C2C 2020 Abstracts

Area 1 - C2C

Full Papers
Paper Nr: 2

Automatic ICD-10 Classification of Diseases from Dutch Discharge Letters


Ayoub Bagheri, Arjan Sammani, Peter M. Van Der Heijden, Folkert W. Asselbergs and Daniel L. Oberski

Abstract: The international classification of diseases (ICD) is a widely used tool to describe patient diagnoses. At University Medical Center Utrecht (UMCU), for example, trained medical coders translate information from hospital discharge letters into ICD-10 codes for research and national disease epidemiology statistics, at considerable cost. To mitigate these costs, automatic ICD coding from discharge letters would be useful. However, this task has proven challenging in practice: it is a multi-label task with a large number of very sparse categories, presented in a hierarchical structure. Moreover, existing ICD systems have been benchmarked only on relatively easier versions of this task, such as single-label performance and performance on the higher “chapter” level of the ICD hierarchy, which contains fewer categories. In this study, we benchmark the state-of-the-art ICD classification systems and two baseline systems on a large dataset constructed from Dutch cardiology discharge letters at UMCU hospital. Performance of all systems is evaluated for both the easier chapter-level ICD codes and single-label version of the task found in the literature, as well as for the lower-level ICD hierarchy and multi-label task that is needed in practice. We find that state-of-the-art methods outperform the baseline for the single-label version of the task only. For the multi-label task, the baselines are not defeated by any state-of-the-art system, with the exception of HA-GRU, which does perform best in the most difficult task on accuracy. We conclude that practical performance may have been somewhat overstated in the literature, although deep learning techniques are sufficiently good to complement, though not replace, human ICD coding in our application.

Paper Nr: 3

Multi-class Semantic Segmentation of Skin Lesions via Fully Convolutional Networks


Manu Goyal, Moi H. Yap and Saeed Hassanpour

Abstract: Melanoma is clinically difficult to distinguish from common benign skin lesions, particularly melanocytic naevus and seborrhoeic keratosis. The dermoscopic appearance of these lesions has huge intra-class variations and high inter-class visual similarities. Most current research is focusing on single-class segmentation irrespective of classes of skin lesions. In this work, we evaluate the performance of deep learning on multi-class segmentation of ISIC-2017 challenge dataset, which consists of 2,750 dermoscopic images. We propose an end-to-end solution using fully convolutional networks (FCNs) for multi-class semantic segmentation to automatically segment the melanoma, seborrhoeic keratosis and naevus. To improve the performance of FCNs, transfer learning and a hybrid loss function are used. We evaluate the performance of the deep learning segmentation methods for multi-class segmentation and lesion diagnosis (with post-processing method) on the testing set of the ISIC-2017 challenge dataset. The results showed that the two-tier level transfer learning FCN-8s achieved the overall best result with Dice score of 78.5% in a naevus category, 65.3% in melanoma, and 55.7% in seborrhoeic keratosis in multi-class segmentation and Accuracy of 84.62% for recognition of melanoma in lesion diagnosis.

Paper Nr: 4

Two-stage Neural-network based Prognosis Models using Pathological Image and Transcriptomic Data: An Application in Hepatocellular Carcinoma Patient Survival Prediction


Zhucheng Zhan, Noshad Hossenei, Olivier Poirion, Maria Westerhoff, Eun-Young Choi, Travers Ching and Lana X. Garmire

Abstract: Pathological images are easily accessible data type with potential as prognostic biomarkers. Here we extend Cox-nnet, a neural network based prognosis method previously used for transcriptomics data, to predict patient survival using hepatocellular carcinoma (HCC) pathological images. Cox-nnet based imaging predictions are more robust and accurate than Cox proportional hazards model. Moreover, using a novel two-stage Cox-nnet complex model, we are able to combine histopathology image and transcriptomics RNA-Seq data to make impressively accurate prognosis predictions, with C-index close to 0.90 and log-ranked p-value of 4e-21 in the testing dataset. This work provides a new, biologically relevant and relatively interpretable solution to the challenge of integrating multi-modal and multiple types of data, particularly for survival prediction.

Paper Nr: 6

Preliminary Evaluation of the Utility of Deep Generative Histopathology Image Translation at a Mid-sized NCI Cancer Center


Joshua J. Levy, Christopher R. Jackson, Aravindhan Sriharan, Brock C. Christensen and Louis J. Vaickus

Abstract: Evaluation of a tissue biopsy is often required for the diagnosis and prognostic staging of a disease. Recent efforts have sought to accurately quantitate the distribution of tissue features and morphology in digitized images of histological tissue sections, Whole Slide Images (WSI). Generative modeling techniques present a unique opportunity to produce training data that can both augment these models and translate histologic data across different intra-and-inter-institutional processing procedures, provide cost-effective ways to perform computational chemical stains (synthetic stains) on tissue, and facilitate the creation of diagnostic aid algorithms. A critical evaluation and understanding of these technologies is vital for their incorporation into a clinical workflow. We illustrate several potential use cases of these techniques for the calculation of nuclear to cytoplasm ratio, synthetic SOX10 immunohistochemistry (IHC, sIHC) staining to delineate cell lineage, and the conversion of hematoxylin and eosin (H&E) stain to trichome stain for the staging of liver fibrosis.

Short Papers
Paper Nr: 5

Semantic Overlapping in Translational Bioinformatics Applied to the Matching between Clinical Trial Eligibility Criteria and Patient Needs


Radmila Juric, Eton Williams and Inhwa Kim

Abstract: Software technologies play an important role in defining clinical trials, their eligibility criteria and recruitment process, in which patient enrol to a trial if they satisfy eligibility criteria. In this research we address the problem of semantic overlapping between eligibility criteria and patient needs through a software architectural model which houses a specific computational model based on reasoning upon the overlapping semantics. The architectural model is deployed using semantic technologies in order to explore the meaning of the relationships between trials, eligibility criteria and patient needs. The novelty is in the reusability and thus converting of the existing conceptual models on deriving eligibility criteria, available in literature, into the proposed OWL model, which can serve any clinical trial and requirements patients may have. This paper is written by computer scientists interested in manipulating semantics of data through computational models using modern software technologies. It serves as an invitation to researchers from the biomedical and translational informatics to debate the future of software support in managing clinical trials.