How RAD-X solves healthcare challenges from the bedside to nationwide imaging infrastructure.
Patients frequently visited different hospitals within the 21-hospital public cluster, yet clinicians could see only the studies stored in their own hospital. At the same time, the cluster needed centralized teleradiology without creating a second unrestricted PACS or copying complete archives to an external provider.
RAD-X PACS was deployed independently in every hospital and connected through a federated architecture. Each site continues to operate locally, while a central repository synchronizes examination metadata and workflow state. When a clinician opens a patient, local studies appear immediately and RAD-X simultaneously discovers relevant priors from the other hospitals. Those priors become available on the same interface within seconds.
A multidimensional authorization model controls who can access or report which examination, at which hospital, from which device, during which period and under which workflow conditions. Claiming, reporting and status changes are synchronized between local and central systems, preventing duplicate work and keeping reports aligned with hospital HIS workflows.
Etlik City Hospital opened in 2022 as one of Europe’s largest hospital campuses, with approximately 4,050 beds, 2,700+ physicians and 14,445 staff. The new hospital consolidated Dışkapı, Onkoloji, Dr. Sami Ulus and Zübeyde Hanım hospitals. The challenge was not simply to move four legacy archives, but to reconcile their patient and examination histories, preserve every study and make the consolidated clinical history available when the new hospital opened.
RAD-X PACS, RIS and VNA became the enterprise-wide imaging platform. Historical studies and images were migrated losslessly, consolidated and enriched with information from the legacy HIS environments so that fragmented records could become a more complete longitudinal history. The archive was also expanded to 11 new modality types — including EEG/EMG, sleep PSG, ophthalmology OCT and elastography — together with non-DICOM clinical content that had previously remained outside the PACS archive.
The migration was planned around a critical operational objective: clinicians entering the new hospital on its first day should not start with an empty archive. Historical imaging from all four predecessor institutions was already available in the RAD-X environment, giving physicians immediate access to the patient history they needed from the beginning of clinical operations.
Nationwide imaging access required thousands of healthcare facilities to participate without replacing their existing PACS and HIS systems. The platform had to interoperate with global and local vendors, reconcile patient identities across institutions and remain flexible enough for national workflows. Because this is mission-critical governmental infrastructure, the Ministry also needed full operational and technical control rather than long-term dependence on a single vendor.
Using RAD-X technologies, a turnkey national platform was developed for the Ministry of Health. International standards such as XDS and HL7 provide a common integration foundation for major global vendors, while flexible integration services address local and country-specific requirements. A Master Patient Index reconciles alternative identifiers, and authorized clinicians can access longitudinal images and reports directly from their clinical workflow, subject to applicable consent and authorization policies.
The application and its delivery environment were designed so the Ministry can build, deploy, operate and maintain the platform independently if required. Customer-controlled continuous integration and deployment environments transfer control of the software lifecycle as well as the production system. The platform runs on Kubernetes, providing scalable container orchestration, deployment flexibility, service isolation and infrastructure portability. Simplex continues to maintain the system, but the architecture deliberately avoids technological lock-in.
Radiation-dose information existed within individual examinations and devices, but there was no practical way to monitor exposure patterns across institutions, equipment and procedures at national scale.
Dose information generated during routine clinical workflows is collected and analyzed centrally. The platform supports real-time monitoring by hospital, device, procedure, modality, patient population and time period, turning everyday examinations into a continuously growing quality dataset.
The aggregated data enabled the establishment and ongoing evaluation of diagnostic reference levels for major ionizing-radiation modalities. Instead of depending only on periodic manual surveys, national quality teams can analyze dose patterns using data produced in everyday clinical practice.
Radiology reimbursement claims were generated from requests and administrative records. Those records alone could not prove that the examination had actually been performed or that it met defined technical and quality conditions.
The national imaging infrastructure was integrated with reimbursement services so that each radiology claim could be validated against the underlying clinical evidence. The platform checks whether images exist, whether the study is genuine, whether a radiology report is present and whether exam-specific quality criteria are satisfied.
A nationally unique accession-number generation mechanism was introduced across participating hospitals, creating a consistent identifier connecting imaging, reporting and administrative workflows.
A physician preparing to request a radiology examination might not know that the patient had recently undergone a similar study in another hospital. This increased radiation exposure, cost and unnecessary use of imaging capacity.
A real-time decision-support service was integrated with hospital information systems. Before a new order is finalized, the HIS can query the service with patient identity, intended examination, department and physician information. Within seconds, relevant priors are returned with direct links to images and reports.
The system moves access to prior information upstream—from the reading room to the ordering moment—so that the clinician can decide whether a new examination is actually necessary before resources and radiation are committed.
In acute neonatal bedside imaging, examinations may need to be performed before a conventional radiology order can be entered. Without a usable modality worklist entry, operators manually registered patients on imaging devices. This created multiple demographic variations of the same newborn and made later reconciliation difficult.
RAD-X was integrated with the Philips intensive-care environment so that inpatient details are continuously received and synchronized. A specialized modality worklist exposes currently admitted ICU patients to connected imaging devices regardless of modality. For devices that require order-based entries, RAD-X dynamically generates the required worklist order.
Patients remain available throughout their ICU stay and disappear from the specialized worklist after discharge. The workflow reduces manual entry during acute cases and helps prevent duplicate or mismatched patient records.
Longitudinal patient records can span years of reports, visits, examinations, diagnoses, clinical notes, laboratory results, structured reports, scanned documents and multidisciplinary decisions. As the record grows, the limiting factor becomes the clinician's time and cognitive capacity to find and process what matters.
The RAD-X AI Chatbot allows clinicians to ask natural-language questions such as "Summarize this patient," "Has the patient undergone surgery on the left elbow?" or "What were the most important findings during the last year?" The chatbot gathers relevant information from the available patient record and returns a concise answer focused on the clinician's question.
Clinical AI should not become a black box. RAD-X responses are designed to include direct references to the underlying evidence, so the clinician can move from a concise answer to the relevant report, examination, pathology result, laboratory finding, clinical note or council document and open the source for verification.
Clinicians increasingly use general-purpose AI tools to summarize complex information, which can encourage sensitive patient data to be copied into external services. RAD-X provides conversational AI inside the hospital-controlled environment, allowing the same workflow without exporting confidential clinical data to public AI systems.
Oncology patients can accumulate exceptionally large and complex records over several years. Their care may involve radiology, pathology, surgery, medical oncology, radiation oncology, nuclear medicine and laboratory medicine. The challenge is not simply retrieving one report; it is understanding how the disease has evolved, what treatments were performed, what decisions were made and how information from different specialties relates over time.
Tumor board and multidisciplinary council reports can be loaded into the patient file like other clinical documents and made available to the chatbot together with radiology, pathology, laboratory and treatment information. This allows future clinicians to reconstruct not only what happened, but also the documented context behind previous decisions.
When an answer refers to a previous finding or decision, the clinician can follow the reference directly to the supporting evidence. For oncology, this creates a practical Ask. Answer. Verify. workflow across a complex, multidisciplinary longitudinal history.
These projects illustrate a simple principle: modern imaging infrastructure is not only about storing and displaying images. It must connect people, systems and institutions while preserving clinical continuity, security and control.
From a single clinical workflow to enterprise and nationwide healthcare infrastructure.