Classification of Cancer Stage from Free-text Histology Reports

Abstract

This article investigates the classification of a patient's lung cancer stage based on analysis of their free-text medical reports. The system uses natural language processing to transform the report text, including identification of UMLS terms and detection of negated findings. The transformed report is then classified using statistical machine learning techniques. A support vector machine is trained for each stage category based on word occurrences in a corpus of histology reports for pathologically staged patients. New reports can be classified according to the most likely stage, allowing the collection of population stage data for analysis of outcomes. While the system could in principle be applied to stage different cancer types, the current work focuses on lung cancer due to data availability. The article presents initial experiments quantifying system performance for T and N staging on a corpus of histology reports from more than 700 lung cancer patients.

Sentence-based classification of free-text breast cancer radiology reports

“…Wilcox and Hripcsak [7] assigned six different clinical conditions to narrative X-ray reports using different machine learning techniques. McCowan et al [8] infer the cancer stages from histology reports using statistical machine learning methods. Goldstein et al [9] compared information retrieval, machine learning and rule-based approaches for automatically assigning ICD-9 codes to radiology reports.…”

Lung carcinoma pigeonholing and vaticination by interspersed approach

“…Iain McCowan [7] proposed techniques for pigeonholing of patient's lung carcinoma stage based on analysis of freetext medical reports. The approach uses natural language processing to transform the report text, UML [8] terms and espials of negated findings.…”

“…At moment the system was applied on lung carcinoma because of data availability. Initial experiments were made to quantify system performance for T and N staging on a corpus of histology reports from more than 700 lung carcinoma patients [7]. At the sensitivity/PPV breakeven point, the system achieves average sensitivity of 0.61 and specificity of 0.88 for T staging and sensitivity of 0.78 and specificity of 0.90 for N staging.…”

Health Services Research and Data Linkages: Issues, Methods, and Directions for the Future

“…Studies have demonstrated some success in using natural language processing (NLP) systems to identify disease. As they become more reliable and widely available, the use of NLPs is likely to greatly enhance the value of text‐based messages (McCowan, Moore, and Fry 2006; McCowan et al 2007; Pakhomov et al 2007; Savova et al 2008).…”