DiffuseRT: predicting likely anatomical deformations of patients undergoing radiotherapy.

Objective. Predicting potential deformations of patients can improve radiotherapy treatment planning. Here, we introduce new deep-learning models that predict likely anatomical changes during radiotherapy for head and neck cancer patients.Approach. Denoising diffusion probabilistic models (DDPMs) were developed to generate fraction-specific anatomical changes based on a reference cone-beam CT (CBCT), the fraction number and the dose distribution delivered. Three distinct DDPMs were developed: (1) theimage modelwas trained to directly generate likely future CBCTs, (2) the deformable vector field (DVF) model was trained to generate DVFs that deform a reference CBCT and (3) thehybrid modelwas trained similarly to the DVF model, but without relying on an external deformable registration algorithm. The models were trained on 9 patients with longitudinal CBCT images (224 CBCTs) and evaluated on 5 patients (152 CBCTs).Results. The generated images mainly exhibited random positioning shifts and small anatomical changes for early fractions. For later fractions, all models predicted weight losses in accordance with the training data. The distributions of volume and position changes of the body, esophagus, and parotids generated with the image and hybrid models were more similar to the ground truth distribution than the DVF model, evident from the lower Wasserstein distance achieved with the image (0.33) and hybrid model (0.30) compared to the DVF model (0.36). Generating several images for the same fraction did not yield the expected variability since the ground truth anatomical changes were only in 76% of the fractions within the 95% bounds predicted with the best model. Using the generated images for robust optimization of simplified proton therapy plans improved the worst-case clinical target volume V95 with 7% compared to optimizing with 3 mm set-up robustness while maintaining a similar integral dose.Significance. The newly developed DDPMs generate distributions similar to the real anatomical changes and have the potential to be used for robust anatomical optimization.

Dose and robustness comparison of nominal, daily and accumulated doses for photon and proton treatment of sinonasal cancer.

INTRODUCTION: The aim was to evaluate and compare the dosimetric effect and robustness towards day-to-day anatomical and setup variations in the delivered dose for photon and proton treatments of sinonasal cancer (SNC) patients. MATERIALS AND METHODS: Photon (VMAT) and proton (IMPT) plans were optimized retrospectively for 24 SNC patients. Synthetic CTs (synCT) were obtained by deforming the planning CT (pCT) to the anatomy of every daily cone-beam CT. Both VMAT and IMPT plans were recalculated on the synCTs. The recalculated daily dose was accumulated over the whole treatment on the pCT. Target coverage and dose to organs and risk (OARs) were evaluated for all patients for the nominal, daily and accumulated dose distribution. RESULTS: In general, dose to OARs farther away from the target, including brain, chiasm and contralateral optic nerve, was lower for proton plans than photon plans. Whereas, OARs in proximity of the target received a lower dose for photon plans. For proton plans, the target coverage (volume of CTV receiving 95% of prescribed dose), V95%, fell below 99% for 9/24 patients in one or more fractions. For photon plans, 4/24 patients had one or more fractions where V95% fell below 99%. For accumulated doses, V95% was below 99% only in two cases, but above 98% for all patients. CONCLUSION: Photon and proton treatment have different strengths regarding OAR sparing. The robustness was high for both treatment modalities. Patient selection for either proton or photon radiation therapy of SNC patients should be based on a case-by-case comparison.

Air variability in maxillary sinus during radiotherapy for sinonasal carcinoma.

INTRODUCTION: The aim was to characterise patterns and predictability of aeration changes in the ipsilateral maxillary sinus during intensity-modulated radiotherapy (IMRT) for sinonasal cancer (SNC), and in a sample evaluate the dosimetric effects of aeration changes for both photon and proton therapy. MATERIALS AND METHODS: The study included patients treated with IMRT for SNC in a single institution in 2009-2017. The volume of air in the ipsilateral maxillary sinus was recorded in 1578 daily cone beam computer tomography (CBCT) from 53 patients. Patterns of changing air volumes were categorised as 'stable', increasing', 'decreasing', or 'erratic'. For the prediction analysis, categorisation was performed based both on the entire treatment course and the first five fractions (F1-5). Photon and proton therapy plans were generated for four patients, the one from each category with the largest aeration variation. Synthetic CT images were generated for each CBCT and all plans were recalculated on the daily synthetic CTs. RESULTS: The absolute volume of air varied considerably during the treatment course, ranging from 0 to 25.9 cm3. Changes within a single participant varied in the range of 0-18.7 cm3. In the categorisation of patterns, most patients had increasing aeration of the sinus. Generally, patterns of aeration could not be predicted from F1-5. Patients categorised as increasing in F1-5 had the best prediction, with 78% predicted correctly as increasing for the entire treatment course. The numeric correlation coefficients for target coverage and air volume were low for 3/4 scenarios (photons 0.03-0.23, protons 0.26-0.48). No straightforward correlation between the dosimetric effect and the volume changes could be detected in the sample test of four patients for neither photon nor proton therapy. CONCLUSION: The variation of aeration was large and unpredictable. No clear dosimetric consequences of the aeration variation were evident for neither IMRT nor proton therapy for the patients investigated.

Image-guided radiotherapy and motion management in lung cancer.

In this review, image guidance and motion management in radiotherapy for lung cancer is discussed. Motion characteristics of lung tumours and image guidance techniques to obtain motion information are elaborated. Possibilities for management of image guidance and motion in the various steps of the treatment chain are explained, including imaging techniques and beam delivery techniques. Clinical studies using different motion management techniques are reviewed, and finally future directions for image guidance and motion management are outlined.

Interobserver delineation variation in lung tumour stereotactic body radiotherapy.

OBJECTIVES: In radiotherapy, delineation uncertainties are important as they contribute to systematic errors and can lead to geographical miss of the target. For margin computation, standard deviations (SDs) of all uncertainties must be included as SDs. The aim of this study was to quantify the interobserver delineation variation for stereotactic body radiotherapy (SBRT) of peripheral lung tumours using a cross-sectional study design. METHODS: 22 consecutive patients with 26 tumours were included. Positron emission tomography/CT scans were acquired for planning of SBRT. Three oncologists and three radiologists independently delineated the gross tumour volume. The interobserver variation was calculated as a mean of multiple SDs of distances to a reference contour, and calculated for the transversal plane (SD(trans)) and craniocaudal (CC) direction (SD(cc)) separately. Concordance indexes and volume deviations were also calculated. RESULTS: Median tumour volume was 13.0 cm(3), ranging from 0.3 to 60.4 cm(3). The mean SD(trans) was 0.15 cm (SD 0.08 cm) and the overall mean SD(cc) was 0.26 cm (SD 0.15 cm). Tumours with pleural contact had a significantly larger SD(trans) than tumours surrounded by lung tissue. CONCLUSIONS: The interobserver delineation variation was very small in this systematic cross-sectional analysis, although significantly larger in the CC direction than in the transversal plane, stressing that anisotropic margins should be applied. This study is the first to make a systematic cross-sectional analysis of delineation variation for peripheral lung tumours referred for SBRT, establishing the evidence that interobserver variation is very small for these tumours.

Methods for estimating the site of origin of locoregional recurrence in head and neck squamous cell carcinoma.

PURPOSE: Methods to estimate the likely origin of recurrences after radiation therapy for head and neck squamous cell carcinoma are compared. METHODS AND MATERIALS: A total of 25 patients meeting the following inclusion criteria were randomly selected: curatively intended intensity-modulated radiotherapy planned on a positron emission tomography-computed tomography (PET/CT) scan during the period 2005-2009; squamous cell carcinoma in the oral cavity, pharynx or larynx; complete clinical response followed by locoregional recurrence; and a CT scan at recurrence before any salvage therapy. Exclusion criteria were previous cancer in the area, surgery prior to radiotherapy, or a synchronous cancer. Three methods of estimating focal points of recurrence origin and two volume overlap methods assigning the recurrences to the most central target volumes encompassing at least 50% or 95% of the recurrence volumes were tested. Treatment planning and recurrence scans were rigid and deformable co-registered in order to transfer focal points to the treatment planning scan. Double determinations of all volumes, points, and co-registrations were made. RESULTS: The volume overlap methods assigned the recurrences to significantly more peripheral target volumes than focal methods (p < 0.0001 for all comparisons of 95% overlap vs. focal methods, p < 0.028 for all comparisons of 50% overlap vs. focal methods). Repeated registrations of the same point had higher reproducibility with deformable registration than with rigid registration (median distance 0.31 vs. 0.35 cm, p = 0.015). No significant differences were observed among the focal methods. CONCLUSION: Significant differences between methods were found which may affect strategies to improve radiotherapy based on pattern of failure analyses.

Evaluation of dose to cardiac structures during breast irradiation.

OBJECTIVE: Adjuvant radiotherapy for breast cancer can lead to late cardiac complications. The highest radiation doses are likely to be to the anterior portion of the heart, including the left anterior descending coronary artery (LAD). The purpose of this work was to assess the radiation doses delivered to the heart and the LAD in respiration-adapted radiotherapy of patients with left-sided breast cancer. METHODS: 24 patients referred for adjuvant radiotherapy after breast-conserving surgery for left-sided lymph node positive breast cancer were evaluated. The whole heart, the arch of the LAD and the whole LAD were contoured. The radiation doses to all three cardiac structures were evaluated. RESULTS: For 13 patients, the plans were acceptable based on the criteria set for all 3 contours. For seven patients, the volume of heart irradiated was well below the set clinical threshold whereas a high dose was still being delivered to the LAD. In 1 case, the dose to the LAD was low while 19% of the contoured heart volume received over 20 Gy. In five patients, the dose to the arch LAD was relatively low while the dose to the whole LAD was considerably higher. CONCLUSION: This study indicates that it is necessary to assess the dose delivered to the whole heart as well as to the whole LAD when investigating the acceptability of a breast irradiation treatment. Assessing the dose to only one of these structures could lead to excessive heart irradiation and thereby increased risk of cardiac complications for breast cancer radiotherapy patients.

Comparison of respiratory surrogates for gated lung radiotherapy without internal fiducials.

An investigation was carried out to compare the ability of two respiratory surrogates to mimic actual lung tumor motion during audio coaching. The investigation employed video clips acquired after patients had had fiducial markers implanted in lung tumors to be used for image-guided stereoscopic radiotherapy. The positions of the markers in the clips were measured within the video frames and used as the standard for tumor volume motion. An external marker was tracked optically during the fluoroscopic acquisitions. An image correlation technique was developed to compute a gating signal from the fluoroscopic images. The correlation gating trace was similar to the optical gating trace in the phase regions of the respiratory cycle used for gating. A cross correlation analysis and comparison of the external optical marker gating with internal fluoroscopic gating was performed. The fluoroscopic image correlation surrogate was found to be superior to the external optical surrogate in the AP-views in four out of six cases. In one of the remaining two cases, the two surrogates performed comparably, while in the last case, the external fiducial trace performed best. It was concluded that fluoroscopic gating based on correlation of native image features in the fluoroscopic images will be adequate for respiratory gating.

Modification of anomalous swelling in multilamellar vesicles induced by alkali halide salts.

By use of small-angle X-ray scattering it is shown that addition of alkali halide salts in small amounts (0-200 mM) shifts the repeat spacing in multilamellar DC13PC vesicles and alters the anomalous swelling behaviour close to the main transition. Both effects follow the Hofmeister series of the ions. We suggest that the shift of repeat spacing can be explained by ion effects on the van der Waals attractive forces between the membranes and on the decay length of the repulsive hydration force. The anomalous swelling is explained in terms of a critical unbinding of the membranes. The proximity of the critical temperature of the unbinding to the main transition temperature can be tuned by varying the concentration and type of salt in the sample.