Clinical Validation of Artificial Intelligence Based Auto-Segmentation of Organs-at-Risk in Total Marrow Irradiation Treatment

Abstract

This study evaluated the survival benefit of asparaginase (ASP)-based versus non-ASP-based regimens in a real-world cohort of patients with early-stage extranodal nasal-type NK/T-cell lymphoma (ENKTCL). Materials/Methods: In the modern era, we identified 376 patients who received combined radiotherapy with either ASP-based (asparaginase, platinum and gemcitabine, n = 286) or non-ASP-based (platinum and gemcitabine, n = 90) regimens. The patients were stratified into low-, intermediate-and high-risk groups using the early stage-adjusted nomogramrevised risk index (ES-NRI). Overall survival (OS) and distant metastasis (DM)-free survival (DMFS) between the chemotherapy regimens were compared using inverse probability of treatment weighting (IPTW) and multivariable analyses. Results: ASP-based (versus non-ASP-based) regimens significantly improved 5-year OS (84.5% vs. 73.2%, P = 0.021) and DMFS (84.4% vs. 74.5%, P = 0.014) for intermediate-and high-risk patients, but not for lowrisk patients. Moreover, ASP-based regimens decreased DM, with 5-year cumulative DM rate of 14.9% for ASP-based regimens compared with 25.1% (P = 0.014) for non-ASP-based regimens. The survival benefit of ASP-based regimens remained consistent after adjusting the confounding variables by IPTW and multivariate analyses. Additional sensitivity analyses confirmed these results. Conclusion: These findings provide an evidence supporting ASP-based regimens as a first-line combined-modality treatment for intermediate-and high-risk early-stage ENKTCL.

Leveraging the Academic Artificial Intelligence Silecosystem to Advance the Community Oncology Enterprise

“…The cumulative energies of the AAI/DS, Center for Informatics, TGen, as well as the efforts of independent COH investigators have helped create a rich resource of AI expertise and maintain a robust portfolio of AI research. Examples of other initiatives at COHthat illustrate the depth and breadth of the AI Silecosystem include the use of AI autosegmentation for patients pending bone marrow transplant irradiation [344][345][346], AI-assisted oncologic drug design [347], expert critical review of clinical AI models [348], AI-based platforms for the evaluation and treatment of lung [349] and breast cancers [350], machine learning enabled pre-surgery physical status scoring [351] and AI-assisted irradiation dose estimation [352].…”