EMERGE and ENGAGE (aducanumab phase 3 trial imaging)

EMERGE and ENGAGE are two phase 3 trials that Biogen ran in parallel with the same protocol to test aducanumab, an antibody against aggregated amyloid beta, in early Alzheimer disease. They enrolled at 348 sites in 20 countries from August 2015 to July 2018 and were stopped in March 2019 after a futility analysis. The results paper reports both trials side by side, and the ARIA analysis pools them into one safety data set, so this entry covers both. The trial MRI captures amyloid-related imaging abnormalities (ARIA) that emerge under treatment. The images are not available outside the sponsor.

Composition

3285 people received at least one dose, 1638 in EMERGE and 1647 in ENGAGE. Participants were 50 to 85 years old with confirmed amyloid pathology and either MCI due to Alzheimer disease (2661) or mild Alzheimer dementia. The mean age was 70.4 years and 1706 were women. Each trial randomized participants 1:1:1 to low-dose aducanumab, high-dose aducanumab or placebo, given every four weeks for 76 weeks.

Acquisition

Brain MRI was acquired at 1.5 T or 3 T at screening and then routinely at weeks 14, 22, 30, 42, 54, 66 and 78, with extra scans when clinically indicated and roughly every four weeks after an ARIA episode until it resolved or stabilised. ARIA-E was read on T2 FLAIR and microbleeds and superficial siderosis on gradient echo (T2*) sequences. 3253 participants had at least one post-baseline MRI. Amyloid PET with florbetapir was done in a substudy at sites able to perform it, and a tau PET substudy ran at selected sites.

Annotations

Central radiologists with ARIA expertise read every MRI and graded ARIA-E by the size of the affected regions and ARIA-H by the number of new microbleeds and areas of siderosis. In the 10 mg/kg group (1029 participants) ARIA-E occurred in 362 (35.2%), new microbleeds in 197 (19.1%) and superficial siderosis in 151 (14.7%); 29 of 1076 placebo participants (2.7%) had ARIA-E. Only these reads and adverse event records exist; no image-level labels are published.

Known limitations

• Biogen's Vivli policy says it does not generally share imaging data, so the scans are not available.
• People with more than four microbleeds or any superficial siderosis at baseline were excluded.
• Vendors and sequence parameters are not published in the sources used here.
• The trials were stopped early; 1812 participants (55.2%) completed the study.