Search

Background Radiomics is the extraction of predefined mathematic features from medical images for the prediction of variables of clinical interest. While some studies report superlative accuracy of radiomic machine learning (ML) models, the published methodology is often incomplete, and the results are rarely validated in external testing data sets. Purpose To characterize the type, prevalence, and statistical impact of methodologic errors present in radiomic ML studies. Materials and Methods Radiomic ML publications were reviewed for the presence of performance-inflating methodologic flaws. Common flaws were subsequently reproduced with randomly generated features interpolated from publicly available radiomic data sets to demonstrate the precarious nature of reported findings. Results In an assessment of radiomic ML publications, the authors uncovered two general categories of data analysis errors: inconsistent partitioning and unproductive feature associations. In simulations, the authors demonstrated that inconsistent partitioning augments radiomic ML accuracy by 1.4 times from unbiased performance and that correcting for flawed methodologic results in areas under the receiver operating characteristic curve approaching a value of 0.5 (random chance). With use of randomly generated features, the authors illustrated that unproductive associations between radiomic features and gene sets can imply false causality for biologic phenomenon. Conclusion Radiomic machine learning studies may contain methodologic flaws that undermine their validity. This study provides a review template to avoid such flaws. © RSNA, 2022

Cervical cancer is a leading cause of cancer mortality, with approximately 90% of the 250,000 deaths per year occurring in low- and middle-income countries (LMIC). Secondary prevention with cervical screening involves detecting and treating precursor lesions; however, scaling screening efforts in LMIC has been hampered by infrastructure and cost constraints. Recent work has supported the development of an artificial intelligence (AI) pipeline on digital images of the cervix to achieve an accurate and reliable diagnosis of treatable precancerous lesions. In particular, WHO guidelines emphasize visual triage of women testing positive for human papillomavirus (HPV) as the primary screen, and AI could assist in this triage task. Published AI reports have exhibited overfitting, lack of portability, and unrealistic, near-perfect performance estimates. To surmount recognized issues, we implemented a comprehensive deep-learning model selection and optimization study on a large, collated, multi-institutional dataset of 9,462 women (17,013 images). We evaluated relative portability, repeatability, and classification performance. The top performing model, when combined with HPV type, achieved an area under the Receiver Operating Characteristics (ROC) curve (AUC) of 0.89 within our study population of interest, and a limited total extreme misclassification rate of 3.4%, on held-aside test sets. Our work is among the first efforts at designing a robust, repeatable, accurate and clinically translatable deep-learning model for cervical screening.

Up to 30% of patients with glioblastoma (GBM) develop venous thromboembolism (VTE) over the course of the disease. Although not as high, the risk for VTE is also increased in patients with meningioma. Direct measurement of peak thrombin generation (TG) allows quantitative assessment of systemic coagulation activation in patients with GBM and meningioma. Our aim was to determine the extent of systemic coagulation activation induced by brain tumors, to measure the shift between pre- and post-operative peak TG in patients with GBM, and to assess the relationship between pre-surgical peak TG and pre-operative brain tumor volume on imaging. Pre- and post-surgical plasma samples were obtained from successive patients with GBM and once from patients with meningioma and healthy age- and sex-matched blood donor controls. TG was measured using the calibrated automated thrombogram (CAT) assay, and tumor volumes were measured in pre-surgical MRI scans. Pre-surgical peak TG was higher in patients with GBM than in controls (288.6 ± 54.1 nM vs 187.1 ± 41.7 nM, respectively, P

Up to 30% of patients with glioblastoma (GBM) develop venous thromboembolism (VTE) over the course of the disease. Although not as high, the risk for VTE is also increased in patients with meningioma. Direct measurement of peak thrombin generation (TG) allows quantitative assessment of systemic coagulation activation in patients with GBM and meningioma. Our aim was to determine the extent of systemic coagulation activation induced by brain tumors, to measure the shift between pre- and post-operative peak TG in patients with GBM, and to assess the relationship between pre-surgical peak TG and pre-operative brain tumor volume on imaging.Pre- and post-surgical plasma samples were obtained from successive patients with GBM and once from patients with meningioma and healthy age- and sex-matched blood donor controls. TG was measured using the calibrated automated thrombogram (CAT) assay, and tumor volumes were measured in pre-surgical MRI scans.Pre-surgical peak TG was higher in patients with GBM than in controls (288.6 ± 54.1 nM vs 187.1 ± 41.7 nM, respectively, P 0.001), and, in the nine patients with GBM and paired data available, peak TG was significantly reduced after surgery (323 ± 38 nM vs 265 ± 52 nM, respectively, P = 0.007). Similarly, subjects with meningioma demonstrated higher peak TG compared to controls (242.2 ± 54.9 nM vs 177.7 ± 57.0 nM, respectively, P 0.001). There was no association between peak TG and pre-operative tumor volume or overall survival.Our results indicate that systemic coagulation activation occurs with both meningioma and GBM, but to a greater degree in the latter. Preoperative peak TG did not correlate with tumor volume, but removal of GBM caused a significant decrease in coagulation activation.

Developments in artificial intelligence have the potential to improve the care of patients with musculoskeletal tumors. We performed a systematic review of the published scientific literature to identify the current state of the art of artificial intelligence applied to musculoskeletal oncology, including both primary and metastatic tumors, and across the radiology, nuclear medicine, pathology, clinical research, and molecular biology literature. Through this search, we identified 252 primary research articles, of which 58 used deep learning and 194 used other machine learning techniques. Articles involving deep learning have mostly involved bone scintigraphy, histopathology, and radiologic imaging. Articles involving other machine learning techniques have mostly involved transcriptomic analyses, radiomics, and clinical outcome prediction models using medical records. These articles predominantly present proof-of-concept work, other than the automated bone scan index for bone metastasis quantification, which has translated to clinical workflows in some regions. We systematically review and discuss this literature, highlight opportunities for multidisciplinary collaboration, and identify potentially clinically useful topics with a relative paucity of research attention. Musculoskeletal oncology is an inherently multidisciplinary field, and future research will need to integrate and synthesize noisy siloed data from across clinical, imaging, and molecular datasets. Building the data infrastructure for collaboration will help to accelerate progress towards making artificial intelligence truly useful in musculoskeletal oncology.

Radiation therapy produces Cherenkov optical emission in tissue, and this light can be utilized to activate molecular probes. The feasibility of sensing luminescence from a tissue molecular oxygen sensor from within a human body phantom was examined using the geometry of the axillary lymph node region. Detection of regions down to 30-mm deep was feasible with submillimeter spatial resolution with the total quantity of the phosphorescent sensor PtG4 near 1 nanomole. Radiation sheet scanning in an epi-illumination geometry provided optimal coverage, and maximum intensity projection images provided illustration of the concept. This work provides the preliminary information needed to attempt this type of imaging in vivo.