Scalable radiotherapy data curation infrastructure for deep-learning based autosegmentation of organs-at-risk: A case study in head and neck cancer.

In this era of patient-centered, outcomes-driven and adaptive radiotherapy, deep learning is now being successfully applied to tackle imaging-related workflow bottlenecks such as autosegmentation and dose planning. These applications typically require supervised learning approaches enabled by relatively large, curated radiotherapy datasets which are highly reflective of the contemporary standard of care. However, little has been previously published describing technical infrastructure, recommendations, methods or standards for radiotherapy dataset curation in a holistic fashion. Our radiation oncology department has recently embarked on a large-scale project in partnership with an external partner to develop deep-learning-based tools to assist with our radiotherapy workflow, beginning with autosegmentation of organs-at-risk. This project will require thousands of carefully curated radiotherapy datasets comprising all body sites we routinely treat with radiotherapy. Given such a large project scope, we have approached the need for dataset curation rigorously, with an aim towards building infrastructure that is compatible with efficiency, automation and scalability. Focusing on our first use-case pertaining to head and neck cancer, we describe our developed infrastructure and novel methods applied to radiotherapy dataset curation, inclusive of personnel and workflow organization, dataset selection, expert organ-at-risk segmentation, quality assurance, patient de-identification, data archival and transfer. Over the course of approximately 13 months, our expert multidisciplinary team generated 490 curated head and neck radiotherapy datasets. This task required approximately 6000 human-expert hours in total (not including planning and infrastructure development time). This infrastructure continues to evolve and will support ongoing and future project efforts.

Impact of physicochemical properties of DNA/PEI complexes on transient transfection of mammalian cells.

Polyethyleneimine (PEI) has been used extensively for transient gene expression (TGE) in mammalian cell cultures. However, the relationship between DNA/PEI complex preparation and their biological activity has not been fully established. Here, a systematic study of DNA/PEI complexes, their physicochemical properties during formation and their transfection efficiency was performed on a virus-like particle (VLP) production platform. The same chemically defined cell culture medium for DNA/PEI complex formation was used as an alternative to simple ionic solutions to minimize changes in complex properties during transfection. Upon formation, an initial concentration of 1E + 10 DNA/PEI complexes/mL underwent partial aggregation with an average size of 300 nm. The participation of NaCl ions in the evolution of complexes was analyzed by X-ray spectroscopy, stressing the relevance of complexing media composition in TGE strategies. After 15 min incubation, 250 complexes plus aggregates per cell were estimated at the time of transfection. Such heterogeneous preparations cannot be easily characterized; subsequently, nanoparticle tracking analysis (NTA) and cryo-electron microscopy were combined to achieve a complete picture of the preparation. Finally, the contribution of each DNA/PEI complex subpopulation was tested by drug inhibition endocytosis. Interestingly, all complexes delivered DNA efficiently and high size aggregates, which enter through macropinocytosis, when inhibited presented a major contribution to transfection efficiency. There is a need to understand the physicochemical factors that participate in DNA delivery protocols. Hence, this study provides new insights into the characterization of DNA/PEI complexes that will assist in more productive and reproducible TGE strategies.

Lipid-based nanovesicles for nanomedicine.

Molecular self-assembly has enabled the fabrication of biologically inspired, advanced nanostructures as lipid-based nanovesicles (L-NVs). The oldest L-NVs, liposomes, have been widely proposed as potential candidates for drug delivery, diagnostic and/or theranostic applications and some liposome-based drug products have already stepped from the lab-bench to the market. This success is attributed to their ability to encapsulate both hydrophobic and/or hydrophilic molecules, efficiently carry and protect them within the body and finally deliver them at the target site. These positive features are also coupled with high biocompatibility. However, liposomes still present some unsolved drawbacks, such as poor colloidal stability, short shelf-life, restricted and expensive conditions of preparation because of the inherent nature of their fundamental constituents (phospholipids). The new tools available in the self-assembly of controlled molecules have significantly advanced the field of L-NV design and synthesis, and non-liposomal L-NVs have been recently developed; this new generation of nanovesicles can represent a paradigm shift in nanomedicine: they may complement liposomes, showing their advantages and overcoming most of their drawbacks. Clearly, being still young, their rocky way to the clinic first and then to the market has just started and it is still long, but they have all the potentialities to reach their objective target. The purpose of this review is to first present the large plethora of L-NVs available, focusing on this new generation of non-liposomal L-NVs and showing their similarities and differences with respect to their ancestors (liposomes). Since the overspread of a nanomaterial to the market is also strongly dependent on the availability of technological-scale preparation methods, we will also extensively review the current approaches exploited for L-NV production. The most cutting-edge approaches based on compressed fluid (CF) technologies will be highlighted here since they show the potential to represent a game-change in the production of L-NVs, favouring their step from the bench to the market. Finally, we will briefly discuss L-NV applications in nanomedicine, looking also for their future perspectives.

Hemodynamic study of intravenous propionyl-L-carnitine in patients with ischemic heart disease and normal left ventricular function.

Propionyl-L-carnitine was given intravenously to ten patients with chronic ischemic heart disease who had normal left ventricular function and had not had a previous myocardial infarction. Subsequently, pulmonary and systemic circulation, left ventricular function, and the relationship between the ventricle and afterload were evaluated. This drug, at a dose of 15 mg/kg, improves ventricular function by easing the load and by enhancing cardiac efficiency. The ejection impedance is reduced with a consequent increase in stroke volume as a result of a) a decrease in systemic and pulmonary resistance and b) an increase in arterial compliance. Arterial pressure is maintained due to an increase in total external heart power. Since the tension time index shows a proportionally smaller increase in the energy requirement, it follows that cardiac efficiency has been improved and ventricle-afterload matching is optimal. These results suggest but do not prove that propionyl-L-carnitine exhibits a positive inotropic property.

Right ventricular dysplasia: angiographic study.

At the moment, the most reliable method for diagnosing right ventricular dysplasia is considered to be angiography. Morphological alterations such as the presence of akinetic/dyskinetic areas, aneurysmatic dilatations and deep anteroapical fissuring, not necessarily associated with an increase in ventricular volume, are understood to be angiographic criteria indicating dysplasia. To verify their diagnostic value, these abnormalities have been evaluated in: (1) 33 patients suspected of having dysplasia because of PVBs with LBBB morphology and with 'borderline' involvement of the right ventricle or without instrumental evidence of cardiac disease (Group A); (2) 16 subjects with no arrhythmia and normal left ventricular angiography, coronary and bioptic findings (Group B); (3) 36 patients with a clinical, angiographic and bioptic diagnosis of dilated idiopathic cardiomyopathy (Group C). In 48.5% of the patients in Group A, angiography showed localized akinesia/dyskinesia (12 patients), small conical outpouchings persisting during systole (10 patients) and apical deep fissuring (two patients). In 81% of these patients, endomyocardial biopsy showed the presence of fibrous and/or adipose tissue in at least 20% of the examined sample. Angiographic abnormalities suggesting dysplasia were found in none of the normal subjects and only in two of the 36 patients with dilated cardiomyopathy (5.5%).