Speech Database of Consonant-based Articulation Assessment Battery plus Sustained Vowels for Oral and Oropharyngeal Cancers.

We present MSDOC (Multimodal Speech Database for Oral and Oropharyngeal Cancers), a comprehensive longitudinal speech database comprising pre- and post-operative audio recordings from 209 patients diagnosed with oral cavity (OC) and oropharyngeal (OPC) cancers. The database includes sustained vowel productions (/a/, /i/, /u/, /o/, /y/, /e/) as well as performances on a standardized 43-item Consonant-based Articulation Assessment Battery (CAAB), specifically developed for Mandarin Chinese. MSDOC is enriched with extensive biological data, including tumor staging, immunoinflammatory indices, patient-reported outcomes, and functional assessments. These biological variables are systematically compiled into audio-level metadata files to ensure traceability and data integration. To the best of our knowledge, MSDOC represents the largest disease-specific speech database for OC and OPC, offering significant potential for applications in patient education, diagnostic support, prognostic modeling, and individualized speech rehabilitation.