Spontaneous speech of patients with gliomas in eloquent areas before and early after surgery

Abstract

Glioma patients often complain about problems in daily conversation. A detailed spontaneous speech analysis could provide more insight in these communicative problems; no previous studies are reported. To select sensitive parameters in spontaneous speech pre- and post-operatively in patients with gliomas in eloquent areas. We included 27 patients and 21 healthy controls. In addition to a naming and category fluency test, spontaneous speech was collected 1 month pre-operatively and 3 months post-operatively, and analysed with the variables: Self-corrections, Repetitions, Lexical Diversity, Incomplete Sentences and Mean Length of Utterance (MLUw). A correlation analysis was performed between the linguistic variables and tumour characteristics (grade, localisation and volume), treatment related factors, and between the linguistic variables and the language tasks. Pre-operatively, patients produced more Incomplete Sentences than the controls (p < 0.001). Post-operatively, patients' utterance length (MLUw) (p < 0.05) was also deviant. The quality of the spontaneous speech was influenced by tumour grade and localisation. There was no influence of tumour volume or treatment-related factors. Pre- and post-operatively, patients' performance on the naming and the fluency task deviated from normal (p < 0.001). The majority of the linguistic variables did not correlate with the language tasks, pointing to a measurement of distinct linguistic aspects. Pre- and post-operatively there was a disorder in naming, category fluency and spontaneous speech, partly influenced by tumour characteristics. A spontaneous speech analysis appeared to be a valuable addition to standardised language tasks. Both measurements are important tools to obtain a complete linguistic profile.