Objectives
To evaluate multidisciplinary team meeting (MDT) performance across a regional tertiary cancer service against the Queensland Multidisciplinary Team Framework, and assess whether artificial intelligence (AI)-assisted document analysis can identify system-level gaps not apparent from individual tumour-stream reviews.
Methods
A retrospective service evaluation examined six Townsville University Hospital cancer MDT reviews—breast, gastrointestinal, genitourinary, head and neck, high-risk skin and lung—against the July 2026 Queensland MDT Framework. A large language model extracted, compared and synthesised findings across nine domains: minimum dataset, membership, case presentation, chairing, coordination, First Nations care, governance, technology and reporting. Outputs were manually reviewed by a senior oncology clinician. Recurrent findings were classified as isolated, tumour-specific or service-wide.
Results
AI-assisted synthesis identified a clinically mature but operationally fragile MDT system. Strengths included regular multidisciplinary review, effective chairing, broad specialist participation, established First Nations and palliative-care pathways, and public–private collaboration. However, structural deficiencies recurred across all six MDTs: incomplete structured minimum datasets; inconsistent documentation of performance status, comorbidity, stage and clinical-trial eligibility; reliance on free-text MOSAIQ workflows; variable referral quality and late additions; inadequately protected radiology and pathology participation; inconsistent quorum monitoring; and suboptimal hybrid-meeting technology. The most significant risk was dependence on one coordinator supporting seven MDTs without reliable trained backfill. AI also detected inconsistencies in meeting frequency, duration and pathway reporting. These findings reframed multiple local problems as a single institutional operating-model deficit.
Conclusions
AI-assisted framework comparison can convert fragmented governance reviews into an actionable regional cancer-service diagnosis. With clinician validation, it may accelerate gap identification, prioritisation and business-case development where analytic capacity is limited. Findings support a unified cancer MDT program, additional coordination capacity, structured digital data capture, protected diagnostic-specialist time and a service-wide quality dashboard. Prospective evaluation should assess implementation fidelity, time savings and impact on equity, timeliness and treatment delivery.