Tech, ExplainedHealthtech and life sciences
Scene

Medical imaging

A chest X-ray read by AI in a minute.

1 X-ray is taken2 AI reads it first3 Radiologist confirms4 TB test ordered5 Eyes are screenedsends the imageputs the flagged X-ray firstsends the confirmed reportrefers him for an eye checkMedicalimaging equipment+AI radiology+PACS and imagingAI platform+Digitalpathology scanner+Retinal camera withAI screening+Molecular test system↗ LABORATORY DIAGNOSTICSHospital EHR↗ HEALTHCARE ITEdge AI↗ GENERATIVEAI APPLICATIONSA farmer with a month-long cough, ata district hospital in BiharA TB test the same day, and an eyecheck for diabetesDuring his six months of TB treatment, his follow-up X-rays are read in the same way.CONCEPTS IN THIS SCENETriageSoftware as a medical device (SaMD)FirmwareDigital twin1 X-ray is taken2 AI readsit first3 Radiologistconfirms4 TBtest ordered5 Eyesare screenedsends the imageputs the flaggedX-ray firstsends theconfirmed reportrefers him foran eye checkMedicalimagingequipment+AI radiology+PACS andimagingAI platform+Digitalpathologyscanner+Retinalcamera withAI screening+Moleculartest system↗ LABORATORYDIAGNOSTICSHospital EHR↗ HEALTHCARE ITEdge AI↗ GENERATIVEAI APPLICATIONSA farmer with amonth-long cough,at a districthospital in BiharA TB test the sameday, and an eyecheck for diabetesDuring his six months of TB treatment, his follow-up X-rays are read in the same way.CONCEPTS IN THIS SCENETriageSoftware as a medical device (SaMD)FirmwareDigital twinA farmer with a month-long cough, at a districthospital in Bihar1 X-ray is takenMedical imaging equipment+↓ sends the image2 AI reads it firstAI radiology+↓ puts the flagged X-ray first3 Radiologist confirmsPACS and imaging AI platform+Digital pathology scanner+↓ sends the confirmed report4 TB test orderedMolecular test system↗ LABORATORY DIAGNOSTICSHospital EHR↗ HEALTHCARE IT↓ refers him for an eye check5 Eyes are screenedRetinal camera with AI screening+Edge AI↗ GENERATIVE AI APPLICATIONS→ A TB test the same day,and an eye check for diabetesDuring his six months of TB treatment, hisfollow-up X-rays are read in the same way.CONCEPTS IN THIS SCENETriageSoftware as a medical device (SaMD)FirmwareDigital twin
A farmer's chest X-ray at a district hospital in Bihar: the scanner makes the image, AI reads it first and flags likely TB, a radiologist confirms it on the imaging platform, a TB test is ordered, and his eyes are screened for diabetic eye disease.
opens a page inside this one a page in the same scene a page in another scene a concept page a label only, with no page

Medical imaging is the use of machines that see inside the body, and of the software that makes and reads their pictures. It is also called radiology, after the X-rays it began with. The main kinds are four. An X-ray passes through the body and shows bones and lungs. A CT scan takes X-rays from all around the body and builds thin slices from them. An MRI scan uses a strong magnet and radio waves instead of X-rays. Ultrasound uses sound waves.

Every image must be read by a specialist doctor. A radiologist reads scans. A pathologist reads thin slices of tissue on glass slides. India has too few of both, especially outside the big cities. So software now does the first reading, and Indian companies build much of it.

Here is a typical case. A farmer has had a cough for a month. He goes to a district hospital in Bihar that has an X-ray machine but no radiologist. A technician takes his chest X-ray on the medical imaging equipment. Software inside the machine turns the raw measurements into a picture. This step is called image reconstruction.

Within a minute, an AI radiology model reads the X-ray. It marks a patch in the upper part of one lung and reports a likely case of tuberculosis (TB). Sorting cases by urgency in this way is called triage. The X-ray and the AI result go into the PACS and imaging AI platform, the system that stores every image and shows it to doctors. A radiologist in another city opens the X-ray there and confirms the finding. Reading scans from a distance like this is called teleradiology. Meanwhile, the hospital's doctor sees the flag and sends the farmer's sputum, the mucus he coughs up, for a TB test the same day.

The farmer also has diabetes. People with diabetes are more likely to get TB, and diabetes can also damage the back of the eye. So a health worker photographs his eyes with a retinal camera with AI screening. The camera checks for diabetic eye disease on the spot, with no eye specialist present.

Tissue samples need the same help. A digital pathology scanner turns a glass slide of tissue into a digital image. A pathologist in another city can then read it on a screen, through the same kind of platform.

AI that reads medical images is regulated as a medical device, because a wrong reading can harm a patient. This idea is called software as a medical device (SaMD). In every case, a doctor still makes the diagnosis. The AI decides only which case a doctor should look at first. Much of the scanners' own software is written in Bengaluru, where the largest scanner makers have big engineering centres.

Tech, Explained is a part of ingrid.fyi. Every page explains one piece of technology in plain words.