Dataset Information

Proof-of-concept study: Homomorphically encrypted data can support real-time learning in personalized cancer medicine.

ABSTRACT: BACKGROUND:The successful introduction of homomorphic encryption (HE) in clinical research holds promise for improving acceptance of data-sharing protocols, increasing sample sizes, and accelerating learning from real-world data (RWD). A well-scoped use case for HE would pave the way for more widespread adoption in healthcare applications. Determining the efficacy of targeted cancer treatments used off-label for a variety of genetically defined conditions is an excellent candidate for introduction of HE-based learning systems because of a significant unmet need to share and combine confidential data, the use of relatively simple algorithms, and an opportunity to reach large numbers of willing study participants. METHODS:We used published literature to estimate the numbers of patients who might be eligible to receive treatments approved for other indications based on molecular profiles. We then estimated the sample size and number of variables that would be required for a successful system to detect exceptional responses with sufficient power. We generated an appropriately sized, simulated dataset (n =?5000) and used an established HE algorithm to detect exceptional responses and calculate total drug exposure, while the data remained encrypted. RESULTS:Our results demonstrated the feasibility of using an HE-based system to identify exceptional responders and perform calculations on patient data during a hypothetical 3-year study. Although homomorphically encrypted computations are time consuming, the required basic computations (i.e., addition) do not pose a critical bottleneck to the analysis. CONCLUSION:In this proof-of-concept study, based on simulated data, we demonstrate that identifying exceptional responders to targeted cancer treatments represents a valuable and feasible use case. Past solutions to either completely anonymize data or restrict access through stringent data use agreements have limited the utility of abundant and valuable data. Because of its privacy protections, we believe that an HE-based learning system for real-world cancer treatment would entice thousands more patients to voluntarily contribute data through participation in research studies beyond the currently available secondary data populated from hospital electronic health records and administrative claims. Forming collaborations between technical experts, physicians, patient advocates, payers, and researchers, and testing the system on existing RWD are critical next steps to making HE-based learning a reality in healthcare.

SUBMITTER: Paddock S

PROVIDER: S-EPMC6894133 | biostudies-literature | 2019 Dec

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:As complex and heterogeneous diseases, cancers require a more tailored therapeutic management than most pathologies. Recent advances in anticancer drug development, including the immuno-oncology revolution, have been too often plagued by unsatisfying patient response rates and survivals. In reaction to this, cancer care has fully transitioned to the "personalized medicine" concept. Numerous tools are now available tools to better adapt treatments to the profile of each patient. They encompass a large array of diagnostic assays, based on biomarkers relevant to targetable molecular pathways. As a subfamily of such so-called companion diagnostics, chemosensitivity and resistance assays represent an attractive, yet insufficiently understood, approach to individualize treatments. They rely on the assessment of a composite biomarker, the ex vivo functional response of cancer cells to drugs, to predict a patient's outcome. Systemic treatments, such as chemotherapies, as well as targeted treatments, whose efficacy cannot be fully predicted yet by other diagnostic tests, may be assessed through these means. The results can provide helpful information to assist clinicians in their decision-making process. We explore here the most advanced functional assays across oncology indications, with an emphasis on tests already displaying a convincing clinical demonstration. We then recapitulate the main technical obstacles faced by researchers and clinicians to produce more accurate, and thus more predictive, models and the recent advances that have been developed to circumvent them. Finally, we summarize the regulatory and quality frameworks surrounding functional assays to ensure their safe and performant clinical implementation. Functional assays are valuable in vitro diagnostic tools that already stand beyond the "proof-of-concept" stage. Clinical studies show they have a major role to play by themselves but also in conjunction with molecular diagnostics. They now need a final lift to fully integrate the common armament used against cancers, and thus make their way into the clinical routine.

Project description:To develop a novel four-dimensional (4D) intensity modulated radiation therapy (IMRT) treatment planning methodology based on dynamic virtual patient models.The 4D model-based planning (4DMP) is a predictive tracking method which consists of two main steps: (1) predicting the 3D deformable motion of the target and critical structures as a function of time during treatment delivery; (2) adjusting the delivery beam apertures formed by the dynamic multi-leaf collimators (DMLC) to account for the motion. The key feature of 4DMP is the application of a dynamic virtual patient model in motion prediction, treatment beam adjustment, and dose calculation. A lung case was chosen to demonstrate the feasibility of the 4DMP. For the lung case, a dynamic virtual patient model (4D model) was first developed based on the patient's 4DCT images. The 4D model was capable of simulating respiratory motion of different patterns. A model-based registration method was then applied to convert the 4D model into a set of deformation maps and 4DCT images for dosimetric purposes. Based on the 4D model, 4DMP treatment plans with different respiratory motion scenarios were developed. The quality of 4DMP plans was then compared with two other commonly used 4D planning methods: maximum intensity projection (MIP) and planning on individual phases (IP).Under regular periodic motion, 4DMP offered similar target coverage as MIP with much better normal tissue sparing. At breathing amplitude of 2 cm, the lung V20 was 23.9% for a MIP plan and 16.7% for a 4DMP plan. The plan quality was comparable between 4DMP and IP: PTV V97 was 93.8% for the IP plan and 93.6% for the 4DMP plan. Lung V20 of the 4DMP plan was 2.1% lower than that of the IP plan and Dmax to cord was 2.2 Gy higher. Under a real time irregular breathing pattern, 4DMP had the best plan quality. PTV V97 was 90.4% for a MIP plan, 88.6% for an IP plan and 94.1% for a 4DMP plan. Lung V20 was 20.1% for the MIP plan, 17.8% for the IP plan and 17.5% for the 4DMP plan. The deliverability of the real time 4DMP plan was proved by calculating the maximum leaf speed of the DMLC.The 4D model-based planning, which applies dynamic virtual patient models in IMRT treatment planning, can account for the real time deformable motion of the tumor under different breathing conditions. Under regular motion, the quality of 4DMP plans was comparable with IP and superior to MIP. Under realistic motion in which breathing amplitude and period change, 4DMP gave the best plan quality of the three 4D treatment planning techniques.

Dataset Information

Proof-of-concept study: Homomorphically encrypted data can support real-time learning in personalized cancer medicine.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets