Stand at the table during a hard case and listen to what the operator is working out. It is rarely where am I. They know where they are. The question that decides the case is what is directly ahead of the tool, and whether the path they are about to commit to is the right one.

That is the question today’s imaging answers least well. The field has spent two decades building tools that show the operator where they already are, and comparatively little that shows where they are going. I want to lay out that gap, why it persisted, why it may be addressable now, and why, as someone who invests in where markets reorder rather than in any one technology, I find it interesting. Note the word gap, not flaw. The tools we have are good. They do one job well. There is a second job the field has not yet filled.

A short history of how we ended up half-blind

It helps to see the lineage, because the blind spot was not a mistake at any single step. It was the result of solving the solvable problems first. Angiography gave a roadmap of the lumen, a shadow of where blood flows, but nothing about the wall or what sits inside an occlusion.

Intravascular ultrasound answered the wall, rendering plaque and calcium and vessel size from inside the artery. Optical coherence tomography sharpened that picture to near-microscopic resolution. Electromagnetic navigation and robotics, in the airway, solved steering a tool out to the periphery. Each was a genuine advance, hard-won.

But trace the direction every one of them looks. Angiography is a projection from outside. IVUS and OCT look sideways at the wall. Navigation reads a map made before the case.

None of them looks forward, in real time, from the tip, in the direction the operator is moving. The field did not decide to ignore the forward view. It built the views it knew how to build, and the forward one was always the hardest, so it kept getting deferred. Twenty years of deferral is how a blind spot becomes invisible.

The moment of truth is usually forward

Here is why the gap is more than academic. In the procedures where imaging should help most, the case is decided at the instant the operator commits to a path they cannot fully see. At the face of a coronary chronic total occlusion, the operator has to enter a blockage without a clear view of where to start, and registries show that an ambiguous entry point predicts failure. In a long calcified occlusion below the knee, staying in the true lumen is the difference between restoring flow and dissecting into a dead end.

At the edge of the lung, the question is whether the tool is in a peripheral lesion or merely near it. These are illustrations, not an exhaustive list. What they share is direction: the decisive moment points forward, and the imaging does not.

In each, operators reach for workarounds, probing with wires, pulling up an aging pre-procedure scan, sampling and hoping the map held. The workaround is the tell. It is a field routing around a blind spot it learned to tolerate because tolerating it was the only option.

Why forward stayed missing, and why now

If a forward view is so obviously useful, two fair questions follow: why did the field spend its energy elsewhere, and why would it be more tractable now than when people reached for it before.

On the first, looking sideways from inside a vessel or an airway is an easier engineering problem than looking ahead from the tip of something small, moving through a tight, often opaque space. The field optimized what it could solve. On the second, and this is the honest part, forward-looking imaging has been attempted and mostly stalled, on tip optics that were not good enough, sensors that were too large, too little compute near the tip, and workflows that asked too much for too little.

That was the idea arriving before the supporting technology. What has changed is the rest of the stack. Sensors have shrunk, real compute can sit close to the tip, and reconstruction has matured. The lesson of the earlier failures is not that the idea cannot work. It is that it could not work yet, which is a different conclusion, and a skeptic who watched those attempts fail should sit with the difference.

What it would, and would not, be

Precision matters more than enthusiasm here, so be precise about the scope of the idea, not its promise. A forward view, if it existed, would be a near-field look at what is at and immediately ahead of a tool tip. It would not be X-ray vision, it would not see through blood, and it would not see down a whole route in advance.

Drawing that boundary clearly is what separates a real engineering question from the overpromises that buried earlier attempts. The operators who matter have seen the overpromises, and precision is how anyone earns their attention back.

And whether such a view would change what happens at those decisive moments is exactly the kind of question that has to be answered with evidence, not asserted. I am describing a gap and an open question, not a result. The work to answer it is ahead, not behind.

Why this would be an instrument problem, and why that interests an investor

Here is the part that holds my attention as someone who allocates capital. If this gap gets filled, it will not be filled by software alone. No model extracts a forward image from a sensor that only looks sideways, or conjures a live view out of a scan taken an hour ago.

A forward view is an instrument problem first. It takes hardware that physically looks the right way and captures data that does not exist in any side-view or pre-op data set.

That is also what would make it defensible. A software product that reads images someone else’s hardware captured inherits that sensor’s blind spots and depends on a data stream it does not own.

A company that owns the capture owns the input, and the input is proprietary from the first frame. In a field where models commoditize and can be rented by anyone, the durable advantage tends to sit with whoever records a stream nobody else can. That is a general lesson about where value accrues, and it is the lens I bring to this gap: the interesting prize is not a better catheter, it is ownership of a view nobody else is recording.

Additive, not a replacement

Whatever fills this gap will earn adoption only if it is additive. The field is not going to put down the side-view probe or stop using the pre-op scan or abandon its navigation platforms, nor should it.

A forward view would sit on top of those, a third input pointed the one way the other two cannot look. That is a far easier path to adoption than a rip-and-replace technology, because it asks an operator to add a view rather than abandon the ones they trust. It is also why I think about this as a potential platform rather than a single device: a genuinely new input, recorded across many cases, is the kind of asset that compounds, while a clever feature is the kind that gets copied.

The honest limits

None of this is magic, and overselling it would be the fastest way to lose the clinicians any of it is meant to serve. A forward view would not soften calcium or straighten a tortuous vessel. It would not replace the operator’s judgment or hands. It would be one input among the ones operators already trust, and its value would have to be proven, case by case and in evidence, rather than claimed.

I find the gap worth watching precisely because it is real, structural, and unglamorous, the kind of thing that reorders a field quietly while everyone is looking at the model on the screen.

The field has spent twenty years getting better at seeing where it is. The interesting question for the next twenty is who, if anyone, adds a trustworthy view of where it is going, on top of everything that already works.