Search

We developed end-to-end deep learning models using whole slide images of adults diagnosed with diffusely infiltrating, World Health Organization (WHO) grade 2 gliomas to predict prognosis and the mutation status of a somatic biomarker, isocitrate dehydrogenase (IDH) 1/2. The models, which utilize ResNet-18 as a backbone, were developed and validated on 296 patients from The Cancer Genome Atlas (TCGA) database. To account for the small sample size, repeated random train/test splits were performed for hyperparameter tuning, and the out-of-sample predictions were pooled for evaluation. Our models achieved a concordance- (C-) index of 0.715 (95% CI: 0.569, 0.830) for predicting prognosis and an area under the curve (AUC) of 0.667 (0.532, 0.784) for predicting IDH mutations. When combined with additional clinical information, the performance metrics increased to 0.784 (95% CI: 0.655, 0.880) and 0.739 (95% CI: 0.613, 0.856), respectively. When evaluated on the WHO grade 3 gliomas from the TCGA dataset, which were not used for training, our models predicted survival with a C-index of 0.654 (95% CI: 0.537, 0.768) and IDH mutations with an AUC of 0.814 (95% CI: 0.721, 0.897). If validated in a prospective study, our method could potentially assist clinicians in managing and treating patients with diffusely infiltrating gliomas. [ABSTRACT FROM AUTHOR]

Purpose: Obesity associated fat infiltration of organ systems is accompanied by organ dysfunction and poor cancer outcomes. Obese women demonstrate variable degrees of fat infiltration of axillary lymph nodes (LNs), and they are at increased risk for node-positive breast cancer. However, the relationship between enlarged axillary nodes and axillary metastases has not been investigated. The purpose of this study is to evaluate the association between axillary metastases and fat-enlarged axillary nodes visualized on mammograms and breast MRI in obese women with a diagnosis of invasive breast cancer. Methods: This retrospective case–control study included 431 patients with histologically confirmed invasive breast cancer. The primary analysis of this study included 306 patients with pre-treatment and pre-operative breast MRI and body mass index (BMI) > 30 (201 node-positive cases and 105 randomly selected node-negative controls) diagnosed with invasive breast cancer between April 1, 2011, and March 1, 2020. The largest visible LN was measured in the axilla contralateral to the known breast cancer on breast MRI. Multivariate logistic regression models were used to assess the association between node-positive status and LN size adjusting for age, BMI, tumor size, tumor grade, tumor subtype, and lymphovascular invasion. Results: A strong likelihood of node-positive breast cancer was observed among obese women with fat-expanded lymph nodes (adjusted OR for the 4th vs. 1st quartile for contralateral LN size on MRI: 9.70; 95% CI 4.26, 23.50; p < 0.001). The receiver operating characteristic curve for size of fat-enlarged nodes in the contralateral axilla identified on breast MRI had an area under the curve of 0.72 for predicting axillary metastasis, and this increased to 0.77 when combined with patient and tumor characteristics. Conclusion: Fat expansion of axillary lymph nodes was associated with a high likelihood of axillary metastases in obese women with invasive breast cancer independent of BMI and tumor characteristics. [ABSTRACT FROM AUTHOR]

Renal cell carcinoma (RCC) is the most common renal cancer in adults. The histopathologic classification of RCC is essential for diagnosis, prognosis, and management of patients. Reorganization and classification of complex histologic patterns of RCC on biopsy and surgical resection slides under a microscope remains a heavily specialized, error-prone, and time-consuming task for pathologists. In this study, we developed a deep neural network model that can accurately classify digitized surgical resection slides and biopsy slides into five related classes: clear cell RCC, papillary RCC, chromophobe RCC, renal oncocytoma, and normal. In addition to the whole-slide classification pipeline, we visualized the identified indicative regions and features on slides for classification by reprocessing patch-level classification results to ensure the explainability of our diagnostic model. We evaluated our model on independent test sets of 78 surgical resection whole slides and 79 biopsy slides from our tertiary medical institution, and 917 surgical resection slides from The Cancer Genome Atlas (TCGA) database. The average area under the curve (AUC) of our classifier on the internal resection slides, internal biopsy slides, and external TCGA slides is 0.98 (95% confidence interval (CI): 0.97–1.00), 0.98 (95% CI: 0.96–1.00) and 0.97 (95% CI: 0.96–0.98), respectively. Our results suggest that the high generalizability of our approach across different data sources and specimen types. More importantly, our model has the potential to assist pathologists by (1) automatically pre-screening slides to reduce false-negative cases, (2) highlighting regions of importance on digitized slides to accelerate diagnosis, and (3) providing objective and accurate diagnosis as the second opinion. [ABSTRACT FROM AUTHOR]

Electronic medical record (EMR) systems provide easy access to radiology reports and offer great potential to support quality improvement efforts and clinical research. Harnessing the full potential of the EMR requires scalable approaches such as natural language processing (NLP) to convert text into variables used for evaluation or analysis. Our goal was to determine the feasibility of using NLP to identify patients with Type 1 Modic endplate changes using clinical reports of magnetic resonance (MR) imaging examinations of the spine. Identifying patients with Type 1 Modic change who may be eligible for clinical trials is important as these findings may be important targets for intervention. Four annotators identified all reports that contained Type 1 Modic change, using N = 458 randomly selected lumbar spine MR reports. We then implemented a rule-based NLP algorithm in Java using regular expressions. The prevalence of Type 1 Modic change in the annotated dataset was 10%. Results were recall (sensitivity) 35/50 = 0.70 (95% confidence interval (C.I.) 0.52–0.82), specificity 404/408 = 0.99 (0.97–1.0), precision (positive predictive value) 35/39 = 0.90 (0.75–0.97), negative predictive value 404/419 = 0.96 (0.94–0.98), and F1-score 0.79 (0.43–1.0). Our evaluation shows the efficacy of rule-based NLP approach for identifying patients with Type 1 Modic change if the emphasis is on identifying only relevant cases with low concern regarding false negatives. As expected, our results show that specificity is higher than recall. This is due to the inherent difficulty of eliciting all possible keywords given the enormous variability of lumbar spine reporting, which decreases recall, while availability of good negation algorithms improves specificity. [ABSTRACT FROM AUTHOR]

We built a natural language processing (NLP) method to automatically extract clinical findings in radiology reports and characterize their level of change and significance according to a radiology-specific information model. We utilized a combination of machine learning and rule-based approaches for this purpose. Our method is unique in capturing different features and levels of abstractions at surface, entity, and discourse levels in text analysis. This combination has enabled us to recognize the underlying semantics of radiology report narratives for this task. We evaluated our method on radiology reports from four major healthcare organizations. Our evaluation showed the efficacy of our method in highlighting important changes (accuracy 99.2%, precision 96.3%, recall 93.5%, and F1 score 94.7%) and identifying significant observations (accuracy 75.8%, precision 75.2%, recall 75.7%, and F1 score 75.3%) to characterize radiology reports. This method can help clinicians quickly understand the key observations in radiology reports and facilitate clinical decision support, review prioritization, and disease surveillance. [ABSTRACT FROM AUTHOR]

Radiology report narrative contains a large amount of information about the patient's health and the radiologist's interpretation of medical findings. Most of this critical information is entered in free text format, even when structured radiology report templates are used. The radiology report narrative varies in use of terminology and language among different radiologists and organizations. The free text format and the subtlety and variations of natural language hinder the extraction of reusable information from radiology reports for decision support, quality improvement, and biomedical research. Therefore, as the first step to organize and extract the information content in a large multi-institutional free text radiology report repository, we have designed and developed an unsupervised machine learning approach to capture the main concepts in a radiology report repository and partition the reports based on their main foci. In this approach, radiology reports are modeled in a vector space and compared to each other through a cosine similarity measure. This similarity is used to cluster radiology reports and identify the repository's underlying topics. We applied our approach on a repository of 1,899,482 radiology reports from three major healthcare organizations. Our method identified 19 major radiology report topics in the repository and clustered the reports accordingly to these topics. Our results are verified by a domain expert radiologist and successfully explain the repository's primary topics and extract the corresponding reports. The results of our system provide a target-based corpus and framework for information extraction and retrieval systems for radiology reports. [ABSTRACT FROM AUTHOR]

Classification of histologic patterns in lung adenocarcinoma is critical for determining tumor grade and treatment for patients. However, this task is often challenging due to the heterogeneous nature of lung adenocarcinoma and the subjective criteria for evaluation. In this study, we propose a deep learning model that automatically classifies the histologic patterns of lung adenocarcinoma on surgical resection slides. Our model uses a convolutional neural network to identify regions of neoplastic cells, then aggregates those classifications to infer predominant and minor histologic patterns for any given whole-slide image. We evaluated our model on an independent set of 143 whole-slide images. It achieved a kappa score of 0.525 and an agreement of 66.6% with three pathologists for classifying the predominant patterns, slightly higher than the inter-pathologist kappa score of 0.485 and agreement of 62.7% on this test set. All evaluation metrics for our model and the three pathologists were within 95% confidence intervals of agreement. If confirmed in clinical practice, our model can assist pathologists in improving classification of lung adenocarcinoma patterns by automatically pre-screening and highlighting cancerous regions prior to review. Our approach can be generalized to any whole-slide image classification task, and code is made publicly available at https://github.com/BMIRDS/deepslide. [ABSTRACT FROM AUTHOR]