The Journal of allergy and clinical immunology · 2026
Etiologic diagnosis of seasonal allergic rhinitis supported by artificial intelligence: The @IT-2020 project.
СтатьяПоддержка исследования: вне правительства СШАНазвание 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.
Matricardi Paolo Maria, Monnati Francesco, Palmieri Luigi, Dramburg Stephanie, Mesonjesi Eris, Sfika Ifigenia, Villella Valeria, Arasi Stefania, Brighetti Maria Antonia, Potapova Ekaterina, Hoffmann Tara Maria, Pelosi Simone, Bregu Blerina, Lame Blerta, Charpin Denis, Delgado Luís, Fonseca João Almeida, Goksel Ozlem, Kalpaklioglu Fusun, Nieto Antonio, Pajno Giovanni Battista, Sackesen Cansin, Travaglini Alessandro, Priftanji Alfred, Salina Gaetano, Tripodi Salvatore, Papadopoulos Nikolaos G
Аннотация
BACKGROUND: A precise etiologic diagnosis of seasonal allergic rhinitis (SAR) is essential for tailored prescription of its only curative treatment, allergen immunotherapy. This is a challenging task in temperate climates, where most patients are polysensitized to multiple pollen with overlapping seasons.
OBJECTIVE: The study aimed to develop a modular, flexible, and validated clinical decision support system (CDSS) generated with artificial intelligence for etiologic diagnosis of SAR.
METHODS: Within the @IT-2020 project, we developed a CDSS for SAR etiologic diagnosis, automated through machine learning (ML). The CDSS includes 3 progressive modules: (a) clinical history and SPT, (b) plus molecular sIgE, (c) plus electronic/environmental diary. Three raters performed ML training, by identifying culprit pollen in 100 SAR patients (Rome, Italy) using guidelines and a Delphi-like process.
RESULTS: Best-performing ML models for each diagnostic module realibly replicated expert diagnosis (area under the receiver operating characteristics curve >95%). Their validity was confirmed by: (A) contextual adaptability, with performance linked to patient complexity; (B) interpretability, as clinical features and sensitization patterns contrtibuted to predictions (SHAP analysis); (C) geographical generalizability, with consistent performance in 92 patients from Tirana (Albania); (D) temporal generalizability, maintaining high performance with reduced monitoring (45 days); (E) AIT prescription adaptability, reproducing gold-standard prescriptions; and (F) human-plus capability, ouperforming 24 physicians in a diagnostic challenge.
CONCLUSION: In this proof-of-concept study, an ML-based modular CDSS reliably replicated raters' diagnosis and AIT prescription in SAR. Further studies should confirm replication and prospectively asses CDSS role in enabling SAR personalized treatment and improving disease control.
OBJECTIVE: The study aimed to develop a modular, flexible, and validated clinical decision support system (CDSS) generated with artificial intelligence for etiologic diagnosis of SAR.
METHODS: Within the @IT-2020 project, we developed a CDSS for SAR etiologic diagnosis, automated through machine learning (ML). The CDSS includes 3 progressive modules: (a) clinical history and SPT, (b) plus molecular sIgE, (c) plus electronic/environmental diary. Three raters performed ML training, by identifying culprit pollen in 100 SAR patients (Rome, Italy) using guidelines and a Delphi-like process.
RESULTS: Best-performing ML models for each diagnostic module realibly replicated expert diagnosis (area under the receiver operating characteristics curve >95%). Their validity was confirmed by: (A) contextual adaptability, with performance linked to patient complexity; (B) interpretability, as clinical features and sensitization patterns contrtibuted to predictions (SHAP analysis); (C) geographical generalizability, with consistent performance in 92 patients from Tirana (Albania); (D) temporal generalizability, maintaining high performance with reduced monitoring (45 days); (E) AIT prescription adaptability, reproducing gold-standard prescriptions; and (F) human-plus capability, ouperforming 24 physicians in a diagnostic challenge.
CONCLUSION: In this proof-of-concept study, an ML-based modular CDSS reliably replicated raters' diagnosis and AIT prescription in SAR. Further studies should confirm replication and prospectively asses CDSS role in enabling SAR personalized treatment and improving disease control.