Systematic reporting of medical kidney biopsies

[Introduction to renal pathology]

In recent decades, nephropathology has developed worldwide as a subspeciality of pathology, which requires special methodological and technical equipment to process the material and specific clinical and pathological expertise to interpret the findings. These special requirements mean that nephropathology is not available at all pathology institutes, but is carried out on a large scale in a few highly specialised centres. The history of nephropathology, or in a narrower sense the specialised histopathological examination of kidney biopsies, began in 1958 with the first use or performance of a kidney biopsy [1]. It thus replaced the practice of urinalysis, which had been common since the Middle Ages, as a diagnostic tool for kidney diseases. Specialised techniques such as immunofluorescence or immunohistology but also electron microscopy are required to assess specific renal changes, for which the examination of renal biopsies is one of the few remaining routine applications today. In Germany and German-speaking countries, the discipline developed thanks to the work of outstanding people in the field of pathology who were primarily involved in this discipline and had the necessary technical and human resources in their laboratories to ensure that these biopsies could be analysed.

Natural Language Processing to extract SNOMED-CT codes from pathological reports

Objective

The use of standardized structured reports (SSR) and suitable terminologies like SNOMED-CT can enhance data retrieval and analysis, fostering large-scale studies and collaboration. However, the still large prevalence of narrative reports in our laboratories warrants alternative and automated labeling approaches. In this project, natural language processing (NLP) methods were used to associate SNOMED-CT codes to structured and unstructured reports from an Italian Digital Pathology Department.

Methods

Two NLP-based automatic coding systems (support vector machine, SVM, and long-short term memory, LSTM) were trained and applied to a series of narrative reports.

Results

The 1163 cases were tested with both algorithms, showing good performances in terms of accuracy, precision, recall, and F1 score, with SVM showing slightly better performances as compared to LSTM (0.84, 0.87, 0.83, 0.82 vs 0.83, 0.85, 0.83, 0.82, respectively). The integration of an explainability allowed identification of terms and groups of words of importance, enabling fine-tuning, balancing semantic meaning and model performance.

Conclusions

AI tools allow the automatic SNOMED-CT labeling of the pathology archives, providing a retrospective fix to the large lack of organization of narrative reports.

Establishing an effective clinical data collecting tool for optimal evaluation of native and allograft renal biopsies

Introduction

Percutaneous kidney biopsy is the gold standard method to reach a precise diagnosis in most medical kidney diseases, which positively impacts patient care by personalizing the treatment. Accurate diagnosis in the pathology report for medical kidney diseases requires clinicopathological correlation, and clinical data is not always reachable to the nephropathologist. This study aimed to create a standardized, paperless requisition form compatible with medical renal biopsies.

Methods

An initial form was prepared for native and allograft renal biopsies according to the current classification of medical kidney diseases. We invited 33 nephropathologists working in Canadian healthcare institutions to answer survey questions about the need to include a particular aspect of clinical information. According to the responses, we modified the experimental form. Eighty nephrologists were asked to complete a clinical data-collecting form given out as PDF files. The time for completing the form and clinicians' satisfaction were assessed.

Results

The experimental form survey was answered by 20 out of 33 nephropathologists (61%) from 14 Canadian healthcare centers. The agreement rate on the questions was from 38.89% to 100.00% (average 83.33% and 77.14% for the native and the allograft section, respectively). Seventeen out of 80 nephrologists and their assistants (21%) responded by completing 22 PDF forms. The time required to finish a PDF form was 10.4 min on average. Nephrologists considered the form time-consuming and suggested making it more clinically relevant. Only seven nephrologists responded to the satisfaction survey; four (57%) were satisfied.

Conclusions

Medical information is critical in renal pathology diagnoses. A uniform paperless clinical data requisition form was evolved through an agreement by Canadian nephropathologists.

Natural Language Processing in Diagnostic Texts from Nephropathology

Introduction: This study investigates whether it is possible to predict a final diagnosis based on a written nephropathological description—as a surrogate for image analysis—using various NLP methods. Methods: For this work, 1107 unlabelled nephropathological reports were included. (i) First, after separating each report into its microscopic description and diagnosis section, the diagnosis sections were clustered unsupervised to less than 20 diagnostic groups using different clustering techniques. (ii) Second, different text classification methods were used to predict the diagnostic group based on the microscopic description section. Results: The best clustering results (i) could be achieved with HDBSCAN, using BoW-based feature extraction methods. Based on keywords, these clusters can be mapped to certain diagnostic groups. A transformer encoder-based approach as well as an SVM worked best regarding diagnosis prediction based on the histomorphological description (ii). Certain diagnosis groups reached F1-scores of up to 0.892 while others achieved weak classification metrics. Conclusion: While textual morphological description alone enables retrieving the correct diagnosis for some entities, it does not work sufficiently for other entities. This is in accordance with a previous image analysis study on glomerular change patterns, where some diagnoses are associated with one pattern, but for others, there exists a complex pattern combination.