Axial versus Axial-Plus-Coronal DWI for Brainstem Infarct Detection: A Blinded Multi-reader Multi-case Study

Acute brainstem infarcts are difficult to detect because of their small size, susceptibility artifacts, and low lesion conspicuity. Although coronal diffusion-weighted imaging (DWI) has been proposed to improve detection, its incremental value beyond standard axial DWI remains uncertain.

To determine whether adding coronal DWI improves diagnostic performance, reader confidence, and recognition of diagnostically challenging acute brainstem infarcts.

This retrospective counterbalanced multi-reader multi-case study included 100 brain MRIs (50 with and 50 without acute brainstem infarction) performed 2022-2026. Six readers (three fellowship-trained neuroradiologists and three neuroradiology fellows) independently interpreted examinations using either axial DWI alone or axial-plus-coronal DWI. Positive cases were confirmed by consensus review of imaging findings and clinical records. Primary outcomes were sensitivity, specificity, and accuracy. Secondary outcomes included reader confidence, identification of cases in which coronal DWI was considered particularly helpful (or desired if unavailable), and subgroup analyses by lesion characteristics and reader experience.

Accuracy was similar for axial-only and axial-plus-coronal DWI (93.3% [95% CI 89.9-95.6] vs 93.0% [89.5-95.4], p=.84), as were sensitivity (87.9% [81.7-92.2] vs 89.3% [CI 83.4-93.3], p=.65) and specificity (98.7% [95.2-99.6] vs 96.6%, [92.4-98.6], p=.24). Attendings demonstrated improved sensitivity with coronal DWI (83.8% [73.8-90.5] to 92.0% [83.6-96.3]);fellows showed no consistent benefit. For infarcts ≤5 mm, sensitivity was 73.0% with axial DWI and 82.5% with axial-plus-coronal DWI (OR, 2.27, 0.79-6.51; p=.13). Confidence was higher with coronal DWI (Δ=0.18, 0.07-0.29; p=0.002). Coronal DWI was flagged as helpful more frequently for positive than negative cases (39.3% vs. 14.4%, OR 3.85, 2.00-7.43), and axial-phase readers more often desired coronal DWI when their interpretation was incorrect (46.2% vs. 17.2%, OR 4.1, 1.03-16.5).

Adding coronal DWI did not significantly improve overall diagnostic performance for acute brainstem infarction in this multi-reader study but significantly increased reader confidence and aided recognition of diagnostically challenging examinations. Subgroup analyses suggested possible benefit for small infarcts.

Coronal DWI may be most valuable as a complementary sequence that reduces diagnostic uncertainty and supports interpretation of subtle brainstem infarcts rather than increasing overall detection accuracy.