Head trauma is one of the most frequent causes of emergency department admissions and requires rapid neuroradiological assessment via head CT imaging. In high-throughput clinical environments, reporting workflows may be delayed and may not always reflect the actual neurological urgency of the patient, potentially postponing the identification of critical findings. We propose an AI-assisted pre-reporting framework for head CT case prioritization, designed as a proof-of-concept aligned with standard neuroradiological practice. Using a retrospective dataset of 150 head CT scans paired with clinical records, the framework preprocesses DICOM volumes, converts volumetric studies into 2D representations compatible with MedGemma, and generates slice-wise preliminary reports that are aggregated at patient level. Validation is conducted under expert neuroradiological supervision, assessing both adherence to a predefined radiological template and the clinical priority inferred from the AI-generated outputs. The goal is to support early identification of high-risk cases and optimize radiological workflows without replacing clinical expertise.
AI-Assisted Pre-Reporting for Head CT Case Prioritization in High-Throughput Healthcare Scenarios / Galassi, S., Durante, D., Filice, F., Bartucci, S., De Rose, E., Pia Zupi, M., Angiulli, F., Calimeri, F., Fassetti, F.. - (2026). (17th ACM Conference on Bioinformatics, Computational Biology, and Health Informatics, ACM-BCB 2026 Rende; Italia ) [10.1145/3807503.3816796].
AI-Assisted Pre-Reporting for Head CT Case Prioritization in High-Throughput Healthcare Scenarios
Simone Bartucci
;
2026
Abstract
Head trauma is one of the most frequent causes of emergency department admissions and requires rapid neuroradiological assessment via head CT imaging. In high-throughput clinical environments, reporting workflows may be delayed and may not always reflect the actual neurological urgency of the patient, potentially postponing the identification of critical findings. We propose an AI-assisted pre-reporting framework for head CT case prioritization, designed as a proof-of-concept aligned with standard neuroradiological practice. Using a retrospective dataset of 150 head CT scans paired with clinical records, the framework preprocesses DICOM volumes, converts volumetric studies into 2D representations compatible with MedGemma, and generates slice-wise preliminary reports that are aggregated at patient level. Validation is conducted under expert neuroradiological supervision, assessing both adherence to a predefined radiological template and the clinical priority inferred from the AI-generated outputs. The goal is to support early identification of high-risk cases and optimize radiological workflows without replacing clinical expertise.| File | Dimensione | Formato | |
|---|---|---|---|
|
Galassi_AI-Assisted_2026.pdf
accesso aperto
Note: https://doi.org/10.1145/3807503.3816796
Tipologia:
Versione editoriale (versione pubblicata con il layout dell'editore)
Licenza:
Creative commons
Dimensione
995.96 kB
Formato
Adobe PDF
|
995.96 kB | Adobe PDF |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.


