Summary
Background
Optical coherence tomography is an imaging method that enables cardiologists to study atheromatous plaques, and to check the implantation and evolution of coronary stents. It is an invasive technique, providing high-resolution (10 μm) in vivo images, but with limitations and artefacts that need to be understood before the field of application can be extended.
Aim
To determine the feasibility and limitations of optical coherence tomography coronary imaging from a single-centre experience.
Methods
We analysed the first 301 optical coherence tomography (version M2, LightLab Imaging) sequences obtained in our department from examination of 73 patients.
Results
Results showed that 92% of sequences for selected lesions were usable, with a mean examination time of 17 min. Only one complication occurred (ventricular fibrillation, reduced by external electroshock). In our registry, sequence quality depended on operator experience (improving after 20 examinations), and was impaired by artefacts, especially in right coronary analysis and in arteries of greater than 3.5 mm calibre.
Conclusions
Proximal coronary occlusion and the distal flush quality currently required for quality imaging should no longer be indispensable with the new generation of optical coherence tomography systems.
Résumé
La tomographie par cohérence optique (OCT) est une imagerie qui permet en cardiologie d’étudier les plaques athéromateuses, de contrôler l’implantation et l’évolution d’endoprothèses coronaires. Cette technique invasive fournit des images in vivo de haute résolution (dix microns) mais comporte des limites et artéfacts qu’il est important de connaître avant d’en étendre le champ d’application. Nous avons analysé les 301 premières séquences OCT (version M2, LightLab Imaging) obtenues dans notre institution à partir des examens de 73 patients. Il ressort de cette expérience que 92 % des séquences sur des lésions sélectionnées sont exploitables au cours d’examens réalisés en moyenne en 17 minutes. Une seule complication a été observée (fibrillation ventriculaire réduite par choc électrique externe). Dans notre registre, la qualité des séquences dépend de l’expérience des opérateurs (meilleure au-delà de 20 examens), est altérée par des artéfacts plus fréquemment retrouvés dans l’analyse de coronaires droites, d’artères de calibres de plus de 3,5 mm. L’occlusion proximale de la coronaire et la qualité du flush distal actuellement nécessaires à l’obtention d’une image de qualité ne devraient plus être indispensables avec la nouvelle génération d’OCT.
Background
Optical coherence tomography (OCT) is a high-resolution imaging method based on the reflection of near-infrared radiation by biological tissue . The light emitted by OCT requires a transparent medium in which to propagate. In coronary imaging this is achieved by saline flush downstream of an occlusion balloon. OCT provides cross-sectional imaging of higher resolution than with intravascular ultrasound (IVUS). The precision of OCT imaging of the more superficial layers of the arterial wall makes it the method of choice for analysis of atheromatous plaque and assessment of coronary stents ; the latter indication applied to most of the present series.
We report the results of the first 73 examinations performed in our department, with 301 automatic pullbacks exploring 112 stents. Analysing the data from this registry disclosed the artefacts and pitfalls of OCT as a method of coronary imaging, enabling us to better define indications and focus research on post-processing to enhance image quality and improve interpretation.
Optical coherence tomography examination
OCT examination was performed by two operators, following a pre-established protocol. It immediately followed coronary angiography, after intravenous injection of 500 mg aspirin and 30 IU/kg of unfractionated heparin. The guide catheter used was a 6-French large lumen model. A long guide catheter was progressively inserted beyond the region of interest (ROI) under X-ray control. The coaxial occlusion balloon was positioned upstream of the ROI before the guide was withdrawn.
The fibre-optic connecting to the console (version M2, LightLab Imaging, Westford, MA, USA) was pre-calibrated manually, introduced into the coaxial balloon and carefully fed downstream of the ROI. The occlusion balloon was inflated, using a dedicated manometer, to between 0.5 and 0.7 atm. Flush was performed manually, using 30 mL of saline in a syringe with a Y-shaped haemostatic connector via the coaxial balloon. When the coronary was flushed with transparent fluid, the OCT image appeared; flush continued during 30 s automatic pullback at 1–2 mm/s. The balloon was deflated at end of pullback. Images were saved in DICOM format to the LightLab console, enabling re-reading and post-real-time measurement. This protocol is similar to ones described elsewhere .
Optical coherence tomography pullback analysis
For each examination, a report detailed the indication, results, limitations and complications, including number of coronaries explored, number of stents studied and number of pullbacks per patient. Examination time was measured from intracoronary introduction of the guide catheter to the end of the last fibre-optic pullback. Procedural complications and hardware problems (e.g., fibre rupture) were inventoried.
For stent analysis, diameters and lengths, time from implantation, rate of complete exploration and number of pullbacks required were recorded.
Concerning pullbacks, artery diameter and site (left anterior descending, circumflex, right coronary artery) were recorded. Pullback quality was assessed by two independent operators, with a third reading in case of disagreement.
Quality criteria
Quality was assessed on predefined scores: 0 equals no usable image; 1 equals less than 50% of pullback images usable; 2 equals greater than 50% of pullback images usable; 3 equals all pullback images usable. An image was defined as usable when the coronary lumen was perfectly distinct from the wall, contrasted, with no break in contour. Pullback acquisition speed was also recorded, as was recourse to proximal occlusion upstream of saline flush.
Artefact identification
Rotation artefacts correspond to false breaks in lumen contour on cross-sectional images. Decentration artefacts are caused by excentric fibre positioning in the lumen, leading to deformation or signal attenuation in the most remote structures and hypersignal in contact structures. Calibre artefacts are caused by an arterial diameter too great for the fibre’s field of exploration to include the entire lumen contour. Flow artefacts comprise all situations of non-optimal lumen transparency (flush defect, incompletely occlusive balloon, or collateral branches). Balloon artefacts occur when the occlusion balloon is present in the most proximal millimetres of pullback, preventing analysis at that point.
Statistical analysis
Quantitative variables are expressed as mean ± standard deviation per group. Comparative analysis used the non-parametric Wilcoxon test quantitative variables, and Fisher’s exact test for qualitative variables. All tests were run bilaterally, with first-order risk set at 5%, and performed on SAS v9 software (SAS Institute, Cary, NC, USA).
Results
Table 1 presents the characteristics of the 73 OCT examinations, grouping separately the first 20 and the following 53. Mean exploration time was 17 min, with more than four pullbacks per examination. Experience enabled exploration time to be shortened by greater than 4 min, despite more sequences being acquired per examination. In all the procedures performed, only one complication arose: a case of ventricular fibrillation at end of acquisition following prolonged occlusion and flush, which was rapidly reduced by external electroshock. Fibre rupture, although always a risk considering the fragility of the fibres, ceased to occur after the 35th examination, and was never of concern for the patient.
| Variable | Total | First 20 patients | Subsequent patients | p |
|---|---|---|---|---|
| Patients ( n ) | 73 | 20 | 53 | |
| Men (%) | 55 (75.3) | 17 (85.0) | 38 (71.7) | 0.36 * |
| Age, years | 57.8 ± 12.1 | 55.1 ± 12.8 | 58.9 ± 11.5 | 0.30 a |
| Indication (%) | ||||
| Stent control | 62 (84.9) | 16 (80.0) | 46 (86.7) | 0.48 * |
| Plaque evaluation | 11 (15.1) | 4 (20.0) | 7 (13.2) | |
| Explored coronary arteries (per patient) | 95 (1.3) | 26 (1.30) | 69 (1.3) | 0.84 a |
| Explored stents (per patient) | 112 (1.5) | 29 (1.45) | 83 (1.6) | 0.89 a |
| Pullbacks (per patient) | 301 (4.1) | 79 (4.0) | 222 (4.2) | 0.25 a |
| Procedure duration (min) | 17.2 ± 7.1 | 20.5 ± 7.4 | 16.0 ± 6.7 | 0.041 a |
| Complications (%) | 1 (1.4) | 0 (0.0) | 1 (1.9) | 0.999 * |
| Fibre rupture (%) | 3 (4.1) | 2 (10) | 1 (1.9) | 0.18 * |
The characteristics of the 112 stents are presented in Table 2 , distinguishing the same two learning-curve examination groups. The stents were mainly drug-eluting stents (DES), mostly explored 6 months after implantation. In 96.4% of cases, the stent (mean length, 19.39 mm) was explored in its entirety, sometimes at the price of iterative pullback (mean number, 1.63 per stent).
| Variable | Total (%) | First 20 patients (%) | Subsequent patients (%) | p |
|---|---|---|---|---|
| Stents explored ( n ) | 112 | 29 | 83 | |
| Type of stent (%) | 0.10 * | |||
| Bare metal | 13 (11.6) | 6 (20.7) | 7 (8.4) | |
| Drug eluting | 99 (88.4) | 23 (79.3) | 76 (91.6) | |
| Time from implantation (days) | ||||
| 0–30 | 14 (12.5) | 7 (24.1) | 7 (8.4) | 0.001 * |
| 140–210 | 90 (80.4) | 17 (58.6) | 73 (88.0) | |
| ≥ 300 | 8 (7.1) | 5 (17.2) | 3 (3.6) | |
| Stent diameter (mm) | 3.00 ± 0.37 | 3.01 ± 0.39 | 2.98 ± 0.37 | 0.26 a |
| Stent length (mm) | 19.39 ± 6.54 | 19.17 ± 7.44 | 19.47 ± 6.24 | 0.79 a |
| Pullbacks per stent | 2.3 ± 1.0 | 2.2 ± 0.9 | 2.3 ± 1.1 | 0.79 a |
| Stents completely explored (%) | 108 (96.4) | 27 (93.1) | 81 (97.6) | 0.28 * |
| Pullbacks required to explore completely ( n ) | 1.6 ± 0.9 | 1.6 ± 0.4 | 1.6 ± 0.9 | 0.9 a |
Data for the 301 pullbacks are presented in Table 3 . Pullback quality was generally good (usable in 92% of cases), and improved with experience (95% for the later examinations). More than half of the good-quality sequences were judged excellent (score 3), and a large majority of the suboptimal sequences (score 2: 43.5% of sequences as a whole) were usable (83% of images per sequence, on average). Taking all acquired images together, 88.8% were usable. It is noteworthy that in the present registry, almost all sequences were acquired with proximal occlusion, and at a pullback speed of 1 mm/s.
| Variable | Total | First 20 patients | Subsequent patients | p |
|---|---|---|---|---|
| Pullbacks | 301 (100) | 79 (26.2) | 222 (73.8) | |
| Usable pullbacks | 277 (92.0) | 66 (83.5) | 211 (95.0) | 0.003 * |
| Unusable pullbacks | 24 (8.0) | 13 (16.5) | 11 (5.0) | |
| Coronaries explored | ||||
| LAD | 161 (53.5) | 36 (45.6) | 125 (56.3) | 0.17 * |
| Cx | 38 (12.6) | 14 (17.7) | 24 (10.8) | |
| RCA | 102 (33.9) | 29 (36.7) | 73 (32.9) | |
| Coronary diameter (mm) | 3.10 ± 0.48 | 3.02 ± 0.54 | 3.14 ± 0.45 | 0.004 a |
| Coronary diameter (mm) | ||||
| < 3.5 | 242 (80.4) | 62 (78.5) | 180 (81.1) | 0.62 * |
| ≥ 3.5 | 59 (19.6) | 17 (21.5) | 42 (18.9) | |
| Pullback quality | ||||
| 0 | 6 (2.0) | 4 (5.1) | 2 (0.9) | 0.003 * |
| 1 | 18 (6.0) | 9 (11.4) | 9 (4.1) | |
| 2 | 131 (43.5) | 37 (46.8) | 94 (42.3) | |
| 3 | 146 (48.5) | 29 (36.7) | 117 (52.7) | |
| OCT frames/pullback | 419.9 ± 79.4 | 422.3 ± 90.2 | 418.9 ± 75.5 | 0.31 a |
| Usable OCT frames/pullback | 373.1 ± 116.7 | 352.1 ± 138.0 | 380.6 ± 107.5 | 0.25 a |
| Non-usable OCT images/pullback | 46.7 ± 79.1 | 70.2 ± 112.6 | 38.4 ± 61.3 | 0.009 a |
| Usable image ratio (%) | 88.8 ± 20.7 | 83.4 ± 27.4 | 90.9 ± 17.1 | 0.006 a |
| Automatic pullback speed (mm/s) | ||||
| 1 | 274 (91.0) | 78 (98.7) | 196 (88.3) | 0.003 * |
| > 1 | 27 (9.0) | 1 (1.3) | 26 (11.7) | |
| Proximal occlusion | ||||
| Yes | 295 (98.0) | 77 (97.5) | 218 (98.2) | 0.66 * |
| No | 6 (2.0) | 2 (2.5) | 4 (1.8) | |
| Rotation artefacts | 18 (6.0) | 4 (5.1) | 14 (6.3) | 0.79 * |
| Decentration artefacts | 93 (30.9) | 28 (35.4) | 65 (29.3) | 0.32 * |
| Coronary calibre artefacts | 45 (15.0) | 14 (17.7) | 31 (14.0) | 0.46 * |
| Flow artefacts | 59 (19.6) | 25 (31.6) | 34 (15.3) | 0.003 * |
| Balloon artefacts | 93 (30.9) | 33 (41.8) | 60 (27.0) | 0.016 * |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
