Although machine learning (ML) has shown promise across disciplines, out-of-sample generalizability is concerning. This is currently addressed by sharing multi-site data, but such centralization is challenging/infeasible to scale due to various limitations. Federated ML (FL) provides an alternative paradigm for accurate and generalizable ML, by only sharing numerical model updates. Here we present the largest FL study to-date, involving data from 71 sites across 6 continents, to generate an automatic tumor boundary detector for the rare disease of glioblastoma, reporting the largest such dataset in the literature (n = 6, 314). We demonstrate a 33% delineation improvement for the surgically targetable tumor, and 23% for the complete tumor extent, over a publicly trained model. We anticipate our study to: 1) enable more healthcare studies informed by large diverse data, ensuring meaningful results for rare diseases and underrepresented populations, 2) facilitate further analyses for glioblastoma by releasing our consensus model, and 3) demonstrate the FL effectiveness at such scale and task-complexity as a paradigm shift for multi-site collaborations, alleviating the need for data-sharing.

Differentiating pseudoprogression from true tumor progression has become a significant challenge in follow-up of diffuse infiltrating gliomas, particularly high grade, which leads to a potential treatment delay for patients with early glioma recurrence. In this study, we proposed to use a multiparametric MRI data as a sequence input for the convolutional neural network with the recurrent neural network based deep learning structure to discriminate between pseudoprogression and true tumor progression. In this study, 43 biopsy-proven patient data identified as diffuse infiltrating glioma patients whose disease progressed/recurred were used. The dataset consists of five original MRI sequences; pre-contrast T1-weighted, post-contrast T1-weighted, T2-weighted, FLAIR, and ADC images as well as two engineered sequences; T1post–T1pre and T2–FLAIR. Next, we used three CNN-LSTM models with a different set of sequences as input sequences to pass through CNN-LSTM layers. We performed threefold cross-validation in the training dataset and generated the boxplot, accuracy, and ROC curve, AUC from each trained model with the test dataset to evaluate models. The mean accuracy for VGG16 models ranged from 0.44 to 0.60 and the mean AUC ranged from 0.47 to 0.59. For CNN-LSTM model, the mean accuracy ranged from 0.62 to 0.75 and the mean AUC ranged from 0.64 to 0.81. The performance of the proposed CNN-LSTM with multiparametric sequence data was found to outperform the popular convolutional CNN with a single MRI sequence. In conclusion, incorporating all available MRI sequences into a sequence input for a CNN-LSTM model improved diagnostic performance for discriminating between pseudoprogression and true tumor progression.

ObjectiveTo investigate the presence of gender disparity in academic involvement during radiology residency and to identify and characterize any gender differences in perceived barriers for conducting research.MethodsAn international call for participation in an online survey was promoted via social media and through multiple international and national radiological societies. A 35-question survey invited radiology trainees worldwide to answer questions regarding exposure and barriers to academic radiology during their training. Gender differences in response proportions were analyzed using either Fisher’s exact or chi-squared tests.ResultsEight hundred fifty-eight participants (438 men, 420 women) from Europe (432), Asia (241), North and South America (144), Africa (37), and Oceania (4) completed the survey. Fewer women radiology residents were involved in research during residency (44.3%, 186/420 vs 59.4%, 260/438; p ≤ 0.0001) and had fewer published original articles (27.9%, 117/420 vs. 40.2%, 176/438; p = 0.001).Women were more likely to declare gender as a barrier to research (24.3%, 102/420 vs. 6.8%, 30/438; p < 0.0001) and lacked mentorship/support from faculty (65%, 273/420 vs. 55.7%, 244/438; p = 0.0055). Men were more likely to declare a lack of time (60.3%, 264/438 vs. 50.7%, 213/420; p = 0.0049) and lack of personal interest (21%, 92/438 vs. 13.6%, 57/420, p = 0.0041) in conducting research.ConclusionFewer women were involved in academic activities during radiology residency, resulting in fewer original published studies compared to their men counterparts. This is indicative of an inherent gender imbalance. Lack of mentorship reported by women radiologists was a main barrier to research.

Ollier disease is a rare condition presenting with enchondromas in an irregular distribution within the medullary cavity of bones. The disease is well known for sarcomatous transformation to chondrosarcomas. It also increases the risk of other malignancies like leukemia, ovarian tumors, and glial tumors. Central nervous system malignancies associated with Ollier disease are thought to arise by somatic IDH mosaicism with their atypical features of distribution, multifocality, and age of onset. We present a case with imaging consistent with diffuse midline glioma in a patient with Ollier disease. We conclude with a brief review of the literature on Ollier Disease with a focus on central nervous system malignancies, tumorigenesis and pathophysiology.