Computational Pathology

Scientific Scope and Overarching Goals

Pathology is the study of the essential processes leading to the development of human diseases, with a specific focus on the molecular, structural, morphological, and functional changes produced in cells and tissues by such processes. In this respect, the scope and scientific goals of Computational and Systems Pathology (CSP) do not differ from those of more classical Pathology disciplines.

The peculiar aspect characterizing CSP is the use of computational approaches and mathematical modeling to the study of the disease process, often leveraging “big‐data” from multiple sources (e.g., digital images, molecular data, etc.). The ultimate goal is to generate diagnostic processes and clinically actionable knowledge to improve patient care and outcomes. Importantly, the scope of CSP extends beyond the mere adoption of computational methods to the study of disease pathogenesis, leveraging and building upon the core competencies of pathology to expand the ability of effectively generating clinically actionable knowledge.

Scientific Domains and Goals

The main scientific domains CSP is concerned with are:

1. digital pathology and image analysis
2. clinical genomics and other omics; and
3. the integration of multi-modality data with clinical information

Digital pathology and image analysis

Technological advances over the past decades have enabled the development of tools, systems, and infrastructure for the massive and parallel digitization of pathology slides with the associated meta‐data, their storage, review, and analysis. At the same time, advances in algorithmics, statistics, mathematics, and computer science have provided the tools for the extraction and analysis of quantifiable information from these images. This has created unprecedented opportunities for the systematic and quantitative analysis of images routinely generated in pathology departments around the world. The adoption of Artificial Intelligence (AI), Machine Learning (ML) – including more data and computation intensive approaches like Deep Learning (DL) – holds the promise to project pathology in the next millennium.

Clinical genomics and other omics

The availability of high-throughput methods to analyze genomic sequences on a global scale has enabled the comprehensive study of the genomic contributions to complex human diseases at the system level. The analysis of whole genomes, exomes, or specific gene panels allows for the detection of mutations and structural anomalies that bear clinical implications. Similar observation can be made for other omics domains (transcriptome, proteome, metabolome, and so on), with unprecedented opportunities for a deeper understanding of disease processes. This body of knowledge, as it essence part of Pathology, is growing at a fast pace and it is at the basis for precision and personalized medicine and health. CSP provides the necessary computational tools and methods to store, organize, analyze, and – most importantly – interpret such data, enabling research, discovery, and clinical use of this knowledge.

Big data integration with clinical information

An overarching goal for current medical care is that it can be tailored to the genomic and molecular profile of the individual. In other words, Medicine is transitioning from treating the “average patient” to seeking to care and cure each individual in a tailored way. This trend has led to precision and personalized medicine approaches like, for instance, the selection of targeted therapies in oncology based on the specific mutation profile of the tumor. Due to the complexity of analyses, and the need for innovative technologies and advanced pathological expertise, this approach to care, however, can be only delivered at lead academic institutions. Hence, the opportunity to deliver better, state of the art clinical care on a large scale and in the community is still missed. By integrating big molecular data, standard clinical laboratory measurements, and digital pathology images, CSP is perfectly positioned to bridge this gap. The goal here is to develop robust, parsimonious models that can forecast patients molecular make-up (e.g., their mutational profiles) and outcome (e.g., their response to targeted treatments) solely based on the integration of digital pathology imaging with routine clinical and laboratory data. While the development of these approaches can only be achieved in leading academic institutions like Weill-Cornell Medicine, the deployment and dissemination of these methods in community and rural hospitals nationwide will finally be enabled. Ultimately, through research in this domain, CSP will foster the democratization of precision and personalized medicine.

Core Missions

In order to achieve the outlined scientific goals, CSP activities will be organized around three Core Missions:

1. Research and technological development
2. Clinical translation, support, and production; and
3. Education and training

Research and technological development

The development of the novel approaches to disease diagnosis and treatment based on quantitative and computational models can be realized only through research and innovation. Hence, a primary core mission for CSP is to establish a strong research program across distinct disease areas (oncology, cardiovascular diseases, neurological disorders, etc, …), integrating data, knowledge, and expertise from classical pathology disciplines with quantitative methods grounded in biostatistics, epidemiology, mathematics, and computer science. This will lead to next-generation biomarkers in the form of predictive models for screening, diagnostics, prognostication, and therapy selection.

Clinical translation, support, and production

The efficient and fast translation to the bed-side and in the network of the next-generation biomarkers and procedures resulting from research and development is a similarly crucial mission for CSP. Hence, all research activities will be coordinated and streamlined for their efficient and fast implementation into clinical-grade procedures ready for patient care use. This mission pertains to defining optimal operational characteristics for the new biomarkers and procedures, but it also extends to regulatory aspects and State and Federal accreditations.

Education and training

The pervasive availability and use of “big data” in science and medicine is ushering into a new era with clear implications for training and education. The next generation of pathologists will need to be well versed in the use of computational approaches and in the interpretation of the results from these methodology. For this reason, a fundamental mission for CSP will be to implement a training curriculum in quantitative approaches to molecular and imaging analysis to complement most classic pathology training.

In conclusion, over the next few years, the CSP Division will implement an innovative research program, develop and support clinical novel procedures and services, and establish a state of the training program focused on the application of computational methods in pathology. It is anticipated that these innovative processes and tasks to be developed and implemented will support the Pathology Department, and hence the whole WCM/NYP enterprise. It is also anticipated that the principal CSP Division focus will be initially on cancer – in order to better support the establishment of an NIH designated cancer center at WCM – and will later to additional disease areas (e.g., infectious diseases, neurological and cardiovascular disorders, metabolic diseases, etc.).

Download the  Computational Pathology Fee Schedule here.