Does FLAIR super-resolution erase or hallucinate small white-matter lesions?

2026-08-06

AI Summary

This study systematically investigates the impact of super-resolution (SR) reconstruction on white matter hyperintensity (WMH) segmentation in clinical FLAIR imaging, where limited resolution due to thick slices poses a critical challenge. Focusing on the risk of erasing genuine small lesions or introducing hallucinated structures, the authors simulate FLAIR data at varying slice thicknesses and evaluate three SR approaches: multi-contrast implicit neural representation (INR), the self-supervised model ECLARE, and trilinear interpolation. Performance is assessed using four WMH segmentation algorithms, including MARS-WMH. Results demonstrate that while SR generally improves lesion detection, it tends to erase small true WMHs—a more significant concern than hallucination. ECLARE achieves the best performance at both 3 mm and 5 mm slice thicknesses, whereas INR offers no advantage over conventional interpolation.

Abstract

White matter hyperintensities (WMH), bright regions on Fluid-attenuated Inversion Recovery (FLAIR) scans are associated with cerebrovascular pathology and neurodegeneration. FLAIR is usually acquired with thick slices in clinical settings, giving it poor through-plane resolution. Super-resolution (SR) is a widely used method for recovering an isotropic volume from an anisotropic scan. Yet whether applying it prior to WMH segmentation preserves lesion content remains unknown: a model may erase small real lesions or hallucinate absent ones. We used 1-mm isotropic high-resolution (HR) FLAIR scans from 29 individuals in the ADNI cohort, each manually segmented for WMH by an expert. Then, we degraded each to simulated 3 and 5 mm through-plane acquisitions. Multi-contrast implicit neural representation (INR), a single-contrast self-supervised model (ECLARE), and cubic interpolation were used to upsample them onto the HR grid. WMH segmentation from a simulated thick slice and the original HR FLAIR set the floor and ceiling, respectively, for the per-lesion analysis. Of four WMH segmentation methods (WMH-SynthSeg, segcsvd, MARS-WMH, TrUE-Net), we ran the analysis under the most sensitive one to small lesions on HR (MARS-WMH) with the evaluation metrics of detection sensitivity, erasure rate (HR-detected lesions lost after reconstruction), and hallucination rate (predicted components absent from both the manual and HR segmentation). The dominant effect of SR was erasure of small real lesions, not hallucination, and it increased with slice thickness, though every reconstruction still improved lesion detection over the raw thick slice. ECLARE recovered small lesion signal best at both thicknesses, while the INR was no better than cubic interpolation.

Problem

• FLAIR
• super-resolution
• white matter hyperintensities
• lesion erasure
• hallucination

Innovation

• super-resolution
• white matter hyperintensities
• FLAIR MRI
• lesion erasure
• ECLARE