Show simple item record

dc.contributor.advisorPorto, Italo <1974>
dc.contributor.advisorAmeri, Pietro <1981>
dc.contributor.authorLuiso, Sofia <2000>
dc.date.accessioned2026-07-09T14:22:45Z
dc.date.available2026-07-09T14:22:45Z
dc.date.issued2026-07-01
dc.identifier.urihttps://unire.unige.it/handle/123456789/16055
dc.description.abstractBackground: Le antracicline sono farmaci antineoplastici associati a danno cardiaco(cancer treatment-related cardiac dysfunction, CTRCD).Algoritmi basati sull’intelligenza artificiale(IA)possono facilitare la stratificazione del rischio e le strategie di prevenzione. Obiettivo: Lo scopo è stato valutare se algoritmo di IA per l’analisi dell’elettrocardiogramma (ECG) predica l’insorgenza di CTRCD subclinica in pazienti trattati con antracicline. Metodi: In 60 pazienti sottoposti a terapia con antracicline,sono state eseguite visite cardiologiche con ECG,dosaggio di NT-proBNP e troponina I ad alta sensibilità (hsTnI) ed ecocardiogramma prima dell’inizio del trattamento(T0)e 3, 6 e 9 mesi dopo (T3,T6 e T9).Qualora nell’arco dei 9 mesi NT-proBNP(>125 ng/L se età<75 anni e>450 ng/L se ≥75 anni)o hsTnI(>18 ng/L)fossero aumentati e/o il global longitudinal strain fosse peggiorato>15% rispetto al T0,i pazienti sono stati classificati come con CTRCD subclinica(CTRCD_Y).L’analisi con IA dei tracciati ECG ha calcolato una probabilità di disfunzione ventricolare sinistra(DVS),confrontata tra pazienti CTRCD_Y e CTRCD_N. Risultati: Al T0, 18(30%)pazienti presentavano già valori di NT-proBNP e/o hsTnI oltre i limiti di normalità e, pertanto,non sono stati considerati.27(45%)appartenevano al gruppo CTRCD_Y e 15(25%)al gruppo CTRCD_N.La probabilità di DVS stimata dall’IA al T0 non era significativamente diversa tra i gruppi CTRCD_Y(68,3±8,7)e CTRCD_N(69,7±10,2).Tuttavia, diminuiva progressivamente e con tendenza quasi significativa nel gruppo CTRCD_N(p =0,08),ma non in quello CTRCD_Y(p =0,69). Conclusioni: L’analisi dell’ECG con IA ha dimostrato limitato valore nell’identificare precocemente la CTRCD subclinica associata a chemioterapia con antracicline.it_IT
dc.description.abstractBackground: Anthracyclines are antineoplastic drugs associated with cardiac damage (cancer treatment-related cardiac dysfunction, CTRCD). Artificial intelligence (AI)-based algorithms may facilitate risk stratification and preventive strategies. Objective: To evaluate whether an AI algorithm for electrocardiogram (ECG) analysis predicts the onset of subclinical CTRCD in patients treated with anthracyclines. Methods: Sixty patients receiving anthracycline therapy underwent cardiological evaluation including ECG, NT-proBNP and high-sensitivity troponin I (hsTnI) measurements, and echocardiography before treatment initiation (T0) and after 3, 6, and 9 months (T3, T6, and T9). Patients were classified as having subclinical CTRCD (CTRCD_Y) if, during the 9-month follow-up, NT-proBNP (>125 ng/L for age <75 years and >450 ng/L for age ≥75 years) or hsTnI (>18 ng/L) increased and/or global longitudinal strain worsened by >15% compared with T0. AI analysis of ECG tracings generated a probability of left ventricular dysfunction (LVD), which was compared between CTRCD_Y and CTRCD_N patients. Results: At T0, 18 (30%) patients already had NT-proBNP and/or hsTnI values above the normal range and were therefore excluded from further analysis. Among the remaining patients, 27 (45%) were classified as CTRCD_Y and 15 (25%) as CTRCD_N. The AI-estimated probability of LVD at T0 was not significantly different between CTRCD_Y (68.3±8.7) and CTRCD_N (69.7±10.2). However, it progressively decreased, with a near-significant trend, in the CTRCD_N group (p=0.08), but not in the CTRCD_Y group (p=0.69). Conclusions: AI-based ECG analysis showed limited value in the early identification of subclinical CTRCD associated with anthracycline-containing chemotherapy.en_UK
dc.language.isoit
dc.rightsinfo:eu-repo/semantics/restrictedAccess
dc.titleStima basata su intelligenza artificiale del rischio di disfunzione ventricolare sinistra in pazienti trattati con antraciclineit_IT
dc.title.alternativeArtificial intellingence-based risk of left ventricular systolic dysfunction in patients treated with anthracyclinesen_UK
dc.typeinfo:eu-repo/semantics/masterThesis
dc.publisher.nameUniversità degli studi di Genova
dc.date.academicyear2025/2026
dc.description.corsolaurea8745 - MEDICINA E CHIRURGIA
dc.description.area6 - MEDICINA E CHIRURGIA
dc.description.department100007 - DIPARTIMENTO DI MEDICINA INTERNA E SPECIALITÀ MEDICHE


Files in this item

This item appears in the following Collection(s)

Show simple item record