Vol.:(0123456789) 1 3 Journal of Neurology https://doi.org/10.1007/s00415-019-09259-3 ORIGINAL COMMUNICATION Aphasia outcome: the interactions between initial severity, lesion size and location Sarah Benghanem 1  · Charlotte Rosso 1,2  · Céline Arbizu 1,3  · Eric Moulton 2  · Didier Dormont 2,4  · Anne Leger 1  · Christine Pires 1  · Yves Samson 1,2 Received: 21 November 2018 / Revised: 20 February 2019 / Accepted: 22 February 2019 © Springer-Verlag GmbH Germany, part of Springer Nature 2019 Abstract Objectives The outcome of aphasia at 3 months is variable in patients with moderate/severe stroke. The aim was to predict 3-month aphasia outcome using prediction models including initial severity in addition to the interaction between lesion size and location at the acute phase. Methods Patients with post-stroke aphasia (assessed by the Aphasia Rapid Test at day 7-ART D7) and MRI performed at day 1 were enrolled (n = 73). Good outcome at 3-months was defined by an Aphasia Handicap Score of 0–2. Each infarct lesion was overlapped with an area of interest in the left temporo-parietal region to compute an intersection index (propor- tion of the critical region damaged by the infarct). We tested ART D7, age, lesion volume, and intersection index as well as a combined variable lesion volume*intersection in a univariate analysis. Then, we performed a multivariate analysis to investigate which variables were independent predictors of good outcome. Results ART at D7, infarct volume, and the intersection index were univariate predictors of good outcome. In the multivari- ate analysis, ART D7 and “volume ≥ 50 ml or intersection index ≥ 20%” correctly classified 89% of the patients (p < 0.0001). When added to the model, the interaction between both variables was significant indicating that the impact of the size or site variable depends on the initial severity of aphasia. Conclusion In patients with initially severe aphasia, large infarct size or critical damage in left temporoparietal junction is associated with poor language outcome at 3 months. Keywords Aphasia · Magnetic resonance imaging · Prognosis Introduction Early prediction of language outcome remains challeng- ing in aphasic stroke patients. Simple and robust progno- sis models are important for patients and their relatives but also for improving the stratification of patients in language therapeutic trials during the first 3 months post stroke, cor- responding to the “spontaneous” recovery period. To date, the best predictor of outcome is the initial severity of aphasia [1–9], whereas other patient-related predictors (age, educa- tion…) are more debated [10, 11]. However, while initially mild aphasias consistently recover, outcome is less predict- able in more severe patients (see for example Fig 1 of Ref. [3]). Some studies [12, 13] have suggested that recovery is proportional to initial severity (about 70% of the maxi- mum possible recovery), but this notion has been recently challenged on theoretical grounds [14, 15]. Lesion size and location are also related to language outcome [10–16]. The Electronic supplementary material The online version of this article (https://doi.org/10.1007/s00415-019-09259-3) contains supplementary material, which is available to authorized users. * Yves Samson yves.samson@aphp.fr 1 APHP-Urgences Cérébro-Vasculaires, Hôpital Pitié- Salpêtrière, 83 Boulevard de l’Hôpital, 75013 Paris, France 2 Inserm U 1127, CNRS UMR 7225, Sorbonne Université, UPMC Univ Paris 06 UMR S 1127, Institut du Cerveau et de la Moelle épinière, ICM, 75013 Paris, France 3 IM2A, Hôpital Pitié-Salpêtrière, Paris, France 4 APHP-Neuroradiology Department, Hôpital Pitié-Salpêtrière, 75013 Paris, France