We use cookies necessary to giving you a better online experience. By using our website you consent to all cookies in accordance with our Privacy Policy.
Yes, I accept
Menu
Search Opener
Vienna, Austria

ESTRO 2023

Session Item

Urology
Poster (Digital)
Clinical
Process mining in PCa Salvage RT patients care path: an AIRO-URO group preliminary analysis
Federico Mastroleo, Italy
PO-1494

Abstract

Process mining in PCa Salvage RT patients care path: an AIRO-URO group preliminary analysis
Authors:

Federico Mastroleo1, Giulia Marvaso1, Riccardo Villa1, Mattia Zaffaroni1, Giulia Corrao1, Maria Giulia Vincini1, Fabio Matrone2, Alessandro Magli3, Giulio Francolini4, Ciro Franzese5, Luca Nicosia6, Francesco Pasqualetti7, Luca Trodella8, Annamaria Vinciguerra9, Salvina Barra10, Giorgia Timon11, Matteo Augugliaro11, Marta Scorsetti5, Barbara Alicja Jereczek-Fossa1, Stefano Arcangeli12, Luca Triggiani13

1IEO, Radiation Oncology, Milan, Italy; 2CRO, Radiation Oncology, Aviano, Italy; 3San Martino Hospital, Radiation Oncology, Belluno, Italy; 4AOU Careggi, Radiation Oncology, Florence, Italy; 5Humanitas Research Hospital, Radiation Oncology, Rozzano, Italy; 6Sacro Cuore-Don Calabria Hospital, Radiation Oncology, Verona, Italy; 7Pisa Hospital, Radiation Oncology, Pisa, Italy; 8Campus Biomedico, Radiation Oncology, Rome, Italy; 9SS. Annunziata Hospital, Radiation Oncology, Chieti, Italy; 10San Martino Hospital, Radiation Oncology, Genova, Italy; 11IRCCS Reggio Emilia, Radiation Oncology, Reggio Emilia, Italy; 12ASST Monza, Radiation Oncology, Monza, Italy; 13Spedali Civili, Radiation Oncology, Brescia, Italy

Show Affiliations
Purpose or Objective

Process mining is a machine learning powerful tool that could help clinicians in better analyzing real-word data to achieve an accurate and efficient workflow prototyping and visualization. The aim of this work is to perform a process mining analysis of the data coming from the AIRO endorsed national-wide survey of salvage radiotherapy (SRT) for prostate cancer patients.

Material and Methods

A custom script has been developed to refine and convert the available data in event log format. The eligible event categories for the analysis were: surgery, biochemical relapse, imaging restaging at biochemical relapse, radiation therapy treatment, progression of the disease, castration resistance development, death, last follow-up. First Order Markov Model has been adopted for the event analysis and mean time to the event has been calculated for each single event category. Process mining analysis has been performed by R v.4.2.1 and pMineR v.045b library. This study has been approved by the ethics committee of IEO and centro cardiologico Monzino of Milan, UID 3576.

Results

This preliminary analysis included 582 prostate cancer patients, from 6 different Italian institutions. A total of 1595 events have been recorded and analyzed with a medium follow-up time of 4.82 years. Considering surgical prostatectomy as starting point of patients’ care path, at the biochemical relapse only 158 patients (26% of the total number of consequent events) underwent restaging by radiological imaging. Mean time from surgery to salvage radiotherapy was 3.25 years. The 39% of the patients who underwent SRT had disease progression in a mean time of 2.73 years. Disease progression involved the development of castration resistance in the 0.1% of the cases with a mean time of 2.39 years and death occurred in the 52% of the patients of this group.



Conclusion

Our preliminary work clearly demonstrates the impact and potential of process mining methodology in analyzing patients’ care path when hundreds of records are involved. The complete accrual of the data from the ongoing AIRO national study and the development of more complex models will help us to gather further stratification and insights on the usage of salvage radiotherapy, while clarifying the role of the different variables involved in prostate cancer disease progression.