Computed tomography synthesis from magnetic resonance imaging using cycle Generative Adversarial Networks with multicenter learning.

Background and Purpose: Addressing the need for accurate dose calculation in MRI-only radiotherapy, the generation of synthetic Computed Tomography (sCT) from MRI has emerged. Deep learning (DL) techniques, have shown promising results in achieving high sCT accuracies. However, existing sCT synthesis methods are often center-specific, posing a challenge to their generalizability. To overcome this limitation, recent studies have proposed approaches, such as multicenter training . Material and methods: The purpose of this work was to propose a multicenter sCT synthesis by DL, using a 2D cycle-GAN on 128 prostate cancer patients, from four different centers. Four cases were compared: monocenter cases, monocenter training and test on another center, multicenter trainings and a test on a center not included in the training and multicenter trainings with an included center in the test. Trainings were performed using 20 patients. sCT accuracy evaluation was performed using Mean Absolute Error, Mean Error and Peak-Signal-to-Noise-Ratio. Dose accuracy was assessed with gamma index and Dose Volume Histogram comparison. Results: Qualitative, quantitative and dose results show that the accuracy of sCTs for monocenter trainings and multicenter trainings using a seen center in the test did not differ significantly. However, when the test involved an unseen center, the sCT quality was inferior. Conclusions: The aim of this work was to propose generalizable multicenter training for MR-to-CT synthesis. It was shown that only a few data from one center included in the training cohort allows sCT accuracy equivalent to a monocenter study.

Practical contouring guidelines with an MR-based atlas of brainstem structures involved in radiation-induced nausea and vomiting.

BACKGROUND AND PURPOSE: The objective of this project was to define consensus guidelines for delineating brainstem substructures (dorsal vagal complex, including the area postrema) involved in radiation-induced nausea and vomiting (RINV). The three parts of the brainstem are rarely delineated, so this study was also an opportunity to find a consensus on this subject. MATERIALS AND METHODS: The dorsal vagal complex (DVC) was identified on autopsy sections and endoscopic descriptions. Anatomic landmarks and boundaries were used to establish radio-anatomic correlations on CT and Magnetic Resonance Imaging (MRI). Additionally, delineation of RINV structures was performed on MRI images and reported on CT scans. Next, guidelines were provided to eight radiation oncologists for delineation guidance of these RINV-related structures on DICOM-RT images of two patients being treated for a nasopharyngeal carcinoma. Interobserver variability was computed. RESULTS: The DVC and the three parts of the brainstem were defined with a concise description of their main anatomic boundaries. The interobserver analysis showed that the DVC, the midbrain, the pons, and the medulla oblongata delineations were reproducible with KI = 0.72, 0.84, 0.94 and 0.89, respectively. The Supplemental Material section provides an atlas of the consensus guidelines projected on 1-mm MR axial slices. CONCLUSIONS: This RINV-atlas was feasible and reproducible for the delineation of RINV structures on planning CT using fused MRI. It may be used to prospectively assess dose-volume relationship for RINV structures and occurrence of nausea vomiting during intracranial or head and neck irradiation.

Proton therapy for locally advanced breast cancer: A systematic review of the literature.

BACKGROUND: Radiation therapy plays a major role in the management of adjuvant breast cancer with nodal involvement, with an iatrogenic increase of cardio-vascular risk. Photon therapy, even with intensity modulation, has the downsides of high mean heart dose and heterogeneous target coverage, particularly in the case of internal mammary irradiation. This systematic review of the literature aims to evaluate proton therapy in locally advanced breast cancer. MATERIAL AND METHODS: PubMed was searched for original full-text articles with the following search terms: «Proton Therapy¼ and «Breast Cancer¼. On-going trials were collected using the words "Breast Cancer" and "Protons". RESULTS: 13 articles met the criteria: 6 with passive proton therapy (Double Scattering), 5 with Pencil Beam Scanning (PBS) and 2 with a combination of both. Proton therapy offered a better target coverage than photons, even compared with intensity modulation radiation therapy (including static or rotational IMRT or tomotherapy). With proton therapy, volumes receiving 95% of the dose were around 98%, with low volumes receiving 105% of the dose. Proton therapy often decreased mean heart dose by a factor of 2 or 3, i.e. 1 Gy with proton therapy versus 3 Gy with conventional 3D, and 6 Gy for IMRT. Lungs were better spared with proton therapy than with photon therapy. Cutaneous toxicity observed with double scattering is improved with PBS. CONCLUSION: Proton therapy reduces mean heart dose in breast cancer irradiation, probably reducing late cardio-vascular toxicity. Large clinical studies will likely confirm a clinical benefit of proton therapy.