WE aimed to develop and validate different machine-learning (ML) prediction models for the complete resposne of oligometastaic gynecological cancer after SBRT. Material and methods: One hundred fifty-seven patients with 272 lesions fro 14 different institutions and treated with SBRT wih radical intent were ncluded. Thirteen datasets including 222 lesions were combined for model training and internal validation purposes, with an 80:20 ratio. the external testing dataset was selected as the fourtheenth Institution with 50 lesions. Lesions that achieved complete response (CR) were defined as responders. Prognostic clinical and dosimetric variables were selected using the LASSO algorithm. Six supervised ML models, including logistic regression (LR), classification and regression tree analysis (CART) and support vector machine (SVM) using four differeny kernels, were trained and tested to predict the complete response of uterine lesions after SBRT: the performance of models was assesed by receiver operating characterisic curves (ROC), area under the curve (AUC) and calibration curves. An explainable approach based on SHapley Additive exPlanation (SHAP) method was deployed to generate individual explanations of model's decision. Results. 63.6% of lesion had a complete response and were used as ground truth for the supervised models. LASSO strongly associated complete response with three variables , namely the lesion volume (PTV), the type of lesions ( Lympho-nodal versus parenchymal), and the biological effective dose (BED10), that were used as imput for M modelling. In the training set, the AUs for compete response were 0.751 (95% CI: 0.716-0.786), 0.766 (95% CI: 0.729-0.802) and 0.800 (95% CI: 0-742-0.857) for the LR, CART and SVM qith a radial basis function kernel, respectively. These models achieve AUC values of 0.727 (95% CI: 0.669-0.815) and 0.771 (95% CI: 0.717-0.824) in the external testing set, demonstrating excellent generalizability.Conclusion: ML models enable a reliable prediction of the treatment response of oligometastatic lesions receiing SBRT. This approach may asist radiation oncologists to tailor more individualized treatment plans for oligometastatic patients.

Machine-learning prediction of treatment response to stereotactic body radiation therapy in oligometastatic gynecological cancer. a multi-istitutional study / Cilla, Savino; Campitelli, Maura; Antonietta Gambacorta, Maria; Michela Rinaldi, Raffaella; Deodato, Francesco; Pezzulla, Donato; Romano, Carmmela; Fodor, Andrei; Laliscia, Concetta; Trippa, Fabio; DE SANCTIS, Vitaliana; Ippolito, Edy; Ferioli, Martina; Titone, Francesca; Russo, Donatella; Balcet, Vittoria; Vicenzi, Lisa; Di Cataldo, Vanessa; Raguso, Arcangela; Giuseppe Morganti, Alessio; Ferrandina, Gabriella; Macchia, Gabriella. - In: RADIOTHERAPY AND ONCOLOGY. - ISSN 0167-8140. - 191:(2024), pp. 1-8. [10.1016/j.radonc.2023.110072]

Machine-learning prediction of treatment response to stereotactic body radiation therapy in oligometastatic gynecological cancer. a multi-istitutional study

Vitaliana De Sanctis;Donatella Russo;
2024

Abstract

WE aimed to develop and validate different machine-learning (ML) prediction models for the complete resposne of oligometastaic gynecological cancer after SBRT. Material and methods: One hundred fifty-seven patients with 272 lesions fro 14 different institutions and treated with SBRT wih radical intent were ncluded. Thirteen datasets including 222 lesions were combined for model training and internal validation purposes, with an 80:20 ratio. the external testing dataset was selected as the fourtheenth Institution with 50 lesions. Lesions that achieved complete response (CR) were defined as responders. Prognostic clinical and dosimetric variables were selected using the LASSO algorithm. Six supervised ML models, including logistic regression (LR), classification and regression tree analysis (CART) and support vector machine (SVM) using four differeny kernels, were trained and tested to predict the complete response of uterine lesions after SBRT: the performance of models was assesed by receiver operating characterisic curves (ROC), area under the curve (AUC) and calibration curves. An explainable approach based on SHapley Additive exPlanation (SHAP) method was deployed to generate individual explanations of model's decision. Results. 63.6% of lesion had a complete response and were used as ground truth for the supervised models. LASSO strongly associated complete response with three variables , namely the lesion volume (PTV), the type of lesions ( Lympho-nodal versus parenchymal), and the biological effective dose (BED10), that were used as imput for M modelling. In the training set, the AUs for compete response were 0.751 (95% CI: 0.716-0.786), 0.766 (95% CI: 0.729-0.802) and 0.800 (95% CI: 0-742-0.857) for the LR, CART and SVM qith a radial basis function kernel, respectively. These models achieve AUC values of 0.727 (95% CI: 0.669-0.815) and 0.771 (95% CI: 0.717-0.824) in the external testing set, demonstrating excellent generalizability.Conclusion: ML models enable a reliable prediction of the treatment response of oligometastatic lesions receiing SBRT. This approach may asist radiation oncologists to tailor more individualized treatment plans for oligometastatic patients.
2024
sbrt; machine learning; predictive models; compete response; gynecological cancer
01 Pubblicazione su rivista::01a Articolo in rivista
Machine-learning prediction of treatment response to stereotactic body radiation therapy in oligometastatic gynecological cancer. a multi-istitutional study / Cilla, Savino; Campitelli, Maura; Antonietta Gambacorta, Maria; Michela Rinaldi, Raffaella; Deodato, Francesco; Pezzulla, Donato; Romano, Carmmela; Fodor, Andrei; Laliscia, Concetta; Trippa, Fabio; DE SANCTIS, Vitaliana; Ippolito, Edy; Ferioli, Martina; Titone, Francesca; Russo, Donatella; Balcet, Vittoria; Vicenzi, Lisa; Di Cataldo, Vanessa; Raguso, Arcangela; Giuseppe Morganti, Alessio; Ferrandina, Gabriella; Macchia, Gabriella. - In: RADIOTHERAPY AND ONCOLOGY. - ISSN 0167-8140. - 191:(2024), pp. 1-8. [10.1016/j.radonc.2023.110072]
File allegati a questo prodotto
File Dimensione Formato  
Cilla_Machine-learning-prediction_2024.pdf

solo gestori archivio

Tipologia: Versione editoriale (versione pubblicata con il layout dell'editore)
Licenza: Tutti i diritti riservati (All rights reserved)
Dimensione 4.09 MB
Formato Adobe PDF
4.09 MB Adobe PDF   Contatta l'autore

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11573/1705070
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 0
  • ???jsp.display-item.citation.isi??? ND
social impact