Diagnostic Value of Pupil Involvement and Yield of Vascular Imaging in Isolated Third Cranial Nerve Palsy: A 10-Year Retrospective Study.

BACKGROUND

Isolated third cranial nerve palsy may result from presumed microvascular ischemia or clinically significant structural lesions. The reliability of pupil involvement for distinguishing non-microvascular causes remains uncertain.

OBJECTIVE

To describe the etiologic spectrum of isolated third cranial nerve palsy and evaluate the diagnostic performance of pupil involvement and the yield of neuroimaging.

METHODS

This retrospective observational study reviewed adult patients with isolated third cranial nerve palsy at Mukdahan Hospital, Thailand, between January 2015 and August 2025. Patients with traumatic, postoperative, recurrent, multiple cranial nerve palsies, or incomplete records were excluded. Clinical features, neuroimaging findings, etiologies, and outcomes were analyzed.

RESULTS

Of 61 records assessed, 51 patients were included. Mean age was 62.3 ± 8.7 years, and 26 patients (51.0%) were male. Presumed microvascular ischemia was the most common etiology (42/51, 82.4%). Non-microvascular causes were identified in 9 patients (17.6%), including aneurysm in 6, nasopharyngeal carcinoma in 1, cavernous sinus meningioma in 1, and invasive aspergillosis in 1. Pupil involvement was more frequent in the non-microvascular group than in the presumed microvascular group (66.7% vs 23.8%, p = 0.020). Sensitivity, specificity, positive predictive value, and negative predictive value were 66.7%, 76.2%, 37.5%, and 91.4%, respectively. In multivariable analysis, pupil involvement remained independently associated with non-microvascular etiology (adjusted OR 8.00, 95% CI 1.49-42.85, p = 0.015). The diagnostic yield of neuroimaging was 18.8% overall and 25.0% among patients undergoing vascular imaging.

CONCLUSION

Pupil involvement was associated with non-microvascular etiology but had only moderate diagnostic performance; pupil sparing did not reliably exclude clinically significant lesions. These findings suggest that vascular imaging may be clinically informative when structural causes cannot be confidently excluded, particularly given that all non-microvascular cases in this cohort were identified through vascular imaging.