Comparison of the sensitivity of different criteria to select lung cancer patients for screening in a cohort of German patients.

INTRODUCTION: Trials of CT-based screening for lung cancer have shown a mortality advantage for screening in North America and Europe. Before introducing a nationwide lung cancer screening program in Germany, it is important to assess the criteria used in international trials in the German population. METHODS: We used data from 3623 lung cancer patients from the data warehouse of the German Center for Lung Research (DZL). We compared the sensitivity of the following lung cancer screening criteria overall and stratified by age and histology: the National Lung Screening Trial (NLST), the Danish Lung Cancer Screening Trial (DLCST), the 2013 and 2021 US Preventive Services Task Force (USPSTF), and an adapted version of the Prostate, Lung, Colorectal, and Ovarian no race model (adapted PLCOm2012) with 6-year risk thresholds of 1.0%/6 year and 1.7%/6 year. RESULTS: Overall, the adapted PLCOm2012 model (1%/6 years), selected the highest proportion of lung cancer patients for screening (72.4%), followed by the 2021 USPSTF (70.0%), the adapted PLCOm2012 (1.7%/6 year) (57.4%), the 2013 USPTF (57.0%), DLCST criteria (48.7%), and the NLST (48.5%). The adapted PLCOm2012 risk model (1.0%/6 year) had the highest sensitivity for all histological types except for small-cell and large-cell carcinomas (non-significant), whereas the 2021 USPTF selected a higher proportion of patients. The sensitivity levels were higher in males than in females. CONCLUSION: Using a risk-based selection score resulted in higher sensitivities compared to criteria using dichotomized age and smoking history. However, gender disparities were apparent in all studied eligibility criteria. In light of increasing lung cancer incidences in women, all selection criteria should be reviewed for ways to close this gender gap, especially when implementing a large-scale lung cancer screening program.

Proof-of-Concept Integration of Heterogeneous Biobank IT Infrastructures into a Hybrid Biobanking Network.

Cross-institutional biobank networks hold the promise of supporting medicine by enabling the exchange of associated samples for research purposes. Various initiatives, such as BBMRI-ERIC and German Biobank Node (GBN), aim to interconnect biobanks for enabling the compilation of joint biomaterial collections. However, building software platforms to facilitate such collaboration is challenging due to the heterogeneity of existing biobank IT infrastructures and the necessary efforts for installing and maintaining additional software components. As a remedy, this paper presents the concept of a hybrid network for interconnecting already existing software components commonly found in biobanks and a proof-of-concept implementation of two prototypes involving four biobanks of the German Biobank Node. Here we demonstrate the successful bridging of two IT systems found in many German biobanks - Samply and i2b2.

Automated Classification of Selected Data Elements from Free-text Diagnostic Reports for Clinical Research.

OBJECTIVES: In the Multiple Myeloma clinical registry at Heidelberg University Hospital, most data are extracted from discharge letters. Our aim was to analyze if it is possible to make the manual documentation process more efficient by using methods of natural language processing for multiclass classification of free-text diagnostic reports to automatically document the diagnosis and state of disease of myeloma patients. The first objective was to create a corpus consisting of free-text diagnosis paragraphs of patients with multiple myeloma from German diagnostic reports, and its manual annotation of relevant data elements by documentation specialists. The second objective was to construct and evaluate a framework using different NLP methods to enable automatic multiclass classification of relevant data elements from free-text diagnostic reports. METHODS: The main diagnoses paragraph was extracted from the clinical report of one third randomly selected patients of the multiple myeloma research database from Heidelberg University Hospital (in total 737 selected patients). An EDC system was setup and two data entry specialists performed independently a manual documentation of at least nine specific data elements for multiple myeloma characterization. Both data entries were compared and assessed by a third specialist and an annotated text corpus was created. A framework was constructed, consisting of a self-developed package to split multiple diagnosis sequences into several subsequences, four different preprocessing steps to normalize the input data and two classifiers: a maximum entropy classifier (MEC) and a support vector machine (SVM). In total 15 different pipelines were examined and assessed by a ten-fold cross-validation, reiterated 100 times. For quality indication the average error rate and the average F1-score were conducted. For significance testing the approximate randomization test was used. RESULTS: The created annotated corpus consists of 737 different diagnoses paragraphs with a total number of 865 coded diagnosis. The dataset is publicly available in the supplementary online files for training and testing of further NLP methods. Both classifiers showed low average error rates (MEC: 1.05; SVM: 0.84) and high F1-scores (MEC: 0.89; SVM: 0.92). However the results varied widely depending on the classified data element. Preprocessing methods increased this effect and had significant impact on the classification, both positive and negative. The automatic diagnosis splitter increased the average error rate significantly, even if the F1-score decreased only slightly. CONCLUSIONS: The low average error rates and high average F1-scores of each pipeline demonstrate the suitability of the investigated NPL methods. However, it was also shown that there is no best practice for an automatic classification of data elements from free-text diagnostic reports.