and death from 7% (fondaparinux) to 19% (bivalirudin). Confidence intervals were mostly overlapping, and results were not influenced by patient population, diagnostic test used, study design, or type of article. DISCUSSION: Effectiveness and safety outcomes were similar among various anticoagulants, and significant factors affecting these outcomes were not identified. These findings support fondaparinux and DOACs as viable alternatives to conventional anticoagulants for treatment of acute HIT in clinical practice.

Utility of the Platelet Function Analyzer in Patients with Suspected Platelet Function Disorders: Diagnostic Accuracy Study.

Introduction The platelet function analyzer (PFA) is widely used as a screening tool for bleeding disorders in various settings. The diagnostic performance regarding platelet function disorders (PFDs), which are among the most common inherited bleeding disorders, is however still elusive. We aimed to assess the diagnostic value of PFA for PFD in clinical practice. Methods Comprehensive clinical and laboratory data of all consecutive patients referred to a specialized outpatient between January 2012 and March 2017 with a suspected bleeding disorder were prospectively recorded. The diagnostic work-up was performed according to a prespecified protocol following current guidelines and platelet function was tested using light transmission aggregometry as well as flow cytometry. Results Five hundred and fifty-five patients were included (median age 43.7 years; interquartile range [IQR] 29.3, 61.7; 66.9% female). Possible PFD was diagnosed in 64 patients (11.5%) and confirmed PFD in 54 patients (9.7%). In patients with confirmed PFD, median closure times were 107 seconds (ADP or adenosine diphosphate; IQR 89, 130) and 169 seconds (EPI; IQR 121, 211). In patients without bleeding disorders, PFA closure times were 96 seconds (ADP; IQR 83, 109) and 137 seconds (EPI; IQR 116, 158). The sensitivity was 19.5% in case of PFA ADP (95%CI 12.6, 30.0; specificity 86.4%; 95% CI 82.4, 89.8), and 44.3% in case of PFA EPI (95% CI 34.9, 53.9; specificity 75.6%; 95% CI 70.8, 79.9). Conclusion The diagnostic performance of PFA for PFD was moderate to poor. Our results do not support the utilization of PFA for screening of PFD in clinical practice.

Diagnostic utility of the ISTH bleeding assessment tool in patients with suspected platelet function disorders.

Essentials The utility of bleeding assessment tools regarding platelet function disorders is still elusive. We studied consecutive patients in a prospective cohort study in a tertiary hospital. Substantially higher scorings were observed in patients with platelet function disorders. Bleeding assessment tools might provide a useful screening tool. BACKGROUND: Bleeding assessment tools (BATs) have been widely implemented in the evaluation of patients with suspected bleeding disorders. However, diagnostic BAT utility regarding platelet function disorders is still elusive. AIM: We aimed to assess the diagnostic value of the International Society on Thrombosis and Haemostasis BAT (ISTH-BAT) for platelet function disorders in clinical practice. METHODS: The clinical characteristics and laboratory data of all consecutive patients with a suspected bleeding disorder referred between January 2012 and March 2017 to an outpatient unit of a university hospital were prospectively collected. The diagnostic evaluation was performed according to current recommendations following a prespecified protocol and platelet function was tested using light transmission aggregometry as well as flow cytometry. RESULTS: Five hundred and fifty-five patients were assessed; 66.9% were female, median age was 43.7 years (interquartile range [IQR] 29.3, 61.7). Confirmed platelet function disorder was diagnosed in 54 patients (9.7%), possible platelet function disorder in 64 patients (11.5%), and other disorders in 170 patients (30.6%). Median scoring of the ISTH-BAT was 2 in patients without a bleeding disorder (IQR 1, 3), 4 in patients with a possible platelet function disorder (2, 7), and 7 in patients with confirmed platelet function disorder (5, 9). Area under the receiver operating characteristic curve (the area under the curve [AUC]) was 0.75 (95% CI 0.70, 0.80). CONCLUSIONS: Presence of a platelet function disorder was associated with substantially higher BAT scorings compared to patients without. Our data suggest that the ISTH-BAT provides a useful screening tool for patients with suspected platelet function disorders.