RESUMEN
BACKGROUND: Deep learning enables accurate high-resolution mapping of cells and tissue structures that can serve as the foundation of interpretable machine-learning models for computational pathology. However, generating adequate labels for these structures is a critical barrier, given the time and effort required from pathologists. RESULTS: This article describes a novel collaborative framework for engaging crowds of medical students and pathologists to produce quality labels for cell nuclei. We used this approach to produce the NuCLS dataset, containing >220,000 annotations of cell nuclei in breast cancers. This builds on prior work labeling tissue regions to produce an integrated tissue region- and cell-level annotation dataset for training that is the largest such resource for multi-scale analysis of breast cancer histology. This article presents data and analysis results for single and multi-rater annotations from both non-experts and pathologists. We present a novel workflow that uses algorithmic suggestions to collect accurate segmentation data without the need for laborious manual tracing of nuclei. Our results indicate that even noisy algorithmic suggestions do not adversely affect pathologist accuracy and can help non-experts improve annotation quality. We also present a new approach for inferring truth from multiple raters and show that non-experts can produce accurate annotations for visually distinctive classes. CONCLUSIONS: This study is the most extensive systematic exploration of the large-scale use of wisdom-of-the-crowd approaches to generate data for computational pathology applications.
Asunto(s)
Neoplasias de la Mama , Colaboración de las Masas , Neoplasias de la Mama/patología , Núcleo Celular , Colaboración de las Masas/métodos , Femenino , Humanos , Aprendizaje AutomáticoRESUMEN
INTRODUCTION: Laser Capture Microdissection (LCM) uses a laser to isolate, or capture, specific cells of interest in a complex heterogeneous tissue section, under direct microscopic visualization. Recently, there has been a surge of publications using LCM for tissue spatial molecular profiling relevant to a wide range of research topics. AREAS COVERED: We summarize the many advances in tissue Laser Capture Proteomics (LCP) using mass spectrometry for discovery, and protein arrays for signal pathway network mapping. This review emphasizes: a) transition of LCM phosphoproteomics from the lab to the clinic for individualized cancer therapy, and b) the emerging frontier of LCM single cell molecular analysis combining proteomics with genomic, and transcriptomic analysis. The search strategy was based on the combination of MeSH terms with expert refinement. EXPERT OPINION: LCM is complemented by a rich set of instruments, methodology protocols, and analytical A.I. (artificial intelligence) software for basic and translational research. Resolution is advancing to the tissue single cell level. A vision for the future evolution of LCM is presented. Emerging LCM technology is combining digital and AI guided remote imaging with automation, and telepathology, to a achieve multi-omic profiling that was not previously possible.