“Most CTO complications are not failures of skill. They are failures of sight, in the few millimeters directly ahead of the wire, where none of our tools look.”

Chronic total occlusion PCI sits among the most demanding work in the interventional suite, and it has gotten markedly better. An algorithmic, anatomy-driven approach to crossing has raised both success and safety over the last decade.^[1] In a dedicated, high-volume program, technical success can climb from roughly 70 percent to the mid-80s as operators move down the learning curve, even as the lesions attempted grow more complex.^[5]

Yet the cases that go wrong rarely fail because the operator lacked skill. They fail because, at one decisive moment, the operator could not confirm what was directly ahead of the wire. That is the uncomfortable pattern in the pitfalls experienced proctors flag most often. Taken one at a time they look like separate problems. Read together, a single root cause keeps surfacing. This piece walks through the six that operators respect most, shows why they reduce to one shared failure, and then asks the question that follows: if most CTO pitfalls are orientation failures at the wire tip, what changes when orientation at the wire tip gets better, and just as important, what does not.

Part One: The Six Pitfalls Every CTO Operator Respects

1.      Losing distal position after the hard part is done. Inadvertent wire retraction or inadequate microcatheter support can surrender hard-won distal position. Hours of effort are negated in a moment, and every recrossing attempt re-exposes the patient to the risk that the first crossing was supposed to retire.

2.      Mistaking an extraplaque path for the true lumen. This is among the most frequently observed pitfalls and the most consequential. The contemporary consensus vocabulary has moved from “sub-intimal” to “extraplaque” for good reason: it names where the wire actually is.^[1] Less experienced operators may read an extraplaque wire path as a side branch without systematic confirmation through orthogonal projections or dual injection. The cost of that misread is perforation and tamponade, and in complication-heavy contexts perforation rates run from roughly 9 to 17 percent.^[6]

3.      Leaving dual injection on the table. Dual injection, contrast in both the antegrade and the collateral-supplying vessel, is close to mandatory in modern CTO work. The EuroCTO consensus is blunt that attempting recanalization without displaying the distal target is “irresponsible,” because dual injection clarifies CTO morphology and, critically, helps elucidate guidewire location during crossing.^[2] And yet it is routinely underused. Dual injection has been documented in only about two thirds of cases in registry experience, with single-catheter angiography providing inadequate information.^[3,4]

4.      Missing the systemic warning signs. Progressive hypotension, a patient becoming less responsive, and operator fatigue in a long case are easy to discount in the focus of crossing. This is where the proctor earns the role, serving as a global safety observer with the authority to abort when the balance of risk and benefit turns. No imaging modality surfaces these signs; they live in the vitals and in the room.

5.      Over-injecting into a compromised distal bed. In the setting of an extraplaque hematoma or a compressed distal true lumen, excess contrast promotes hydraulic dissection, which is one reason consensus recommends side-hole guides to prevent forceful intraplaque injection.^[2] It also loads a renal cost onto a patient who often has no margin to spare, a point Part Five takes up in detail.

6.      Staying on a failing plan too long. Reluctance to pivot from the initial crossing strategy, despite stagnation or rising risk, lengthens procedures and compounds hazard until a proctor intervenes. The discipline to change approach early is one of the hardest to teach and the easiest to lose under pressure. It is no accident that the global algorithm builds in explicit stop rules, discussed below.^[1]

Part Two: One Thread Runs Through All Six

Strip these down and most reduce to the same question, asked at the worst possible time: where is the wire, and what is in front of it?

Losing distal position, exiting into the extraplaque space, leaning on contrast for a roadmap, refusing to pivot. Each is, at its core, a decision made with incomplete information about the path ahead. The tools we reach for to fill that gap each leave part of it open. Fluoroscopy gives a shadow, a two-dimensional projection of a three-dimensional problem. Dual injection gives an excellent borrowed map, but it is a map of the lumen on either side of the occlusion, it fades within seconds, and it costs contrast to refresh.^[2] Intravascular ultrasound, which belongs in this conversation and which many of the best operators rightly love, is built into the consensus crossing algorithm for proximal-cap ambiguity and for reentry.^[1] But IVUS, like OCT, principally characterizes the vessel wall around and beside the catheter rather than the territory directly ahead of the tip.

So the operator fills the remaining gap with experience and inference, and the best operators are remarkably good at it. But inference under fatigue, against the clock, in a patient who is quietly deteriorating, is exactly the condition under which these six pitfalls bite. The skill is real. The blind spot is structural, and it sits on the forward axis, ahead of the wire, where none of our current orientation tools look directly.

Part Three: What “Looking Forward” Actually Means, and Where Its Physics Allow It

It is worth being precise, because the imprecise version of this claim deserves the skepticism it gets. Forward viewing does not mean a clear cinematic view down a blood-filled distal vessel. Light does not behave that way in blood. Near-infrared light is strongly scattered and absorbed by red blood cells, which is exactly why optical coherence tomography requires the blood field to be cleared with a flush before it can image at all.^[10,11] Any honest account of an optical forward view has to start by conceding that constraint rather than papering over it.

The defensible claim is narrower and more interesting. The place a forward-viewing modality is most useful is not the open vessel beyond the occlusion but the occlusion itself, the territory the wire is crossing right now, where the medium ahead of the tip is plaque, fibrous tissue, calcium, and organized material rather than flowing blood. In that setting the relevant questions are concrete and answerable: is the wire tracking intraplaque or has it entered the extraplaque space; what is the character of the proximal cap the wire is about to engage; what tissue sits immediately ahead of the tip. Those are the orientation questions the six pitfalls are really asking, and they live in precisely the near-field, tissue-filled space where a forward view has a physical right to operate.

Framed that way, the value is sharpest in exactly the cases that test even master operators, the ones where current tools leave the most ambiguity. The blunt or ambiguous proximal cap with no clear stump. The in-stent CTO, where the path is constrained but the entry is unclear. The heavily calcified cap where the operator cannot tell penetration from deflection. The long occlusion where reentry has to be timed against a distal cap. The consensus algorithm already routes these to special handling, IVUS for proximal-cap ambiguity, move-the-cap and dissection-reentry techniques for the rest.^[1] A forward view is aimed at the same hard cases, on the one axis those other tools do not cover. This is not a beginner’s aid that happens to help experts. It is an expert’s tool that also, secondarily, gives less experienced operators a scaffold, a point Part Seven returns to with appropriate caution.

Part Four: It Changes Which Tool You Reach For, and What That Saves

The six pitfalls are about not getting lost. There is a second advantage to seeing clearly along the path, and it is about not guessing. A chronic total occlusion is not one substance. Histopathology describes it as a sequence: a younger occlusion is largely organized thrombus threaded with microchannels, while an older one is dense fibrocalcific plaque, often with no microchannels left at all.^[15] What the wire, and then the balloon and the stent, should do depends entirely on which of those sits in front of the tip, and angiography cannot reliably tell the operator which it is.

Reading the tissue ahead turns one blind decision into three informed ones. If the material is soft and carries a microchannel, the task is to find and track that opening; microchannels on the order of 100 to 500 microns are a recognized crossing pathway, and a wire does best when it finds the path of least resistance through them rather than forcing a new one or drifting into the extraplaque space.^[16] If the lesion is laden with organized thrombus, the sequence is different, and may favor clearing or managing that thrombus before modification and stenting rather than crushing it beneath a stent.^[17] And if the obstruction is dense calcium, no balloon will durably expand it and a stent dropped into it will underexpand; the calcium has to be modified first, by atherectomy or intravascular lithotripsy, which is precisely the therapy-selection decision that intravascular calcium assessment already informs through validated calcium-scoring.^[18,19]

Each of those is a fork, and taking the wrong branch is expensive in the most literal sense. The balloon that will not expand because calcium was treated as fibrous tissue. The stent deployed into underprepared calcium that underexpands and sets up restenosis, thrombosis, and a return trip to the laboratory.^[18] The crossing attempt that runs long because a microchannel was there to be tracked and was missed. The costly inefficiency in a complex PCI is rarely the imaging step. It is the wasted motion of a strategy chosen blind, and the rework that follows when the blind choice was wrong.

This is where the clinician’s view and the laboratory’s economics align. A cath lab is a fixed-cost engine whose output is measured in time: it operates as a profit center with high material and labor costs, and its contribution margin is generated per hour of laboratory time.^[20] Inefficient use of that limited resource raises cost and can compromise care.^[21] Seeing the lesion and reaching for the right tool the first time returns time and avoids rework, and in a per-hour-margin model, recovered capacity is where the economics actually live. The honest version of this argument is careful about its limits. Raw procedure time is driven far more by the nature of the case than by any single device,^[21] so minutes saved should be measured and demonstrated, not assumed. Intravascular imaging adds its own steps, cost, and sometimes contrast, and it is unevenly reimbursed, a point Part Nine returns to. Above all, the efficiency case must never become a volume case. The value is doing the indicated procedure more efficiently and getting it right the first time, never doing more procedures for their own sake. Throughput is a reward for good decisions, not a reason to make a bad one.

Part Five: It Changes What Orientation Costs the Patient

In a long CTO case, contrast is not a consumable. It is a budget the patient pays from, and the account does not refill during the procedure. The relationship between volume and harm is not vague. The risk of contrast-associated acute kidney injury rises roughly 12 percent for every additional 100 mL administered, and consensus quality guidance asks operators to keep the ratio of contrast volume to creatinine clearance under three, and ideally under two.^[7] In patients with reduced renal reserve a volume-to-clearance ratio above two is an independent predictor of injury, and in the highest-risk patients even small volumes near 30 mL can do damage.^[8,9] The global crossing algorithm encodes the same physiology into a hard stop: abandon the procedure when contrast exceeds roughly three times the glomerular filtration rate.^[1]

Here is the part that should bother us. A meaningful share of that budget is spent not on diagnosis or final assessment but simply on staying oriented: a dual injection to refresh a fading roadmap, a confirmatory look to reassure ourselves the wire is where we think it is. The patients drawn to CTO work are often precisely those least able to absorb the cost, diabetic, with chronic kidney disease, sometimes with a single remaining territory worth opening. So a sensible operator rations the very orientation the situation most demands. The registry finding that dual injection is used in only about two thirds of cases is partly a discipline gap, but it is also this rationing made visible.^[3,4]

The shift a forward view offers is not more orientation. It is orientation whose cost is not paid in contrast. A read that comes from imaging the tissue ahead of the tip, rather than from another bolus of dye, lets the operator keep their bearings without drawing down the renal budget to do it. The downstream effects compound: fewer reflex injections into a hematoma, less hydraulic dissection from forceful injection into a compromised bed, and a clearer, lower-cost basis for the hardest decision in the case, which is the subject of Part Six. The claim worth defending is exactly this narrow one. Forward viewing does not make a CTO a low-contrast procedure on its own; angiography still governs the case. What it can reduce is the orientation tax, the contrast spent purely to keep one’s bearings.

Part Six: It Changes When a Pivot Becomes Defensible

The technical skills get the attention in CTO work. But ask an experienced proctor where cases are won and lost, and the answer is rarely hardware. It is the decision to abandon a plan that has stopped working before the procedure forces the issue. Reluctance to pivot is one of the most cited pitfalls, and one of the hardest to teach, because while it is happening it does not look like a mistake. It looks like persistence.

Several forces pull the same way. Sunk cost, after an hour invested in one approach. Momentum, the belief the next wire will be the one. Fatigue, which narrows attention to the task in front of the operator and away from whether it is still the right task. And underneath all of it, the evidence to justify the switch is usually ambiguous: a fluoroscopic shadow and an inference about trajectory. On that thin basis, restarting from a different approach is an act of judgment under real doubt, and doubt is what makes people wait for one more attempt. The global algorithm tries to externalize that judgment with explicit ceilings, stop at roughly three hours of procedure time, at contrast beyond three times the GFR, or at an air kerma radiation dose above five Gray.^[1] Those ceilings exist precisely because operators, left to their own momentum, tend to run past them.

This is where a forward view earns a place in the decision, not by making the call, but by sharpening the evidence behind it. An operator who can directly interrogate whether the path is advancing intraplaque toward the distal true lumen, or merely burrowing further into the extraplaque space, has an objective basis for the pivot. Seeing that a trajectory is not converging turns a judgment call into an evidence-based one. That is easier to make, easier to make early, and easier to defend, both to a proctor and to oneself. Earlier is the whole point: better orientation does not change the eventual decision so much as let the operator reach it before another thirty minutes and another wave of contrast prove the same thing the hard way.

Part Seven: It Changes How Fast Expertise Can Be Built, If We Are Careful

CTO work is a craft, transmitted person to person across hundreds of cases under the eye of a proctor. That apprenticeship has produced remarkable operators, and the consensus algorithm was written in part to make the craft teachable across regions.^[1] But it has a bottleneck: only so many proctors, only so many cases, and only so many hours in which judgment can be built one repetition at a time. The pitfalls do not fall evenly. The most consequential misread, taking an extraplaque path for the true lumen, is the one less experienced operators make most, because the inference required to avoid it is the last thing expertise installs.

A forward view offers a more direct read on exactly that inference, where is this path going, and is it going there through plaque or out of it. Used well, it does not hand a trainee a finished judgment. It scaffolds the judgment they are building, and it gives a proctor a shared, objective reference to teach from rather than a felt sense that is hard to put into words. That is augmentation in its most honest form, and it is genuinely valuable. But it is also where the argument is most dangerous, and an honest piece has to say why.

Part Eight: The Honest Boundary, Automation Bias and Deskilling

The failure mode every thoughtful operator already fears has a name, two names. The first is automation bias: the tendency to trust an automated readout uncritically, even when it is wrong, producing both errors of commission, acting on a bad cue, and errors of omission, failing to act because the system did not prompt it.^[12,13] The second is deskilling, the gradual erosion of independent judgment when a tool is allowed to stand in for it, and its more troubling cousin in trainees, never-skilling, where the underlying competence is never built in the first place.^[14]

These are not hypothetical. When incorrect machine advice was injected into image interpretation, reader performance was impaired across every level of experience, from novice to expert.^[12] Under time pressure, the exact condition of a long CTO case, decision-makers abandoned a measurable share of their initially correct judgments when a confident system disagreed with them.^[13] A forward-viewing tool dropped into the highest-pressure moment of an interventional procedure inherits every one of these risks. If it is designed or taught as an oracle, it will not raise the floor; it will quietly lower the ceiling, and it will do the most damage to the least experienced operator, the very person it was supposed to help.

So the boundary has to be drawn in ink, not pencil. Forward viewing is augmentation, not replacement. It does not retire the proctor, the dual injection, the orthogonal view, or the operator’s read on a patient who is quietly deteriorating, because that last signal, the systemic warning sign of Pitfall Four, lives in the vitals and the gestalt where no imaging modality belongs. It is complementary to IVUS and to dual injection, covering the forward axis they do not, never a substitute for tools the field already trusts. And it has to be built and taught in a way that actively resists automation bias, that keeps the operator critically engaged rather than passively monitoring a screen. The right design philosophy and the right teaching posture are not afterthoughts here. They are the difference between a tool that builds judgment and one that erodes it.

Part Nine: What This Does Not Settle

Two honest admissions, because the argument is stronger for making them. First, no essay changes practice. Interventionalists of consequence do not adopt a modality because the case for it reads well; they adopt it because of a registry, a randomized signal, or a proctored case where they watched it work with their own hands. This piece is meant to frame a category and sharpen a conversation, not to substitute for the proof point that adoption actually requires.

Second, the orientation argument is, in principle, already won. Everyone agrees that knowing where the wire is and what lies ahead of it is good. The real fight over any new tool is fought on two questions this piece has not resolved, and that every operator asks privately: how much time does it add per case, and what does it cost, in dollars and in workflow friction, against reimbursement that rarely rewards intravascular imaging well.^[1] A modality that improves orientation but adds meaningful time or expense at the table will be judged on that trade, not on the elegance of its rationale. Those are the honest open questions, and they deserve their own piece.

The Narrow, Durable Case

Most CTO pitfalls are orientation failures at the wire tip. A modality built to improve orientation at the wire tip, on the one axis our existing tools do not cover, and within the tissue-filled space where its physics actually permit it to work, is aimed at the right target. It earns adoption one careful case at a time, by making the operator’s existing discipline easier to execute, never by promising to think for them. The moment any tool is sold as a substitute for judgment is the moment it begins to erode the very safety culture these six pitfalls exist to protect. The case for looking forward is narrower, and more durable, than a slogan. That is exactly why it is worth making carefully.

Field Notes is a personal series on the craft of interventional work and the tools that support it. Views are the author’s own, and nothing here is medical advice or a product claim.

References

1. Wu EB, Brilakis ES, Mashayekhi K, et al. Global Chronic Total Occlusion Crossing Algorithm: JACC State-of-the-Art Review. J Am Coll Cardiol. 2021. Consensus of 121 to 125 operators from 50 countries; defines extraplaque techniques and stop thresholds (3 h, contrast > 3x GFR, air kerma > 5 Gy). jacc.org/doi/10.1016/j.jacc.2021.05.055

2. Galassi AR, Werner GS, Boukhris M, et al. Percutaneous recanalisation of chronic total occlusions: 2019 consensus document from the EuroCTO Club. EuroIntervention. 2019. Dual injection described as mandatory; single-guide recanalization without displaying the distal target characterized as irresponsible. eurointervention.pcronline.com

3. Galassi AR, et al. Percutaneous Treatment of Coronary Chronic Total Occlusion Part 2: Technical Approach. Interventional Cardiology Review. 2017. Contralateral injection used in 62% of a EuroCTO Club series; dual injection in 78% of a North American series. pmc.ncbi.nlm.nih.gov/articles/PMC5808555

4. CTO PCI in 2014 beyond the J-CTO Score. Dual-injection coronary angiography performed in only about two thirds of cases; single-catheter angiography provides inadequate information. pmc.ncbi.nlm.nih.gov/articles/PMC4561179

5. Contemporary Strategies and Outcomes of Dedicated CTO PCI Programs: A Prospective Multicentre Registry. Technical success improved from 70.0% to 85.6% over a dedicated program; stable major complication rate of 4.7%. pmc.ncbi.nlm.nih.gov/articles/PMC8670896

6. Retrograde CTO PCI via ipsilateral collaterals (PROGRESS-CTO data). Perforation rates ranged from approximately 9.0% to 17.3% across comparison groups. pubmed.ncbi.nlm.nih.gov/38563074

7. SCAI Quality Initiatives for Prevention of Contrast-Induced Acute Kidney Injury. Risk of CI-AKI rises approximately 12% per 100 mL of contrast; keep contrast-volume to creatinine-clearance ratio under 3, ideally under 2. scai.org

8. When Prevention is Truly Better than Cure: Contrast-Associated AKI in PCI. A contrast-volume to creatinine-clearance ratio above 2 is an independent predictor in patients with eGFR < 30; risk minimal below 100 mL or 5 mL/kg. pmc.ncbi.nlm.nih.gov/articles/PMC9461685

9. Contrast-Induced AKI: Specialty-Specific Protocols. Volume of no more than 100 mL preferable for eGFR < 60; even ~30 mL may cause AKI in very high-risk patients; volume-to-clearance ratio above 3.7 associated with increased risk. pmc.ncbi.nlm.nih.gov/articles/PMC2664588

10. Optical coherence tomography to guide percutaneous coronary intervention. EuroIntervention. 2024. OCT requires blood clearance for image acquisition; penetration depth limited to 1 to 2 mm. pmc.ncbi.nlm.nih.gov/articles/PMC11443254

11. Maddali VR, Koganti S. Role of Optical Coherence Tomography in Coronary Interventions. 2024. Blood strongly scatters and attenuates light, so OCT requires a field of view free of blood. journals.sagepub.com/doi/10.1177/26324636231224598

12. Dratsch T, et al. Automation Bias in Mammography: The Impact of Artificial Intelligence BI-RADS Suggestions on Reader Performance. Radiology. 2023. Incorrect machine advice impaired readers across all experience levels; early reliance raises a deskilling risk. pubs.rsna.org/doi/full/10.1148/radiol.222176

13. Automation Bias in AI-Assisted Medical Decision-Making under Time Pressure in Computational Pathology. 2024. Under time pressure, decision-makers abandoned a measurable share of initially correct judgments when faced with erroneous automated advice. arXiv:2411.00998

14. Artificial intelligence in medicine: a scoping review of the risk of deskilling and loss of expertise among physicians. 2026. Overreliance erodes independent reasoning (deskilling) and can prevent skill formation in trainees (never-skilling). sciencedirect.com/science/article/pii/S2949820126000123

15. Sakakura K, et al. Fundamental Wire Technique and Current Standard Strategy of PCI for CTO With Histopathological Insights. JACC Cardiovasc Interv. 2011. CTOs progress from fresh thrombus to organized thrombus with microchannels to fibrotic plaque with dense calcium; older occlusions show dense calcium without microchannels. jacc.org/doi/10.1016/j.jcin.2011.06.011

16. Carlino M, et al. CTO recanalization by intraocclusion injection of contrast: the microchannel technique. Catheter Cardiovasc Interv. 2008. Most CTOs contain intraluminal microchannels of roughly 100 to 500 microns that can provide a guidewire crossing pathway. onlinelibrary.wiley.com/doi/10.1002/ccd.21396

17. Chronic Total Occlusion, histopathology overview. Recanalized channels within organized thrombus enable guidewire crossing, which should be followed by thrombus removal before plaque modification and stenting. sciencedirect.com (Chronic Total Occlusion topic overview)

18. Riley RF, et al. Long-term outcomes of intravascular lithotripsy-facilitated stenting (pooled Disrupt CAD III and IV). J Soc Cardiovasc Angiogr Interv. 2023. Stents placed in inadequately prepared calcific lesions can underexpand or malappose, raising rates of stent thrombosis, restenosis, and target vessel failure. jscai.org

19. Lesion-specific assessment of coronary artery calcium to predict stent underexpansion. Intravascular imaging-based calcium scoring (e.g., OCT angle, thickness, and length; IVUS superficial calcium criteria) directs lesion-preparation strategy, including atherectomy or lithotripsy. pmc.ncbi.nlm.nih.gov/articles/PMC11832659

20. Profitability and portfolio analysis of cardiac catheterization laboratory costs over five years. Cath labs operate as profit centers with high material and labor variable costs; contribution margin is calculated per hour of laboratory time. pmc.ncbi.nlm.nih.gov/articles/PMC5094137

21. Reed GW, et al. Operational Efficiency and Effective Management in the Catheterization Laboratory: JACC Review Topic of the Week. J Am Coll Cardiol. 2018. Inefficient use of the lab raises cost and can compromise care; raw procedure time is heavily influenced by the nature of the case rather than modifiable system variables. jacc.org/doi/10.1016/j.jacc.2018.08.2179