Comparison of Machine-Learning Algorithms for the Prediction of Current Procedural Terminology (CPT) Codes from Pathology Reports

Applications of Large Language Models in Pathology

Large language models (LLMs) are transformer-based neural networks that can provide human-like responses to questions and instructions. LLMs can generate educational material, summarize text, extract structured data from free text, create reports, write programs, and potentially assist in case sign-out. LLMs combined with vision models can assist in interpreting histopathology images. LLMs have immense potential in transforming pathology practice and education, but these models are not infallible, so any artificial intelligence generated content must be verified with reputable sources. Caution must be exercised on how these models are integrated into clinical practice, as these models can produce hallucinations and incorrect results, and an over-reliance on artificial intelligence may lead to de-skilling and automation bias. This review paper provides a brief history of LLMs and highlights several use cases for LLMs in the field of pathology.

TCGA-Reports: A machine-readable pathology report resource for benchmarking text-based AI models

In cancer research, pathology report text is a largely untapped data source. Pathology reports are routinely generated, more nuanced than structured data, and contain added insight from pathologists. However, there are no publicly available datasets for benchmarking report-based models. Two recent advances suggest the urgent need for a benchmark dataset. First, improved optical character recognition (OCR) techniques will make it possible to access older pathology reports in an automated way, increasing the data available for analysis. Second, recent improvements in natural language processing (NLP) techniques using artificial intelligence (AI) allow more accurate prediction of clinical targets from text. We apply state-of-the-art OCR and customized post-processing to report PDFs from The Cancer Genome Atlas, generating a machine-readable corpus of 9,523 reports. Finally, we perform a proof-of-principle cancer-type classification across 32 tissues, achieving 0.992 average AU-ROC. This dataset will be useful to researchers across specialties, including research clinicians, clinical trial investigators, and clinical NLP researchers.

Artificially Intelligent Billing in Spine Surgery: An Analysis of a Large Language Model

STUDY DESIGN

Retrospective cohort study.

OBJECTIVES

This study assessed the effectiveness of a popular large language model, ChatGPT-4, in predicting Current Procedural Terminology (CPT) codes from surgical operative notes. By employing a combination of prompt engineering, natural language processing (NLP), and machine learning techniques on standard operative notes, the study sought to enhance billing efficiency, optimize revenue collection, and reduce coding errors.

METHODS

The model was given 3 different types of prompts for 50 surgical operative notes from 2 spine surgeons. The first trial was simply asking the model to generate CPT codes for a given OP note. The second trial included 3 OP notes and associated CPT codes to, and the third trial included a list of every possible CPT code in the dataset to prime the model. CPT codes generated by the model were compared to those generated by the billing department. Model evaluation was performed in the form of calculating the area under the ROC (AUROC), and area under precision-recall curves (AUPRC).

RESULTS

The trial that involved priming ChatGPT with a list of every possible CPT code performed the best, with an AUROC of .87 and an AUPRC of .67, and an AUROC of .81 and AUPRC of .76 when examining only the most common CPT codes.

CONCLUSIONS

ChatGPT-4 can aid in automating CPT billing from orthopedic surgery operative notes, driving down healthcare expenditures and enhancing billing code precision as the model evolves and fine-tuning becomes available.

Benchmark Pathology Report Text Corpus with Cancer Type Classification

In cancer research, pathology report text is a largely un-tapped data source. Pathology reports are routinely generated, more nuanced than structured data, and contain added insight from pathologists. However, there are no publicly-available datasets for benchmarking report-based models. Two recent advances suggest the urgent need for a benchmark dataset. First, improved optical character recognition (OCR) techniques will make it possible to access older pathology reports in an automated way, increasing data available for analysis. Second, recent improvements in natural language processing (NLP) techniques using AI allow more accurate prediction of clinical targets from text. We apply state-of-the-art OCR and customized post-processing to publicly available report PDFs from The Cancer Genome Atlas, generating a machine-readable corpus of 9,523 reports. We perform a proof-of-principle cancer-type classification across 32 tissues, achieving 0.992 average AU-ROC. This dataset will be useful to researchers across specialties, including research clinicians, clinical trial investigators, and clinical NLP researchers.

Improving Diagnostics with Deep Forest Applied to Electronic Health Records

An electronic health record (EHR) is a vital high-dimensional part of medical concepts. Discovering implicit correlations in the information of this data set and the research and informative aspects can improve the treatment and management process. The challenge of concern is the data sources' limitations in finding a stable model to relate medical concepts and use these existing connections. This paper presents Patient Forest, a novel end-to-end approach for learning patient representations from tree-structured data for readmission and mortality prediction tasks. By leveraging statistical features, the proposed model is able to provide an accurate and reliable classifier for predicting readmission and mortality. Experiments on MIMIC-III and eICU datasets demonstrate Patient Forest outperforms existing machine learning models, especially when the training data are limited. Additionally, a qualitative evaluation of Patient Forest is conducted by visualising the learnt representations in 2D space using the t-SNE, which further confirms the effectiveness of the proposed model in learning EHR representations.

Towards Medical Billing Automation: NLP for Outpatient Clinician Note Classification

Objectives

Our primary objective was to develop a natural language processing approach that accurately predicts outpatient Evaluation and Management (E/M) level of service (LoS) codes using clinicians' notes from a health system electronic health record. A secondary objective was to investigate the impact of clinic note de-identification on document classification performance.

Methods

We used retrospective outpatient office clinic notes from four medical and surgical specialties. Classification models were fine-tuned on the clinic notes datasets and stratified by subspecialty. The success criteria for the classification tasks were the classification accuracy and F1-scores on internal test data. For the secondary objective, the dataset was de-identified using Named Entity Recognition (NER) to remove protected health information (PHI), and models were retrained.

Results

The models demonstrated similar predictive performance across different specialties, except for internal medicine, which had the lowest classification accuracy across all model architectures. The models trained on the entire note corpus achieved an E/M LoS CPT code classification accuracy of 74.8% (CI 95: 74.1-75.6). However, the de-identified note corpus showed a markedly lower classification accuracy of 48.2% (CI 95: 47.7-48.6) compared to the model trained on the identified notes.

Conclusion

The study demonstrates the potential of NLP-based document classifiers to accurately predict E/M LoS CPT codes using clinical notes from various medical and procedural specialties. The models' performance suggests that the classification task's complexity merits further investigation. The de-identification experiment demonstrated that de-identification may negatively impact classifier performance. Further research is needed to validate the performance of our NLP classifiers in different healthcare settings and patient populations and to investigate the potential implications of de-identification on model performance.

Can a Novel Natural Language Processing Model and Artificial Intelligence Automatically Generate Billing Codes From Spine Surgical Operative Notes?

STUDY DESIGN

Retrospective cohort.

OBJECTIVE

Billing and coding-related administrative tasks are a major source of healthcare expenditure in the United States. We aim to show that a second-iteration Natural Language Processing (NLP) machine learning algorithm, XLNet, can automate the generation of CPT codes from operative notes in ACDF, PCDF, and CDA procedures.

METHODS

We collected 922 operative notes from patients who underwent ACDF, PCDF, or CDA from 2015 to 2020 and included CPT codes generated by the billing code department. We trained XLNet, a generalized autoregressive pretraining method, on this dataset and tested its performance by calculating AUROC and AUPRC.

RESULTS

The performance of the model approached human accuracy. Trial 1 (ACDF) achieved an AUROC of .82 (range: .48-.93), an AUPRC of .81 (range: .45-.97), and class-by-class accuracy of 77% (range: 34%-91%); trial 2 (PCDF) achieved an AUROC of .83 (.44-.94), an AUPRC of .70 (.45-.96), and class-by-class accuracy of 71% (42%-93%); trial 3 (ACDF and CDA) achieved an AUROC of .95 (.68-.99), an AUPRC of .91 (.56-.98), and class-by-class accuracy of 87% (63%-99%); trial 4 (ACDF, PCDF, CDA) achieved an AUROC of .95 (.76-.99), an AUPRC of .84 (.49-.99), and class-by-class accuracy of 88% (70%-99%).

CONCLUSIONS

We show that the XLNet model can be successfully applied to orthopedic surgeon's operative notes to generate CPT billing codes. As NLP models as a whole continue to improve, billing can be greatly augmented with artificial intelligence assisted generation of CPT billing codes which will help minimize error and promote standardization in the process.

Natural Language Processing in Pathology: Current Trends and Future Insights

Natural language processing (NLP) plays a key role in advancing health care, being key to extracting structured information from electronic health reports. In the last decade, several advances in the field of pathology have been derived from the application of NLP to pathology reports. Herein, a comprehensive review of the most used NLP methods for extracting, coding, and organizing information from pathology reports is presented, including how the development of tools is used to improve workflow. In addition, this article discusses, from a practical point of view, the steps necessary to extract data and encode natural language information for its analytical processing, ranging from preprocessing of text to its inclusion in complex algorithms. Finally, the potential of NLP-based automatic solutions for improving workflow in pathology and their further applications in the near future is highlighted.

Automated Generation of Synoptic Reports from Narrative Pathology Reports in University Malaya Medical Centre Using Natural Language Processing

Pathology reports represent a primary source of information for cancer registries. University Malaya Medical Centre (UMMC) is a tertiary hospital responsible for training pathologists; thus narrative reporting becomes important. However, the unstructured free-text reports made the information extraction process tedious for clinical audits and data analysis-related research. This study aims to develop an automated natural language processing (NLP) algorithm to summarize the existing narrative breast pathology report from UMMC to a narrower structured synoptic pathology report with a checklist-style report template to ease the creation of pathology reports. The development of the rule-based NLP algorithm was based on the R programming language by using 593 pathology specimens from 174 patients provided by the Department of Pathology, UMMC. The pathologist provides specific keywords for data elements to define the semantic rules of the NLP. The system was evaluated by calculating the precision, recall, and F1-score. The proposed NLP algorithm achieved a micro-F1 score of 99.50% and a macro-F1 score of 98.97% on 178 specimens with 25 data elements. This achievement correlated to clinicians’ needs, which could improve communication between pathologists and clinicians. The study presented here is significant, as structured data is easily minable and could generate important insights.