Articulatory and acoustic signals recorded from 2 patients who underwent the resection of a tongue carcinoma were analyzed longitudinally. The data were recorded before and after intra-oral surgery (1 month and 3 months after).Lingual movements were recorded with ultrasound. We focused on the lingual stop consonants and the lingual sibilants of French. The analysis was based on spectral moments (COG, skewness, kurtosis) and VOT. Recording ultrasound data requires a calibration which measures the position of the probe in reference to the subject's head. This is classically done with an optical system. However such an experimental setup cannot be easily used with patients in a clinical context. Hence, the analysis of the articulatory data required elaborating a methodology enabling to compensate a posteriori for the absence of calibration.. It was based on 2 measurements: (1) the variation rate of the area of the tongue section in the mid-sagittal plane, which is a way to assess tongue mobility, and the height of the highest point of the tongue, which is a way to assess tongue deformation. A perceptual evaluation of the consonants in pre- and post-surgery conditions was also carried out by speech therapists,. This serves as a reference to evaluate the discrimination capacity of the proposed acoustic and articulatory analysis. Results Acoustics: For both subjects, the Skewness is the most efficient spectral parameter to account for the results of the perceptual evaluation for sibilants in context /i/ and /a/. VOT is relevant to characterize articulation changes in [ka] and in [ga]. Articulatory data: Bad perception scores in post-surgery conditions were associated with a a low lingual mobility for a patient and with a low capacity to shape the tongue for the other one. Compensation: Differences still exists between the consonants in pre- and in post-operation conditions. Three months is not a sufficient delay to recover the articulation on these consonants.