Frontiers in nutrition · 2026
Multidimensional modeling and stratification of nutritional risk in non-small cell lung cancer based on artificial intelligence-derived CT body composition phenotypes.
СтатьяНазвание и аннотация: Перевод в очереди
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Xiu Dianhui, Wang Yanjing, Sun Xiaosong, Cheng Kailiang, Ye Zhen, Liu Lin, Cheng Min
Аннотация
OBJECTIVE: To develop and temporally validate a multidimensional nutritional risk model integrating artificial intelligence (AI)-derived CT body composition phenotypes for patients with non-small cell lung cancer (NSCLC).
METHODS: This single-center retrospective cohort included 656 patients with pathologically confirmed NSCLC, including 432 in the development cohort and 224 in the temporal validation cohort. Baseline chest CT, clinical-nutritional variables, and immune-inflammatory markers were collected before first-line treatment. AI-based automated body composition analysis was used to extract CT phenotypes. The primary outcome was a 90-day nutrition-related adverse clinical trajectory. Clinical-nutritional, imaging, and integrated models were developed and compared.
RESULTS: The incidence of short-term nutrition-related adverse clinical trajectories was 34.3% in the development cohort and 32.6% in the validation cohort. Percentage weight loss, prognostic nutritional index, systemic immune-inflammation index, skeletal muscle index, and intermuscular adipose tissue volume were independent predictors. The integrated model achieved the best performance, with AUCs of 0.842 and 0.816 in the development and validation cohorts, respectively, and showed superior calibration and clinical utility. Higher risk scores were also associated with worse overall survival and progression-free survival.
CONCLUSION: Integrating AI-derived CT body composition phenotypes with nutritional and inflammatory indicators improved nutritional risk stratification in NSCLC and may support earlier risk-adapted nutritional management.
METHODS: This single-center retrospective cohort included 656 patients with pathologically confirmed NSCLC, including 432 in the development cohort and 224 in the temporal validation cohort. Baseline chest CT, clinical-nutritional variables, and immune-inflammatory markers were collected before first-line treatment. AI-based automated body composition analysis was used to extract CT phenotypes. The primary outcome was a 90-day nutrition-related adverse clinical trajectory. Clinical-nutritional, imaging, and integrated models were developed and compared.
RESULTS: The incidence of short-term nutrition-related adverse clinical trajectories was 34.3% in the development cohort and 32.6% in the validation cohort. Percentage weight loss, prognostic nutritional index, systemic immune-inflammation index, skeletal muscle index, and intermuscular adipose tissue volume were independent predictors. The integrated model achieved the best performance, with AUCs of 0.842 and 0.816 in the development and validation cohorts, respectively, and showed superior calibration and clinical utility. Higher risk scores were also associated with worse overall survival and progression-free survival.
CONCLUSION: Integrating AI-derived CT body composition phenotypes with nutritional and inflammatory indicators improved nutritional risk stratification in NSCLC and may support earlier risk-adapted nutritional management.