Abstract Book

ESTRO 37

S559

Conclusion The 5-mm-thick Y-90 solution seems appropriate for the suggested system. To deliver 3 Gy with the suggested system, treatment time of 18.5 min would be required. PO-1003 Feasibility of a machine learning QA system for failure detection in IORT with a mobile accelerator. R. Ayala 1 , R. Fabregat 2 , M. Alarcia 1 , J. Vilanova 1 , M.J. García 1 , G. Ruiz 1 , S. Gómez 1 , R. Jiménez 1 , M.A. López 1 1 Hospital General Universitario Gregorio Marañón, Servicio de Dosimetría y Radioprotección, Madrid, Spain 2 Hospital Universitario Marqués de Valdecilla, Servicio de Radiofísica y Protección Radiológica, Santander, Spain Purpose or Objective The aim of this work was to investigate the applicability of a failure detection system using information from log files and using machine learning tools to detect anomalies. Log files analysis has been a topic of interest for QA in conventional linacs but has not been addressed for mobile ones probably due to a more complicated process to obtain them. We focused on machine items that are involved in the correct functioning of the system. Drifts in those item values will presumably correspond with future linac failures or unexpected behaviors. In intraoperative radiotherapy (IORT) linac availability is a critical problem since the patient cannot wait for the machine to be repaired. A system that can anticipate possible failures is therefore of extreme importance. Material and Methods Linac This study has been performed with a Sordina Liac 12MeV model, a mobile electron linac with four energies: 6, 8, 10 and 12MeV. It is provided with 7 circular PMMA applicators of different diameter and 4 different beveled endings (0º, 15º, 30º and 45º). Log files The Liac accelerator does not have a control system PC interface nor a record and verify connection but Sordina has developed a software to monitor several machine parameters via RS232 cable. It is then possible to save a text file containing dosimetric information as well as some inner parameters of the machine. An in- house software has been developed to read log files, visualize and store them for further analysis (figure 1).

The flexible pouch is located inside the patient-specific mold which is solid, therefore, when the liquid radioisotope is injected into the flexible pouch, the pouch is inflated and form shape according to the mold shape. For the treatment, [atient’s head is scanned with a 3D scanner. With this information, solid patient-specific mold is fabricated with a 3D printer. Inside the mold, the thin flexible pouch is located and connected with a tube for injecting the liquid radioisotope. When injecting the liquid radioisotope through the tube from the liquid radioisotope tank to the flexible pouch inside the solid mold, the pouch fills inside the mold. The liquid radioisotope stays in the pouch to deliver prescription doses to the scalp. To examine which liquid radioisotope would be appropriate for this system, we performed Monte Carlo simulation using GEANT4 version 10.3 patch- 2. The physics list of the simulation was QGSP_BIC_LIV. The cutoff range was set to 0.001 mm. The energy deposition was save to each voxel of which dimension was 0.2 mm 0.2 mm 0.2 mm. Liquid radioisotopes of P- 32, Sr-89 and Y-90 were tested in this study. The water- equivalent rectangular-shaped phantoms (target volumes) were covered with each of the liquid isotopes (thickness of 10 mm) contained in the pouch. After evaluating the percent depth doses (PDDs) in the phantoms by each liquid radioisotope, we selected the best appropriate liquid radioisotope for the suggested system. After that, we acquired PDDs of the selected radioisotope varying the thicknesses from 1 mm to 5 mm with spherical phantoms with radii of 77 mm (77Sph_phantom) and 91 mm (91Sph_phantom). Results The PDD of the Y-90 at the depth of 2 mm in the phantom was 1.97 times higher than that of the P-32 and 2.63 times higher than that of the Sr-89. The PDD of the Y-90 was less than 1% at the depth of 6 mm. The PDD values of the Y-90 in the 77Sph_phantom and the 91Sph_phantom at the depth of 2 mm were 17.4% and 16.9%, respectively. As increasing the thickness of the Y-90 from 1 mm to 5 mm, the dose rates at the depth of 2 mm increased from 0.0015 Gy/s to 0.0027 Gy/s (Fig. 2).

SVM A one-class support vector machine is an unsupervised learning model that is able to trace the boundaries of a set of parameters, a new set of data that would be analyzed with the SVM will be categorized as either in or out of the boundaries. In this case we have chosen the scikit-learn python implementation with a radial basis function kernel (RBF). K(x,x') = exp( -γ ||x-x'||²)

No noticeable differences in the dose rates were observed between the thickness of 4 mm and 5 mm of Y- 90.