The Scan Already Knows. PVCT Makes It Visible.

Big things happened in medical imaging these past few weeks. We have been watching it closely, not as outside observers, but with the keen interest of a company that has spent the last two years building something in this space, although from a very different angle.

An Ode to the First Step

Alibaba's DAMO Academy published a model called RADAR in the journal Science and we all got the opportunity to take a close look at the open-sourced system. In plain terms: RADAR looks at a CT scan of the abdomen and tells a doctor what's wrong with it; this means automatically flagging up to 146 different conditions (across 18 organs) from a single scan.

RADAR is reported to have either matched or outperformed 23 of 26 practicing radiologists who manually evaluated the same scans. Used as a support tool, it helped radiologists catch more diagnoses while working faster.

As a company building AI for clinical applications, we recognize that this is not an incremental advance. Medical imaging AI has largely relied on narrow, single-purpose models, each designed for a specific condition or task. A single model capable of interpreting an entire organ system at specialist level, trained from existing radiology reports rather than extensive manual annotations, represents a meaningful shift. We make this observation not as a courtesy, but because the distinction matters to the point that follows.

So, RADAR helps us find what’s wrong in a scan, or the anomalies in the patient who underwent the scan. We focus on the question that follows: “What does this anomaly actually look like in this patient, in three dimensions like in the real world?

That is the problem our company is built to solve. RADAR was never designed to do this. It can tell a radiologist, “There is a mass here.” But it cannot show, in vivid three dimensional views, what that mass actually looks like within the patient, how it relates to surrounding vessels and structures, or reliably depict what a surgeon would actually encounter during a procedure. That spatial understanding still has to be reconstructed mentally by the doctor, one two-dimensional slice at a time.

Cresendos PVCT: New pathway from an existing crossroad

Our platform, Cresendos Predictive Virtual Clinical Twin, takes the same kind of CT or MRI data which RADAR reads, and does something drastically different with it. Instead of scanning the images for findings, it uses the data to reconstruct the patient's actual anatomy as an interactive, patient-specific 3D model - a digital twin of the anatomy. This is something a clinician can rotate, explore, and walk through from any angle, on a standard monitor, a tablet, or in VR.

Every one of the hundreds of flat slices that a radiologist would normally flip through, one at a time, is combined into a single, coherent three-dimensional shape depicting the true geometry of a vessel curving around a tumor, the exact angle at which a fracture line cuts through bone, and more. Every detail becomes visible at once, rather than having to be mentally reassembled slice by slice.

And most importantly, no new physical infrastructure is required for this in-depth visualization. The scan files we already have contain this information. PVCT reads that data more effectively, plugs the gap between what the scan contains and what can be readily understood, and unlocks valuable insights.

Why now? What makes this significant in healthcare today?

Desktop 3D planning and specialty-specific VR platforms already exist, but they typically require manual segmentation, dedicated workflows, or specialty-specific tools. What sets PVCT apart is the combination: fully automated 3D models in minutes, across organ systems, using the CT/MRI infrastructure hospitals already have, with no new hardware or dedicated workstation. That combination, and not 3D visualization alone, is what “like never before” points to.

Precise, patient-specific 3D anatomy changes what's possible before and during a procedure. It means sizing an implant or a valve correctly before the operating room, not discovering a mismatch mid-procedure. It means a tumor board actually seeing the same spatial relationship between a mass and the vessels around it, instead of describing it to each other in words. And because the model can be explored together in real time regardless of location, a doctor in a smaller town can bring a specialist in a major city into the same 3D space. Not a phone call, not a description, but the same model, live in front of their eyes.

Where we actually are today: It works!

We'd rather say this plainly than let the headline imply more than is true: PVCT is a working prototype. We've engaged more than 25 clinicians across radiology, cardiology, and emergency medicine, and the response has been clear: over 85% expressed strong interest in a 3D-first approach to medical imaging. Structured clinical validation is the work ahead of us now, not behind us.

Detection, visualization, and what comes after

Zoom out, and it becomes clear that RADAR and PVCT are two points on the same trajectory, not exactly direct competitors. On is AI that reads a scan and tells you what's there. One is AI that rebuilds the data to show you exactly what that means in three-dimensional space. Eventually, further down the road, once the evidence confirms it, PVCT will help anticipate how it changes over time.

Detection, visualization, prediction. We are grateful to RADAR for proof that the first step works at expert level. We're building the second and third.

Built in India, for the world we are preparing for

We’re building PVCT from India, with a simple belief: access to specialist-level anatomical insight shouldn’t depend on a hospital’s location. A single, capable platform spanning departments can help bring that capability to hospitals beyond major cities. PVCT is being designed as a multi-organ platform from the start, built for India’s regulatory and data landscape, not adapted to it later.

RADAR showed the world that AI can read a scan like an expert. We're teaching it to show you the anatomy like never before, not because no one has shown anatomy in 3D before us, but because no one has made it this automated, this complete, and this easy for a hospital to simply pick up and start using.

Zabound Technologies Private Limited

No-301, 1st Floor, KP Towers,
Arcot Road, Vadapalani,
Chennai - 600026

+91 - 9884466143
contact@cresendos.com

© Zabound Technologies Private Limited & Zabound Services LLP - 2026

Zabound Services LLP

TC 24/3088-2, Ushasandhya Building,
Kaudiar, Trivandrum 695003