TECH2030
Insights
Case Study · 4 min readAI-assisted

Dongtan City Hospital's AI MRI Breaks Even in Just 15 Months

Dongtan City Hospital cut AI MRI scan times to about 5-8 minutes and recouped its investment in just 15 months.

Dongtan City Hospital's AI MRI Breaks Even in Just 15 Months

Key Takeaways

  • MRI scan time: 20-30 minutes → about 5-8 minutes
  • One machine's results became the basis for expansion
  • The first project targeted work where time directly equals throughput
  • Finding tasks measured in minutes

Kim Mi-young, administrative director of Dongtan City Hospital in Hwaseong, Gyeonggi Province, said, "For the AI MRI, the break-even point on the investment cost was reached in just 15 months." According to the hospital, MRI scan times that previously averaged 20 to 30 minutes have been shortened to about 5 to 8 minutes, depending on the body part and imaging method.

Dongtan City Hospital is a 180-bed regional hospital specializing in spine and joint surgery. GE Healthcare's AI MRI uses deep learning-based image reconstruction technology to reduce noise and produce clearer images. As scan times decreased, patients spent less time enduring the confined scanning space, and the hospital gained the capacity to examine more patients within the same time frame.

In what order did Dongtan City Hospital expand its AI adoption?

Based on the results confirmed with MRI, the hospital gradually expanded the scope of its AI applications.

  1. Imaging exams — Expanded to assist with CT scans, ultrasound, and chest X-ray interpretation. For CT scans, a camera above the examination table recognizes the patient's body and helps position it to match the area being scanned.
  2. Inpatient wards (end of 2025) — Adopted "thynC," an inpatient monitoring system distributed and sold domestically by Daewoong Pharmaceutical. Sensors such as patches or rings display ECG, heart rate, respiration rate, oxygen saturation, blood pressure, and body temperature on a central monitor at the ward entrance, and alert medical staff if abnormalities are detected.
  3. Operating rooms (second half of 2026) — Adopted Stryker's "Mako SmartRobotics," a surgical assistance system for joint replacement procedures. It renders pre-operative CT scans in 3D to analyze joint structure and the degree of deformity.
  4. Rehabilitation — AI-equipped devices assess patients' gait patterns and foot pressure distribution to track changes before and after surgery and rehabilitation.
  5. Under review: medical records — A system in which AI recognizes conversations between doctors and patients and, rather than producing a simple summary, generates records formatted to match the hospital's chart fields.

Hospital staff noted that shorter MRI scan times created more flexibility in scheduling and exam operations, and that thynC reduced the burden of repetitive measurement and documentation, allowing nurses to focus more on patient observation and care. The hospital has been designated an official GE Healthcare reference site for the Asia-Korea-Australia (AKA) region, and according to the hospital, officials from countries such as India and Thailand regularly visit to learn about equipment operation.

Why could a 180-bed hospital move so quickly?

Dongtan City Hospital opened as a clinic in 2010, expanded and relocated to its main building in 2017, and built a new wing in 2025, rapidly expanding its clinical and examination spaces. This expansion also increased the organization's receptiveness to AI adoption.

A hospital official explained that large hospitals often have complex decision-making structures, which can slow the adoption and validation of new technologies, whereas small and mid-sized hospitals can make decisions quickly, allowing them to rapidly adopt proven technologies and generate real-world case results.

Implications for Korean Enterprise AX

What stands out in this case is the sequence. The hospital first confirmed results with a single piece of equipment where processing time could be measured numerically, then used that evidence to expand into imaging exams overall, inpatient wards, operating rooms, and rehabilitation. The results were also presented on two levels: the operational metric of scan time and the financial metric of a 15-month break-even point.

Organizations selecting their first AX project should consider applying the following criteria in order.

  • Is it a task whose time requirement is measured in minutes?
  • Is it a task where reduced time directly translates into increased processing volume?
  • Can the results be measured in both operational and financial terms?

It's also worth noting that the goal of the medical records system is not conversation summarization but documentation formatted to match the hospital's chart fields. For organizations considering automating meeting minutes or consultation records, a more practical evaluation criterion is whether all fields in the company's own format are filled in completely, rather than whether the summary reads naturally.

This week, identify one task in your organization where reduced time would directly increase processing volume, and start by checking whether the time it currently takes is actually being recorded.

Source: "Installing AI MRI recoups investment in just 15 months" [Health Hankyoreh]

If you found this helpful, share it.
Related servicesAX Transformation
Enterprise AX
Want to apply this to your own operations?
Request an AX consultation