AI Volumetric Assessment Detects Internal Regrowth in Vestibular Schwannomas After Radiosurgery

AI volumetric segmentation after vestibular schwannoma radiosurgery tracked internal regrowth with median Dice 0.85 vs expert reads; internal regrowth was just 4.7% of cases but accounted for 27% of radiographic failures (170 MRI scans, 85 patients).

• Retrospective single-center study (2001–2023) compared a lab-developed CNN segmentation tool with expert manual segmentation on 170 MRI scans from 85 patients with unilateral vestibular schwannoma after stereotactic radiosurgery.
• Internal regrowth was rare (4.7% of cases) but overrepresented among radiographic failures (27%); within this subset AI achieved median Dice 0.85 and concordance correlation coefficient 0.87, with median percent difference 9.29% (range 3.57–20.78%).
• AI classification of radiographic failure (≥20% volume increase) matched expert assessment in the internal regrowth subset; the authors describe the work as preliminary.

Automated summary

RadPigeon summaries are original and for information only. They are not clinical advice.