Integrated digital pathology at scale: A solution for clinical diagnostics and cancer research at a large academic medical center

Abstract

Objective: Broad adoption of digital pathology (DP) is still lacking, and examples for DP connecting diagnostic, research, and educational use cases are missing. We blueprint a holistic DP solution at a large academic medical center ubiquitously integrated into clinical workflows; researchapplications including molecular, genetic, and tissue databases; and educational processes.

Materials and methods: We built a vendor-agnostic, integrated viewer for reviewing, annotating, sharing, and quality assurance of digital slides in a clinical or research context. It is the first homegrown viewer cleared by New York State provisional approval in 2020 for primary diagnosis and remote sign-out during the COVID-19 (coronavirus disease 2019) pandemic. We further introduce an interconnected Honest Broker for BioInformatics Technology (HoBBIT) to systematically compile and share large-scale DP research datasets including anonymized images, redacted pathology reports, and clinical data of patients with consent.

Results: The solution has been operationally used over 3 years by 926 pathologists and researchers evaluating 288 903 digital slides. A total of 51% of these were reviewed within 1 month after scanning. Seamless integration of the viewer into 4 hospital systems clearly increases the adoption of DP. HoBBIT directly impacts the translation of knowledge in pathology into effective new health measures, including artificial intelligence-driven detection models for prostate cancer, basal cell carcinoma, and breast cancer metastases, developed and validated on thousands of cases.

Conclusions: We highlight major challenges and lessons learned when going digital to provide orientation for other pathologists. Building interconnected solutions will not only increase adoption of DP, but also facilitate next-generation computational pathology at scale for enhanced cancer research.

Keywords: artificial intelligence; computational pathology; digital pathology; honest broker, pathology; whole slide imaging.

Figure 1.

Overview of the digital and computational pathology landscape. Multiple scanner models (Leica AT2, Leica GT450, 3DHistech P1000, Philips UFS) digitize clinical and research slides and store them in our whole slide images (WSI) database. Metadata are sent to the vendor’s image management systems (IMS), forwarding them to the anatomic pathology laboratory information system (AP-LIS). The AP-LIS also hosts pathology reports and further clinical information. The MSK Viewer is connected to the WSI database and the IMS. Thus, it can visualize any digital image selected by image ID enabling integration into many hospital applications, such as CoPath, MPath, Precision Pathology Biobanking Center (PPBC), cBioPortal, research projects and education portal. For the clinical apps, authentication is leveraged with an authentication token, and protected health information (PHI) can be displayed. For research apps, PHI is hidden, and annotations can be gathered and shared across the systems. The Honest Broker for BioInformatics Technology (HoBBIT) combines data from digital slides with their meta and clinical information from the IMS, MPath, and CoPath together with patient consent information from the institutional database (IDB). It therewith provides large research datasets upon request by searching by clinical information, redacting pathology reports and de-identifying image data. Datasets are then stored on the high-performance computing (HPC) cluster, and a viewer project is optionally generated to visualize and/or annotate the images. DB: database.

Figure 2.

Histogram of the digitally reviewed whole slide images (WSI) ordered by their access time after scanning. Each bar contains the number of WSI first accessed in the corresponding month after scanning of the slides. Note that this plot only includes digitally reviewed WSI, and not all scanned WSI. Gray vertical lines indicate years. Cumulative percentages of accessed WSI are given. As a reading example, 51% of the digitally reviewed WSI were accessed within 1 month after scanning, 61% within 2 months, and 91% within 12 months.

Figure 3.

Screenshot of the MSK Viewer in a clinical context. Shown is a prostate needle biopsy with a measurement of the cancer extend. On the left side, the parts and slides of the case are listed together with the macro and label image for the pathologist to verify the scanned tissue and correct slide to be shown (enlarged on mouseover, protected health information blinded for publication). The viewer provides standard tools such as magnifier, overview image, and measurement tools, and can be opened on any computer or laptop within the hospital’s network.

Figure 4.

Scanning effort at Memorial Sloan Kettering Cancer Center over time. (A) The overall inventory contains 3.6 million digital slides and continuesto grow. (B) Currently, 140 000 slides are digitized every month for both routine diagnostic and archival scanning.

Figure 5.

Number of (A) distinct users and (B) distinct slides per month processed by the digital slide viewer. Our platform launched 2017 as a research tool for annotations of individual slides in specific research projects. Starting 2018, it became the main viewer for clinical cases in the anatomic pathology laboratory information system. In 2019, more systems have been connected to the MSK Viewer, including cBioPortal and MPath. Each of those systems increased the number of users steadily, as pathologists and researchers use the viewer for different tasks. After approval of the digital workflow from New York State regulatory bodies in mid 2020, significantly more slides have been accessed via CoPath. In total over the last 3 academic years, 926 users (from CoPath: 455, cBioPortal: 449, Projects: 291, MPath: 50, and Precision Pathology Biobanking Center [PPBC]: 18) accessed 288 903 whole slide images in the viewer.