Radiology

The Documentation Gap: What Actually Gets Interventional Radiology Procedures Paid in 2026

Practical analysis by Elina Sabilova, CPC — Billing Department, WCH Service Bureau

Editorial note: This article synthesizes documentation considerations for five frequently reviewed interventional radiology procedures — UFE, PAE, GAE, HAE, and lower-extremity revascularization — as of mid-2026, including the CPT restructuring that took effect January 1, 2026. It is for informational and educational purposes only and does not constitute legal, coding, or compliance advice. Payer-specific and code-specific requirements should be verified against the current CPT manual and payer policy before billing.

The Real Audit Risk Isn’t the Procedure — It’s the Note

As office-based labs and freestanding IR suites take on a growing share of embolization and endovascular volume, one question comes up in claim reviews more than any other: does the procedure note actually say what the CPT codes claim happened? Not “was the treatment appropriate” — that’s rarely the point of dispute — but whether each vessel, each side, and each step was described specifically enough for the code to stand on its own. In our review of claim denials and downcoding patterns, that gap between what was done and what was documented is consistently one of the more common drivers of coding disputes in interventional radiology.

That gap tends to open in the same five procedure categories, for recognizable reasons. Walking through them side by side is more useful than reading any one in isolation, because the failure pattern repeats — and because 2026 brought a genuine structural change to one of them that makes old habits actively wrong, not just imprecise.

Uterine Fibroid Embolization: Separating the Embolization Code from the Catheterization Question

UFE documentation starts with the basics — symptomatic fibroids, failed conservative therapy, a negative pregnancy test, prior imaging on record — and CPT 37243 covers the embolization/occlusion itself, inclusive of the radiological supervision and interpretation and intraprocedural imaging guidance needed to complete it.

Where notes get muddled is in conflating that embolization code with the separate question of selective catheter placement. Selective catheterization can be separately reported when it’s actually performed and documented as such, and whether treatment is unilateral or bilateral is a clinical finding, not a documentation formality — but it needs to be described accurately rather than assumed. A note that states plainly which side(s) were catheterized, to what order, and why treatment was unilateral or bilateral gives the coder something defensible to work from. A note that’s vague on that point forces a coder to guess, and guessing is what invites a payer to ask questions later.

The other detail worth flagging: routine angiographic imaging performed as part of a planned UFE is generally bundled into 37243 and doesn’t independently support 75736. Where a diagnostic angiogram is genuinely a separate and distinct diagnostic study — meeting the documentation and coding requirements for separate reporting — the note should reflect the actual clinical circumstances that made it so, rather than treating it as an automatic add-on to every case.

Prostatic Artery Embolization: Staged Cases Need Their Own Story

PAE has no single dedicated CPT code of its own; it’s reported using 37243 and the applicable selective catheterization codes based on what was actually performed. Documentation still needs to establish medical necessity for the BPH diagnosis and failure of conservative therapy — IPSS and QOL scores are useful supporting documentation when available, though not a universal requirement across payers — plus a catheter pathway traceable to each individually named prostatic artery, embolic agent and endpoint tied to each side treated.

Where PAE notes most often fall short is in staged treatment. If the case is split across two sessions, the note has to say so explicitly: which side was treated today, and whether the other side is planned for a future session or whether this session completes a two-stage bilateral treatment. A final impression that just says “PAE performed” without stating laterality and stage status leaves a reviewer to infer details that should have been stated outright.

Genicular Artery Embolization: A Newer Code Family That Punishes Vague Language

GAE is newer to most coding workflows than the other four, and — helped along by a fresh SIR position statement on GAE published in May 2026 — the documentation expectations are still settling. The baseline is familiar: laterality, functional limitation, failure of conservative treatment, imaging findings, and, increasingly common though not universally required, a WOMAC score.

The detail that most reliably separates a defensible GAE note from a weak one is specificity at the vessel level: each genicular artery treated should be named individually rather than folded into “the genicular arteries were embolized.” And one phrase is worth banning from GAE templates outright — “nonselective catheter placement” has no business appearing in a note that’s billing selective catheterization codes. A single leftover phrase like that can undercut an otherwise well-documented case.

Hemorrhoid Artery Embolization: Where Templates Go to Get Contaminated

HAE notes typically involve a longer catheterization sequence than the other four — initial catheterization of the inferior mesenteric artery, followed by the third-order branches actually catheterized (right and left anterior and posterior superior rectal arteries, and middle rectal arteries when they’re part of the case), each described individually with laterality and technique. That level of granularity reflects how the anatomy is typically approached and reported; it’s a documentation practice worth following closely, though the specific branches described in any given note should match what was actually catheterized in that case rather than a fixed template applied regardless of anatomy.

HAE is also, not coincidentally, the category where template contamination shows up most often. Because HAE protocols are structurally similar to UFE protocols — both are embolization procedures built around a branching arterial tree — leftover UFE language has a way of surviving into HAE notes: uterine-specific phrasing, fibroid-vascularity references, or symptom language that doesn’t match a hemorrhage indication. A reviewer who spots one stray sentence from the wrong template tends to read the rest of the note more skeptically, whether or not anything else is actually wrong.

Lower-Extremity Revascularization: A Genuinely New Code Set, Not Just New Habits

This is the one category on this list where the change isn’t about documentation discipline — it’s about an entirely new coding structure that took effect January 1, 2026. CPT deleted the familiar 37220–37235 range and replaced it with 46 new codes, 37254–37299, organized around vascular territory rather than the old code-per-vessel logic. There are now four territories — iliac, femoral/popliteal, tibial/peroneal, and a newly defined inframalleolar territory covering the dorsalis pedis and plantar arteries, including the pedal arch — and within each territory, codes are further split by technology used (angioplasty, stent, atherectomy, and combinations) and by lesion complexity, with a straightforward lesion defined as a stenosis and a complex lesion as an occlusion.

Within each territory, the code selected depends on which technology was used — angioplasty alone, angioplasty with stent, angioplasty with atherectomy, angioplasty with atherectomy and stent, or intravascular lithotripsy, which the 2026 restructuring also carved out as its own reportable component — combined with whether the lesion was straightforward or complex, plus a matching add-on code for an additional vessel treated in the same territory. The organizing question a note now needs to answer isn’t just “which vessel,” but “which territory, which technology (or combination of technologies), and which complexity” — and that has to be documented per vessel, since add-on codes can’t be used for additional lesions within the same vessel.

Practically, this means lower-extremity notes written on autopilot using pre-2026 habits will misfire on the new code set even when the underlying clinical work is described accurately. Independent wire crossing and lesion characterization per vessel still matter, but the note also now needs to make the technology used and lesion complexity legible — including whether the lesion is a stenosis or occlusion and which intervention or combination of interventions was performed — because that’s what the 2026 code selection actually turns on.

IVUS (37252/37253) sits alongside this, largely unchanged in principle: each vessel examined should be named, diagnostic findings documented, and the record should reflect that IVUS served a genuine diagnostic purpose rather than device sizing alone. Ultrasound-guided access (76937) has its own documentation expectations — real-time guidance, a retained permanent image, and a patency statement — and it’s worth noting that CMS at one point proposed bundling 76937 into the NCCI edits for these procedures before reversing course after specialty society comment, with ACR confirming it continues to be separately reportable when properly documented. That back-and-forth is a useful reminder that even settled-looking codes can move, and that documentation habits built around “this is how it’s always been billed” are worth rechecking against the current-year rules rather than assumed.

The Pattern Underneath the Pattern

Read across all five categories and a handful of failure modes keep resurfacing: placeholder text (“VERIFY”) left in a signed note; a final impression that doesn’t match the procedural detail above it — bilateral in the summary, unilateral in the body; vessels described collectively (“multiple branches catheterized”) instead of individually; angioplasty written up as independent therapy when it actually treated the same lesion a stent addressed, rather than a genuinely separate one; and template language from an entirely different procedure surviving a copy-paste.

Most of these aren’t failures of clinical detail — the procedure was, in nearly every case, performed and performed appropriately. They’re failures of specificity: the note didn’t say, vessel by vessel and side by side, what the code actually requires it to say.

★ The Takeaway Across UFE, PAE, GAE, and HAE, the coding support comes down to describing each vessel, side, and stage in its own sentence with its own endpoint, rather than in collective or templated language. For lower-extremity revascularization, 2026 raises the bar further: the note now also has to make the territory, technology, and lesion-complexity determination explicit, because that’s what the new 37254–37299 code family is actually built around. Practices scaling up OBL-based IR volume this year are better served auditing their note templates against both standards now — field by field, procedure by procedure — than finding the gaps the way a payer will.

Sources:

• Society of Interventional Radiology. “Coding Q&A: New and Revised Codes for 2026.” IR Quarterly, sirweb.org.
• Society for Cardiovascular Angiography and Interventions. “2026 Lower Extremity Revascularization (LER) Codes” and “Coding Guidelines for Peripheral Interventions 2026.” scai.org.
• American College of Radiology. “2026 Interventional Radiology Coding Update.” acr.org.
• American Medical Association. CPT® 2026 Professional Edition and CPT Changes 2026.
• Centers for Medicare & Medicaid Services. National Correct Coding Initiative Policy Manual, 2026. cms.gov