A high-performance method of deep learning for prostate MR-only radiotherapy planning using an optimized Pix2Pix architecture.

PURPOSE: The first aim was to generate and compare synthetic-CT (sCT) images using a conditional generative adversarial network (cGAN) method (Pix2Pix) for MRI-only prostate radiotherapy planning by testing several generators, loss functions, and hyper-parameters. The second aim was to compare the optimized Pix2Pix model with five other architectures (bulk-density, atlas-based, patch-based, U-Net, and GAN). METHODS: For 39 patients treated by VMAT for prostate cancer, T2-weighted MRI images were acquired in addition to CT images for treatment planning. sCT images were generated using the Pix2Pix model. The generator, loss function, and hyper-parameters were tuned to improve sCT image generation (in terms of imaging endpoints). The final evaluation was performed by 3-fold cross-validation. This method was compared to five other methods using the following imaging endpoints: the mean absolute error (MAE) and mean error (ME) between sCT and reference CT images (rCT) of the whole pelvis, bones, prostate, bladder, and rectum. For dose planning analysis, the dose-volume histogram metric differences and 3D gamma analysis (local, 1 %/1 mm) were calculated using the sCT and reference CT images. RESULTS: Compared with the other architectures, Pix2Pix with Perceptual loss function and generator ResNet 9 blocks showed the lowest MAE (29.5, 107.7, 16.0, 13.4, and 49.1 HU for the whole pelvis, bones, prostate, bladder, and rectum, respectively) and the highest gamma passing rates (99.4 %, using the 1 %/1mm and 10 % dose threshold criterion). Concerning the DVH points, the mean errors were -0.2% for the planning target volume V95%, 0.1 % for the rectum V70Gy, and -0.1 % for the bladder V50Gy. CONCLUSION: The sCT images generated from MRI data with the Pix2Pix architecture had the lowest image errors and similar dose uncertainties (in term of gamma pass-rate and dose-volume histogram metric differences) than other deep learning methods.

[What do we need to deliver "online" adapted radiotherapy treatment plans?] / Que faut-il pour faire de la radiothérapie adaptative « online ¼ ?

During the joint SFRO/SFPM session of the 2019 congress, a state of the art of adaptive radiotherapy announced a strong impact in our clinical practice, in particular with the availability of treatment devices coupled to an MRI system. Three years later, it seems relevant to take stock of adaptive radiotherapy in practice, and especially the "online" strategy because it is indeed more and more accessible with recent hardware and software developments, such as coupled accelerators to a three-dimensional imaging device and algorithms based on artificial intelligence. However, the deployment of this promising strategy is complex because it contracts the usual time scale and upsets the usual organizations. So what do we need to deliver adapted treatment plans with an "online" strategy?

Deep learning methods to generate synthetic CT from MRI in radiotherapy: A literature review.

PURPOSE: In radiotherapy, MRI is used for target volume and organs-at-risk delineation for its superior soft-tissue contrast as compared to CT imaging. However, MRI does not provide the electron density of tissue necessary for dose calculation. Several methods of synthetic-CT (sCT) generation from MRI data have been developed for radiotherapy dose calculation. This work reviewed deep learning (DL) sCT generation methods and their associated image and dose evaluation, in the context of MRI-based dose calculation. METHODS: We searched the PubMed and ScienceDirect electronic databases from January 2010 to March 2021. For each paper, several items were screened and compiled in figures and tables. RESULTS: This review included 57 studies. The DL methods were either generator-only based (45% of the reviewed studies), or generative adversarial network (GAN) architecture and its variants (55% of the reviewed studies). The brain and pelvis were the most commonly investigated anatomical localizations (39% and 28% of the reviewed studies, respectively), and more rarely, the head-and-neck (H&N) (15%), abdomen (10%), liver (5%) or breast (3%). All the studies performed an image evaluation of sCTs with a diversity of metrics, with only 36 studies performing dosimetric evaluations of sCT. CONCLUSIONS: The median mean absolute errors were around 76 HU for the brain and H&N sCTs and 40 HU for the pelvis sCTs. For the brain, the mean dose difference between the sCT and the reference CT was <2%. For the H&N and pelvis, the mean dose difference was below 1% in most of the studies. Recent GAN architectures have advantages compared to generator-only, but no superiority was found in term of image or dose sCT uncertainties. Key challenges of DL-based sCT generation methods from MRI in radiotherapy is the management of movement for abdominal and thoracic localizations, the standardization of sCT evaluation, and the investigation of multicenter impacts.

PET and MRI guided adaptive radiotherapy: Rational, feasibility and benefit.

Adaptive radiotherapy (ART) corresponds to various replanning strategies aiming to correct for anatomical variations occurring during the course of radiotherapy. The goal of the article was to report the rational, feasibility and benefit of using PET and/or MRI to guide this ART strategy in various tumor localizations. The anatomical modifications defined by scanner taking into account tumour mobility and volume variation are not always sufficient to optimise treatment. The contribution of functional imaging by PET or the precision of soft tissue by MRI makes it possible to consider optimized ART. Today, the most important data for both PET and MRI are for lung, head and neck, cervical and prostate cancers. PET and MRI guided ART appears feasible and safe, however in a very limited clinical experience. Phase I/II studies should be therefore performed, before proposing cost-effectiveness comparisons in randomized trials and before using the approach in routine practice.

Implementation of an optimization method for parotid gland sparing during inverse planning for head and neck cancer radiotherapy.

PURPOSE: To guide parotid gland (PG) sparing at the dose planning step, a specific model based on overlap between PTV and organ at risk (Moore et al.) was developed and evaluated for VMAT in head-and-neck (H&N) cancer radiotherapy. MATERIALS AND METHODS: One hundred and sixty patients treated for locally advanced H&N cancer were included. A model optimization was first performed (20 patients) before a model evaluation (110 patients). Thirty cases were planned with and without the model to quantify the PG dose sparing. The inter-operator variability was evaluated on one case, planned by 12 operators with and without the model. The endpoints were PG mean dose (Dmean), PTV homogeneity and number of monitor units (MU). RESULTS: The PG Dmean predicted by the model was reached in 89% of cases. Using the model significantly reduced the PG Dmean: -6.1±4.3Gy. Plans with the model showed lower PTV dose homogeneity and more MUs (+10.5% on average). For the inter-operator variability, PG dose volume histograms without the optimized model were significantly different compared to those with the model; the Dmean standard deviation for the ipsilateral PG decreased from 2.2Gy to 1.2Gy. For the contralateral PG, this value decreased from 2.9Gy to 0.8Gy. CONCLUSION: During the H&N inverse planning, the optimized model guides to the lowest PG achievable mean dose, allowing a significant PG mean dose reduction of -6.1Gy. Integrating this method at the treatment-planning step significantly reduced the inter-patient and inter-operator variabilities.

[Overview of adaptive radiotherapy in 2019: From implementation to clinical use]. / État des lieux de la radiothérapie adaptative en 2019 : de la mise en place à l'utilisation clinique.

Intensity modulated radiotherapy combined with image guided radiotherapy has led to increase the precision of external beam radiotherapy. However, intra or inter-fraction anatomical variations are frequent during the treatment course and can cause under-dosing of the target volume and/or over-dosing of the organs at risk. Several adaptive radiotherapy (ART) strategies can be defined to compensate these anatomical variations. The purpose of this article is to provide an overview of available ART strategies: offline, online, hybrid (library of treatment plans) or in real-time, while considering the arrival of MR-Linac devices in radiotherapy departments. The tools required to these ART strategies such as auto-segmentation, deformable image registration, calculation of the daily dose or dose accumulation, are also described. Implementing an ART strategy requires a rigorous quality assurance process, at each stage and on the entire workflow, as well as prior organization and training from of all the trades. A strong multidisciplinary involvement is finally required in order to ensure ART treatments.

[Adaptive radiotherapy: Strategies and benefits depending on tumor localization]. / Radiothérapie adaptative : stratégies et bénéfices selon les localisations tumorales.

Adaptive radiotherapy (ART) is a complexe image-guided radiotherapy modality that comprises multiple planning to account for anatomical variations occurring during irradiation. Schematically, two strategies of RTA can be distinguished and combined according to tumor locations. One or more replanning can be proposed to correct systematic variations such as tumor shrinkage. A library of treatment plans with day-to-day plan selection from cone-beam CT imaging can also be proposed to correct random variations such as uterine motion or bladder/rectum volume changes. Because of strong anatomical variations occurring during irradiation, RTA appears therefore particularly justified in head and neck, lung, bladder, cervical and rectum and pancreas tumors, and to a lesser extent for prostate tumors and other digestive tumors. For these tumor locations, ART provides a fairly clear dosimetric benefit but a clinical benefit not yet formally demonstrated. ART cannot be proposed in a routine practice but must be evaluated medico-economically in the context of prospective trials. A rigorous quality control must be associated.

A density assignment method for dose monitoring in head-and-neck radiotherapy.

BACKGROUND AND PURPOSE: During head-and-neck (H&N) radiotherapy, the parotid glands (PGs) may be overdosed; thus, a tool is required to monitor the delivered dose. This study aimed to assess the dose accuracy of a patient-specific density assignment method (DAM) for dose calculation to monitor the dose to PGs during treatment. PATIENTS AND METHODS: Forty patients with H&N cancer received an intensity modulated radiation therapy (IMRT), among whom 15 had weekly CTs. Dose distributions were calculated either on the CTs (CTref), on one-class CTs (1C-CT, water), or on three-class CTs (3C-CT, water-air-bone). The inter- and intra-patient DAM uncertainties were evaluated by the difference between doses calculated on CTref and 1C-CTs or 3C-CTs. PG mean dose (Dmean) and spinal cord maximum dose (D2%) were considered. The cumulated dose to the PGs was estimated by the mean Dmean of the weekly CTs. RESULTS: The mean (maximum) inter-patient DAM dose uncertainties for the PGs (in cGy) were 23 (75) using 1C-CTs and 12 (50) using 3C-CTs (p ≤ 0.001). For the spinal cord, these uncertainties were 118 (245) and 15 (67; p ≤ 0.001). The mean (maximum) DAM dose uncertainty between cumulated doses calculated on CTs and 3C-CTs was 7 cGy (45 cGy) for the PGs. Considering the difference between the planned and cumulated doses, 53% of the ipsilateral and 80% of the contralateral PGs were overdosed by +3.6 Gy (up to 8.2 Gy) and +1.9 Gy (up to 5.2 Gy), respectively. CONCLUSION: The uncertainty of the three-class DAM appears to be clinically non-significant (<0.5 Gy) compared with the PG overdose (up to 8.2 Gy). This DAM could therefore be used to monitor PG doses and trigger replanning.

[External beam radiotherapy cone beam-computed tomography-based dose calculation]. / Calcul de dose de radiothérapie à partir de tomographies coniques : état de l'art.

In external beam radiotherapy, the dose planning is currently based on computed tomography (CT) images. A relation between Hounsfield numbers and electron densities (or mass densities) is necessary for dose calculation taking heterogeneities into account. In image-guided radiotherapy process, the cone beam CT is classically used for tissue visualization and registration. Cone beam CT for dose calculation is also attractive in dose reporting/monitoring perspectives and particularly in a context of dose-guided adaptive radiotherapy. The accuracy of cone beam CT-based dose calculation is limited by image characteristics such as quality, Hounsfield numbers consistency and restrictive sizes of volume acquisition. The analysis of the literature identifies three kinds of strategies for cone beam CT-based dose calculation: establishment of Hounsfield numbers versus densities curves, density override to regions of interest, and deformable registration between CT and cone beam CT images. Literature results show that discrepancies between the reference CT-based dose calculation and the cone beam CT-based dose calculation are often lower than 3%, regardless of the method. However, they can also reach 10% with unsuitable method. Even if the accuracy of the cone beam CT-based dose calculation is independent of the method, some strategies are promising but need improvements in the automating process for a routine implementation.