Adapted large language models can outperform medical experts in clinical text summarization.

Analyzing vast textual data and summarizing key information from electronic health records imposes a substantial burden on how clinicians allocate their time. Although large language models (LLMs) have shown promise in natural language processing (NLP) tasks, their effectiveness on a diverse range of clinical summarization tasks remains unproven. Here we applied adaptation methods to eight LLMs, spanning four distinct clinical summarization tasks: radiology reports, patient questions, progress notes and doctor-patient dialogue. Quantitative assessments with syntactic, semantic and conceptual NLP metrics reveal trade-offs between models and adaptation methods. A clinical reader study with 10 physicians evaluated summary completeness, correctness and conciseness; in most cases, summaries from our best-adapted LLMs were deemed either equivalent (45%) or superior (36%) compared with summaries from medical experts. The ensuing safety analysis highlights challenges faced by both LLMs and medical experts, as we connect errors to potential medical harm and categorize types of fabricated information. Our research provides evidence of LLMs outperforming medical experts in clinical text summarization across multiple tasks. This suggests that integrating LLMs into clinical workflows could alleviate documentation burden, allowing clinicians to focus more on patient care.

Clinical Text Summarization: Adapting Large Language Models Can Outperform Human Experts.

Sifting through vast textual data and summarizing key information from electronic health records (EHR) imposes a substantial burden on how clinicians allocate their time. Although large language models (LLMs) have shown immense promise in natural language processing (NLP) tasks, their efficacy on a diverse range of clinical summarization tasks has not yet been rigorously demonstrated. In this work, we apply domain adaptation methods to eight LLMs, spanning six datasets and four distinct clinical summarization tasks: radiology reports, patient questions, progress notes, and doctor-patient dialogue. Our thorough quantitative assessment reveals trade-offs between models and adaptation methods in addition to instances where recent advances in LLMs may not improve results. Further, in a clinical reader study with ten physicians, we show that summaries from our best-adapted LLMs are preferable to human summaries in terms of completeness and correctness. Our ensuing qualitative analysis highlights challenges faced by both LLMs and human experts. Lastly, we correlate traditional quantitative NLP metrics with reader study scores to enhance our understanding of how these metrics align with physician preferences. Our research marks the first evidence of LLMs outperforming human experts in clinical text summarization across multiple tasks. This implies that integrating LLMs into clinical workflows could alleviate documentation burden, empowering clinicians to focus more on personalized patient care and the inherently human aspects of medicine.

Targeted treatment of severe vascular malformations harboring PIK3CA and TEK mutations with alpelisib is highly effective with limited toxicity.

This was a prospective cohort study of eighteen patients with large and debilitating vascular malformations with one or more major systemic complications. In all patients, we discovered activating alterations in either TEK or PIK3CA. Based on these findings, targeted treatment using the PI3K inhibitor alpelisib was started with regular check-ups, therapy duration varied from 6 to 31 months. In all patients, marked improvement in quality of life was observed. We observed radiological improvement in fourteen patients (two of them being on combination with either propranolol or sirolimus), stable disease in 2 patients. For 2 patients, an MRI scan was not available as they were shortly on treatment, however, a clinically visible response in size reduction or structure regression, together with pain relief was observed. In patients with elevated D-dimer levels before alpelisib administration, a major improvement was noted, suggesting its biomarker role. We observed overall very good tolerance of the treatment, documenting a single patient with grade 3 hyperglycemia. Patients with size reduction were offered local therapies wherever possible. Our report presents a promising approach for the treatment of VMs harboring different targetable TEK and PIK3CA gene mutations with a low toxicity profile and high efficacy.

Cerebrovascular Complications of COVID-19 Disease in Children: A Single-Center Case Series.

This work presents a case series of four children diagnosed with severe cerebrovascular disease in association with recent severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection, yet no patient from the group met typical diagnostic criteria for multisystem inflammatory syndrome in children. Our aim was to highlight the possible vascular involvement and coagulopathies associated with SARS-CoV-2 infection in the pediatric population. Further data are needed to better understand the pathophysiological basis of this condition in children and to ensure its optimal management.

A Newly Observed Mutation of the ABCA3 Gene Causing Lethal Respiratory Failure of a Full-Term Newborn: A Case Report.

Respiratory distress syndrome caused by a secondary surfactant deficiency is one of the most common diagnoses requiring admission to the Neonatal Intensive Care Unit. We illustrate the case of a term female newborn without prenatal and peripartal risks. There had been significant signs of respiratory distress 4 h after delivery. The condition gradually worsened to the point of needing oscillatory ventilation. The most common infectious and non-infectious causes were excluded. Considering the course of illness, a congenital surfactant deficiency was suspected. There nevertheless was no significant improvement after administration of surfactant. Following a short period of palliative care, the child died at 34 days of age due to respiratory failure. DNA diagnostics revealed compound heterozygosity of ABCA3 functional mutations leading to the p.Pro147Leu and p.Pro246Leu exchanges. The second identified mutation of ABCA3 c.737C>T had not to date been described in connection with primary surfactant deficiency.

[From Xrays to PET/MR, and then? - Future imaging in pediatric radiology]. / Vom Röntgen zum PET/MRT, und dann? ­ Zukunftsweisende Bildgebung in der Kinderradiologie.

Significant changes can be expected in modern pediatric radiology. New imaging techniques are progressively added to basic modalities like Xrays and ultrasound. This essay summarizes recent advances and technical innovations in pediatric radiology, which are supposed to gain further importance in the future. Thus, CT dose reduction techniques including artificial intelligence as well as advances in the fields of magnetic resonance and molecular imaging are presented. KEY POINTS: · Technical innovations will lead to significant changes in pediatric radiology.. · CT dose reduction is crucial for pediatric patient collectives.. · New MR-techniques will lower the need for sedation and contrast media application.. · Functional MR-imaging might gain further importance in patients with chronic lung disease.. · Molecular imaging enables detection, characterization and quantification of molecular processes in tumors.. CITATION FORMAT: · Staatz G, Daldrup-Link HE, Herrmann J et al. From Xrays to PET/MR, and then? - Future imaging in pediatric radiology. Fortschr Röntgenstr 2019; 191: 357 - 366.

Contrast Media Administration in Coronary Computed Tomography Angiography - A Systematic Review.

Background Various different injection parameters influence enhancement of the coronary arteries. There is no consensus in the literature regarding the optimal contrast media (CM) injection protocol. The aim of this study is to provide an update on the effect of different CM injection parameters on the coronary attenuation in coronary computed tomographic angiography (CCTA). Method Studies published between January 2001 and May 2014 identified by Pubmed, Embase and MEDLINE were evaluated. Using predefined inclusion criteria and a data extraction form, the content of each eligible study was assessed. Initially, 2551 potential studies were identified. After applying our criteria, 36 studies were found to be eligible. Studies were systematically assessed for quality based on the validated Quality Assessment of Diagnostic Accuracy Studies (QUADAS)-II checklist. Results Extracted data proved to be heterogeneous and often incomplete. The injection protocol and outcome of the included publications were very diverse and results are difficult to compare. Based on the extracted data, it remains unclear which of the injection parameters is the most important determinant for adequate attenuation. It is likely that one parameter which combines multiple parameters (e. g. IDR) will be the most suitable determinant of coronary attenuation in CCTA protocols. Conclusion Research should be directed towards determining the influence of different injection parameters and defining individualized optimal IDRs tailored to patient-related factors (ideally in large randomized trials). Key points · This systematic review provides insight into decisive factors on coronary attenuation.. · Different and contradicting outcomes are reported on coronary attenuation in CCTA.. · One parameter combining multiple parameters (IDR) is likely decisive in coronary attenuation.. · Research should aim at defining individualized optimal IDRs tailored to individual factors.. · Future directions should be tailored towards the influence of different injection parameters.. Citation Format · Mihl C, Maas M, Turek J et al. Contrast Media Administration in Coronary Computed Tomography Angiography - A Systematic Review. Fortschr Röntgenstr 2017; 189: 312 - 325.

Case report: rapid and durable response to PDGFR targeted therapy in a child with refractory multiple infantile myofibromatosis and a heterozygous germline mutation of the PDGFRB gene.

BACKGROUND: Infantile myofibromatosis belongs to a family of soft tissue tumors. The majority of these tumors have benign behavior but resistant and malignant courses are known, namely in tumors with visceral involvement. The standard of care is surgical resection. Observations suggest that low dose chemotherapy is beneficial. The treatment of resistant or relapsed patients with multifocal disease remains challenging. Patients that harbor an actionable mutation in the kinase domain are potential subjects for targeted tyrosine kinase inhibitor therapy. CASE PRESENTATION: An infant boy with inborn generalized infantile myofibromatosis that included bone, intracranial, soft tissue and visceral involvement was treated according to recent recommendations with low dose chemotherapy. The presence of a partial but temporary response led to a second line of treatment with six cycles of chemotherapy, which achieved a partial response again but was followed by severe toxicity. The generalized progression of the disease was observed later. Genetic analyses were performed and revealed a PDGFRB gene c.1681C>A missense heterozygous germline mutation, high PDGFRß phosphokinase activity within the tumor and the heterozygous germline Slavic Nijmegen breakage syndrome 657del5 mutation in the NBN gene. Targeted treatment with sunitinib, the PDGFRß inhibitor, plus low dose vinblastine led to an unexpected and durable response without toxicities or limitations to daily life activities. The presence of the Slavic NBN gene mutation limited standard chemotherapy dosing due to severe toxicities. Sister of the patient suffred from skull base tumor with same genotype and histology. The same targeted therapy led to similar quick and durable response. CONCLUSION: Progressive and resistant incurable infantile myofibromatosis can be successfully treated with the new approach described herein. Detailed insights into the biology of the patient's tumor and genome are necessary to understand the mechanisms of activity of less toxic and effective drugs except for up to date population-based chemotherapy regimens.

Automated Tube Voltage Selection for Radiation Dose Reduction in CT Angiography Using Different Contrast Media Concentrations and a Constant Iodine Delivery Rate.

OBJECTIVE: The purpose of this study was to systematically investigate radiation dose reduction using automated tube voltage selection during CT angiography (CTA) and to evaluate the impact of contrast medium (CM) injection protocols on dose reduction. MATERIALS AND METHODS: A circulation phantom containing the thoracic and abdominal vasculature was used. Four different concentrations of CM (iopromide 300 and 370 mg I/mL and iomeprol 350 and 400 mg I/mL) were administered while maintaining an identical iodine delivery rate (1.8 g I/s) and total iodine load (20.0 g). Three different scanning protocols for CTA of the thoracoabdominal aorta were used: protocol A, no dose modulation; protocol B, automated tube current modulation (CARE Dose4D); and protocol C, automated tube voltage selection (CARE kV). The dose-length product was recorded to calculate the effective dose. Attenuation values (in Hounsfield units), image noise levels, and signal-to-noise ratios (SNRs) in six predefined intravascular sites (three thoracic and three abdominal) were measured by two readers. All values were analyzed using the Kruskal-Wallis test and two-way ANOVA. RESULTS: There was a significant reduction in the effective dose (in millisieverts) for protocols B (mean ± SD, 2.03 ± 0.1 mSv) and C (1.00 ± 0.0 mSv) compared with protocol A (4.34 ± 0.0 mSv). The dose was reduced by 53% for protocol B and by 77% for protocol C. No significant differences were found in the effective dose among the different CM injection protocols within the scanning protocols; all p values were > 0.05. The attenuation values and SNRs were comparable among all the different CM injection protocols; all p values were > 0.05. CONCLUSION: A large radiation dose reduction (77%) can be achieved using automated tube voltage selection independent of the CM injection protocol.

Feasibility of low contrast media volume in CT angiography of the aorta.

OBJECTIVES: Using smaller volumes of contrast media (CM) in CT angiography (CTA) is desirable in terms of cost reduction and prevention of contrast-induced nephropathy (CIN). The purpose was to evaluate the feasibility of low CM volume in CTA of the aorta. METHODS: 77 patients referred for CTA of the aorta were scanned using a standard MDCT protocol at 100 kV. A bolus of 50 ml CM (Iopromide 300 mg Iodine/ml) at a flow rate of 6 ml/s was applied (Iodine delivery rate IDR = 1.8 g/s; Iodine load 15 g) followed by a saline bolus of 40 ml at the same flow rate. Scan delay was determined by the test bolus method. Subjective image quality was assessed and contrast enhancement was measured at 10 anatomical levels of the aorta. RESULTS: Diagnostic quality images were obtained for all patients, reaching a mean overall contrast enhancement of 324 ± 28 HU. Mean attenuation was 350 ± 60 HU at the thoracic aorta and 315 ± 83 HU at the abdominal aorta. CONCLUSIONS: A straightforward low volume CM protocol proved to be technically feasible and led to CTA examinations reaching diagnostic image quality of the aorta at 100 kV. Based on these findings, the use of a relatively small CM bolus can be incorporated into routine clinical imaging.

Influence of contrast media viscosity and temperature on injection pressure in computed tomographic angiography: a phantom study.

PURPOSE: Iodinated contrast media (CM) in computed tomographic angiography is characterized by its concentration and, consecutively, by its viscosity. Viscosity itself is directly influenced by temperature, which will furthermore affect injection pressure. Therefore, the purposes of this study were to systematically evaluate the viscosity of different CM at different temperatures and to assess their impact on injection pressure in a circulation phantom. MATERIALS AND METHODS: Initially, viscosity of different contrast media concentrations (240, 300, 370, and 400 mgI/mL) was measured at different temperatures (20°C-40°C) with a commercially available viscosimeter. In the next step, a circulation phantom with physical conditions was used. Contrast media were prepared at different temperatures (20°C, 30°C, 37°C) and injected through a standard 18-gauge needle. All other relevant parameters were kept constant (iodine delivery rate, 1.9 g I/s; total amount of iodine, 15 g I). Peak flow rate (in milliliter per second) and injection pressure (psi) were monitored. Differences in significance were tested using the Kruskal-Wallis test (Statistical Package for the Social Sciences). RESULTS: Viscosities for iodinated CM of 240, 300, 370, and 400 mg I/mL at 20°C were 5.1, 9.1, 21.2, and 28.8 mPa.s, respectively, whereas, at 40°C, these were substantially lower (2.8, 4.4, 8.7, and 11.2 mPa.s). In the circulation phantom, mean (SD) peak pressures for CM of 240 mg I/mL at 20°C, 30°C, and 37°C were 107 (1.5), 95 (0.6), and 92 (2.1) psi; for CM of 300 mg I/mL, 119 (1.5), 104 (0.6), and 100 (3.6) psi; for CM of 370 mg I/mL, 150 (0.6), 133 (4.4), and 120 (3.5) psi; and for CM of 400 mg I/mL, 169 (1.0), 140 (2.1), and 135 (2.9) psi, respectively, with all P values less than 0.05. CONCLUSIONS: Low concentration, low viscosity, and high temperatures of CM are beneficial in terms of injection pressure. This should also be considered for individually tailored contrast protocols in daily routine scanning.