Article Text

Download PDFPDF

P-025 Large scale, CT evaluation can improve access to imaging studies within multi-center stroke trials
  1. B Jankowitz1,
  2. J Davies2,
  3. J Day3,
  4. S Chowdhry4,
  5. C Schirmer5,
  6. N Levkovitz6,
  7. O Bibas7,
  8. E Blanc7,
  9. G Pradilla8
  1. 1Neurosurgery, Cooper Neurological Institute, Camden, NJ
  2. 2Neurosurgery, UBNS, Buffalo, NY
  3. 3Neurosurgery, UAMS, LIttle Rock, AR
  4. 4Neurosurgery, NorthShore, glenview, IL
  5. 5Neurosurgery, Geisinger, Wilkes Barre, PA
  6.,, Israel, Israel
  7., Israel, Israel
  8. 8Neurosurgery, Emory, Atlanta, GA


Introduction Artificial intelligence (AI) can automate the detection and triage of Intracerebral Hemorrhage (ICH). Early Minimally invasive Removal of IntraCerebral Hemorrhage (ENRICH) is a RCT evaluating the efficacy of minimally invasive surgery (MIS) for ICH. AI ENRICH is a prospective trial operating within and in parallel to the ENRICH trial that utilizes an AI application, Viz RECRUIT ICH Volume, to identify and segment ICH to quickly identify potentially eligible subjects.

Methods Non-contrast CT scans performed at 5 participating US hospitals were evaluated specifically for a parenchymal hemorrhage by Viz RECRUIT ICH Volume. Participating health care professionals downloaded a phone application that allowed users to be notified for any hemorrhage ≥ 5 mL. Time metrics included the onset of CT scan, phone alert, and user recognition of that alert.

Results Over a combined period of 374 days, 24,137 CT scans were evaluated by the Viz ICH VOLUME application. Of these, 817 CT scans were determined to contain an ICH yielding a total of 154 patients that met ENRICH criteria for trial inclusion (30-80 mL). The median time from CT scan to cell phone notification was 2.6 minutes. The median time from cell phone notification to the user viewing the CT scan was 2.95 minutes.

Conclusions Viz ICH VOLUME can screen large numbers of CT scans and send alerts within minutes to medical professionals searching for clinical trial candidates in time-sensitive environments. Studies relying on radiographic selection criteria may benefit from automated screening.

Disclosures B. Jankowitz: 2; C; Stryker, Medtronic. J. Davies: None. J. Day: None. S. Chowdhry: None. C. Schirmer: None. N. Levkovitz: 5; C; O. Bibas: 5; C; E. Blanc: 5; C; G. Pradilla: None.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.