Machine learning classification of surgical pathology reports and chunk recognition for information extraction noise reduction

Background and aims: Machine learning techniques for the text mining of cancer-related clinical documents have not been sufficiently explored. Here some techniques are presented for the pre-processing of free-text breast cancer pathology reports, with the aim of facilitating the extraction of information relevant to cancer staging.

Materials and methods: The first technique was implemented using the freely available software RapidMiner to classify the reports according to their general layout: ‘semi-structured’ and ‘unstructured’. The second technique was developed using the open source language engineering framework GATE and aimed at the prediction of chunks of the report text containing information pertaining to the cancer morphology, the tumour size, its hormone receptor status and the number of positive nodes. The classifiers were trained and tested respectively on sets of 635 and 163 manually classified or annotated reports, from the Northern Ireland Cancer Registry.

Results: The best result of 99.4% accuracy – which included only one semi-structured report predicted as unstructured – was produced by the layout classifier with the k nearest algorithm, using the binary term occurrence word vector type with stopword filter and pruning. For chunk recognition, the best results were found using the PAUM algorithm with the same parameters for all cases, except for the prediction of chunks containing cancer morphology. For semi-structured reports the performance ranged from 0.97 to 0.94 and from 0.92 to 0.83 in precision and recall, while for unstructured reports performance ranged from 0.91 to 0.64 and from 0.68 to 0.41 in precision and recall. Poor results were found when the classifier was trained on semi-structured reports but tested on unstructured.

Conclusions: These results show that it is possible and beneficial to predict the layout of reports and that the accuracy of prediction of which segments of a report may contain certain information is sensitive to the report layout and the type of information sought.

ALK translocation is associated with ALK immunoreactivity and extensive signet-ring morphology in primary lung adenocarcinoma.

BACKGROUND: ALK rearrangement is particularly observed in signet-ring sub-type adenocarcinoma. Since fluorescence in situ hybridization (FISH) is not suitable for mass screening, we aimed to characterize the predictive utility of tumour morphology and ALK immunoreactivity to identify ALK rearrangement, in a primary lung adenocarcinoma dataset enriched for signet-ring morphology, compared with that of other morphology. METHODS: 7 adenocarcinomas from diagnostic archives reported with signet-ring morphology were assessed and compared with 11 adenocarcinomas without signet-ring features over the same time period. Growth patterns were reviewed, ALK expression was assessed by standard immunohistochemistry using ALK1 clone and Envision detection (Dako), and ALK rearrangement was assessed by FISH (Abbott Molecular). Associations between groups and predictive utility of tumour morphology and ALK expression using FISH as gold standard were calculated. RESULTS: 2 excision lung biopsy cases with pure (100%) signet-ring morphology and solid patterns demonstrated diffuse moderate cytoplasmic ALK immunoreactivity (2+) and harboured ALK rearrangements (p=0.007), unlike 5 mixed-signet-ring and 11 non-signet-ring adenocarcinomas, which showed negative or 1+ immunoreactivity; and did not harbour ALK rearrangements (p>0.1). ALK expression was not associated with ALK copy number. 6 of 7 cases with signet ring morphology stained for TTF-1. Pure signet-ring morphology and moderate ALK expression were both associated with ALK rearranged tumours. CONCLUSION: ALK rearrangement is strongly associated with ALK immunoreactivity, and was seen only in tumours with pure signet-ring morphology and solid growth pattern. Tumour morphology, growth pattern and ALK immunoreactivity appear to be good indicators of ALK rearrangement, with TTF-1 positivity aiding in proving primary pulmonary origin.