Artificial intelligence (AI) and machine learning (ML) are rapidly transforming clinical decision support systems (CDSSs) in intensive care units (ICUs), where vast amounts of real-time data present both an opportunity and a challenge for timely clinical decision-making. Here, we trace the evolution of machine intelligence in critical care. This technology has been applied across key ICU domains such as early warning systems, sepsis management, mechanical ventilation, and diagnostic support. We highlight a transition from rule-based systems to more sophisticated machine learning approaches, including emerging frontier models. While these tools demonstrate strong potential to improve predictive performance and workflow efficiency, their implementation remains constrained by concerns around transparency, workflow integration, bias, and regulatory challenges. Ensuring the safe, effective, and ethical use of AI in intensive care will depend on validated, human-centered systems supported by transdisciplinary collaboration, technological literacy, prospective evaluation, and continuous monitoring.

Intracranial multimodal monitoring (iMMM) is increasingly used in neurocritical care, but a lack of standardization hinders its evidence-based development. Here, we devised core outcome sets (COS) and reporting guidelines to harmonize iMMM practices and research. An open, decentralized, three-round Delphi consensus study involved experts between December 2023 and June 2024. Items—spanning three domains: (i) patient characteristics, (ii) practices, and (iii) outcomes—with ≥ 75% agreement were classified as strong agreement, while those with 50–75% were reconsidered in subsequent rounds, requiring ≥ 66% for moderate agreement. An international, multidisciplinary panel comprised 58 neurocritical physicians and researchers with low attrition (12%). They were predominantly from Western regions (96%), actively involved in iMMM (82%), at least weekly (72.4%), with more than 10 years of specific experience (57%). Of the 127 items assessed for inclusion in COS and reporting guidelines, 45 (35.4%) reached strong and 8 (6.3%) moderate agreement. Main strong agreement items were: (i) demographics: age (98%) and sex/gender (90%); comorbidities: coagulation/platelet disorders (95%); initial scoring: Glasgow Coma Scale (97%) and pathology-specific scores (90%); active treatments: antithrombotics (95%) (ii) clinical practice: iMMM implantation indications (98%) and iMMM-guided interventions (91%); surgical practice: targeting strategies (97%) and concomitant external ventricular drainage (97%); technical details: recording modalities (98%); (iii) monitoring parameters: duration (97%) and triggered interventions (95%); standardized outcome reporting (93%); surgical complications (e.g., postoperative intracranial hemorrhages, CNS infections, and probe misplacement, all > 90%) and adverse events (accidental dislodgement, probe breakage, and technical malfunctions, all > 90%). This consensus establishes foundational COS and reporting guidelines for iMMM in neurocritical care. These harmonization tools can enhance research quality, comparability, and reproducibility, facilitating evidence-based practices for this emerging technology. However, challenges remain in developing purpose-specific guidelines and adapting them to diverse clinical and research settings.

Background/Objectives: Artificial intelligence (AI), particularly large language models (LLMs), has demonstrated versatility in various applications but faces challenges in specialized domains like neurology. This study evaluates a specialized LLM’s capability and trustworthiness in complex neurological diagnosis, comparing its performance to neurologists in simulated clinical settings. Methods: We deployed GPT-4 Turbo (OpenAI, San Francisco, CA, US) through Neura (Sciense, New York, NY, US), an AI infrastructure with a dual-database architecture integrating “long-term memory” and “short-term memory” components on a curated neurological corpus. Five representative clinical scenarios were presented to 13 neurologists and the AI system. Participants formulated differential diagnoses based on initial presentations, followed by definitive diagnoses after receiving conclusive clinical information. Two senior academic neurologists blindly evaluated all responses, while an independent investigator assessed the verifiability of AI-generated information. Results: AI achieved a significantly higher normalized score (86.17%) compared to neurologists (55.11%, p < 0.001). For differential diagnosis questions, AI scored 85% versus 46.15% for neurologists, and for final diagnosis, 88.24% versus 70.93%. AI obtained 15 maximum scores in its 20 evaluations and responded in under 30 s compared to neurologists’ average of 9 min. All AI-provided references were classified as relevant with no hallucinatory content detected. Conclusions: A specialized LLM demonstrated superior diagnostic performance compared to practicing neurologists across complex clinical challenges. This indicates that appropriately harnessed LLMs with curated knowledge bases can achieve domain-specific relevance in complex clinical disciplines, suggesting potential for AI as a time-efficient asset in clinical practice.

Objective: To evaluate predictive diagnostic performance of open- and closed-source large language models (LLMs) in emergency medicine, addressing the urgent need for innovative clinical decision support tools amid rising patient volumes and staffing shortages. Materials and Methods: We generated 2370 AI-driven diagnostic predictions (Top-5 diagnoses from each of 6 model pipelines per patient), using data from 79 real-world emergency department cases collected consecutively during a 24-hour peak influx period at a tertiary care center. Pipelines combined open- and closed-source embedding models (text-embedding-ada-002, MXBAI) with foundational models (GPT-4, Llama3, and Qwen2) grounded via retrieval-augmented generation using emergency medicine textbooks. Models’ predictions were assessed against reference diagnoses established by expert consensus. Results: All pipelines achieved comparable diagnostic match rates (62.03%-72.15%). Diagnostic performance was significantly influenced by case characteristics: match rates were notably higher for specific versus unspecific diagnoses (85.53% vs 31.41%, P < .001) and surgical versus medical cases (79.49% vs 56.25%, P < .001). Open-source models demonstrated markedly superior sourcing capabilities compared to GPT-4-based combinations (P < 1.4e-12), with MBXAI/Qwen2 pipeline achieving perfect citation verification. Discussion: Diagnostic accuracy primarily depended on case characteristics rather than the choice of model pipeline, highlighting fundamental AI alignment challenges in clinical reasoning. Low performance in unspecific diagnoses underscores inherent complexities in clinical definitions rather than technological shortcomings alone. Conclusion: Open-source LLM pipelines provide enhanced sourcing capabilities, crucial for transparent clinical decision-making and interpretability. Further research should expand knowledge bases to include hospital guidelines and regional epidemiology, while exploring on-premises solutions to better align with privacy regulations and clinical integration.

Brain tissue oxygenation (PbtO2)-guided therapy can improve the neurological outcome of traumatic brain injury (TBI) patients. With several Phase-III ongoing studies, most of the existing evidence is based on before-after cohort studies and a phase-II randomized trial. The aim of this study was to assess the effectiveness of PbtO2-guided therapy in a single-center cohort. We performed a retrospective analysis of consecutive severe TBI patients admitted to our center who received either intracranial pressure (ICP) guided therapy (from January 2012 to February 2016) or ICP/PbtO2-guided therapy (February 2017 to December 2019). A genetic matching was performed based on covariates including demographics, comorbidities, and severity scores on admission. Intracranial hypertension (IH) was defined as ICP > 20 mmHg for at least 5 min. Brain hypoxia (BH) was defined as PbtO2 < 20 mmHg for at least 10 min. IH and BH were targeted by specific interventions. Mann–Whitney U and Fisher’s exact tests were used to assess differences between groups. A total of 35 patients were matched in both groups: significant differences in the occurrence of IH (ICP 85.7% vs. ICP/PbtO2 45.7%, p < 0.01), ICU length of stay [6 (3–13) vs. 16 (9–25) days, p < 0.01] and Glasgow Coma Scale at ICU discharge [10 (5–14) vs. 13 (11–15), p = 0.036] were found. No significant differences in ICU mortality and Glasgow Outcome Scales at 3 months were observed. This study suggests that the role of ICP/PbtO2-guided therapy should await further confirmation in well-conducted large phase III studies.