Imagine typing an IP address into your browser and a stranger’s CT scan opens in front of you. No login, no password, no warning. Next to the image — the patient’s name, date of birth, examination date, facility name and the physician’s name. This is not a movie scenario. It is the everyday reality of an internet in which more than 1.19 billion medical images are publicly accessible through misconfigured PACS servers.
Medical imaging systems are today the most internet-exposed — and most underestimated — part of a hospital’s infrastructure. In this article we explain why the DICOM protocol is vulnerable almost by design, how attackers find unsecured servers and what a facility that has not checked its own exposure really risks.
What Is PACS and Why Is the DICOM Protocol Insecure by Design?
PACS (Picture Archiving and Communication System) is a system that archives and shares diagnostic images — computed tomography (CT), magnetic resonance imaging (MRI), X-rays and ultrasound. Equipment sends examinations to it, and physicians retrieve them at reporting workstations. The glue of this entire ecosystem is a single standard: DICOM (Digital Imaging and Communications in Medicine).
And this is where the problem begins. DICOM emerged in the 1980s and 1990s, at a time when the radiology network was physically cut off from the outside world. The standard’s designers made an assumption that was reasonable then and dangerous now: since every device on the network is physically controlled by the hospital, everyone is trusted. As a result:
- DICOM does not enforce authentication. A connection to a PACS node often requires neither a username nor a password.
- DICOM does not encrypt transmission by default. Data — including full patient data — travels across the network in cleartext. Anyone on the same network segment can read it.
- DICOM carries personal data inside the file itself. Each image is not just pixels. It is also a set of “tags” with the patient’s name, date of birth, examination number, procedure description and facility details.
As long as radiology was an island, these shortcomings did not matter. But the digitalization of healthcare, teleradiology, remote reporting and migration to the cloud connected the former islands to the public internet. A standard designed for an offline network found itself online — with its entire lack of protection.
Where Do More Than a Billion Medical Images on the Internet Come From?
The scale of the phenomenon was first brought to attention in 2019. Researchers found hundreds of PACS servers accessible from the internet, exposing tens of millions of patient records. Subsequent analyses showed that the problem is not going away — on the contrary, it is growing.
| Finding | Scale |
|---|---|
| Medical images publicly accessible online (updated report) | approx. 1.19 bn |
| DICOM servers reachable from the internet (Shodan scan, 2025) | approx. 3,627 |
| Unsecured medical images detected by researchers (2024) | approx. 45 mn |
| Countries with exposed systems detected | over 50 |
The most common cause is not sophisticated at all. It is a configuration error: a PACS server made available for teleradiology without a VPN tunnel, a DICOM node exposed “for a moment” for testing and never hidden behind a firewall, a simple web interface of an image viewer running over plain HTTP, with no encryption and no password. Some exposed systems allow not only viewing but also downloading complete examinations, and even uploading your own files.
📚 Read also: A cyberattack scenario on healthcare — how it unfolds and how to defend
How Does an Attacker Find an Unsecured PACS Server?
Here comes the most uncomfortable truth: finding an exposed PACS server requires no hacking skills. A search engine is enough.
Services such as Shodan continuously scan the entire internet and catalog the devices connected to it. An attacker — or a security auditor — can within seconds search for systems responding on ports typical for DICOM (104, 11112) or web interfaces of well-known image viewers. The result is a ready-made list of targets along with their location and often the facility’s name.
It gets even simpler from there. Since DICOM does not enforce authentication, the next step is often simply… to connect. Standard, freely available diagnostic tools allow you to query a node, list the stored examinations and download them. No password cracking — because there are no passwords to crack.
That is why “invisibility” alone is not a security strategy. If a server is reachable from the internet, you should assume it has already been indexed.
What Do the Patient and the Facility Really Risk?
A leak of images is only the start of the list of consequences. An unsecured PACS opens several distinct risk vectors.
A personal data breach. Medical data is a special category of personal data, subject to the strictest protection. DICOM tags contain a full set of patient data, so the exposure of examinations is a fully fledged GDPR breach — with an obligation to report it to the supervisory authority and a risk of a financial penalty, not to mention the loss of patient trust.
Diagnosis manipulation. If an attacker can not only read but also write data into the imaging system, a scenario far more dangerous than a leak appears. Security research has shown that it is possible to forge a medical image — to add or remove a lesion on a scan. The consequence is a misdiagnosis and incorrect treatment.
Files that are both an image and a program at once. One notable vulnerability (CVE-2019-11687) demonstrated that a DICOM file can be crafted so that it is simultaneously a valid medical image and a valid Windows executable. Such a “dual-personality file” opens in a viewer like an ordinary examination, and when executed — it runs malicious code. A perfect carrier for hiding malware in legitimate-looking documentation.
Remote takeover of the server. PACS software and DICOM libraries have their own, “classic” vulnerabilities. The popular DCMTK library accumulated more than a dozen CVEs between 2020 and 2025, including a flaw that allows code execution via a crafted DICOM file (CVE-2024-47796). In the popular Orthanc server, in turn, a critical vulnerability was found (CVE-2025-0896, CVSS 9.8): with remote access enabled, authentication was not active by default. A compromised PACS server is not just a leak — it is a foothold for a further attack on the hospital network, all the way to ransomware.
Hospitals in the Crosshairs
This is not a distant problem. In 2025, Poland ranked first in the world in the number of detected ransomware attacks, accounting for around 6% of all registered incidents of this type globally.
The healthcare sector is on the front line. It is enough to look at the start of 2026, when medical facilities fell victim to attacks almost in series — in one period four incidents were recorded within 25 days. After part of its server resources were encrypted, one regional hospital had to switch to “paper-based” work, and the perpetrators demanded a ransom counted in millions of dollars. The scale of incidents in healthcare is growing dramatically year over year.
An unsecured imaging server is doubly dangerous in this context: it is at the same time a ready-made data leak and an open door into the network, through which ransomware enters.
NIS2 and the KSC Act — Security Testing Is Already Mandatory
The legal environment is changing too. The NIS2 directive and the amendment to the act on the National Cybersecurity System (KSC) that implements it classify most hospitals and many other healthcare entities as essential or important entities. This means specific, legally enforceable obligations:
- Regular security testing — in practice at least once a year — and reviews of security policies.
- Risk and incident management, including an obligation to report serious incidents within strict time frames.
- Personal responsibility of management — the facility’s director is accountable for implementing cybersecurity obligations.
- Registration in the register of entities — essential and important entities have until 3 October 2026 to do so.
In other words: checking whether imaging systems are not leaking to the internet stops being good practice and becomes an element of legal compliance.
📚 See the details: NIS2 for the healthcare sector — requirements for hospitals and clinics
How to Secure PACS and DICOM? A 7-Step Checklist
The good news is that PACS server exposure is a solvable problem — and one that does not require replacing the entire system. Here are seven actions, from the most urgent to the strategic.
- Check your own exposure. Search the facility’s IP addresses in the Shodan search engine and verify ports 104 and 11112. If a server responds — you have a problem to solve immediately.
- Hide PACS behind a firewall. No imaging server should be directly reachable from the internet. Remote access and teleradiology only through a VPN.
- Enforce authentication. Configure access control wherever the standard allows it — and verify that remote access is definitely not running “through an open door”.
- Enable TLS encryption. DICOM transmission and the web interfaces of image viewers must run over an encrypted connection, never over plain HTTP.
- Update the software. Patch the PACS server, DICOM viewers and libraries (DCMTK, Orthanc and others) — these are the ones accumulating new CVEs.
- Segment the medical network. Separate imaging systems, IoMT devices and the administrative network so that the compromise of one element does not mean the compromise of the whole.
- Commission a medical systems penetration test. Only a controlled attack from the outside and the inside will show how your facility looks from an attacker’s perspective — and which vulnerabilities to fix first.
Summary
An unsecured PACS server is one of the most serious and most frequently overlooked risks in healthcare. The DICOM protocol does not protect data on its own — it was designed in a different era, for networks that no longer exist. The responsibility for closing this gap therefore rests entirely with the facility: with configuration, segmentation, encryption and regular verification.
The worst scenario is to learn about your own exposure from an attacker — or from the supervisory authority after a leak. The best one is to check it yourself, before anyone else does.
If you want to be sure that your facility’s HIS, PACS and LIS systems are genuinely isolated from the internet and resistant to attack, see our medical systems penetration testing service. We test them with methods safe for 24/7 hospital operation and deliver a report with remediation priorities.
