RESUMEN
Invasive fungal infections (IFIs) are particularly dangerous to high-risk patients with haematological malignancies and are responsible for excessive mortality and delays in cancer therapy. Surveillance of IFI in clinical settings offers an opportunity to identify potential risk factors and evaluate new therapeutic strategies. However, manual surveillance is both time- and resource-intensive. As part of a broader project aimed to develop a system for automated IFI surveillance by leveraging electronic medical records, we present our approach to detecting evidence of IFI in the key diagnostic domain of histopathology. Using natural language processing (NLP), we analysed cytology and histopathology reports to identify IFI-positive reports. We compared a conventional bag-of-words classification model to a method that relies on concept-level annotations. Although the investment to prepare data supporting concept annotations is substantial, extracting targeted information specific to IFI as a pre-processing step increased the performance of the classifier from the PR AUC of 0.84 to 0.92 and enabled model interpretability. We have made publicly available the annotated dataset of 283 reports, the Cytology and Histopathology IFI Reports corpus (CHIFIR), to allow the clinical NLP research community to further build on our results.
Asunto(s)
Infecciones Fúngicas Invasoras , Humanos , Infecciones Fúngicas Invasoras/epidemiología , Registros Electrónicos de Salud , Procesamiento de Lenguaje Natural , Factores de RiesgoRESUMEN
AIM: Our study aim was to investigate the degree of adherence to international recommendations for cutaneous melanoma pathology reports at the population level by a EUROCARE high resolution study. METHODS: The availability of nine characteristics - predominant cell type, tumour-infiltrating lymphocytes, mitotic index, histological subtype, growth phase, Clark level, Breslow thickness, ulceration, and sentinel-node biopsy - was examined on pathology reports of a random sample of 636 cases diagnosed in 2003-2005 in seven Italian cancer registries: Biella, Ferrara, Firenze, Latina, Ragusa, Reggio Emilia, Romagna. The odds of having (versus not having) information for all four core characteristics (last four listed above) were estimated. RESULTS: Sentinel node biopsy was available most often, followed by Clark level, Breslow thickness, histological subtype and ulceration. Information on all nine characteristics was more often available in Biella and Ferrara (northern Italy) than elsewhere. Information on all four core items was available for 78% of cases. Odds of four-core-item availability were higher (than mean) in Biella and lower in Latina (centre) and Ragusa (south). CONCLUSIONS: The availability of information important for staging and management was good overall on pathology reports, but varied with geography. It is likely to be improved by wider dissemination of reporting guidelines and adoption of a standardised synoptic reporting system.