Diagnostic Value of Deep Learning-Based Rapid MRI in Preoperative Evaluation of Acute Cholecystitis

Fast Abdominopelvic MRI With Deep Learning-Based Reconstruction Algorithm: Image Quality Evaluation and Application in Patients With Common Abdominal and Pelvic Diseases

This is a single-center, prospective observational study at Peking Union Medical College Hospital. The main goal is to evaluate the image quality and clinical value of a new fast abdominopelvic MRI sequence that uses deep learning-based reconstruction.

The study will include adult patients (18 years or older) who need an MRI scan of the abdomen or pelvis as part of their routine care. These patients will receive both the standard MRI sequence and the new fast sequence, with the goal of comparing the two. The fast sequence is designed to be shorter (5-8 minutes) while maintaining or improving image quality, which can help reduce patient discomfort and motion-related artifacts.

Radiologists will assess the quality of the images from both sequences. The study will also look at how well the fast sequence can detect common abdominal and pelvic diseases, such as gallstones, inflammation, and tumors. This study does not involve new drugs, invasive procedures, or changes to standard care. No additional risks to patients are expected beyond those of a standard MRI scan.

Inclusion Criteria:

• Age ≥ 18 years
• Clinical indication for abdominopelvic MRI examination
• Able to understand and sign the informed consent form
• Able to lie supine and cooperate with breath-holding/respiratory commands during MRI

Exclusion Criteria:

• Presence of absolute contraindications to MRI (e.g., pacemaker, ferromagnetic foreign body, severe claustrophobia)
• Inability to complete the MRI scan due to severe pain, instability, or altered mental status
• Pregnant or breastfeeding women (unless explicitly approved by protocol)
• Previous abdominal/pelvic surgery with contraindications to MRI sequences