A novel external beam radiotherapy method for cervical cancer patients using virtual straight or bending boost areas; an in-silico feasibility study.

AIM: To investigate the potential role of a novel spatially fractionated radiation therapy (SFRT) method where heterogeneous dose patterns are created in target areas with virtual rods, straight or curving, of variable position, diameter, separation and alignment personalised to a patient's anatomy. The images chosen for this study were CT scans acquired for the external beam part of radiotherapy. METHODS: Ten patients with locally advanced cervical cancer were retrospectively investigated with SFRT. The dose prescription was 30 Gy in 5 fractions to 90% target volume coverage. Peak-and-valley (SFRT_1) and peak-only (SFRT_2) strategies were applied to generate the heterogeneous dose distributions. The planning objectives for the target (CTV) were D90% ≥ 30 Gy, V45Gy ≥ 50-55% and V60Gy ≥ 30%. The planning objectives for the organs at risk (OAR) were: D2cm3 ≤ 23.75 Gy, 17.0 Gy, 19.5 Gy, 17.0 Gy for the bladder, rectum, sigmoid and bowel, respectively. The plan comparison was performed employing the quantitative analysis of the dose-volume histograms. RESULTS: The D2cm3 was 22.4 ± 2.0 (22.6 ± 2.1) and 13.9 ± 2.9 (13.2 ± 3.0) for the bladder and the rectum for SFRT_1 (SFRT_2). The results for the sigmoid and the bowel were 2.6 ± 3.1 (2.8 ± 3.0) and 9.1 ± 5.9 (9.7 ± 7.3), respectively. The hotspots in the target volume were V45Gy = 43.1 ± 7.5% (56.6 ± 5.6%) and V60Gy = 15.4 ± 5.6% (26.8 ± 6.6%) for SFRT_1 (SFRT_2). To account for potential uncertainties in the positioning, the dose prescription could be escalated to D90% = 33-35 Gy to the CTV without compromising any constraints to the OARs CONCLUSION: In this dosimetric study, the proposed novel planning technique for boosting the cervix uteri was associated with high-quality plans, respecting constraints for the organs at risk and approaching the level of dose heterogeneity achieved with routine brachytherapy. Based on a sample of 10 patients, the results are promising and might lead to a phase I clinical trial.

Using federated data sources and Varian Learning Portal framework to train a neural network model for automatic organ segmentation.

PURPOSE: In this study we trained a deep neural network model for female pelvis organ segmentation using data from several sites without any personal data sharing. The goal was to assess its prediction power compared with the model trained in a centralized manner. METHODS: Varian Learning Portal (VLP) is a distributed machine learning (ML) infrastructure enabling privacy-preserving research across hospitals from different regions or countries, within the framework of a trusted consortium. Such a framework is relevant in the case when there is a high level of trust among the participating sites, but there are legal restrictions which do not allow the actual data sharing between them. We trained an organ segmentation model for the female pelvic region using the synchronous data distributed framework provided by the VLP. RESULTS: The prediction performance of the model trained using the federated framework offered by VLP was on the same level as the performance of the model trained in a centralized manner where all training data was pulled together in one centre. CONCLUSIONS: VLP infrastructure can be used for GPU-based training of a deep neural network for organ segmentation for the female pelvic region. This organ segmentation instance is particularly difficult due to the high variation in the organs' shape and size. Being able to train the model using data from several clinics can help, for instance, by exposing the model to a larger range of data variations. VLP framework enables such a distributed training approach without sharing protected health information.

Creation of knowledge-based planning models intended for large scale distribution: Minimizing the effect of outlier plans.

Knowledge-based planning (KBP) can be used to estimate dose-volume histograms (DVHs) of organs at risk (OAR) using models. The task of model creation, however, can result in estimates with differing accuracy; particularly when outlier plans are not properly addressed. This work used RapidPlan™ to create models for the prostate and head and neck intended for large-scale distribution. Potential outlier plans were identified by means of regression analysis scatter plots, Cook's distance, coefficient of determination, and the chi-squared test. Outlier plans were identified as falling into three categories: geometric, dosimetric, and over-fitting outliers. The models were validated by comparing DVHs estimated by the model with those from a separate and independent set of clinical plans. The estimated DVHs were also used as optimization objectives during inverse planning. The analysis tools lead us to identify as many as 7 geometric, 8 dosimetric, and 20 over-fitting outliers in the raw models. Geometric and over-fitting outliers were removed while the dosimetric outliers were replaced after re-planning. Model validation was done by comparing the DVHs at 50%, 85%, and 99% of the maximum dose for each OAR (denoted as V50, V85, and V99) and agreed within -2% to 4% for the three metrics for the final prostate model. In terms of the head and neck model, the estimated DVHs agreed from -2.0% to 5.1% at V50, 0.1% to 7.1% at V85, and 0.1% to 7.6% at V99. The process used to create these models improved the accuracy for the pharyngeal constrictor DVH estimation where one plan was originally over-estimated by more than twice. In conclusion, our results demonstrate that KBP models should be carefully created since their accuracy could be negatively affected by outlier plans. Outlier plans can be addressed by removing them from the model and re-planning.

Continuum model of cell adhesion and migration.

The motility of cells crawling on a substratum has its origin in a thin cell organ called lamella. We present a 2-dimensional continuum model for the lamella dynamics of a slowly migrating cell, such as a human keratinocyte. The central components of the model are the dynamics of a viscous cytoskeleton capable to produce contractile and swelling stresses, and the formation of adhesive bonds in the plasma cell membrane between the lamella cytoskeleton and adhesion sites at the substratum. We will demonstrate that a simple mechanistic model, neglecting the complicated signaling pathways and regulation processes of a living cell, is able to capture the most prominent aspects of the lamella dynamics, such as quasi-periodic protrusions and retractions of the moving tip, retrograde flow of the cytoskeleton and the related accumulation of focal adhesion complexes in the leading edge of a migrating cell. The developed modeling framework consists of a nonlinearly coupled system of hyperbolic, parabolic and ordinary differential equations for the various molecular concentrations, two elliptic equations for cytoskeleton velocity and hydrodynamic pressure in a highly viscous two-phase flow, with appropriate boundary conditions including equalities and inequalities at the moving boundary. In order to analyse this hybrid continuum model by numerical simulations for different biophysical scenarios, we use suitable finite element and finite volume schemes on a fixed triangulation in combination with an adaptive level set method describing the free boundary dynamics.