Optimizing infant magnetic resonance imaging efficiency: Deep learning-assisted feed-and-wrap technique versus...

Infant MRI is limited by motion and the frequent reliance on general anesthesia (GA), which suppresses motion but increases procedural risk, resource use, and turnover time. A sedation-free alternative is the feed-and-wrap (FW) technique-natural sleep supported by swaddling and noise control-which avoids anesthesia but can be limited by motion and variable success. Deep-learning (DL)-based image reconstruction shortens acquisitions and improves image quality, potentially strengthening the FW pathway (DL-FW) by reducing scan time and repeat sequences.

Our study aimed to investigate whether, compared with GA, DL-FW reduces MRI turnover time in infants younger than 4 months.

In this single-center retrospective study, we included consecutive infants aged ≤4 months who underwent brain MRI during the study period and met predefined criteria. Infants underwent either DL-FW or GA according to institutional practice. The primary endpoint was the turnover time of the MRI room, defined as the interval from the starting point to the end of the procedure. Times are summarized as median (IQR) and mean±SD; groups were compared using a two-sided Mann-Whitney U test (α=0.05). The between-group shift was estimated with the Hodges-Lehmann (HL) estimator and 95% confidence interval (CI).

Forty-eight infants were analyzed (DL-FW n=22; GA n=26). Turnover time was shorter with DL-FW (23 min [21-27], mean±SD 26.8±11.3, range 14-52) versus GA (30 min [27-38], 32.6±9.8, 19-58), and the difference was significant (U=166; z= - 2.48; P=0.013). The HL estimator indicated that the GA turnover time exceeded that of DL-FW by 6 min (95% CI 2-11), corresponding to a 7-min reduction in group medians (~23% relative to GA). No data were missing.

In infants aged ≤4 months, a DL-FW pathway with an infant MRI stabilizer was associated with significantly shorter MRI room turnover time than GA was, supporting the use of DL-FW as an anesthesia-sparing approach that may improve workflow and safety.