HealthMyne provides radiologists and oncology clinicians with an Imaging Informatics platform that brings together imaging and electronic health record (EHR) information for use in evidence-based analytics and decision support. At the heart of the system is a powerful image analysis engine that generates information such as tumor size, Lung-RADS categories for use in lung cancer screening, and other advanced quantitative metrics derived from the imaging study.
At HealthMyne, our display technology transforms workflow that is purely case-focused into one that is also
correlative and evidence-based through convenient comparison of relevant findings.
Exams, test results, treatment details, and health status for each patient are viewable in an intuitive timeline-based representation. This provides a longitudinal perspective, and clear visual insights, into a patient’s continuum of care. The timeline could include graphical representations of laboratory tests, imaging procedures, disease diagnosis and courses of therapy, along with quantitative values and plots from imaging and non-image-based analyses.
An example of a longitudinal representation features a plot of tumor size relative to the duration of a course of radiotherapy. Icons denote the dates of follow-up CT scans from which tumor size was determined. Scans can then be examined by clicking on the icon and opening a viewer. Elements of the time line, such as analytics derived from follow-up exams are added as they become available. With this information all in one place, the need to refer back to paper and other electronic information sources while reading and reporting on imaging exams is eliminated.
Unfortunately, a significant amount of data that must be quantified and captured to enable Precision Medicine remains trapped in the pixels of medical images. The most straightforward of these quantitative metrics are size measurements of pathological tissues such as tumors. More advanced metrics, such as those based on texture of the image pixels within a tumor, are representative of tissue structure such as heterogeneity in density. This in turn may indicate the presence of sub-populations of resistant cells, or variations in blood supply, within the tumor. The extraction and clinical use of image features such as size, shape and texture is referred to as quantitative imaging, or radiomics.
HealthMyne reinterprets the pixel data as a set of quantitative metrics that can be assessed in conjunction with other information about the patient. Core functionality includes a suite of exam reading tools, with automated segmentation of selected lesions, nodules or tumors. Image analytics can be performed for segmented volumes, deriving properties such as nodule size and growth and adding these to the patient record. In a screening context, nodules can be assigned standardized properties such as a Lung-RADS category or Fleischner Society size range. Automatic registration of current and prior exams aids in nodule identification as well as in visualizing and quantifying changes such as nodule progression or tumor response.
Novel search and cohort creation tools have been implemented to quickly define groups of patients according to selected properties such as demographics, medical history, nodule progression, tumor response or selected radiomic features. For example, high and low risk lung cancer cohorts can be created according to smoking history and other demographics. Cohorts are defined by filters, each narrowing the scope of cohort membership.
Nodule progression and other metrics can be analyzed across cohorts, providing a latitudinal perspective of imaging analytics information. Additionally, if outcome data is included in the information derived from the EHR (e.g. whether a nodule was determined to be malignant or whether a patient survived a certain time past diagnosis), searches can then reveal the rate of a particular outcome for selected cohorts. Outcomes themselves can be used to define cohorts and metrics for a given patient can be compared with those patients in the cohort. By analyzing cohorts of past patients with known diagnoses and outcomes, clinicians can make more informed decisions about the prognosis and care of a current patient.
With the ability to evaluate data derived across patient cohorts in a database, each and every patient within the enterprise becomes part of evidence-based analysis. Radiologists will be able to tap findings across all patients, rather than the 3% of cancer patients in clinical trials.
Reporting & Sharing
Providing the radiologist with new evidence-based insights and analytical tools is certainly an important step forward, but the biggest impact is realized when the radiologist is also given a seamless mechanism to efficiently share those added insights with the rest of the clinical care team. HealthMyne offers a unique and effective EHR integration pathway for communicating their quantitative findings and well-informed impressions.
Another radiologist workflow challenge involves the ever-increasing reporting requirements. At HealthMyne, high priority is being given to the development of specific capabilities such as automated determination of Lung-RADS categories and assembly of structured reports according to templates such as those arrived at by expert groups within the RSNA reporting initiative.
HealthMyne is incorporating radiomic analysis as part of its image analytics toolkit. Many additional features of a nodule can be quantified using radiomic feature extraction algorithms. Features are quantifiable characteristics from within categories such as size, shape, intensity and texture, any of which may be useful biomarkers for prediction of outcome. The goal is to allow customers to participate in the discovery and use of these biomarkers. A mathematical combination of selected feature values can form a signature that is a more robust prognostic indicator. Our approach is to provide a platform that will accommodate selected radiomic features and signatures on a modular basis.