Anna M. Dinkla

Deep MR to CT Synthesis Using Unpaired Data

MR-only radiotherapy treatment planning requires accurate MR-to-CT synthesis. Current deep learning methods for MR-to-CT synthesis depend on pairwise aligned MR and CT training images of the same patient. However, misalignment between paired images could lead to errors in synthesized CT images. To overcome this, we propose to train a generative adversarial network (GAN) with unpaired MR and CT images. A GAN consisting of two synthesis convolutional neural networks (CNNs) and two discriminator CNNs was trained with cycle consistency to transform 2D brain MR image slices into 2D brain CT image slices and vice versa. Brain MR and CT images of 24 patients were analyzed. A quantitative evaluation showed that the model was able to synthesize CT images that closely approximate reference CT images, and was able to outperform a GAN model trained with paired MR and CT images.

A comparison of inverse optimization algorithms for HDR/PDR prostate brachytherapy treatment planning

PURPOSE Graphical optimization (GrO) is a common method for high-dose-rate/pulsed-dose-rate (PDR) prostate brachytherapy treatment planning. New methods performing inverse optimization of the dose distribution have been developed over the past years. The purpose is to compare GrO and two established inverse methods, inverse planning simulated annealing (IPSA) and hybrid inverse treatment planning and optimization (HIPO), and one new method, enhanced geometric optimization-interactive inverse planning (EGO-IIP), in terms of speed and dose-volume histogram (DVH) parameters. METHODS AND MATERIALS For 26 prostate cancer patients treated with a PDR brachytherapy boost, an experienced treatment planner optimized the dose distributions using four different methods: GrO, IPSA, HIPO, and EGO-IIP. Relevant DVH parameters (prostate-V100%, D90%, V150%; urethra-D(0.1cm3) and D(1.0cm3); rectum-D(0.1cm3) and D(2.0cm3); bladder-D(2.0cm3)) were evaluated and their compliance to the constraints. Treatment planning time was also recorded. RESULTS All inverse methods resulted in shorter planning time (mean, 4-6.7 min), as compared with GrO (mean, 7.6 min). In terms of DVH parameters, none of the inverse methods outperformed the others. However, all inverse methods improved on compliance to the planning constraints as compared with GrO. On average, EGO-IIP and GrO resulted in highest D90%, and the IPSA plans resulted in lowest bladder D2.0cm3 and urethra D(1.0cm3). CONCLUSIONS Inverse planning methods decrease planning time as compared with GrO for PDR/high-dose-rate prostate brachytherapy. DVH parameters are comparable for all methods.

Improved tumour control probability with MRI-based prostate brachytherapy treatment planning

Abstract Backgroun. Due to improved visibility on MRI, contouring of the prostate is improved compared to CT. The aim of this study was to quantify the benefits of using MRI for treatment planning as compared to CT-based planning for temporary implant prostate brachytherapy. Material and methods. CT and MRI image data of 13 patients were used to delineate the prostate and organs at risk (OARs) and to reconstruct the implanted catheters (typically 12). An experienced treatment planner created plans on the CT-based structure sets (CT-plan) and on the MRI-based structure sets (MRI-plan). Then, active dwell-positions and weights of the CT-plans were transferred to the MRI-based structure sets (CT-planMRI-contours) and resulting dosimetric parameters and tumour control probabilities (TCPs) were studied. Results. For the CT-planMRI-contours a statistically significant lower target coverage was detected: mean V100 was 95.1% as opposed to 98.3% for the original plans (p < 0.01). Planning on CT caused cold-spots that influence the TCP. MRI-based planning improved the TCPs by 6–10%, depending on the parameters of the radiobiological model used for TCP calculation. Basing the treatment plan on either CT- or MRI-delineations does not influence plan quality. Conclusion. Evaluation of CT-based treatment planning by transferring the plan to MRI reveals underdosage of the prostate, especially at the base side. Planning on MRI can prevent cold-spots in the tumour and improves the TCP.

Deviations from the planned dose during 48 hours of stepping source prostate brachytherapy caused by anatomical variations

BACKGROUND AND PURPOSE To determine the uncertainties in planned dose associated with catheter and organ movement during 48 hours of stepping source prostate brachytherapy. MATERIAL AND METHODS Pulsed-dose rate (PDR) prostate brachytherapy as a boost is given in 24 pulses every 2 hours, making the total treatment last 48 hours. The entire treatment is based on one plan, created on the planning CT (CT1). Two follow-up CTs (CT2 and CT3) were acquired; halfway through the treatment and at the end of treatment. On these repeat scans the catheters were reconstructed and PTV and OARs were delineated. The original treatment plan was calculated on the repeat CTs. Target coverage V(100%), D(90), dose to 2cm(3) (D2cm(3)) of the rectum and bladder and dose to 0.1cm(3) of the urethra were recorded from the recalculated DVHs. RESULTS On the two repeat CTs the V100% decreased -1.5% and -2.3% as compared to the planning CT. For the rectum D2cm(3), the average increase was 14.8% (CT1-CT2) and 17.3% (CT1-CT3). Increase in bladder D2cm(3) was on average 23.1% (CT1-CT2) and 24.8% (CT1-CT3). For the urethra D0.1cm(3) an average decrease of -2% (CT1-CT2) and -3.2% (CT2-CT3) was observed. CONCLUSIONS Changes in target coverage during treatment were small and considered clinically irrelevant. However, an overall increase in dose to the OARs was found as compared to the planned dose, which should be taken into account during treatment planning.

Novel tools for stepping source brachytherapy treatment planning: Enhanced geometrical optimization and interactive inverse planning

PURPOSE Dose optimization for stepping source brachytherapy can nowadays be performed using automated inverse algorithms. Although much quicker than graphical optimization, an experienced treatment planner is required for both methods. With automated inverse algorithms, the procedure to achieve the desired dose distribution is often based on trial-and-error. METHODS A new approach for stepping source prostate brachytherapy treatment planning was developed as a quick and user-friendly alternative. This approach consists of the combined use of two novel tools: Enhanced geometrical optimization (EGO) and interactive inverse planning (IIP). EGO is an extended version of the common geometrical optimization method and is applied to create a dose distribution as homogeneous as possible. With the second tool, IIP, this dose distribution is tailored to a specific patient anatomy by interactively changing the highest and lowest dose on the contours. RESULTS The combined use of EGO-IIP was evaluated on 24 prostate cancer patients, by having an inexperienced user create treatment plans, compliant to clinical dose objectives. This user was able to create dose plans of 24 patients in an average time of 4.4 min/patient. An experienced treatment planner without extensive training in EGO-IIP also created 24 plans. The resulting dose-volume histogram parameters were comparable to the clinical plans and showed high conformance to clinical standards. CONCLUSIONS Even for an inexperienced user, treatment planning with EGO-IIP for stepping source prostate brachytherapy is feasible as an alternative to current optimization algorithms, offering speed, simplicity for the user, and local control of the dose levels.

SU‐E‐T‐326: Repeated CT‐Scans in Pulsed Doserate Prostate Brachytherapy: Assessment of Deviations from the Treatment Plan

Purpose: Estimation of deviations between planned and delivered dose in Pulsed Doserate (PDR) brachytherapy for prostate cancer.Methods: A boost of 28.8 Gy is given with PDR brachytherapy in addition to 46 Gy delivered with External Beam RT. Brachytherapy is given in 24 pulses of 120 cGy, with an interpulse period time of 2.0 hours, resulting in a treatment time of over 46 hours. For 31 patients, additional CT‐scans were made apart from the Treatment PlanningCT, i.e., one at 24 hours after start of PDR treatment and one shortly before finishing PDR treatment. On the second and third CT, the brachytherapy catheters were newly reconstructed and the treating physician delineated the PTV and organs at risk. Dwell positions and dwell times as used for the original Treatment Plan were imported into the newly reconstructed catheters and the dose distribution was recalculated. Plan comparison parameters were prostate V100 and D90 and rectum and bladder D2cc. Results: Averaged over 3 CT scans and all patients, the prostate V100 decreased 1.2% and D90 decreased 2.7%. For rectum, D2cc was within the tolerance dose (96 cGy/pulse) for all patients on the planning CT, but exceeded the tolerance dose on scan #2 in 7/31 patients with maximally 46% and in scan #3 in 5/25 patients with maximally 29%. Also for bladder D2cc was within the tolerance dose (120 cGy/pulse) for all patients on the planning CT. Here, the tolerance dose was exceeded on scan #2 in 2/31 patients with maximally 14% and in scan #3 in 3/25 patients with maximally 28%. Conclusions: In PDR prostate brachytherapy the relatively long treatment time has no clinically relevant deteriorating effect on the dosimetric quality of the treatment. PTV dose hardly deviates from planned dose, while OAR tolerance doses are rarely exceeded and only in small volumes.