bg
Medicine and healthcare
14:56, 17 November 2026
views
1

3D Lung Model: AI to Help Surgeons Plan Operations

Russia has begun pilot testing an AI module that builds a three-dimensional model of the lungs from computed tomography (CT) scans. The system is designed to help surgeons plan procedures more precisely and assess in advance how surgery could affect a patient's respiratory function.

The challenge in any operation is to remove diseased tissue while causing as little damage as possible to healthy tissue. This is especially important in lung surgery because the lungs have a porous, spongy structure. Every piece of healthy tissue removed means losing alveoli that will no longer participate in breathing. If too much tissue is removed, the patient may develop respiratory failure.

That is why surgeons treating lung tumors must consider several factors at once: the exact location of the growth, how close it is to the bronchi and blood vessels, and which parts of the organ can be preserved so the patient can continue to breathe normally. The more accurately a surgeon can understand the individual anatomy of a patient's lungs, the better prepared they will be for the operation. Conventional CT scans, however, give physicians a series of two-dimensional images. They must study those images carefully and mentally piece them together to form a complete picture.

Seeing the Full Anatomy

The new AI module is designed to make that task easier. It was developed by specialists at Sechenov University in partnership with RTK Radiology. The software analyzes CT scans and automatically identifies the main anatomical structures in the images: the lungs, their lobes and segments, the bronchi, arteries and veins, as well as tumors.

Using this data, the system creates an interactive 3D model. Surgeons can view it from different angles, examine the tumor's position in relation to blood vessels and bronchi, and map out the proposed resection boundaries. This kind of visualization is particularly important in complex cases where a lesion is located close to vital structures.

Another distinctive feature of the module is its ability to estimate respiratory function after resection, or the surgical removal of part of a lung. The algorithm calculates how much of the organ will remain after surgery and how that compares with the lungs' total volume. The results are saved in a report that can be added to the patient's electronic medical record.

Preserving the Ability to Breathe

The importance of these calculations for patients is hard to overstate. The lungs supply the body with oxygen, so preserving the ability to breathe fully is essential after surgery. The new AI module is intended to help surgeons perform operations while sacrificing as little healthy lung tissue as possible.

The technology is now undergoing a further evaluation at Sechenov University's Institute of Personalized Oncology and Clinic of Hospital Surgery No. 2. Researchers plan to test the system on de-identified CT scans from at least 30 patients. They will then compare the automated analysis with assessments by medical experts. The evaluation will examine not only how accurately the algorithm identifies individual anatomical structures, but also whether it produces consistent results when the same scan is analyzed repeatedly. This will help determine how useful the 3D model is for preoperative planning.

The module has already undergone initial testing at the university's clinics. At that stage, specialists assessed how accurately the system identified lung tissue, bronchi and blood vessels. Testing showed a Dice coefficient of 93% for lung tissue segmentation. The system identified 96% of the bronchial tree's branching structure and 97% of the vascular branching structure.

A Tool for Surgeons

“Our priority is to make sure that the AI-powered solution can quickly identify abnormalities, analyze their location in relation to the anatomical structures of the lungs, and predict a patient's respiratory function after surgery. This gives surgeons additional information for preoperative planning and can help reduce tissue damage during resection,” said Marina Sekacheva, director of Sechenov University's Institute of Personalized Oncology.

Andrey Dolgov, deputy technical director at RTK Radiology, highlighted another important aspect of the project: the division of responsibilities between the software and the medical professional. AI handles image processing and 3D model generation. The physician evaluates the results, considers them in light of the patient's clinical picture, and decides on the treatment strategy.

From Pilot Testing to Clinical Use

The technology is expected to enter clinical practice in 2027. The project is significant for Russia's medical IT sector because the software is being developed in close collaboration with clinicians and tailored to a specific surgical need.

In the longer term, tools like this could become part of the digital preparation for complex operations. Their ultimate value lies in helping patients retain what matters most after major surgery: the ability to breathe.

Creating personalized 3D models allows us not only to visualize a patient's anatomical features but also to literally ‘rehearse’ the operation. This is especially valuable in complex cases: we can work through every step, identify potential difficulties, and choose the optimal surgical approach before we even pick up a scalpel.
quote

like
heart
fun
wow
sad
angry
Latest news
Important
Recommended
previous
next