Invited Commentary: Evaluation of Horner Syndrome in the MRI Era

This Invited Commentary discusses the following article: BACKGROUND: To identify the etiologies of adult Horner syndrome (HS) in the MRI era using a targeted evaluation approach and to assess the value and yield of targeted imaging. A retrospective chart review was performed of 200 adult outpatients with HS, confirmed with cocaine eyedrop testing. Patients were divided into subgroups based on the presence or absence of symptoms and those who did or did not receive additional testing with hydroxyamphetamine drops. Imaging was obtained based on pharmacologic localization and/or clinical evaluation. The etiology of HS and the yield of imaging were determined in all subgroups. Imaging showed causative lesions in 24 of 179 (12.84%) imaged patients with HS, and 13 (69.0%) were determined "idiopathic." Of the patients who underwent testing with hydroxyamphetamine drops (132 patients), 86 had a postganglionic localization with an imaging yield of 8.1%, and 46 had preganglionic cause with an imaging yield of 21.7%. Fifty-three patients (26.5%) never noticed ptosis/anisocoria before examination, and the imaging yield in this subgroup was 2.8%. Eighteen of the 200 patients (9.0%) had serious pathology, including carotid artery dissection, brain, or neck mass, and 6 of these (31.6%) had acute symptoms and/or pain. HS is most often idiopathic with serious pathology being relatively infrequent. When determining etiology, the absence of symptoms is not predictive of the pathology. However, acute onset of symptoms and/or pain are possible indicators for serious pathology. Localizing the lesion using hydroxyamphetamine drops whenever obtainable and available is still an efficient way to target imaging evaluation.

Invited Commentary Invited Commentary: Evaluation of Horner Syndrome in the MRI Era Aki Kawasaki, MD This Invited Commentary discusses the following article: Sadaka A, Schockman SL, Golnik KC. Evaluation of Horner syndrome in the MRI era. J Neuroophthalmol. 2017;37:268-272. Background: To identify the etiologies of adult Horner syndrome (HS) in the MRI era using a targeted evaluation approach and to assess the value and yield of targeted imaging. Methods: A retrospective chart review was performed of 200 adult outpatients with HS, confirmed with cocaine eyedrop testing. Patients were divided into subgroups based on the presence or absence of symptoms and those who did or did not receive additional testing with hydroxyamphetamine drops. Imaging was obtained based on pharmacologic localization and/or clinical evaluation. The etiology of HS and the yield of imaging were determined in all subgroups. Results: Imaging showed causative lesions in 24 of 179 (12.84%) imaged patients with HS, and 13 (69.0%) were determined "idiopathic." Of the patients who underwent testing with hydroxyamphetamine drops (132 patients), 86 had a postganglionic localization with an imaging yield of 8.1%, and 46 had preganglionic cause with an imaging yield of 21.7%. Fifty-three patients (26.5%) never noticed ptosis/anisocoria before examination, and the imaging yield in this subgroup was 2.8%. Eighteen of the 200 patients (9.0%) had serious pathology, including carotid artery dissection, brain, or neck mass, and 6 of these (31.6%) had acute symptoms and/or pain. Conclusion: HS is most often idiopathic with serious pathology being relatively infrequent. When determining etiology, the absence of symptoms is not predictive of the pathology. However, acute onset of symptoms and/or pain are possible indicators for serious pathology. Localizing the lesion using hydroxyamphetamine drops whenever obtainable and available is still an efficient way to target imaging evaluation. H ow to image a patient with Horner syndrome is a question that has generated a wide range of responses (1-3). The article by Sadaka et al (4) provides yet another response to the question, but distinguishes itself by making cost consciousness an important consideration. In their retrospective series, the authors examined the imaging yield of 200 adult patients with cocaine-confirmed Horner syndrome who were evaluated by one neuroophthalmologist. The advantage of such single examiner/ single institute study is a consistent approach to patient evaluation, such that over time a large, yet fairly homogeneous, database is acquired. Specifically, if a probable cause of Horner syndrome could be defined from history alone, then neuroimaging was not performed. These probable causes were 1) longstanding anisocoria and ptosis proven by old photographs that suggested a congenital origin, 2) onset Department of Neuro-Ophthalmology, University of Lausanne, Hôpital Ophtalmique Jules Gonin, Lausanne, Switzerland. The author reports no conflicts of interest. Address correspondence to Aki Kawasaki, MD, Department of Neuro-Ophthalmology, University of Lausanne, Hôpital Ophtalmique Jules Gonin, Avenue de France 15, Lausanne, 1004, Lausanne, Switzerland; E-mail: [e-mail] 52 of anisocoria/ptosis immediately after a neck procedure indicating a traumatic etiology, and 3) transient ptosis/ anisocoria associated with cluster headaches. For 179 patients, MRI without supplemental angiographic sequences was performed: the nonuse of MRA was a choice in this series, not an oversight, and motivated by cost reduction. Overall, MRI revealed a lesion of the oculosympathetic pathway in 24 of 179 (12.8%) patients with Horner syndrome. The authors then examined the imaging yield in a subgroup of 132 patients for whom MRI had been directed by pharmacologic localization using topical 1% hydroxyamphetamine eye drops (5). Those patients found to have a postganglionic Horner syndrome (third order neuron involved) received only brain MRI (targeted imaging), whereas those with a preganglionic Horner syndrome (first and second order neuron involved) had MRI of the entire oculosympathetic pathway (nontargeted imaging). In patients with a postganglionic Horner syndrome, MRI revealed a causative lesion in 8.1% (7/86 patients); those with a preganglionic Horner syndrome had a yield that was nearly 3 times higher of 21.7% (10/46 patients). These findings strengthen the notion that a preganglionic Horner Kawasaki: J Neuro-Ophthalmol 2018; 38: 52-53 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Invited Commentary syndrome is more likely to be due to a structural lesion than a postganglionic Horner syndrome, which tends to be idiopathic (1,6). The differential imaging yield also supports the authors' primary message that imaging the entire oculosympathetic pathway in all patients with Horner syndrome is not cost-effective. However, one important question persists: should angiography (computed tomography angiography or MRA or digital subtraction) have been included to more thoroughly search for an underlying arterial dissection, be it vertebral or carotid? In this series, without the use of MRA, 7 patients with arterial dissection were detected. Four experienced sudden-onset ptosis/anisocoria with pain and the other 3 had additional neurologic findings. These patients with suspected dissection underwent urgent neuroimaging, without having hydroxyamphetamine eye drop testing. In another recently published series (3), 88 patients with isolated Horner syndrome underwent a prespecified imaging protocol that included MRA. Seven patients (8%) were found to have dissection. Whether this higher prevalence of arterial dissection is due to better detection by inclusion of angiographic sequences or whether it is a difference in the patient population studied is not clear from the published data. In the absence of such comparative information, the choice to include or later add angiography remains an individualized one, best guided by the degree of clinical suspicion. In the study by Sadaka et al, there was a subgroup of 47 patients who may represent the scenario that most clinicians encounter: individuals with an acquired, nontraumatic Horner syndrome who may or may not have signs and symptoms suggesting an arterial dissection and who do not undergo hydroxyamphetamine testing to aid in localization. These patients underwent MRI of the entire oculosympathetic pathway that detected a causative lesion in 9.5% of cases. Kawasaki: J Neuro-Ophthalmol 2018; 38: 52-53 This brings us to a final question: do most clinicians who see patients with Horner syndrome perform hydroxyamphetamine testing? I would hazard a guess and say "no," primarily because there is very limited availability of hydroxyamphetamine eye drops. It traditionally has been available as a custom order from compounding pharmacies, but these independent businesses are one-by-one closing due to competition by large pharmaceutical distributors. Do most clinicians who see patients with Horner syndrome follow a "scan everything" protocol? Likewise, I would think not. For many clinicians, it remains the careful history and clinical examination which guide the decision of who to scan, when to scan, where to scan, and how to scan. And for those lucky enough to still have access to hydroxyamphetamine eye drops, Sadaka et al have shown their utility in avoiding overscanning patients with a postganglionic Horner syndrome. REFERENCES 1. Davagnanam I, Fraser CL, Miszkiel K, Daniel CS, Plant GT. Adult Horner's syndrome: a combined clinical, pharmacological, and imaging algorithm. Eye. 2013;27:291-298. 2. Almog Y, Gepstein R, Kesler A. Diagnostic value of imaging in Horner syndrome in adults. J Neuroophthalmol. 2010;30:7-11. 3. Beebe JD, Kardon RH, Matthew MJ. The yield of diagnostic imaging in patients with isolated Horner syndrome. Neurol Clin. 2017;35:145-151. 4. Sadaka A, Schockman SL, Golnik KC. Evaluation of Horner syndrome in the MRI Era. J Neuroophthamol. 2017;37:268- 272. 5. Cremer SA, Thompson HS, Digre KB, Kardon RH. Hydroxyamphetamine mydriasis in Horner's syndrome. Am J Ophthalmol. 1990;110:71-76. 6. Reede DL, Garcon E, Smoker WRK, Kardon RH. Horner's syndrome clinical and radiographic evaluation. Neuroimag Clin N Am. 2008;18:369-385. 53 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited.