Searching Full-Text Anatomic Pathology Reports Using Business Intelligence Software

PIRO: A web-based search platform for pathology reports, leveraging large language models to generate discrete searchable insights

Pathologists rely on access to historical diagnostic case texts for research, education, and peer learning. However, many laboratory information systems (LIS), including Epic Beaker, lack optimized search tools tailored to pathology-specific text queries. To address this need, we developed PIRO (Pathology Information Retrieval Optimizer), a web-based platform enabling efficient text searches of diagnostic archives. Built using FastAPI, Angular, and Apache Solr, PIRO supports both basic and advanced search functionalities, faceted filtering, and data extraction, while ensuring compliance with institutional privacy protocols. PIRO's capabilities extend to case cohort building, search result export, and secure access control within the institutional network. In an 8-month study, we observed significantly higher PIRO adoption rates (67 %) among pathologists compared to Epic Beaker's SlicerDicer (9 %), underscoring PIRO's usability and relevance. Additionally, we implemented a large language model (LLM) to annotate reports with a "Malignancy Risk" label, enhancing search precision and enabling future expansion of automated annotations. Ongoing work focuses on integrating PIRO with our digital pathology platform, enabling direct access to digital slides from case results. PIRO's adaptable design makes it applicable across institutions, advancing search and retrieval efficiency in pathology archives and enhancing support for pathology research and education.

The Development of a Comprehensive Clinicopathologic Registry for Glomerular Diseases Using Natural Language Processing

Background

Glomerulonephritis (GN) represents a common cause of chronic kidney disease, and treatment to slow or prevent progression of GN is associated with significant morbidity. Large patient registries have improved the understanding of risk stratification, treatment selection, and definitions of treatment response in GN, but can be resource-intensive, with incomplete patient capture.

Objective

To describe the creation of a comprehensive clinicopathologic registry for all patients undergoing kidney biopsy in Manitoba, using natural language processing software for data extraction from pathology reports, as well as to describe cohort characteristics and outcomes.

Design

Retrospective population-based cohort study.

Setting

Tertiary care center in the province of Manitoba.

Patients

All patients undergoing a kidney biopsy in the province of Manitoba from 2002 to 2019.

Measurements

Descriptive statistics are presented for the most common glomerular diseases, along with outcomes of kidney failure and mortality for the individual diseases.

Methods

Data from native kidney biopsy reports from January 2002 to December 2019 were extracted into a structured database using a natural language processing algorithm employing regular expressions. The pathology database was then linked with population-level clinical, laboratory, and medication data, creating a comprehensive clinicopathologic registry. Kaplan-Meier curves and Cox models were constructed to assess the relationship between type of GN and outcomes of kidney failure and mortality.

Results

Of 2421 available biopsies, 2103 individuals were linked to administrative data, of which 1292 had a common glomerular disease. The incidence of yearly biopsies increased almost 3-fold over the study period. Among common glomerular diseases, immunoglobulin A (IgA) nephropathy was the most common (28.6%), whereas infection-related GN had the highest proportions of kidney failure (70.3%) and all-cause mortality (42.3%). Predictors of kidney failure included urine albumin-to-creatinine ratio at the time of biopsy (adjusted hazard ratio [HR] = 1.43, 95% confidence interval [CI] = 1.24-1.65), whereas predictors of mortality included age at the time of biopsy (adjusted HR = 1.05, 95% CI = 1.04-1.06) and infection-related GN (adjusted HR = 1.85, 95% CI = 1.14-2.99, compared with the reference category of IgA nephropathy).

Limitations

Retrospective, single-center study with a relatively small number of biopsies.

Conclusions

Creation of a comprehensive glomerular diseases registry is feasible and can be facilitated through the use of novel data extraction methods. This registry will facilitate further epidemiological research in GN.

A Review on Electronic Health Record Text-Mining for Biomedical Name Entity Recognition in Healthcare Domain

Biomedical-named entity recognition (bNER) is critical in biomedical informatics. It identifies biomedical entities with special meanings, such as people, places, and organizations, as predefined semantic types in electronic health records (EHR). bNER is essential for discovering novel knowledge using computational methods and Information Technology. Early bNER systems were configured manually to include domain-specific features and rules. However, these systems were limited in handling the complexity of the biomedical text. Recent advances in deep learning (DL) have led to the development of more powerful bNER systems. DL-based bNER systems can learn the patterns of biomedical text automatically, making them more robust and efficient than traditional rule-based systems. This paper reviews the healthcare domain of bNER, using DL techniques and artificial intelligence in clinical records, for mining treatment prediction. bNER-based tools are categorized systematically and represent the distribution of input, context, and tag (encoder/decoder). Furthermore, to create a labeled dataset for our machine learning sentiment analyzer to analyze the sentiment of a set of tweets, we used a manual coding approach and the multi-task learning method to bias the training signals with domain knowledge inductively. To conclude, we discuss the challenges facing bNER systems and future directions in the healthcare field.

Development of an interactive web dashboard to facilitate the reexamination of pathology reports for instances of underbilling of CPT codes

Current Procedural Terminology Codes is a numerical coding system used to bill for medical procedures and services and crucially, represents a major reimbursement pathway. Given that pathology services represent a consequential source of hospital revenue, understanding instances where codes may have been misassigned or underbilled is critical. Several algorithms have been proposed that can identify improperly billed CPT codes in existing datasets of pathology reports. Estimation of the fiscal impacts of these reports requires a coder (i.e., billing staff) to review the original reports and manually code them again. As the re-assignment of codes using machine learning algorithms can be done quickly, the bottleneck in validating these reassignments is in this manual re-coding process, which can prove cumbersome. This work documents the development of a rapidly deployable dashboard for examination of reports that the original coder may have misbilled. Our dashboard features the following main components: (1) a bar plot to show the predicted probabilities for each CPT code, (2) an interpretation plot showing how each word in the report combines to form the overall prediction, and (3) a place for the user to input the CPT code they have chosen to assign. This dashboard utilizes the algorithms developed to accurately identify CPT codes to highlight the codes missed by the original coders. In order to demonstrate the function of this web application, we recruited pathologists to utilize it to highlight reports that had codes incorrectly assigned. We expect this application to accelerate the validation of re-assigned codes through facilitating rapid review of false-positive pathology reports. In the future, we will use this technology to review thousands of past cases in order to estimate the impact of underbilling has on departmental revenue.