Coregistration-fusion MRI review identified missed radiologic disease activity in approximately 1 in 5 people with relapsing-remitting multiple sclerosis (RRMS) whose scans had originally been reported as stable using conventional side-by-side comparison, according to a retrospective cohort study published in the Multiple Sclerosis Journal.

Investigators evaluated 314 people with RRMS who had 3 consecutive brain MRI scans: a previous MRI, an index MRI originally reported as stable using conventional side-by-side comparison, and a subsequent follow-up MRI. The index MRI was performed during a period before semiautomated coregistration-fusion had been adopted into the institution’s routine clinical workflow.

Coregistration-fusion was applied to 3-dimensional fusion-attenuated inversion recovery (FLAIR) images to compare the index MRI with the previous MRI. The method automatically coregistered prior and current FLAIR images and generated color-coded overlays, with new or enlarging lesions highlighted for radiologist review. Investigators then assessed whether lesions missed on the index MRI were associated with future radiologic disease activity on the subsequent MRI and future clinical disease activity, defined as relapse and/or confirmed disability progression after the index scan.

Key results include the following:

  • The cohort included 314 people with RRMS, contributing 942 MRI studies.
  • Coregistration-fusion identified 183 new lesions that had been missed using conventional side-by-side comparison.
  • Overall, 65 of 314 individuals (20.7%) whose index MRIs were originally reported as stable were reclassified as having missed new lesions.
  • The median number of missed new lesions was 2 per person, and the median lesion size was 4 mm.
  • Most missed lesions were 5 mm or smaller, and missed lesions were most commonly located in the deep white matter, followed by the periventricular and subcortical regions.
  • Future radiologic disease activity occurred in 70.8% of individuals with missed new lesions compared with 16.1% of those confirmed as stable by coregistration-fusion (P<.001).
  • Future clinical disease activity occurred in 20.0% of individuals with missed new lesions compared with 10.0% of those confirmed as stable (P=.028).
  • In adjusted analyses, missed new lesions independently predicted future radiologic disease activity (adjusted odds ratio, 10.43; 95% CI, 5.31 to 20.46; P<.001) and future clinical disease activity (adjusted odds ratio, 2.49; 95% CI, 1.10 to 5.65; P=.029).
  • Each additional missed lesion at the index MRI was associated with a 20.5% increase in the expected number of new lesions on the subsequent MRI (incidence rate ratio, 1.205; 95% CI, 1.184 to 1.226; P<.001).

The authors suggested that coregistration-fusion may help identify subtle radiologic disease activity that can be missed with conventional side-by-side MRI comparison. Because new lesions on follow-up MRI can influence treatment reassessment and disease-modifying therapy optimization, the findings may be clinically relevant for patients whose scans are interpreted as stable but who may still have evidence of ongoing inflammatory activity.

The study’s findings should be interpreted with caution. The analysis was retrospective and conducted at a single academic tertiary referral multiple sclerosis center, which may limit generalizability. All MRI comparisons used the same MRI platform and 3D-FLAIR sequence, so performance across other scanners or acquisition protocols was not evaluated. The study was not designed as a formal diagnostic accuracy analysis of coregistration-fusion, and false-positive lesion detection was not systematically assessed. The number of individuals with missed new lesions was modest, and clinical events were relatively infrequent, limiting analyses of relapse and disability progression separately.

Source: Suthiphosuwan S, Lim TR, Raufdeen F, et al. Missed radiologic disease activity by conventional side-by-side MRI comparison is associated with future disease activity in multiple sclerosis: adopting semiautomated coregistration-fusion into clinical practice. Multiple Sclerosis Journal. 2026;32(9):938-949. doi:10.1177/13524585261453816