IBIgor Bessieres

Igor Bessieres

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Igor Bessieres has a total of 38 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Impact of phantom design on fetal radiation exposure assessment in pregnant women undergoing head and neck 6MV photon radiotherapy

    by Magali Edouard, Axel Martin, Michel Chea, Catherine Jenny, Yoann Ristic, Miray Razanajatovo, Aurélie Isambert, Igor Bessières, Christelle Huet

  • Open Access

    A case report of radiotherapy for head and neck cancer in a pregnant woman: dosimetric optimization and in vivo dosimetry

    by Axel Martin, David Thibouw, Laurent Delcoudert, Léone Aubignac, Karine Peignaux, Christelle Huet, Magali Edouard, Jérémy Baude, Igor Bessieres

  • Open Access

    In-depth understanding of the multimodal deformable image registration tool of a 0.35 T MR-linac

    by Pierre Bugnon, Igor Bessieres, Alain Lalande, Aurélie Petitfils, Louis Marage, Léone Aubignac, Jérémy Baude, Ilyas Achag, Julien Boudet

  • Open Access

    Benchmarking MapRT and first clinical experience: A novel solution for collision‐free non‐coplanar treatment planning

    by Mathieu Gonod, Ilyas Achag, Jad Farah, Léone Aubignac, Igor Bessieres

    In recent years, complex re‐irradiations and stereotactic treatments have triggered the use of non‐coplanar treatments for better dose conformality, entailing risks of collision between the machine and the patient, couch, or immobilization device. To ensure the plans deliverability without collisions, time‐consuming actions are typically performed, including dry runs, in‐room couch rotations, and beam configuration tests during planning. To overcome these challenges, a new tool called MapRT (VisionRT Ltd., London, UK) was developed. MapRT predicts a clearance map based on a patients' 3D model (acquired with dedicated cameras at the CT simulation) and pre‐established machine models. This work evaluates the accuracy of MapRT using a 30 × 35 × 40 cm3 phantom and 64 gantry/couch collision coordinates on a Truebeam Linac (Varian, Palo Alto, USA). Collision coordinates were recorded for gantry and couch rotations. The agreement of real collision coordinates and MapRT's predictions was evaluated for different buffer margins around the couch/patient models customizable in MapRT. Results of the first clinical implementation of MapRT were also reported. With no buffer margin, MapRT's predictions and experimental collision coordinates showed small average differences but with large standard deviations for gantry (‐0.5°±6.2°) and couch (‐0.1°±4.8°) collision coordinates. When excluding the kV imaging components, these values were of ‐0.8°±3.5° for gantry and 0.4°±4.4° for couch. Finally, a 3 cm buffer margin allows for 100% accurate predictions by MapRT of gantry‐to‐phantom and gantry‐to‐couch collisions. Among the ∼900 treatment plans checked with MapRT, 22 collisions could be avoided while another 6 plans still incurred a collision but these are mainly due to users' oversights. MapRT easily predict collisions in complex treatment planning. This work demonstrated its reliability using a 3 cm buffer margin. MapRT is a promising tool for increasing security, time saving and workflow improvement.


  • Open Access

    Deep Learning Based on EfficientNet for Multiorgan Segmentation of Thoracic Structures on a 0.35 T MR-Linac Radiation Therapy System

    by Mohammed Chekroun, Youssef Mourchid, Igor Bessières, Alain Lalande

    The advent of the 0.35 T MR-Linac (MRIdian, ViewRay) system in radiation therapy allows precise tumor targeting for moving lesions. However, the lack of an automatic volume segmentation function in the MR-Linac’s treatment planning system poses a challenge. In this paper, we propose a deep-learning-based multiorgan segmentation approach for the thoracic region, using EfficientNet as the backbone for the network architecture. The objectives of this approach include accurate segmentation of critical organs, such as the left and right lungs, the heart, the spinal cord, and the esophagus, essential for minimizing radiation toxicity during external radiation therapy. Our proposed approach, when evaluated on an internal dataset comprising 81 patients, demonstrated superior performance compared to other state-of-the-art methods. Specifically, the results for our approach with a 2.5D strategy were as follows: a dice similarity coefficient (DSC) of 0.820 ± 0.041, an intersection over union (IoU) of 0.725 ± 0.052, and a 3D Hausdorff distance (HD) of 10.353 ± 4.974 mm. Notably, the 2.5D strategy surpassed the 2D strategy in all three metrics, exhibiting higher DSC and IoU values, as well as lower HD values. This improvement strongly suggests that our proposed approach with the 2.5D strategy may hold promise in achieving more precise and accurate segmentations when compared to the conventional 2D strategy. Our work has practical implications in the improvement of treatment planning precision, aligning with the evolution of medical imaging and innovative strategies for multiorgan segmentation tasks.


  • Open Access

    Automatic Multiorgan Segmentation in Pelvic Region with Convolutional Neural Networks on 0.35 T MR-Linac Images

    by Emmanouil Koutoulakis, Louis Marage, Emmanouil Markodimitrakis, Leone Aubignac, Catherine Jenny, Igor Bessieres, Alain Lalande

    MR-Linac is a recent device combining a linear accelerator with an MRI scanner. The improved soft tissue contrast of MR images is used for optimum delineation of tumors or organs at risk (OARs) and precise treatment delivery. Automatic segmentation of OARs can contribute to alleviating the time-consuming process for radiation oncologists and improving the accuracy of radiation delivery by providing faster, more consistent, and more accurate delineation of target structures and organs at risk. It can also help reduce inter-observer variability and improve the consistency of contouring while reducing the time required for treatment planning. In this work, state-of-the-art deep learning techniques were evaluated based on 2D and 2.5D training strategies to develop a comprehensive tool for the accurate segmentation of pelvic OARs dedicated to 0.35 T MR-Linac. In total, 103 cases with 0.35 T MR images of the pelvic region were investigated. Experts considered and contoured the bladder, rectum, and femoral heads as OARs and the prostate as the target volume. For the training of the neural network, 85 patients were randomly selected, and 18 were used for testing. Multiple U-Net-based architectures were considered, and the best model was compared using both 2D and 2.5D training strategies. The evaluation of the models was performed based on two metrics: the Dice similarity coefficient (DSC) and the Hausdorff distance (HD). In the 2D training strategy, Residual Attention U-Net (ResAttU-Net) had the highest scores among the other deep neural networks. Due to the additional contextual information, the configured 2.5D ResAttU-Net performed better. The overall DSC were 0.88 ± 0.09 and 0.86 ± 0.10, and the overall HD was 1.78 ± 3.02 mm and 5.90 ± 7.58 mm for 2.5D and 2D ResAttU-Net, respectively. The 2.5D ResAttU-Net provides accurate segmentation of OARs without affecting the computational cost. The developed end-to-end pipeline will be merged with the treatment planning system for in-time automatic segmentation.


  • Open Access

    Online adaptive radiotherapy and dose delivery accuracy: A retrospective analysis

    by Igor Bessieres, Olivier Lorenzo, Aurélie Bertaut, Aurélie Petitfils, Léone Aubignac, Julien Boudet

    Purpose: With online adaptive radiotherapy (ART), patient‐specific quality assurance (PSQA) testing cannot be performed prior to delivery of the adapted treatment plan. Consequently, the dose delivery accuracy of adapted plans (i.e., the ability of the system to interpret and deliver the treatment as planned) are not initially verified. We investigated the variation in dose delivery accuracy of ART on the MRIdian 0.35 T MR‐linac (Viewray Inc., Oakwood, USA) between initial plans and their respective adapted plans, by analyzing PSQA results.Methods: We considered the two main digestive localizations treated with ART (liver and pancreas). A total of 124 PSQA results acquired with the ArcCHECK (Sun Nuclear Corporation, Melbourne, USA) multidetector system were analyzed. PSQA result variations between the initial plans and their respective adapted plans were statistically investigated and compared with the variation in MU number.Results: For the liver, limited deterioration in PSQA results was observed, and was within the limits of clinical tolerance (Initial = 98.2%, Adapted = 98.2%, p = 0.4503). For pancreas plans, only a few significant deteriorations extending beyond the limits of clinical tolerance were observed and were due to specific, complex anatomical configurations (Initial = 97.3%, Adapted = 96.5%, p = 0.0721). In parallel, we observed an influence of the increase in MU number on the PSQA results.Conclusion: We show that the dose delivery accuracy of adapted plans, in terms of PSQA results, is preserved in ART processes on the 0.35 T MR‐linac. Respecting good practices, and minimizing the increase in MU number can help to preserve the accuracy of delivery of adapted plans as compared to their respective initial plans.


  • Characterisation of a split gradient coil design induced systemic imaging artefact on 0.35 T MR-linac systems

    by Louis Marage, Paul-Michael Walker, Julien Boudet, Pierre Fau, Pierre Debuire, Emmanuelle Clausse, Aurélie Petitfils, Léone Aubignac, Stanislas Rapacchi, Igor Bessieres

    Objective . The aim of this work was to highlight and characterize a systemic ‘star-like’ artefact inherent to the low field 0.35 T MRIdian MR-linac system, a magnetic resonance guided radiotherapy device. This artefact is induced by the original split gradients coils design. This design causes a surjection of the intensity gradient in Z (or head-feet) direction. This artefact appears on every sequence with phase encoding in the head-feet direction. Approach . Basic gradient echo sequence and clinical mandatory bSSFP sequence were used. Three setups using manufacturer provided QA phantoms were designed: two including the linearity control grid used for the characterisation and a third including two homogeneity control spheres dedicated to the artefact management in a more clinical like situation. The presence of the artefact was checked in four different MRidian sites. The tested parameters based on the literature were: phase encoding orientation, slab selectivity, excitation bandwidth (BW RF ), acceleration factor (R) and phase/slab oversampling (PO/SO). Main results . The position of this artefact is constant and reproducible over the tested MRIdian sites. The typical singularity saturated dot or star is visible even with the 3D slab-selection enabled. A management is proposed by decreasing the BW RF , the R in head-feet direction and increasing the PO/SO. The oversampling can be optimized using a formula to anticipate the location of artefact in the field of view. Significance . The star-like artefact has been well characterised. A manageable solution comes at the cost of acquisition time. Observed in clinical cases, the artefact may degrade the images used for the RT planning and repositioning during the treatment unless corrected.