Smartphone imaging repository: a novel method for creating a CT image bank

Area of Science:

• Medical Imaging
• Clinical Trials
• Emergency Medicine

Background:

• Traditional methods for capturing and transmitting medical images in clinical trials, such as compact discs or manual file transfers, are time-consuming and labor-intensive.
• Existing electronic file transfer methods require significant staff time and dedicated equipment within radiology departments.

Purpose of the Study:

• To evaluate the feasibility and effectiveness of using smartphone camera video-derived images for capturing brain computed tomography (CT) scans in traumatic intracranial hemorrhage.
• To assess the time efficiency, image quality, and diagnostic accuracy of smartphone-based image acquisition compared to traditional Picture Archiving and Communication Systems (PACS).

Main Methods:

• Smartphone videos of deidentified brain CT scans were captured in the emergency department and emailed or uploaded for central review.
• Eight scans with various types of intracranial hemorrhages were used, with data acquired by ten users on seven different smartphones.
• Key metrics included time to capture, hematoma volume (ABC/2), Marshall scale, midline shift, and image quality (contrast-to-noise ratio, CNR).

Main Results:

• High intraclass correlation coefficients (ICCs) were observed for lesion volume (0.973), midline shift (0.998), and Marshall score (0.973), indicating excellent agreement.
• No significant difference in lesion conspicuity was found between smartphone-derived images and PACS images, as assessed by CNR.
• The mean time to capture and email video data was rapid, averaging 60.1 seconds.

Conclusions:

• Standard smartphones can produce sufficient video image quality for clinical trial imaging repositories.
• Smartphone video capture and transfer is significantly faster than traditional methods.
• Diagnostic measurements derived from smartphone-acquired images show no significant difference compared to those from PACS.