PAIP 2020: Microsatellite instability prediction in colorectal cancer.

Microsatellite instability (MSI) refers to alterations in the length of simple repetitive genomic sequences. MSI status serves as a prognostic and predictive factor in colorectal cancer. The MSI-high status is a good prognostic factor in stage II/III cancer, and predicts a lack of benefit to adjuvant fluorouracil chemotherapy in stage II cancer but a good response to immunotherapy in stage IV cancer. Therefore, determining MSI status in patients with colorectal cancer is important for identifying the appropriate treatment protocol. In the Pathology Artificial Intelligence Platform (PAIP) 2020 challenge, artificial intelligence researchers were invited to predict MSI status based on colorectal cancer slide images. Participants were required to perform two tasks. The primary task was to classify a given slide image as belonging to either the MSI-high or the microsatellite-stable group. The second task was tumor area segmentation to avoid ties with the main task. A total of 210 of the 495 participants enrolled in the challenge downloaded the images, and 23 teams submitted their final results. Seven teams from the top 10 participants agreed to disclose their algorithms, most of which were convolutional neural network-based deep learning models, such as EfficientNet and UNet. The top-ranked system achieved the highest F1 score (0.9231). This paper summarizes the various methods used in the PAIP 2020 challenge. This paper supports the effectiveness of digital pathology for identifying the relationship between colorectal cancer and the MSI characteristics.

Development and operation of a digital platform for sharing pathology image data.

BACKGROUND: Artificial intelligence (AI) research is highly dependent on the nature of the data available. With the steady increase of AI applications in the medical field, the demand for quality medical data is increasing significantly. We here describe the development of a platform for providing and sharing digital pathology data to AI researchers, and highlight challenges to overcome in operating a sustainable platform in conjunction with pathologists. METHODS: Over 3000 pathological slides from five organs (liver, colon, prostate, pancreas and biliary tract, and kidney) in histologically confirmed tumor cases by pathology departments at three hospitals were selected for the dataset. After digitalizing the slides, tumor areas were annotated and overlaid onto the images by pathologists as the ground truth for AI training. To reduce the pathologists' workload, AI-assisted annotation was established in collaboration with university AI teams. RESULTS: A web-based data sharing platform was developed to share massive pathological image data in 2019. This platform includes 3100 images, and 5 pre-processing algorithms for AI researchers to easily load images into their learning models. DISCUSSION: Due to different regulations among countries for privacy protection, when releasing internationally shared learning platforms, it is considered to be most prudent to obtain consent from patients during data acquisition. CONCLUSIONS: Despite limitations encountered during platform development and model training, the present medical image sharing platform can steadily fulfill the high demand of AI developers for quality data. This study is expected to help other researchers intending to generate similar platforms that are more effective and accessible in the future.

PAIP 2019: Liver cancer segmentation challenge.

Pathology Artificial Intelligence Platform (PAIP) is a free research platform in support of pathological artificial intelligence (AI). The main goal of the platform is to construct a high-quality pathology learning data set that will allow greater accessibility. The PAIP Liver Cancer Segmentation Challenge, organized in conjunction with the Medical Image Computing and Computer Assisted Intervention Society (MICCAI 2019), is the first image analysis challenge to apply PAIP datasets. The goal of the challenge was to evaluate new and existing algorithms for automated detection of liver cancer in whole-slide images (WSIs). Additionally, the PAIP of this year attempted to address potential future problems of AI applicability in clinical settings. In the challenge, participants were asked to use analytical data and statistical metrics to evaluate the performance of automated algorithms in two different tasks. The participants were given the two different tasks: Task 1 involved investigating Liver Cancer Segmentation and Task 2 involved investigating Viable Tumor Burden Estimation. There was a strong correlation between high performance of teams on both tasks, in which teams that performed well on Task 1 also performed well on Task 2. After evaluation, we summarized the top 11 team's algorithms. We then gave pathological implications on the easily predicted images for cancer segmentation and the challenging images for viable tumor burden estimation. Out of the 231 participants of the PAIP challenge datasets, a total of 64 were submitted from 28 team participants. The submitted algorithms predicted the automatic segmentation on the liver cancer with WSIs to an accuracy of a score estimation of 0.78. The PAIP challenge was created in an effort to combat the lack of research that has been done to address Liver cancer using digital pathology. It remains unclear of how the applicability of AI algorithms created during the challenge can affect clinical diagnoses. However, the results of this dataset and evaluation metric provided has the potential to aid the development and benchmarking of cancer diagnosis and segmentation.

Automated extraction of Biomarker information from pathology reports.

BACKGROUND: Pathology reports are written in free-text form, which precludes efficient data gathering. We aimed to overcome this limitation and design an automated system for extracting biomarker profiles from accumulated pathology reports. METHODS: We designed a new data model for representing biomarker knowledge. The automated system parses immunohistochemistry reports based on a "slide paragraph" unit defined as a set of immunohistochemistry findings obtained for the same tissue slide. Pathology reports are parsed using context-free grammar for immunohistochemistry, and using a tree-like structure for surgical pathology. The performance of the approach was validated on manually annotated pathology reports of 100 randomly selected patients managed at Seoul National University Hospital. RESULTS: High F-scores were obtained for parsing biomarker name and corresponding test results (0.999 and 0.998, respectively) from the immunohistochemistry reports, compared to relatively poor performance for parsing surgical pathology findings. However, applying the proposed approach to our single-center dataset revealed information on 221 unique biomarkers, which represents a richer result than biomarker profiles obtained based on the published literature. Owing to the data representation model, the proposed approach can associate biomarker profiles extracted from an immunohistochemistry report with corresponding pathology findings listed in one or more surgical pathology reports. Term variations are resolved by normalization to corresponding preferred terms determined by expanded dictionary look-up and text similarity-based search. CONCLUSIONS: Our proposed approach for biomarker data extraction addresses key limitations regarding data representation and can handle reports prepared in the clinical setting, which often contain incomplete sentences, typographical errors, and inconsistent formatting.

Development of biomarker positivity analysis system for cancer diagnosis based on clinical data.

In Korea, there were 224,000 new cases of cancer and 75,334 deaths caused by cancer in 2013, which was three times more than the number of death caused by heart disease, the second leading cause of death. This study proposes a biomarker positivity analysis system based on clinical data, for personalized diagnosis and therapy of cancer. Data of 78,912 cases were obtained from immunopathology and surgical pathology reports. Data on sex, age, organ, diagnosis, and biomarkers were entered into a database. To verify the reliability of the clinical data, an additional 50,450 cases from positivity-related research papers were added. The proposed biomarker positivity analysis system makes it possible to extract and combine information for searching. The positivity values are in graphical and tabular format for ease of use. With a link to the internal network of the hospital, real-time pathology reports are available. Twenty-five pathology specialists are chosen as subjects to further confirm the reliability of this system; primary assessment results demonstrate a satisfaction level of 4.7 out of 5 and a concordance rate of 79% with positive data under the same conditions as reported in the literature. In the present study, analysis methods and platforms using large volumes of clinical and literature data are developed for cancer prognoses. It is expected that these tools will benefit both healthcare professionals and non-professionals involved in cancer diagnosis and treatment.

Prognosis of Glioblastoma With Oligodendroglioma Component is Associated With the IDH1 Mutation and MGMT Methylation Status.

Glioblastoma (GBM) with oligodendroglioma component (GBMO) is a newly described GBM subtype in the 2007 World Health Organization classification. However, its biological and genetic characteristics are largely unknown. We investigated the clinicopathological and molecular features of 34 GBMOs and compared the survival rate of these patients with those of patients with astrocytoma, oligodendroglioma, anaplastic oligoastrocytoma (AOA), and conventional GBMs in our hospital. GBMO could be divided into two groups based on the presence of an IDH1 mutation. The IDH1 mutation was more frequently found in secondary GBMO, which had lower frequencies of EGFR amplification but higher MGMT methylation than the wild type IDH1 group, and patients with mutant IDH1 GBMO were on average younger than those with wild-type IDH1. Therefore, GBMO is a clinically and molecularly heterogeneous subtype, largely belonging to a proneural and classical subtype of GBM. The survival rate of GBMO patients itself was worse than that of AOA patients but not significantly better than that of conventional GBM patients. GBMO survival was independent of the dominant histopathological subtype i.e., astrocyte-dominant or oligodendroglioma -dominant, but it was significantly associated with the IDH1 mutation and MGMT methylation status. Therefore, GBMO should be regarded as a separate entity from AOA and must be classified as a subtype of GBM. However, further study is needed to determine whether it is a pathologic variant or a pattern of GBM because GBMO has a similar prognosis to conventional GBMs.

Clinicopathologic analysis of pituitary adenoma: a single institute experience.

Pituitary adenoma (PA) is a common benign neuroendocrine tumor; however, the incidence and proportion of hormone-producing PAs in Korean patients remain unknown. Authors analyzed 506 surgically resected and pathologically proven pituitary lesions of the Seoul National University Hospital from 2006 to 2011. The lesions were categorized as: PAs (n = 422, 83.4%), Rathke's cleft cysts (RCCs) (n = 54, 10.6%), inflammatory lesions (n = 8, 1.6%), meningiomas (n = 4), craniopharyngiomas (n = 4), granular cell tumors (n = 1), metastatic renal cell carcinomas (n = 2), germinomas (n = 1), ependymomas (n = 1), and unsatisfactory specimens (n = 9, 1.8%). PAs were slightly more prevalent in women (M: F = 1:1.17) with a mean age of 48.8 yr (9-80 yr). Immunohistochemical analysis revealed that prolactin-producing PAs (16.6%) and growth hormone-producing adenomas (9.2%) were the most common functional PAs. Plurihormonal PAs and nonfunctioning (null cell) adenomas were found in 14.9% and 42.4% of patients with PAs, respectively. The recurrence rate of PAs was 11.1%, but nearly 0% for the remaining benign lesions such as RCCs. 25.4% of patients with PAs were treated by gamma-knife after surgery due to residual tumors or regrowth of residual tumor. In conclusion, the pituitary lesions and the proportions of hormone-producing PAs in Korean patients are similar to those of previous reports except nonfunctioning (null cell) PAs, which are unusually frequent.

Study on user interface of pathology picture archiving and communication system.

OBJECTIVES: It is necessary to improve the pathology workflow. A workflow task analysis was performed using a pathology picture archiving and communication system (pathology PACS) in order to propose a user interface for the Pathology PACS considering user experience. METHODS: An interface analysis of the Pathology PACS in Seoul National University Hospital and a task analysis of the pathology workflow were performed by observing recorded video. Based on obtained results, a user interface for the Pathology PACS was proposed. RESULTS: Hierarchical task analysis of Pathology PACS was classified into 17 tasks including 1) pre-operation, 2) text, 3) images, 4) medical record viewer, 5) screen transition, 6) pathology identification number input, 7) admission date input, 8) diagnosis doctor, 9) diagnosis code, 10) diagnosis, 11) pathology identification number check box, 12) presence or absence of images, 13) search, 14) clear, 15) Excel save, 16) search results, and 17) re-search. And frequently used menu items were identified and schematized. CONCLUSIONS: A user interface for the Pathology PACS considering user experience could be proposed as a preliminary step, and this study may contribute to the development of medical information systems based on user experience and usability.

Comparison of visibility measurement techniques for forklift truck design factors.

This study applied the light bulb shadow test, a manikin vision assessment test, and an individual test to a forklift truck to identify forklift truck design factors influencing visibility. The light bulb shadow test followed the standard of ISO/DIS 13564-1 for traveling and maneuvering tests with four test paths (Test Nos. 1, 3, 4, and 6). Digital human and forklift truck models were developed for the manikin vision assessment test with CATIA V5R13 human modeling solutions. Six participants performed the individual tests. Both employed similar parameters to the light bulb shadow test. The individual test had better visibility with fewer numbers and a greater distribution of the shadowed grids than the other two tests due to eye movement and anthropometric differences. The design factors of load backrest extension, lift chain, hose, dashboard, and steering wheel should be the first factors considered to improve visibility, especially when a forklift truck mainly performs a forward traveling task in an open area.

Automation of gross photography using a remote-controlled digital camera system.

CONTEXT: Conventional gross photography requires a series of tedious and time-consuming steps, including taking, developing, labeling, sorting, filing, and tracking numerous photographs. OBJECTIVE: To describe how to automate the gross photographic process by way of controlling a digital camera remotely. DESIGN: After defining the requirements of automation regarding gross photography, a remote control board, foot switch, barcode system, and image retrieval system were devised. SETTING: The surgical pathology laboratory of a university medical center with a commercially available megapixel digital camera. RESULTS: The digital camera zoom and shutter were controlled remotely by a foot switch. A large portion of the gross photographic process, including specimen number labeling, image downloading, labeling, sorting, filing, and tracking, were automated. In addition, the elimination of several manual specimen-processing steps, along with not having to wait for the developing and mounting of conventional 35-mm film, reduced the entire time span required in conventional gross photography from 2 to 5 days, to a few minutes. It was also possible to review the gross images at the time of microscopic sign-out. CONCLUSIONS: The automation of gross photography using a remote-controlled digital camera changes the conventional gross workflow markedly. We found use of a remote-controlled gross photography system to be practical, convenient, and efficient.