1838P - AI-based smart oncology follow-up system: Prospective application testing and enhancement of clinical efficacy

Background

Follow-up care is essential in impacting treatment efficacy and survival outcomes. Traditional approaches are limited by outdated data management, the imbalance of medical resources and inefficient resource allocation. The arrival of the AI era has made it possible to resolve such situations.

Methods

The smartMTB, an AI model, integrates ESMO, NCCN, CACA, CSCO guidelines, and data from 100,000 precision oncology cases and 50,000 follow-ups. It uses designed prompts to analyze patient data for personalized care and clinical trial options and integrates a risk prediction model via API to enhance its predictive and planning functions.

Results

SmartMTB demonstrated exceptional efficiency in interpreting patient reports, achieving an 60% (from an average of 4.7 min to 1.9 min, n=454) reduction in physicians' time expenditure, exhibiting performance across several metrics: accuracy (96.9 ± 10.3), comprehensiveness (97.4 ± 9.9), intelligibility (95.7 ± 12.6). SmartMTB exhibited a positive predictive value of 79% (53/67) for patient relapse progression, and a negative predictive value of 95% (43/41). AI-assisted follow-up, by concentrating on pivotal subjects, has demonstrated a 140% (25% vs 60%, n=124) increase in effectiveness compared to traditional manual methods. The re-examination rate increased by approximately 135% (23% vs 54%, n=124). Under the surveillance of SmartMTB, two patients initially misdiagnosed were accurately diagnosed with the system's assistance. Additionally, three patients modified their existing treatment plans based on recommendations from SmartMTB. The SmartMTB platform successfully matched 142 patients with suitable clinical trials. To date, six patients have been enrolled in these trials; three of them have achieved significant disease remission, while the treatment efficacy for the remaining three is yet to be assessed.

Conclusions

SmartMTB has pioneered a novel follow-up paradigm that formulates efficient and all-encompassing follow-up strategies. This innovative model markedly enhances physician efficiency while offering patients dynamically updated care manuals, thereby increasing the benefit rate.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

Origimed.

Funding

Has not received any funding.

Disclosure

All authors have declared no conflicts of interest.