With over 85 million CT scans performed annually in the United States, creating tumor-related reports is a challenging and time-consuming task for radiologists. To address this need, we present RadGPT, an Anatomy-Aware Vision-Language AI Agent for generating detailed reports from CT scans. RadGPT first segments tumors, including benign cysts and malignant tumors, and their surrounding anatomical structures, then transforms this information into both structured reports and narrative reports. These reports provide tumor size, shape, location, attenuation, volume, and interactions with surrounding blood vessels and organs. Extensive evaluation on unseen hospitals shows that RadGPT can produce accurate reports, with high sensitivity/specificity for small tumor (<2 cm) detection: 80/73% for liver tumors, 92/78% for kidney tumors, and 77/77% for pancreatic tumors. For large tumors, sensitivity ranges from 89% to 97%. The results significantly surpass the state-of-the-art in abdominal CT report generation. RadGPT generated reports for 17 public datasets. Through radiologist review and refinement, we have ensured the reports' accuracy, and created the first publicly available image-text 3D medical dataset, comprising over 1.8 million text tokens and 2.7 million images from 9,262 CT scans, including 2,947 tumor scans/reports of 8,562 tumor instances. Our reports can: (1) localize tumors in eight liver sub-segments and three pancreatic sub-segments annotated per-voxel; (2) determine pancreatic tumor stage (T1-T4) in 260 reports; and (3) present individual analyses of multiple tumors--rare in human-made reports. Importantly, 948 of the reports are for early-stage tumors.

MOTIVATION: Studying protein isoforms is an essential step in biomedical research; at present, the main approach for analyzing proteins is via bottom-up mass spectrometry proteomics, which return peptide identifications, that are indirectly used to infer the presence of protein isoforms. However, the detection and quantification processes are noisy; in particular, peptides may be erroneously detected, and most peptides, known as shared peptides, are associated to multiple protein isoforms. As a consequence, studying individual protein isoforms is challenging, and inferred protein results are often abstracted to the gene-level or to groups of protein isoforms. RESULTS: Here, we introduce IsoBayes, a novel statistical method to perform inference at the isoform level. Our method enhances the information available, by integrating mass spectrometry proteomics and transcriptomics data in a Bayesian probabilistic framework. To account for the uncertainty in the measurement process, we propose a two-layer latent variable approach: first, we sample if a peptide has been correctly detected (or, alternatively filter peptides); second, we allocate the abundance of such selected peptides across the protein(s) they are compatible with. This enables us, starting from peptide-level data, to recover protein-level data; in particular, we: (i) infer the presence/absence of each protein isoform (via a posterior probability), (ii) estimate its abundance (and credible interval), and (iii) target isoforms where transcript and protein relative abundances significantly differ. We benchmarked our approach in simulations, and in two multi-protease real datasets: our method displays good sensitivity and specificity when detecting protein isoforms, its estimated abundances highly correlate with the ground truth, and can detect changes between protein and transcript relative abundances. AVAILABILITY AND IMPLEMENTATION: IsoBayes is freely distributed as a Bioconductor R package, and is accompanied by an example usage vignette.

How can we test AI performance? This question seems trivial, but it isn't. Standard benchmarks often have problems such as in-distribution and small-size test sets, oversimplified metrics, unfair comparisons, and short-term outcome pressure. As a consequence, good performance on standard benchmarks does not guarantee success in real-world scenarios. To address these problems, we present Touchstone, a large-scale collaborative segmentation benchmark of 9 types of abdominal organs. This benchmark is based on 5,195 training CT scans from 76 hospitals around the world and 5,903 testing CT scans from 11 additional hospitals. This diverse test set enhances the statistical significance of benchmark results and rigorously evaluates AI algorithms across various out-of-distribution scenarios. We invited 14 inventors of 19 AI algorithms to train their algorithms, while our team, as a third party, independently evaluated these algorithms on three test sets. In addition, we also evaluated pre-existing AI frameworks---which, differing from algorithms, are more flexible and can support different algorithms—including MONAI from NVIDIA, nnU-Net from DKFZ, and numerous other open-source frameworks. We are committed to expanding this benchmark to encourage more innovation of AI algorithms for the medical domain.

We introduce an improved version of the principal path method, an algorithm conceived to find smooth paths between objects in space. Some key steps of the algorithm have been changed, making the solution intrinsically local and preventing it from being attracted by a global manifold. Judiciously performing the initialization step with the Dijkstra algorithm and a proper metric, the functional now only performs a final refinement of the initial solution. Hence the algorithm is stabler as the space of possible solutions has been considerably reduced with respect to the original method. We tested the proposed algorithm in 2D toy data sets (to understand the behaviour) and in high-dimensional data sets. Compared to the previous version of the algorithm, we obtained significantly stabler and more realistic generated samples.

Feature selection is a prominent activity when dealing with classification/regression problems in biological and omics data. Despite the effort devoted to this issue theoretically, feature selection stability within and across methods is often overlooked in practice. This is a compelling issue because a unique or at least stable answer is needed in clinical scenarios. Here, we analyse in detail a multiomics small sample data set, the Oxford Street II data set, and discuss how existing methods perform in terms of the usual metrics but also in terms of intra and inter feature selection stability. To mitigate the observed instability, we propose a simple unsupervised feature prefiltering, achieving promising results.

SUMMARY: A primary problem in high-throughput genomics experiments is finding the most important genes involved in biological processes (e.g. tumor progression). In this applications note, we introduce spathial, an R package for navigating high-dimensional data spaces. spathial implements the Principal Path algorithm, which is a topological method for locally navigating on the data manifold. The package, together with the core algorithm, provides several high-level functions for interpreting the results. One of the analyses we propose is the extraction of the genes that are mainly involved in the progress from one state to another. We show a possible application in the context of tumor progression using RNA-Seq and single-cell datasets, and we compare our results with two commonly used algorithms, edgeR and monocle3, respectively. AVAILABILITY AND IMPLEMENTATION: The R package spathial is available on the Comprehensive R Archive Network (https://cran.r-project.org/web/packages/spathial/index.html) and on GitHub (https://github.com/erikagardini/spathial). It is distributed under the GNU General Public License (version 3). SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

The fatty acid amide hydrolase (FAAH) regulates the endocannabinoid system cleaving primarily the lipid messenger anandamide. FAAH has been well characterized over the years and, importantly, it represents a promising drug target to treat several diseases, including inflammatory-related diseases and cancer. But its enzymatic mechanism for lipid selection to specifically hydrolyze anandamide, rather than similar bioactive lipids, remains elusive. Here, we clarify this mechanism in FAAH, examining the role of the dynamic paddle, which is formed by the gating residues Phe432 and Trp531 at the boundary between two cavities that form the FAAH catalytic site (the "membrane-access" and the "acyl chain-binding" pockets). We integrate microsecond-long MD simulations of wild type and double mutant model systems (Phe432Ala and Trp531Ala) of FAAH, embedded in a realistic membrane/water environment, with mutagenesis and kinetic experiments. We comparatively analyze three fatty acid substrates with different hydrolysis rates (anandamide > oleamide > palmitoylethanolamide). Our findings identify FAAH's mechanism to selectively accommodate anandamide into a multi-pocket binding site, and to properly orient the substrate in pre-reactive conformations for efficient hydrolysis that is interceded by the dynamic paddle. Our findings therefore endorse a structural framework for a lipid selection mechanism mediated by structural flexibility and gating residues between multiple binding cavities, as found in FAAH. Based on the available structural data, this exquisite catalytic strategy for substrate specificity seems to be shared by other lipid-degrading enzymes with similar enzymatic architecture. The mechanistic insights for lipid selection might assist de-novo enzyme design or drug discovery efforts.

The steady increase of computing power at lower and lower cost enables molecular dynamics simulations to investigate the process of protein folding with an explicit treatment of water molecules. Such simulations are typically done with well known computational chemistry codes like CHARMM. Desktop grids such as the United Devices MetaProcessor are highly attractive platforms, since scavenging for unused machines on Intra- and Internet delivers compute power that is almost free. However, the predominant programming paradigm for current desktop grids is pure task parallelism and might not fit the needs for protein folding simulations with explicit water molecules. A short overall turn-around time of a simulation remains highly important for research productivity, but the need for an accurate model and long simulation time-scales leads to tasks that are too large for optimal scheduling on a desktop grid. To address this problem, we introduce a combination of task- and data parallelism as a well suitable computing paradigm for protein folding investigations on grid platforms. As a proof of concept, we design and implement a simple system for protein folding simulations based on the notion of combined task and data parallelism with clustered workers. Clustered workers are machines grouped into small clusters according to network and CPU performance criteria and act as super-nodes within a desktop grid, permitting the utilization of data parallelism in addition to the task parallelism. We integrate our new paradigm into the existing software environment of the United Devices MetaProcessor. For a test protein, we reach a better quality of the folding calculations than we reached using just task parallelism on distributed systems.