Challenges drive the state-of-the-art of automated medical image analysis. The quantity of public training data that they provide can limit the performance of their solutions. Public access to the training methodology for these solutions remains absent. This study implements the Type Three (T3) challenge format, which allows for training solutions on private data and guarantees reusable training methodologies. With T3, challenge organizers train a codebase provided by the participants on sequestered training data. T3 was implemented in the STOIC2021 challenge, with the goal of predicting from a computed tomography (CT) scan whether subjects had a severe COVID-19 infection, defined as intubation or death within one month. STOIC2021 consisted of a Qualification phase, where participants developed challenge solutions using 2000 publicly available CT scans, and a Final phase, where participants submitted their training methodologies with which solutions were trained on CT scans of 9724 subjects. The organizers successfully trained six of the eight Final phase submissions. The submitted codebases for training and running inference were released publicly. The winning solution obtained an area under the receiver operating characteristic curve for discerning between severe and non-severe COVID-19 of 0.815. The Final phase solutions of all finalists improved upon their Qualification phase solutions.

Transcatheter aortic valve implantation (TAVI) is a minimally-invasive method for the treatment of aortic valve stenosis in patients with high surgical risk. Despite the success of TAVI, side effects such as paravalvular leakages can occur postoperatively. The goal of this project is to quantitatively analyze the co-occurrence of this complication and several potential risk factors such as stent shape after implantation, implantation height, amount and distribution of calcifications, and contact forces between stent and surrounding structure. In this paper, we present a two-dimensional visualization (stent maps), which allows (1) to comprehensively display all these aspects from CT data and mechanical simulation results and (2) to compare different datasets to identify patterns that are typical for adverse effects. The area of a stent map represents the surface area of the implanted stent - virtually straightened and uncoiled. Several properties of interest, like radial forces or stent compression, are displayed in this stent map in a heatmap-like fashion. Important anatomical landmarks and calcifications are plotted to show their spatial relation to the stent and possible correlations with the color-coded parameters. To provide comparability, the maps of different patient datasets are spatially adjusted according to a corresponding anatomical landmark. Also, stent maps summarizing the characteristics of different populations (e.g. with or without side effects) can be generated. Up to this point several interesting patterns have been observed with our technique, which remained hidden when examining the raw CT data or 3D visualizations of the same data. One example are obvious radial force maxima between the right and non-coronary valve leaflet occurring mainly in cases without leakages. These observations confirm the usefulness of our approach and give starting points for new hypotheses and further analyses. Because of its reduced dimensionality, the stent map data is an appropriate input for statistical group evaluation and machine learning methods.