Data privacy mechanisms are essential for rapidly scaling medical training databases to capture the heterogeneity of patient data distributions toward robust and generalizable machine learning systems. In the current COVID-19 pandemic, a major focus of artificial intelligence (AI) is interpreting chest CT, which can be readily used in the assessment and management of the disease. This paper demonstrates the feasibility of a federated learning method for detecting COVID-19 related CT abnormalities with external validation on patients from a multinational study. We recruited 132 patients from seven multinational different centers, with three internal hospitals from Hong Kong for training and testing, and four external, independent datasets from Mainland China and Germany, for validating model generalizability. We also conducted case studies on longitudinal scans for automated estimation of lesion burden for hospitalized COVID-19 patients. We explore the federated learning algorithms to develop a privacy-preserving AI model for COVID-19 medical image diagnosis with good generalization capability on unseen multinational datasets. Federated learning could provide an effective mechanism during pandemics to rapidly develop clinically useful AI across institutions and countries overcoming the burden of central aggregation of large amounts of sensitive data.

No abstract

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.

BACKGROUND AND PURPOSE:We evaluated modifications to our contrast-enhanced MR imaging grading system for symptomatic patients with suspected nasopharyngeal carcinoma, aimed at improving discrimination of early-stage cancer and benign hyperplasia. We evaluated a second non-contrast-enhanced MR imaging grading system for asymptomatic patients from nasopharyngeal carcinoma plasma screening programs. MATERIALS AND METHODS:Dedicated nasopharyngeal MR imaging before (plain scan system) and after intravenous contrast administration (current and modified systems) was reviewed in patients from a nasopharyngeal carcinoma-endemic region, comprising 383 patients with suspected disease without nasopharyngeal carcinoma and 383 patients with nasopharyngeal carcinoma. The modified and plain scan systems refined primary tumor criteria, added a nodal assessment, and expanded the system from 4 to 5 grades. The overall combined sensitivity and specificity of the 3 systems were compared using the extended McNemar test (a x 2 value x 2 ð2Þ . 5.99 indicates significance). RESULTS:The current, modified, and plain scan MR imaging systems yielded sensitivities of 99.74%, 97.91%, and 97.65%, respectively, and specificities of 63.45%, 89.56% and 86.42%, respectively. The modified system yielded significantly better performance than the current (x 2 ð2Þ ¼ 122) and plain scan (x 2 ð2Þ ¼ 6.1) systems. The percentages of patients with nasopharyngeal carcinoma in grades 1-2, grade 3, and grades 4-5 for the modified and plain scan MR imaging systems were 0.42% and 0.44%; 6.31% and 6.96%; and 90.36% and 87.79%, respectively. No additional cancers were detected after contrast administration in cases of a plain scan graded 1-2. CONCLUSIONS:We propose a modified MR imaging grading system that improves diagnostic performance for nasopharyngeal carcinoma detection. Contrast was not valuable for low MR imaging grades, and the plain scan shows potential for use in screening programs. ABBREVIATION: NPC ¼ nasopharyngeal carcinoma