A PACS and imaging AI platform is the system in which a hospital stores its medical images and its radiologists read them, with the results of AI models shown beside each image. PACS stands for picture archiving and communication system. It is the hospital's store and viewer for every X-ray, CT, MRI and ultrasound image. An imaging AI platform adds one more job: it runs AI models from many different companies on each scan, inside the same viewer.
A PACS does three things. It stores every image, often for many years, in DICOM, the standard file format for medical images. It shows the images in a viewer, where the radiologist can scroll through slices, zoom, adjust the brightness and measure. And it keeps the worklist, the queue of scans waiting to be read. Because the images are stored in one place, a radiologist can read them from another city. Reading scans from a distance is called teleradiology. Many Indian radiologists do it, for hospitals in India and abroad.
A hospital rarely buys all its AI from one company. One company may make the best TB model, and another the best stroke model. Without a platform, each model would need its own connection to every scanner and its own screen. The AI platform is connected once. It sends each new scan to the right model: chest X-rays to the TB and lung model, and head CTs to the bleeding model. It collects each result, shows it in the viewer and moves flagged scans up the worklist.
Imagine a group of 12 diagnostic centres around Ludhiana. They take about 1,500 scans a day, and 9 radiologists read them, some from home. At 2 p.m., a head CT arrives from one centre. The platform sends it to the bleeding model, which reports a likely bleed in under two minutes. The platform moves the scan to the top of the worklist and alerts the radiologist on duty. She opens it on her medical display, confirms the bleed and phones the doctor who ordered the scan. The platform also records whether she agreed with the model. Over months, these records show the group how accurate each model is on its own patients.
The images come from medical imaging equipment, and the AI comes from AI radiology companies. The platform can also store the slide images from a digital pathology scanner. For difficult cases, the radiologist can open 3D medical imaging. Images are shared with other hospitals through medical record retrieval. Drug trials send their scans to central imaging reading.
A PACS is different from a hospital EHR. The EHR holds the patient's record and the written report. The PACS holds the images themselves, which are far larger.