A pathologist-AI collaboration framework for enhancing diagnostic accuracies and efficiencies.

In pathology, the deployment of artificial intelligence (AI) in clinical settings is constrained by limitations in data collection and in model transparency and interpretability. Here we describe a digital pathology framework, nuclei.io, that incorporates active learning and human-in-the-loop real-time feedback for the rapid creation of diverse datasets and models. We validate the effectiveness of the framework via two crossover user studies that leveraged collaboration between the AI and the pathologist, including the identification of plasma cells in endometrial biopsies and the detection of colorectal cancer metastasis in lymph nodes. In both studies, nuclei.io yielded considerable diagnostic performance improvements. Collaboration between clinicians and AI will aid digital pathology by enhancing accuracies and efficiencies.

A rare neuromyelitis optica mimic: Primary CNS histiocytic sarcoma.

Primary central nervous system (CNS) histiocytic sarcoma is a rare hematolymphoid malignancy with features of mature histiocytes and carries a poor prognosis. We describe a unique case in which a 50-year-old woman presented with recurrent acute brainstem syndrome, area postrema syndrome, and myelitis with corresponding magnetic resonance imaging (MRI) lesions meeting diagnostic criteria for seronegative neuromyelitis optica spectrum disorder (NMOSD). Despite initial improvement with steroids and plasma exchange, she experienced recurrent symptoms over 10 months referable to new and persistently enhancing lesions. At autopsy, neuropathology revealed a diffusely infiltrative primary CNS histiocytic sarcoma. This case represents a rare clinicoradiologic mimic of NMOSD, underscoring the importance of evaluation for infiltrative diseases in cases of atypical seronegative NMOSD.

A deep-learning classifier identifies patients with clinical heart failure using whole-slide images of H&E tissue.

Over 26 million people worldwide suffer from heart failure annually. When the cause of heart failure cannot be identified, endomyocardial biopsy (EMB) represents the gold-standard for the evaluation of disease. However, manual EMB interpretation has high inter-rater variability. Deep convolutional neural networks (CNNs) have been successfully applied to detect cancer, diabetic retinopathy, and dermatologic lesions from images. In this study, we develop a CNN classifier to detect clinical heart failure from H&E stained whole-slide images from a total of 209 patients, 104 patients were used for training and the remaining 105 patients for independent testing. The CNN was able to identify patients with heart failure or severe pathology with a 99% sensitivity and 94% specificity on the test set, outperforming conventional feature-engineering approaches. Importantly, the CNN outperformed two expert pathologists by nearly 20%. Our results suggest that deep learning analytics of EMB can be used to predict cardiac outcome.

Lipid Rafts Assemble Dynein Ensembles.

New work by Rai et al. identifies a novel mechanism regulating phagosome transport in cells: the clustering of dynein motors into lipid microdomains, leading to enhanced unidirectional motility. Clustering may be especially important for dynein, a motor that works most efficiently in teams.

α-Tubulin Tyrosination and CLIP-170 Phosphorylation Regulate the Initiation of Dynein-Driven Transport in Neurons.

Motor-cargo recruitment to microtubules is often the rate-limiting step of intracellular transport, and defects in this recruitment can cause neurodegenerative disease. Here, we use in vitro reconstitution assays with single-molecule resolution, live-cell transport assays in primary neurons, computational image analysis, and computer simulations to investigate the factors regulating retrograde transport initiation in the distal axon. We find that phosphorylation of the cytoskeletal-organelle linker protein CLIP-170 and post-translational modifications of the microtubule track combine to precisely control the initiation of retrograde transport. Computer simulations of organelle dynamics in the distal axon indicate that while CLIP-170 primarily regulates the time to microtubule encounter, the tyrosination state of the microtubule lattice regulates the likelihood of binding. These mechanisms interact to control transport initiation in the axon in a manner sensitive to the specialized cytoskeletal architecture of the neuron.

LC3 binding to the scaffolding protein JIP1 regulates processive dynein-driven transport of autophagosomes.

Autophagy is essential for maintaining cellular homeostasis in neurons, where autophagosomes undergo robust unidirectional retrograde transport along axons. We find that the motor scaffolding protein JIP1 binds directly to the autophagosome adaptor LC3 via a conserved LIR motif. This interaction is required for the initial exit of autophagosomes from the distal axon, for sustained retrograde transport along the midaxon, and for autophagosomal maturation in the proximal axon. JIP1 binds directly to the dynein activator dynactin but also binds to and activates kinesin-1 in a phosphorylation-dependent manner. Following JIP1 depletion, phosphodeficient JIP1-S421A rescues retrograde transport, while phosphomimetic JIP1-S421D aberrantly activates anterograde transport. During normal autophagosome transport, residue S421 of JIP1 may be maintained in a dephosphorylated state by autophagosome-associated MKP1 phosphatase. Moreover, binding of LC3 to JIP1 competitively disrupts JIP1-mediated activation of kinesin. Thus, dual mechanisms prevent aberrant activation of kinesin to ensure robust retrograde transport of autophagosomes along the axon.