Research · Apr 19, 2026

DCB in chronic total occlusion: the order of the questions

The debate on DCB in chronic total occlusion (CTO) is gaining ground. In my reading, the comparison with DES is being made before two upstream questions have been settled.

Question 1: does CTO PCI change hard outcomes?

DECISION-CTO was neutral for MACE. EuroCTO showed a quality-of-life benefit but was underpowered for clinical events. ORBITA-CTO was needed to show symptom benefit. The question remains open.

Question 2: does DCB work beyond small vessels and restenosis?

REC-CAGEFREE I failed noninferiority against up-front DES in de novo lesions, on 2-year DoCE. SELUTION DeNovo met noninferiority at 1 year (presented at TCT 2025, not yet published at the time). Favourable DCB results remain concentrated in small vessels and in-stent restenosis.

The pooled DCB-in-CTO data

All of it is observational, retrospective and confounded. The Gurgoglione meta-analysis reports TVF with an IRR of 0.56 (0.28–1.10; p = 0.09) and cardiac death with an IRR of 0.55 (0.27–1.10; p = 0.09). Neither is significant.

The cardiac death signal is the most telling one. PCI versus optimal medical therapy does not change mortality in stable coronary disease: ISCHEMIA, COURAGE and BARI-2D showed this. Swapping one revascularization device for another and reducing cardiac death by about 45% at 1–3 years is not biologically plausible. I attribute the signal to residual confounding: sicker patients, longer lesions and worse anatomy get DES. If the mortality signal is confounded, the TVF signal from the same dataset is subject to the same confounding.

DES in CTO is already a heterogeneous comparator

The variability of CTO PCI is not only procedural. DES performs differently depending on whether the wire crossed in the intraplaque or the subintimal space. Song et al. imaged 219 successful CTO recanalizations with IVUS: 52% were subintimal. Even with wire escalation, where operators assume intraplaque crossing, 28% tracked subintimally. Angiography does not reliably show which plane the wire is in.

A trial that randomizes DCB against DES in all-comer CTO estimates an average effect over populations in which the control arm itself behaves differently.

The case for less metal

Less metal has its own rationale, particularly in CTO. Avoiding long overlapping stents (full metal jacket), late malapposition and impaired vasomotion is desirable. This argument does not depend on the questions above.

The sequence I would propose

First restrict the device question to a more controllable anatomic subset. I consider the IMAGINATION approach the best one: intraplaque crossing only, scoring balloon preparation, PCB, and IVUS and FFR endpoints.

The DCB-in-CTO question deserves rigorous study. I think we are getting ahead of ourselves.