Excimer laser angioplasty for acute coronary occlusion: a stratified meta-analysis of efficacy against aspiration thrombectomy and conventional PCI Authors: Fouad Hanna, Ali Dway, Sara T Zaki, Mazen Negmeldin Yassin, Abdelrahman El-Helbawy, Luiz F Ybarra, Habib Rehman Khan Published: 2026 ABSTRACT AbstractPrimary percutaneous coronary intervention (PCI) achieves epicardial reperfusion in most STEMI patients, yet microvascular obstruction persists in up to 60% of patients. Excimer laser coronary angioplasty (ELCA) vaporizes thrombus in situ and may reduce distal embolization, but the evidence base has not been systematically synthesized. This systematic review and meta-analysis (PROSPERO CRD420261422463) included comparative studies of adjunctive ELCA versus aspiration thrombectomy (Stratum A) or PCI alone (Stratum B) in acute coronary occlusion. Primary outcomes were final TIMI-3 flow and myocardial blush grade (MBG) 3; secondary outcomes were short-term mortality, MACCE, and slow-flow/no-reflow. A random-effects model with Hartung-Knapp-Sidik-Jonkman confidence intervals was applied to all outcomes. Certainty was assessed with GRADE. Ten studies (1 RCT, 9 observational) were included, from a total enrolled population exceeding 3,500. In Stratum A, no outcome reached significance: MBG-3 (OR 3.57, 95% CI 0.07–185.10), mortality (OR 0.31, 0.02–4.04), MACCE (OR 0.22, 0.04–1.26), TIMI-3 flow (OR 1.58, 0.67–3.75) and slow-flow/no-reflow (OR 0.78, 0.22–2.78). In Stratum B, using each study’s propensity-matched data, no outcome differed significantly (TIMI-3 OR 0.88, 0.38–2.03; MBG-3 OR 1.06, 0.13–8.43; slow-flow/no-reflow OR 0.93, 0.29–3.02; mortality OR 0.44, 0.05–3.80). Composite endpoints were not pooled across incompatible follow-up horizons, and all outcomes were of very low certainty. Adjunctive ELCA-containing strategies during primary PCI were not associated with improved angiographic or short-term clinical outcomes against either comparator. Multicenter randomized trials are required before recommending clinical adoption.Graphical abstractClinical Rationale, Methodological Framework, and Outcomes of Adjunctive Excimer Laser Coronary Angioplasty (ELCA) in Acute Coronary Occlusion. This graphical abstract summarizes the clinical unmet need, intervention methodology, and pooled quantitative evidence evaluating the efficacy of adjunctive ELCA during primary percutaneous coronary intervention (PCI). The synthesis encompasses 10 comparative studies (1 RCT, 9 observational) comprising over 3,500 patients, pre-specified into two comparator strata (mechanical aspiration thrombectomy vs. PCI alone using matched data). CI: confidence interval; ELCA: excimer laser coronary angioplasty; GLMM: generalized linear mixed model; GRADE: Grading of Recommendations Assessment, Development and Evaluation; HK-SJ: Hartung-Knapp-Sidik-Jonkman; MACCE: major adverse cardiovascular and cerebrovascular events; MBG: myocardial blush grade; MVO: microvascular obstruction; NS: not significant; OR: odds ratio; PCI: percutaneous coronary intervention; PRISMA: Preferred Reporting Items for Systematic Reviews and Meta-Analyses; PROSPERO: International Prospective Register of Systematic Reviews; RCT: randomized controlled trial; STEMI: ST-elevation myocardial infarction; TIMI: Thrombolysis in Myocardial Infarction. Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1007/s10103-026-05032-x.Keywords: Aspiration thrombectomy, Excimer laser coronary angioplasty, Myocardial infarction, PCI, STEMI, Unstable angina Introduction Primary percutaneous coronary intervention (PCI) is the standard of care for emergent reperfusion in ST-elevation myocardial infarction (STEMI); more than 600,000 PCI procedures, including primary PCI for STEMI, are performed annually in the United States [1]. Normalized epicardial flow is achieved in approximately 90% of patients [2]. Despite a high rate of epicardial success, up to 60% experience suboptimal myocardial reperfusion, distal atherothrombotic embolization, and ischemia–reperfusion injury, which lead to microvascular obstruction (MVO) and, in turn, to adverse left ventricular remodeling, heart failure, and death [3–5]. This gap between epicardial patency and tissue reperfusion is the unmet need in STEMI management. Manual aspiration thrombectomy improved myocardial blush grade (MBG) and ST-segment resolution in the TAPAS trial but did not reduce mortality or reinfarction in larger randomized controlled trials (RCTs) [6]. Consequently, the 2021 ACC/AHA/SCAI revascularization guideline, the 2023 ESC ACS guidelines, and the 2025 ACC/AHA/ACEP/NAEMSP/SCAI ACS guideline assign it a Class 3 (No Benefit) recommendation with Level A evidence [7–9]. Glycoprotein IIb/IIIa inhibitors are recommended mainly as bailout rather than routine therapy [9]. Excimer laser coronary angioplasty (ELCA) uses pulsed ultraviolet energy at 308 nm through fiber-optic catheters, producing photochemical ablation, photothermal vaporization of intracellular water, and photomechanical disruption by expanding vapor bubbles [10, 11]. Unlike aspiration thrombectomy, ELCA is proposed to vaporize thrombus in situ, which would reduce distal microembolization and the stroke risk associated with catheter withdrawal through the aorta; these are mechanistic rationales shown experimentally rather than established clinical effects [10, 12]. The clinical evidence is mainly from small, retrospective studies, with no systematic review available. This systematic review and meta-analysis asked whether adjunctive ELCA during primary PCI for acute coronary occlusion improves angiographic outcomes and reduces short-term events against two pre-specified comparators: PCI with manual aspiration thrombectomy (Stratum A) and PCI alone without thrombus management (Stratum B), since combining different comparators would produce an uninterpretable summary. Because ELCA is almost always deployed within a broader procedural strategy, the comparisons are between ELCA-containing and comparator strategies. Methods Study design and registration This systematic review and meta-analysis followed the PRISMA 2020 guidelines and the Cochrane Handbook framework [13, 14] and was prospectively registered on PROSPERO (CRD420261422463). Eligibility criteria Population, intervention, comparator, and outcomes (PICO) Eligible studies enrolled adults (≥ 18 years) with acute coronary occlusion, including STEMI, non-ST-elevation myocardial infarction (NSTEMI) or unstable angina, undergoing primary or urgent PCI. The intervention was ELCA as adjunctive therapy during PCI with any catheter size, fluence or repetition rate. The comparator was pre-specified in two independent strata: Stratum A, mechanical aspiration thrombectomy; and Stratum B, conventional or balloon-based PCI alone without dedicated thrombus management. Primary outcomes were final Thrombolysis in Myocardial Infarction (TIMI) 3 flow and MBG 3. Secondary outcomes were slow-flow/no-reflow (intraprocedural or on final angiogram), short-term all-cause or cardiovascular mortality (in-hospital or ≤ 30 days), and MACE/MACCE as defined by each study. Inclusion and exclusion criteria Studies were eligible if they were RCTs or cohort studies with at least one ELCA arm and one non-ELCA control arm, reported at least one pre-specified outcome, and were available in English, with no restriction on publication date. Single-arm registries, animal or in vitro experiments, case reports or series, case–control studies, and secondary publications were excluded. Information sources and search strategy A pre-specified search was conducted across PubMed/MEDLINE, Scopus, and Web of Science from database inception to 8 June 2026, using the complete database-specific Boolean strings reported in Supplementary Table S1. Google Scholar was searched as a supplementary source, and backward and forward citation searching were performed. No language or date restrictions were applied beyond the availability of an English full text Supplementary Table S2. Data extraction Two reviewers independently extracted standardized data using a pre-piloted form; disagreements were resolved by consensus. Extracted study data comprised author, year, country, design, enrollment period, sample size per arm, ELCA catheter specifications, comparator stratum, endpoint definitions, outcome ascertainment, follow-up duration, statistical adjustment, and funding source. Patient-level characteristics and procedural co-interventions were extracted per arm for every study. Risk of bias assessment Risk of bias was independently assessed by two reviewers using study design-specific tools: the Cochrane Risk of Bias 2 (RoB 2) tool for RCTs[15] and the Risk of Bias in Non-randomized Studies of Interventions (ROBINS-I) tool for observational studies [16]. Discrepancies were resolved by consensus. Statistical analysis Choice of statistical model Statistical model choice was based on the expected distribution of true effects rather than by the number of contributing studies: a random-effects model with Hartung–Knapp–Sidik–Jonkman confidence intervals is reported for every outcome, including those with fewer than four studies. τ2 was estimated by restricted maximum likelihood, and the ad hoc variance correction of Röver, Knapp and Friede was applied so that the Hartung–Knapp interval is never narrower than the classical random-effects interval [17]. All outcomes were binary; study-level log odds ratios (OR) with 95% confidence intervals (CI) were computed from each 2 × 2 table and pooled by inverse variance under the random-effects model. When a study contributed a zero-event cell, a continuity correction of 0.5 was applied to all four of its cells before pooling; studies with zero events in both arms were non-estimable and excluded from the pooled estimate while remaining listed in the tables. To mitigate rare event bias, a pre-specified sensitivity analysis used a generalized linear mixed model (GLMM; binomial family, random intercept logistic link) without continuity correction. All analyses were performed in R using the metafor package [18, 19]. Adjusted estimate and influence analyses When a study reported adjusted analyses (propensity matched or multivariable), we used those estimates rather than crude event counts to reduce confounding. Accordingly, the Stratum B analyses use the propensity matched datasets of Masuda 2022 (71 pairs), Matsuda 2025/STELA (368 pairs), and Iiya 2025 (180 pairs). Composite endpoints were not pooled when the contributing studies ascertained events over incompatible horizons (in-hospital versus approximately two years); the individual study estimates are reported instead. Leave-one-out influence analysis was performed by refitting the pooled model after omitting each contributing study in turn, using the same random effects Hartung–Knapp specification as the primary analysis [20]. Because several Stratum B estimates received more than 80% of their weight from the STELA registry, it is reported in full for every outcome to which that registry contributed Fig. S1. Further pre-specified sensitivity analyses excluded studies at serious risk of bias (ROBINS-I), historical control studies, and studies defining slow flow intraprocedurally rather than on final angiography, and restricted the analysis to STEMI only and to non-overlapping cohorts Supplementary Table S3. Heterogeneity assessment Statistical heterogeneity was quantified using Cochran's Q test (P < 0.10) and the I2 statistic (< 25%, low; 25–50%, moderate; 50–75%, substantial; > 75%, considerable) derived from the random effects fit, and τ2 was also reported. Certainty of evidence Certainty of evidence for each pooled outcome was rated using the Grading of Recommendations, Assessment, Development and Evaluations (GRADE) approach [21] and summarized in GRADEpro GDT evidence profiles for each comparator stratum, by two reviewers independently. Because the evidence was predominantly non randomized, all outcomes began at low certainty. Certainty was then rated down across the five domains: risk of bias when studies received serious overall ratings; inconsistency when I2 exceeded 50%, when τ2 indicated meaningful between study variance, or when point estimates were widely dispersed, with confidence interval overlap alone not used as a criterion; indirectness when populations, comparators or outcome definitions differed from the review question; and imprecision on the optimal information size, the number of events and whether the 95% CI crossed the no effect line or an appreciable benefit or harm threshold, with sparse event outcomes and those dominated by a single registry downgraded accordingly. Publication bias could not be tested because fewer than ten studies contributed to every outcome. The rationale for each downgrade is recorded in the evidence profile footnotes. Results Study selection A total of 1,411 records were identified; after removal of 279 duplicates, 1,132 were screened by title and abstract, and 20 full texts were retrieved. Ten comparative studies met all pre-specified eligibility criteria and were included in the qualitative and quantitative synthesis [12, 22–30] (Fig. 1). Preferred reporting items for systematic reviews and meta-analyses flow chart of literature search Study characteristics and methodological quality The evidence base comprises ten comparative studies enrolling patients with acute coronary syndrome (ACS), predominantly STEMI. Only one RCT was identified (LaserAMI, n = 27), the remaining nine being retrospective. Three applied propensity score matching (PSM) (Masuda 2022; the multicenter STELA registry, Matsuda 2025; and Iiya 2025), two relied on era-based group assignment (Kujiraoka 2023; Shibata 2019), and one used age- and sex matched controls (Shishikura 2013). The pooled estimates were dominated by STELA (2593 patients; 368 matched pairs), and eight of ten studies originated from Japanese single centers. The comparator was pre-specified into two strata: Stratum A compared ELCA with manual aspiration thrombectomy (Shishikura 2013; Shibata 2019; Arai 2023; Shimojo 2023), whereas Stratum B compared ELCA with conventional or balloon-based PCI without dedicated thrombus management (Ilkay 2005; Dörr 2007; Masuda 2022; Kujiraoka 2023; Matsuda 2025/STELA; Iiya 2025). Six studies enrolled STEMI exclusively, three admitted non-STEMI presentations (Shishikura 2013; Arai 2023; Iiya 2025), and Masuda 2022 enrolled both. Follow-up was predominantly confined to the index hospitalization; only STELA (median 815 days) and Iiya 2025 (approximately two years) extended beyond discharge Table 1. Composite outcomes were analyzed separately for each follow-up period because definitions and ascertainment windows varied Supplementary Table S4. Table 1Characteristics of studies included in the systematic review and meta-analysis of excimer laser coronary angioplasty/atherectomy (ELCA)Study (identifier)CountryDesign & settingPopulation and diagnosisSample size (ELCA vs comparator)ComparatorELCA device & techniqueEndpoint ascertainmentFollow-upStratum A — ELCA versus manual aspiration thrombectomy Shishikura 2013 (Circ J)Japan (Takarazuka)Retrospective single-center; age- and sex-matchedSTEMI 40/50 (80.0%) vs 41/48 (85.4%); NSTEMI 2 (4.0%) vs 4 (8.3%); UAP 1 (2.0%) vs 0; recent MI 7 (14.0%) vs 3 (6.3%)ELCA 50 vs aspiration 48 (analyzed 98)Manual aspiration thrombectomy (TVAC); no distal protectionCVX-300, 308 nm, 135 ns, 200 mJ/pulse; 1.4/1.7/2.0 mm; 45 → 60 mJ/mm2 at 25 → 40 Hz; 0.5 mm/sTIMI, MBG, cTFC (3 observers); 30–60 min ECG STR; WMSIIn-hospital/short-term (WMSI to 2 wk) Shibata 2019 (Int J Cardiovasc Imaging)Japan (Ogaki)Retrospective single-center; era-based (ELCA available from February 2016)STEMI 32/32 (100%) vs 40/40 (100%); all initial TIMI 0/1; OBT < 6 hELCA 32 vs non-ELCA 40 (analyzed 72/316)Conventional PCI incl. manual aspiration (used in 85.0% of comparator and 40.6% of ELCA arms)Spectranetics ELCA before balloon; catheter size not specifiedFinal TIMI; myocardial salvage index by 123I-BMIPP/⁹⁹ᵐTc-tetrofosmin SPECTIn-hospital (MACCE); SPECT 3–7 d Arai 2023 (Lasers Med Sci)Japan (Tokyo)Retrospective single-center comparativeSTEMI 47/48 (98%) vs 42/50 (84%); NSTEMI 1 vs 7; UAP 0 vs 1; all culprit TIMI 0 within 24 hELCA 48 vs TA 50 (analyzed 98/113)Manual thrombus aspiration (Thrombuster II)Spectranetics; 0.9 mm in 47/48 (98%), 1.4 mm in 1; saline 2–3 mL/sTIMI grade, MBG, corrected TIMI frame count (visual)In-hospital (to discharge) Shimojo 2023 (Int J Cardiovasc Imaging)Japan (Ogaki)Retrospective single-center comparativeSTEMI 63/63 (100%) vs 80/80 (100%)ELCA 63 vs aspiration 80 (analyzed 143/571)Manual aspiration thrombectomySpectranetics; 0.5 mm/s; retrograde saline at discretion; size NR; IVUS 100%Peak CK/CK-MB; final TIMI; MSI & LVEF by SPECT; MACCE (ARC)In-hospital (MACCE); SPECT 3–10 d & 3 moStratum B — ELCA versus PCI alone Ilkay 2005 (Angiology)Turkey (Elazığ)Comparative, sequential treatment groups; single-centerSTEMI 36/36 (100%) vs 44/44 (100%)ELCA 36 vs PTCA 44 (analyzed 80)Balloon angioplasty (PTCA) + stentXeCl CVX-300 (Spectranetics), 308 nm, 135 ns; 1.4/1.7 mm; 45 mJ/mm2 at 25 Hz; saline 7–15 mL flush; 0.5 mm/sTIMI flow; QCA MLD; 90-min ST-segment resolutionIn-hospital; 90-min STR Dörr 2007 (LaserAMI)Germany (Greifswald)Prospective open-label RCT; blinded angiographic core-lab; part-funded by Spectranetics IncSTEMI 14/14 (100%) vs 13/13 (100%); presenting < 12 h; cardiogenic shock excludedRandomized 27 (ELCA 14 vs control 13); all analyzedBalloon angioplasty + BMS aloneCVX-300 (Spectranetics), 308 nm, 135 ns, 165 mJ/pulse; 0.9/1.4/1.7 mm; 25 Hz at 45 mJ/mm2; saline-flush; 0.2–0.5 mm/sQCA, TIMI, cTFC, MBG (blinded core-lab); 90-min ECG STRIn-hospital/procedural (to discharge) Masuda 2022 (Lasers Med Sci)Japan (Tokyo CCU Network)Retrospective single-centre; PS-matched + multivariable logistic regressionSTEMI + NSTEMI (mixed; split not reported per arm)Pre-PSM 104 vs 89; post-PSM 71 vs 71Conventional PCI alone (incl. aspiration, distal protection)Small-diameter, low-emission ELCA (0.9 mm); fluence/rate escalated in 10.9%In-hospital 30-day mortality; TIMI flow (site-reported)In-hospital (30-day) Kujiraoka 2023 (Lasers Med Sci)Japan (Tokyo, Hiroo)Retrospective single-center; era-based (ELCA April 2015–March 2019 vs conventional April 2009–March 2012)STEMI 167/167 (100%) vs 123/123 (100%); OBT 0–72 hELCA 167 vs conventional 123 (analyzed 290/319)Conventional PCI incl. aspiration (pre-ELCA era)Spectranetics ELCA; fluence/rate escalated as needed; then BMS or DESFinal TIMI, MBG, intra-procedural slow-flow/no-reflow (visual)In-hospital/procedural Matsuda 2025 (STELA; UMIN-CTR 000043818)Japan (multicenter)Multicenter retrospective registry; PS-matched (368 pairs); Cox regression; blinded event adjudication and central angiographic core labSTEMI 427/427 (100%) vs 2,166/2,166 (100%); primary PCI ≤ 24 hTotal 2,593 (ELCA 427 vs non-ELCA 2,166); PSM 736 (368 pairs)Mixed conventional PCI without ELCA0.9 mm (82.9%) or 1.4–1.7 mm (17.1%); Spectranetics/PhilipsQCA TIMI, cTFC, MBG; baseline TIMI thrombus grade; TV-MACEMedian 815 d (390–1385) for TV-MACE; angiographic at procedure Iiya 2025 (Catheter Cardiovasc Interv)Japan (Hiratsuka)Retrospective single-center observational; PS-matched (180 pairs)ACS, approximately 70% STEMI before matchingPre-PSM 256 vs 330; post-PSM 180 vs 180Mixed conventional PCI without ELCA0.9 mm (96%), 1.4 mm (4%), 1.7 mm (< 1%); single pass; mean 72 ± 46 s; pre-balloonMACE (cardiac death, non-fatal MI, TLR); slow-flow, perforation, distal embolization; IVUS-guidedUp to ≈2 y (Kaplan–Meier)Studies are grouped by comparator and ordered chronologically. Sample size is ELCA versus comparator; for propensity-matched studies both the pre- and post-matching cohorts are shown. Dörr 2007 was supported in part by Spectranetics IncACS, acute coronary syndrome; ARC, Academic Research Consortium; BMIPP, β-methyl-iodophenyl pentadecanoic acid; BMS, bare-metal stent; CK, creatine kinase; cTFC, corrected TIMI frame count; DES, drug-eluting stent; ELCA, excimer laser coronary angioplasty/atherectomy; IVUS, intravascular ultrasound; LVEF, left ventricular ejection fraction; MACCE, major adverse cardiac and cerebrovascular events; MACE, major adverse cardiac events; MBG, myocardial blush grade; MI, myocardial infarction; MLD, minimal lumen diameter; MSI, myocardial salvage index; NR, not reported; NSTEMI, non–ST-elevation myocardial infarction; OBT, onset-to-balloon time; PCI, percutaneous coronary intervention; PS, propensity score; PSM, propensity score matching; PTCA, percutaneous transluminal coronary angioplasty; QCA, quantitative coronary angiography; RCT, randomized controlled trial; SPECT, single-photon emission computed tomography; STEMI, ST-elevation myocardial infarction; STR, ST-segment resolution; TA, thrombus aspiration; TIMI, Thrombolysis In Myocardial Infarction; TLR, target-lesion revascularization; TV-MACE, target-vessel MACE; TVAC, TransVascular Aspiration Catheter; UAP, unstable angina pectoris; WMSI, wall-motion score index Characteristics of studies included in the systematic review and meta-analysis of excimer laser coronary angioplasty/atherectomy (ELCA) Studies are grouped by comparator and ordered chronologically. Sample size is ELCA versus comparator; for propensity-matched studies both the pre- and post-matching cohorts are shown. Dörr 2007 was supported in part by Spectranetics Inc ACS, acute coronary syndrome; ARC, Academic Research Consortium; BMIPP, β-methyl-iodophenyl pentadecanoic acid; BMS, bare-metal stent; CK, creatine kinase; cTFC, corrected TIMI frame count; DES, drug-eluting stent; ELCA, excimer laser coronary angioplasty/atherectomy; IVUS, intravascular ultrasound; LVEF, left ventricular ejection fraction; MACCE, major adverse cardiac and cerebrovascular events; MACE, major adverse cardiac events; MBG, myocardial blush grade; MI, myocardial infarction; MLD, minimal lumen diameter; MSI, myocardial salvage index; NR, not reported; NSTEMI, non–ST-elevation myocardial infarction; OBT, onset-to-balloon time; PCI, percutaneous coronary intervention; PS, propensity score; PSM, propensity score matching; PTCA, percutaneous transluminal coronary angioplasty; QCA, quantitative coronary angiography; RCT, randomized controlled trial; SPECT, single-photon emission computed tomography; STEMI, ST-elevation myocardial infarction; STR, ST-segment resolution; TA, thrombus aspiration; TIMI, Thrombolysis In Myocardial Infarction; TLR, target-lesion revascularization; TV-MACE, target-vessel MACE; TVAC, TransVascular Aspiration Catheter; UAP, unstable angina pectoris; WMSI, wall-motion score index Baseline demographic, clinical, and angiographic characteristics The ELCA and comparator groups were comparable in age, sex, and cardiovascular risk factors. However, ELCA was preferentially used in high thrombus burden lesions, and several studies restricted enrolment to lesions with initial TIMI flow grade 0/1 or angiographic high grade thrombus. In STELA’s multivariable model, TIMI thrombus grade 4/5 independently predicted ELCA use (OR 1.48, 95% CI 1.05–2.08), as did larger reference vessel diameter and younger age. Confounding by indication is therefore measurable, not merely postulated Table 2. Table 2Baseline demographic, clinical, and angiographic characteristics of the ELCA and comparator groupsStudyAge, yFemale, n/N (%)BMI, kg/m2RVD, mmLesion length, mmHigh thrombus burdenLoading antiplateletHypertension, n/N (%)Diabetes mellitus, n/N (%)Prior MI / CADStratum A — ELCA versus manual aspiration thrombectomyShishikura 2013E: 67.0 ± 9.8C: 66.1 ± 12.9E: 9/50 (18.0%)C: 12/48 (25.0%)NRE: 3.1 ± 0.6C: 3.2 ± 0.6E: 25.4 ± 14.1C: 18.8 ± 7.6E: High (NR as n/N)C: NRE: 100%C: 100%E: 29/50 (58.0%)C: 26/48 (54.2%)E: 16/50 (32.0%)C: 15/48 (31.3%)E: MI 6/50 (12.0%)C: MI 2/48 (4.2%)Shibata 2019E: 62.9 ± 12.4C: 69.8 ± 11.0E: 6/32 (18.8%)C: 13/40 (32.5%)E: 23.9 ± 4.6C: 24.4 ± 4.4NRNRE: All TIMI 0/1C: All TIMI 0/1E: 32/32 (100%)C: 16/40 (40.0%)E: 20/32 (62.5%)C: 22/40 (55.0%)E: 4/32 (12.5%)C: 13/40 (32.5%)NRArai 2023E: 67.6 ± 2.0C: 67.7 ± 2.2E: 12/48 (25.0%)C: 16/50 (32.0%)NRE: 3.03 ± 0.88C: 2.73 ± 0.93NRE: All TIMI 0C: All TIMI 0E: 100%C: 100%E: 22/48 (45.8%)C: 30/50 (60.0%)E: 11/48 (22.9%)C: 14/50 (28.0%)NRShimojo 2023E: 67.6 ± 11.1C: 68.8 ± 12.0E: 11/63 (17.5%)C: 18/80 (22.5%)E: 24.7 ± 3.3C: 23.7 ± 3.0NRNRE: STEMI culpritC: STEMI culpritE: 100%C: 100%E: 47/63 (74.6%)C: 60/80 (75.0%)E: 20/63 (31.7%)C: 26/80 (32.5%)NRStratum B — ELCA versus PCI aloneIlkay 2005E: 50.4 ± 9.8C: 57.8 ± 12.2E: 4/36 (11.1%)C: 5/44 (11.4%)NRNRNRE: STEMI culpritC: STEMI culpritNRE: 12/36 (33.3%)C: 10/44 (22.7%)E: 5/36 (13.9%)C: 8/44 (18.2%)NRDörr 2007 (LaserAMI)E: 59.2 ± 10.6C: 57.8 ± 9.2E: 1/14 (7.1%)C: 1/13 (7.7%)E: 30.4 ± 5.0C: 27.9 ± 3.0E: 2.9 ± 0.2C: 2.9 ± 0.4E: 18.4 ± 4.9C: 16.0 ± 4.5E: 11/14 (78.6%)C: 10/13 (76.9%)E: 14/14 (100%)C: 13/13 (100%)E: 2/14 (14.3%)C: 3/13 (23.1%)E: 1/14 (7.1%)C: 1/13 (7.7%)NRMasuda 2022E: 67.9 ± 14.2C: 68.5 ± 14.7E: 16/71 (22.5%)C: 20/71 (28.2%)NRNRNRE: High (NR as n/N)C: High (NR as n/N)E: 15/71 (21.1%) *C: 20/71 (28.2%) *E: 39/71 (54.9%)C: 46/71 (64.8%)E: 20/71 (28.2%)C: 24/71 (33.8%)NRKujiraoka 2023E: 66.1 ± 14.0C: 65.5 ± 14.1E: 39/167 (23.4%)C: 27/123 (22.0%)E: 23.7 ± 3.9C: 23.9 ± 3.9NRNRE: High (long OBT)C: High (long OBT)E: 138/167 (82.1%) †C: 0/123 (0%) †E: 96/167 (57.1%)C: 72/123 (58.5%)E: 46/167 (27.3%)C: 34/123 (27.6%)NRMatsuda 2025 (STELA)E: 66.4 ± 13.2C: 68.5 ± 12.7E: 89/427 (20.8%)C: 509/2166 (23.5%)E: 23.8 ± 4.1C: 23.8 ± 3.9E: 2.76 ± 0.76C: 2.64 ± 0.82E: 13.4 ± 7.0C: 13.4 ± 8.4E: 371/427 (86.9%)C: 1829/2166 (84.6%)NRE: 237/427 (55.5%)C: 1336/2166 (61.7%)E: 121/427 (28.3%)C: 704/2166 (32.5%)E: ISL 19/427 (4.4%)C: ISL 76/2166 (3.5%)Iiya 2025E: 67 ± 12C: 68 ± 11E: 39/180 (21.7%)C: 39/180 (21.7%)E: 24.3 ± 3.8C: 24.6 ± 3.7E: 3.1 ± 0.6C: 3.1 ± 0.5E: 28.1 ± 13.0C: 26.1 ± 12.2E: 154/180 (85.6%)C: 162/180 (90.0%)NRE: 114/180 (63.3%)C: 104/180 (57.8%)E: 66/180 (36.7%)C: 64/180 (35.6%)E: MI 9/180 (5.0%)C: MI 13/180 (7.2%)Upper value (E) is the ELCA group, lower value (C) the comparator. Continuous variables are mean ± SD, categorical variables n/N (%). *Masuda 2022 reported any antiplatelet use rather than a loading dose. †Kujiraoka 2023 reported prasugrel loading, with a marked between-era imbalance. Values are from the matched cohorts for Masuda 2022 and Iiya 2025 and from the total cohort for Matsuda 2025CAD, coronary artery disease; ELCA, excimer laser coronary angioplasty/atherectomy; ISL, in-stent lesion; MI, myocardial infarction; NR, not reported; OBT, onset-to-balloon time; PCI, percutaneous coronary intervention; RVD, reference vessel diameter; STEMI, ST-elevation myocardial infarction; TIMI, Thrombolysis In Myocardial Infarction Baseline demographic, clinical, and angiographic characteristics of the ELCA and comparator groups E: 67.0 ± 9.8 C: 66.1 ± 12.9 E: 9/50 (18.0%) C: 12/48 (25.0%) E: 3.1 ± 0.6 C: 3.2 ± 0.6 E: 25.4 ± 14.1 C: 18.8 ± 7.6 E: High (NR as n/N) C: NR E: 100% C: 100% E: 29/50 (58.0%) C: 26/48 (54.2%) E: 16/50 (32.0%) C: 15/48 (31.3%) E: MI 6/50 (12.0%) C: MI 2/48 (4.2%) E: 62.9 ± 12.4 C: 69.8 ± 11.0 E: 6/32 (18.8%) C: 13/40 (32.5%) E: 23.9 ± 4.6 C: 24.4 ± 4.4 E: All TIMI 0/1 C: All TIMI 0/1 E: 32/32 (100%) C: 16/40 (40.0%) E: 20/32 (62.5%) C: 22/40 (55.0%) E: 4/32 (12.5%) C: 13/40 (32.5%) E: 67.6 ± 2.0 C: 67.7 ± 2.2 E: 12/48 (25.0%) C: 16/50 (32.0%) E: 3.03 ± 0.88 C: 2.73 ± 0.93 E: All TIMI 0 C: All TIMI 0 E: 100% C: 100% E: 22/48 (45.8%) C: 30/50 (60.0%) E: 11/48 (22.9%) C: 14/50 (28.0%) E: 67.6 ± 11.1 C: 68.8 ± 12.0 E: 11/63 (17.5%) C: 18/80 (22.5%) E: 24.7 ± 3.3 C: 23.7 ± 3.0 E: STEMI culprit C: STEMI culprit E: 100% C: 100% E: 47/63 (74.6%) C: 60/80 (75.0%) E: 20/63 (31.7%) C: 26/80 (32.5%) E: 50.4 ± 9.8 C: 57.8 ± 12.2 E: 4/36 (11.1%) C: 5/44 (11.4%) E: STEMI culprit C: STEMI culprit E: 12/36 (33.3%) C: 10/44 (22.7%) E: 5/36 (13.9%) C: 8/44 (18.2%) E: 59.2 ± 10.6 C: 57.8 ± 9.2 E: 1/14 (7.1%) C: 1/13 (7.7%) E: 30.4 ± 5.0 C: 27.9 ± 3.0 E: 2.9 ± 0.2 C: 2.9 ± 0.4 E: 18.4 ± 4.9 C: 16.0 ± 4.5 E: 11/14 (78.6%) C: 10/13 (76.9%) E: 14/14 (100%) C: 13/13 (100%) E: 2/14 (14.3%) C: 3/13 (23.1%) E: 1/14 (7.1%) C: 1/13 (7.7%) E: 67.9 ± 14.2 C: 68.5 ± 14.7 E: 16/71 (22.5%) C: 20/71 (28.2%) E: High (NR as n/N) C: High (NR as n/N) E: 15/71 (21.1%) * C: 20/71 (28.2%) * E: 39/71 (54.9%) C: 46/71 (64.8%) E: 20/71 (28.2%) C: 24/71 (33.8%) E: 66.1 ± 14.0 C: 65.5 ± 14.1 E: 39/167 (23.4%) C: 27/123 (22.0%) E: 23.7 ± 3.9 C: 23.9 ± 3.9 E: High (long OBT) C: High (long OBT) E: 138/167 (82.1%) † C: 0/123 (0%) † E: 96/167 (57.1%) C: 72/123 (58.5%) E: 46/167 (27.3%) C: 34/123 (27.6%) E: 66.4 ± 13.2 C: 68.5 ± 12.7 E: 89/427 (20.8%) C: 509/2166 (23.5%) E: 23.8 ± 4.1 C: 23.8 ± 3.9 E: 2.76 ± 0.76 C: 2.64 ± 0.82 E: 13.4 ± 7.0 C: 13.4 ± 8.4 E: 371/427 (86.9%) C: 1829/2166 (84.6%) E: 237/427 (55.5%) C: 1336/2166 (61.7%) E: 121/427 (28.3%) C: 704/2166 (32.5%) E: ISL 19/427 (4.4%) C: ISL 76/2166 (3.5%) E: 67 ± 12 C: 68 ± 11 E: 39/180 (21.7%) C: 39/180 (21.7%) E: 24.3 ± 3.8 C: 24.6 ± 3.7 E: 3.1 ± 0.6 C: 3.1 ± 0.5 E: 28.1 ± 13.0 C: 26.1 ± 12.2 E: 154/180 (85.6%) C: 162/180 (90.0%) E: 114/180 (63.3%) C: 104/180 (57.8%) E: 66/180 (36.7%) C: 64/180 (35.6%) E: MI 9/180 (5.0%) C: MI 13/180 (7.2%) Upper value (E) is the ELCA group, lower value (C) the comparator. Continuous variables are mean ± SD, categorical variables n/N (%). *Masuda 2022 reported any antiplatelet use rather than a loading dose. †Kujiraoka 2023 reported prasugrel loading, with a marked between-era imbalance. Values are from the matched cohorts for Masuda 2022 and Iiya 2025 and from the total cohort for Matsuda 2025 CAD, coronary artery disease; ELCA, excimer laser coronary angioplasty/atherectomy; ISL, in-stent lesion; MI, myocardial infarction; NR, not reported; OBT, onset-to-balloon time; PCI, percutaneous coronary intervention; RVD, reference vessel diameter; STEMI, ST-elevation myocardial infarction; TIMI, Thrombolysis In Myocardial Infarction Procedural co-interventions and in-hospital/short-term clinical and angiographic outcomes Procedural co-interventions were unevenly reported and, when reported, unevenly distributed. Aspiration was more frequent in the comparator arms of Shibata 2019 (40.6% vs 85.0%) and Kujiraoka 2023 (29.9% vs 83.7%), and IABP use was imbalanced in opposite directions in Shishikura 2013 and Shibata 2019 but balanced in STELA. Glycoprotein IIb/IIIa inhibitors, intracoronary vasodilators, and fibrinolysis were mostly unreported, and loading antiplatelet regimens differed in the two era-based studies Table 3; Supplementary Table S5. Table 3Procedural co-interventions in the ELCA and comparator groupsStudyAspiration thrombectomyDistal protectionIntracoronary vasodilatorGPIIb/IIIa inhibitorIntracoronary/pre-procedural fibrinolyticMechanical circulatory supportLoading antiplateletStent/deviceStratum A — ELCA versus manual aspiration thrombectomy Shishikura 20130/50 (0%) vs 48/48 (100%)0/50 (0%) vs 0/48 (0%)NR0% vs 0% (not given in either arm)0% vs 0% (not given)IABP 11/50 (22.0%) vs 4/48 (8.3%); PCPS 0/50 vs 1/48 (2.1%)Clopidogrel 300 mg or aspirin 243 mg: 100% vs 100%BMS 31/46 (67.4%) vs 44/45 (97.8%); DES 15/46 (32.6%) vs 1/45 (2.2%) Shibata 201913/32 (40.6%) vs 34/40 (85.0%)2/32 (6.3%) vs 2/40 (5.0%)NRNRNRIABP 5/32 (15.6%) vs 17/40 (42.5%); PCPS excluded by designPrasugrel 32/32 (100%) vs 16/40 (40.0%)NR Arai 20230/48 (0%) vs 50/50 (100%)0/48 (0%) vs 0/50 (0%) (exclusion criterion)NRNR0% vs 0% (exclusion criterion; 15 t-PA patients excluded)NRAspirin 200 mg + clopidogrel 300 mg or prasugrel 20 mg: 100% vs 100%DES 44/48 (91.7%) vs 39/50 (78.0%); DCB 2 vs 1; POBA 2 vs 6 Shimojo 20230/63 (0%) vs 80/80 (100%)0/63 (0%) vs 2/80 (2.5%)NRNRNRNR100% vs 100%DES 58/63 (92.1%) vs 72/80 (90.0%)Stratum B — ELCA versus PCI alone Ilkay 2005NRNRNRNRNRNRNRStent in all (type NR) Dörr 20070% vs 0% (not used)0/14 (0%) vs 0/13 (0%)NRTirofiban 14/14 (100%) vs 13/13 (100%)0% vs 0%NR (cardiogenic shock was an exclusion criterion)Aspirin 500 mg + clopidogrel 300 mg + heparin 70 U/kg: 100% vs 100%BMS 14/14 (100%) vs 13/13 (100%) Masuda 2022Incl. in conventional Rx (both arms); n NRNRNRNRNRNRNRDES era; breakdown NR Kujiraoka 202350/167 (29.9%) vs 103/123 (83.7%)Named as part of "conventional methods" but not quantifiedNitroprusside 54/167 (32.3%) vs 41/123 (33.3%),NRPre-thrombolysis 20/167 (11.2%) vs 19/123 (15.4%)NRPrasugrel 138/167 (82.1%) vs 0/123 (0%); clopidogrel 28/167 (16.7%) vs 122/123 (99.2%)BMS or DES (mix; NR). Direct stenting 123/167 (73.6%) vs 36/123 (29.3%) Matsuda 2025 (STELA)208/427 (48.7%) vs 1195/2166 (85.2%)NRNRNRNRIABP before PCI 20/427 (4.7%) vs 106/2166 (4.9%); ECMO 7/427 (1.6%) vs 31/2166 (1.4%)NRDES era; breakdown NR Iiya 202513/180 (7.2%) vs 15/180 (8.3%)2/180 (1.1%) vs 2/180 (1.1%)NRNRNRNRNRDES 163/180 (90.6%); DCB 23/180 (12.8%) vs DES 155/180 (86.1%); DCB 29/180 (16.1%)Values are ELCA arm versus comparator arm. NR, not reported. BMS, bare-metal stent; DCB, drug-coated balloon; DES, drug-eluting stent; ELCA, excimer laser coronary angioplasty/atherectomy; NR, not reported; PCI, percutaneous coronary intervention; POBA, plain balloon angioplasty Procedural co-interventions in the ELCA and comparator groups Values are ELCA arm versus comparator arm. NR, not reported. BMS, bare-metal stent; DCB, drug-coated balloon; DES, drug-eluting stent; ELCA, excimer laser coronary angioplasty/atherectomy; NR, not reported; PCI, percutaneous coronary intervention; POBA, plain balloon angioplasty Final TIMI grade 3 flow was comparable or modestly higher with ELCA, and MBG 3 favored ELCA wherever it was reported. In STELA’s propensity-matched cohort, however, final TIMI-3 flow (84.2% vs 89.4%) and MBG 3 (73.4% vs 76.8%) were numerically lower with ELCA, and corrected TIMI frame count and peak CK-MB were significantly worse (26.0 vs 24.0, P = 0.012; 241 vs 193 mg/dL, P = 0.011). The slow-flow endpoint varied across studies: Kujiraoka 2023 included transient intraprocedural flow limitation reversed by nitroprusside or bailout, whereas the other studies recorded only final angiogram events, which explains the much higher rates in that study. Procedurally, ELCA appeared safe, and laser-specific complications were not elevated Table 4. Table 4In-hospital/short-term outcomes and angiographic outcomes in the ELCA and comparator groupsStudyFinal TIMI-3 flow, n/N (%)Myocardial blush grade 3, n/N (%)Slow-flow/no-reflow, n/N (%)In-hospital/short-term mortality, n/N (%)In-hospital MACE/MACCE, n/N (%)ELCACompELCACompELCACompELCACompELCACompStratum A — ELCA versus manual aspiration thrombectomy Shishikura 201343/50 (86.0%)33/48 (68.8%)38/50 (76.0%)26/48 (54.2%)1/50 (2.0%)3/48 (6.2%)0/50 (0.0%)3/48 (6.2%)0/50 (0.0%)5/48 (10.4%) Shibata 201924/32 (75.0%)26/40 (65.0%)NRNR11/32 (34.4%)9/40 (22.5%)0/32 (0.0%)1/40 (2.5%)0/32 (0.0%)2/40 (5.0%) Arai 202340/48 (83.3%)41/50 (82.0%)30/48 (62.5%)13/50 (26.0%)1/48 (2.1%)4/50 (8.0%)3/48 (6.2%)8/50 (16.0%)3/48 (6.2%)10/50 (20.0%) Shimojo 202358/63 (92.1%)73/80 (91.2%)NRNR15/63 (23.8%)25/80 (31.2%)NRNR0/63 (0.0%)2/80 (2.5%)Stratum B — ELCA versus PCI alone Ilkay 200533/36 (91.7%)40/44 (90.9%)NRNRNRNR0/36 (0.0%)0/44 (0.0%)0/36 (0.0%)0/44 (0.0%) Dörr 2007 (LaserAMI)NRNR7/14 (50.0%)8/13 (61.5%)0/14 (0.0%)0/13 (0.0%)NRNRNRNR Masuda 202270/71 (98.6%)67/71 (94.4%)NRNRNRNR2/71 (2.8%)10/71 (14.1%)NRNR Kujiraoka 2023147/167 (88.0%)107/123 (87.0%)133/167 (79.6%)81/123 (65.9%)52/167 (31.1%)52/123 (42.3%)NRNRNRNR Matsuda 2025 (STELA)358/427 (83.8%)1811/2166 (83.6%)311/427 (72.8%)1522/2166 (70.3%)69/427 (16.2%)355/2166 (16.4%)18/427 (4.2%)169/2166 (7.8%)54/427 (12.6%)301/2166 (13.9%) Iiya 2025NRNRNRNR37/180 (20.6%)43/180 (23.9%)2/180 (1.1%)6/180 (3.3%)3/180 (1.7%)20/180 (11.1%)Event rates are n/N (%) per arm, calculated from the reported numerator and denominator. Shibata 2019 is grouped with the aspiration comparator because its comparator arm incorporated manual aspirationStudy-specific caveats. Dörr 2007: TIMI/MBG reported as group means; mortality and MACE not reported. Ilkay 2005: no perforation, death, stroke, CABG, reintervention or infarction in either arm (entered as 0). Masuda 2022: in-hospital 30-day mortality, matched cohort. Kujiraoka 2023: slow-flow was intra-procedural, summed across onset-to-balloon subgroups. Matsuda 2025: MBG-3 derived by subtraction; cardiovascular death used for mortality; MACE is target-vessel MACE over a median of 815 days. Iiya 2025: maximum CK and MACE are pre-matching; cardiac death used for mortalityCABG, coronary artery bypass grafting; CK, creatine kinase; ELCA, excimer laser coronary angioplasty/atherectomy; MACCE, major adverse cardiac and cerebrovascular events; MACE, major adverse cardiac events; MBG, myocardial blush grade; NR, not reported; PCI, percutaneous coronary intervention; TIMI, Thrombolysis In Myocardial Infarcti In-hospital/short-term outcomes and angiographic outcomes in the ELCA and comparator groups Event rates are n/N (%) per arm, calculated from the reported numerator and denominator. Shibata 2019 is grouped with the aspiration comparator because its comparator arm incorporated manual aspiration Study-specific caveats. Dörr 2007: TIMI/MBG reported as group means; mortality and MACE not reported. Ilkay 2005: no perforation, death, stroke, CABG, reintervention or infarction in either arm (entered as 0). Masuda 2022: in-hospital 30-day mortality, matched cohort. Kujiraoka 2023: slow-flow was intra-procedural, summed across onset-to-balloon subgroups. Matsuda 2025: MBG-3 derived by subtraction; cardiovascular death used for mortality; MACE is target-vessel MACE over a median of 815 days. Iiya 2025: maximum CK and MACE are pre-matching; cardiac death used for mortality CABG, coronary artery bypass grafting; CK, creatine kinase; ELCA, excimer laser coronary angioplasty/atherectomy; MACCE, major adverse cardiac and cerebrovascular events; MACE, major adverse cardiac events; MBG, myocardial blush grade; NR, not reported; PCI, percutaneous coronary intervention; TIMI, Thrombolysis In Myocardial Infarcti Risk of bias within studies The single RCT was rated as having some concerns: it lacked allocation concealment, was open-label, had a small sample with baseline imbalances, no pre-registered protocol, and was partly funded by the device manufacturer. On ROBINS-I, the nine non-randomized studies were rated moderate (three) or serious (six) risk of bias, primarily due to confounding: studies with robust adjustment received a moderate rating, whereas those using minimal matching or historical comparators were rated serious. Detection bias was moderate for subjective outcomes assessed by unblinded investigators, and the absence of pre-registered protocols increased reporting bias risk (Fig. 2). Risk of bias assessment of non-randomized studies using Risk of Bias in Non-randomized Studies of Interventions Certainty of evidence (GRADE assessment) Certainty was very low for all five outcomes in both strata Tables 5 and 6, driven by the observational design, small samples with wide CI, and few contributing studies. In the aspiration stratum the apparent mortality and MACCE differences rest on very few events and cannot be separated from confounding, because every contributing study is at serious risk of bias. Table 5GRADE evidence profile Stratum A: ELCA versus manual aspiration thrombectomyOutcome№ of participants (studies)Relative effect (95% CI)Anticipated absolute effects: risk with aspiration → risk with ELCA (95% CI)Certainty (GRADE)Reasons for downgradingFinal TIMI-3 flow411 (4 non-randomized studies)OR 1.58 (0.67 to 3.75)794 per 1,000 → 859 per 1,000 (720 to 935)⨁◯◯◯ Very lowRisk of bias (− 2)a; imprecision (− 1)bMyocardial blush grade 3196 (2 non-randomized studies)OR 3.57 (0.07 to 185.10)398 per 1,000 → 702 per 1,000 (44 to 992)⨁◯◯◯ Very lowRisk of bias (− 2)a; imprecision (− 2)cSlow-flow/no-reflow411 (4 non-randomized studies)OR 0.78 (0.22 to 2.78)188 per 1,000 → 153 per 1,000 (48 to 392)⨁◯◯◯ Very lowRisk of bias (− 2)a; inconsistency (− 1)d; imprecision (− 1)bIn-hospital/≤ 30-day mortality268 (3 non-randomized studies)OR 0.31 (0.02 to 4.04)87 per 1,000 → 29 per 1,000 (2 to 278)⨁◯◯◯ Very lowRisk of bias (− 2)a; imprecision (− 2)eIn-hospital MACCE411 (4 non-randomized studies)OR 0.22 (0.04 to 1.26)87 per 1,000 → 21 per 1,000 (4 to 107)⨁◯◯◯ Very lowRisk of bias (− 2)a; imprecision (− 2)eRandom effects, REML between-study variance, Hartung–Knapp confidence intervals with the ad hoc variance correction. No fixed-effect estimate is reported. Absolute effects are derived from the comparator-arm risk and the pooled odds ratioaRisk of bias, − 2. All four studies are serious on ROBINS-I, in every case for confounding: matching on age and sex only (Shishikura 2013), no adjustment (Arai 2023, Shimojo 2023) and era-based historical controls with an antiplatelet imbalance (Shibata 2019)bImprecision, − 1. The 95% CI crosses the null and includes appreciable benefit and harmcImprecision, − 2. Two studies and 196 patients contribute; the interval spans a 14-fold reduction to a 185-fold increasedInconsistency, − 1. I2 = 31%, τ2 = 0.200eImprecision, − 2. Three ELCA events in total, two studies with zero ELCA events, and 62–71% of the weight from Arai 2023Indirectness was not downgraded. Publication bias could not be assessed formally (fewer than ten studies per outcome)CI, confidence interval; MACCE, major adverse cardiovascular and cerebrovascular events; OR, odds ratio; TIMI, Thrombolysis In Myocardial InfarctionGRADE certainty: high, very confident the true effect is close to the estimate; moderate, likely close but possibly substantially different; low, confidence limited; very low, little confidence in the estimate GRADE evidence profile Stratum A: ELCA versus manual aspiration thrombectomy Random effects, REML between-study variance, Hartung–Knapp confidence intervals with the ad hoc variance correction. No fixed-effect estimate is reported. Absolute effects are derived from the comparator-arm risk and the pooled odds ratio aRisk of bias, − 2. All four studies are serious on ROBINS-I, in every case for confounding: matching on age and sex only (Shishikura 2013), no adjustment (Arai 2023, Shimojo 2023) and era-based historical controls with an antiplatelet imbalance (Shibata 2019) bImprecision, − 1. The 95% CI crosses the null and includes appreciable benefit and harm cImprecision, − 2. Two studies and 196 patients contribute; the interval spans a 14-fold reduction to a 185-fold increase dInconsistency, − 1. I2 = 31%, τ2 = 0.200 eImprecision, − 2. Three ELCA events in total, two studies with zero ELCA events, and 62–71% of the weight from Arai 2023 Indirectness was not downgraded. Publication bias could not be assessed formally (fewer than ten studies per outcome) CI, confidence interval; MACCE, major adverse cardiovascular and cerebrovascular events; OR, odds ratio; TIMI, Thrombolysis In Myocardial Infarction GRADE certainty: high, very confident the true effect is close to the estimate; moderate, likely close but possibly substantially different; low, confidence limited; very low, little confidence in the estimate Table 6GRADE evidence profile Stratum B: ELCA versus PCI alone (propensity-matched data)Outcome№ of participants (studies)Relative effect (95% CI)Anticipated absolute effects: risk with PCI alone → risk with ELCA (95% CI)Certainty (GRADE)Reasons for downgradingFinal TIMI-3 flow1,247 (4 non-randomized studies)OR 0.88 (0.38 to 2.03)896 per 1,000 → 883 per 1,000 (766 to 946)⨁◯◯◯ Very lowRisk of bias (− 2)a,b; imprecision (− 1)cMyocardial blush grade 31,044 (3 non-randomized studies)OR 1.06 (0.13 to 8.43)743 per 1,000 → 754 per 1,000 (273 to 961)⨁◯◯◯ Very lowRisk of bias (− 2)a,b; inconsistency (− 2)d; imprecision (− 1)cSlow-flow/no-reflow1,386 (3 estimable studies; Dörr 2007 listed, non-estimable)OR 0.93 (0.29 to 3.02)201 per 1,000 → 190 per 1,000 (68 to 432)⨁◯◯◯ Very lowRisk of bias (− 2)a,b; inconsistency (− 1)e; imprecision (− 1)cShort-term mortality1,238 (3 non-randomized studies)OR 0.44 (0.05 to 3.80)53 per 1,000 → 24 per 1,000 (3 to 176)⨁◯◯◯ Very lowRisk of bias (− 2)a,b; imprecision (− 2)fComposite endpoints (MACE/MACCE)Not pooled — 368 matched pairs (STELA) and 180 matched pairs (Iiya 2025), reported separatelySTELA OR 1.03 (0.66 to 1.59); Iiya 2025 OR 0.14 (0.04 to 0.46)STELA 122 → 125 per 1,000; Iiya 2025 111 → 17 per 1,000⨁◯◯◯ Very lowNot pooled: incompatible ascertainment windows and component setsgModel as in Table 5, applied to each study's propensity-matched dataset: Masuda 2022 (71 pairs), Matsuda 2025/STELA (368 pairs), Iiya 2025 (180 pairs)aRisk of bias, − 2. Three studies are moderate on ROBINS-I (Masuda 2022, Matsuda 2025/STELA, Iiya 2025) and two serious (Ilkay 2005, Kujiraoka 2023). Residual confounding persists within the matched cohorts: in STELA, TIMI thrombus grade 4/5 independently predicted ELCA use (OR 1.48, 95% CI 1.05–2.08)bCohort overlap, judged under risk of bias. STELA's sites include the institutions of Kujiraoka 2023, Masuda 2022 and Iiya 2025, with overlapping enrolment windows, so patients may be counted more than once. Excluding those studies did not change any conclusioncImprecision, − 1. The 95% CI crosses the null and encompasses appreciable benefit and harmdInconsistency, − 2. I2 = 80%, τ2 = 0.391, Q P = 0.009, with direction reversal between Kujiraoka 2023 (OR 2.03, 95% CI 1.19–3.45) and matched STELA (OR 0.83, 95% CI 0.60–1.17)eInconsistency, − 1. I2 = 74%, τ2 = 0.167, Q P = 0.018. Restricting to final-angiogram definitions does not resolve it (I2 = 71%). Within matched STELA, slow-flow was more frequent with ELCA (OR 1.53, 95% CI 1.00–2.36)fImprecision, − 2. Fifty-one deaths in total and an interval crossing the null; within matched STELA alone the estimate is close to null (OR 0.82, 95% CI 0.40–1.68)gNot pooled: STELA reports target-vessel MACE over a median of 815 days, Iiya 2025 target-lesion MACE to approximately two years, and Ilkay 2005 an in-hospital composite with zero events. Pooling gave I2 = 89%Indirectness was not downgraded. Publication bias could not be assessed formallyCI, confidence interval; MACE, major adverse cardiovascular events; MACCE, major adverse cardiovascular and cerebrovascular events; OR, odds ratio; PCI, percutaneous coronary intervention; TIMI, Thrombolysis In Myocardial InfarctionGRADE certainty as in Table 5 GRADE evidence profile Stratum B: ELCA versus PCI alone (propensity-matched data) Model as in Table 5, applied to each study's propensity-matched dataset: Masuda 2022 (71 pairs), Matsuda 2025/STELA (368 pairs), Iiya 2025 (180 pairs) aRisk of bias, − 2. Three studies are moderate on ROBINS-I (Masuda 2022, Matsuda 2025/STELA, Iiya 2025) and two serious (Ilkay 2005, Kujiraoka 2023). Residual confounding persists within the matched cohorts: in STELA, TIMI thrombus grade 4/5 independently predicted ELCA use (OR 1.48, 95% CI 1.05–2.08) bCohort overlap, judged under risk of bias. STELA's sites include the institutions of Kujiraoka 2023, Masuda 2022 and Iiya 2025, with overlapping enrolment windows, so patients may be counted more than once. Excluding those studies did not change any conclusion cImprecision, − 1. The 95% CI crosses the null and encompasses appreciable benefit and harm dInconsistency, − 2. I2 = 80%, τ2 = 0.391, Q P = 0.009, with direction reversal between Kujiraoka 2023 (OR 2.03, 95% CI 1.19–3.45) and matched STELA (OR 0.83, 95% CI 0.60–1.17) eInconsistency, − 1. I2 = 74%, τ2 = 0.167, Q P = 0.018. Restricting to final-angiogram definitions does not resolve it (I2 = 71%). Within matched STELA, slow-flow was more frequent with ELCA (OR 1.53, 95% CI 1.00–2.36) fImprecision, − 2. Fifty-one deaths in total and an interval crossing the null; within matched STELA alone the estimate is close to null (OR 0.82, 95% CI 0.40–1.68) gNot pooled: STELA reports target-vessel MACE over a median of 815 days, Iiya 2025 target-lesion MACE to approximately two years, and Ilkay 2005 an in-hospital composite with zero events. Pooling gave I2 = 89% Indirectness was not downgraded. Publication bias could not be assessed formally CI, confidence interval; MACE, major adverse cardiovascular events; MACCE, major adverse cardiovascular and cerebrovascular events; OR, odds ratio; PCI, percutaneous coronary intervention; TIMI, Thrombolysis In Myocardial Infarction GRADE certainty as in Table 5 Pooled results Pooled effect estimates across all pre-specified outcomes are summarized in Table 7 and illustrated as forest plots in Figs. 3, 4 and 5. Table 7Summary of pooled effect estimates by comparator stratumOutcomeStudies (n)ELCA events/total (%)Comparator events/total (%)Pooled OR (95% CI)PI2 (%)τ2Cochran’s Q PWeight of largest studyStratum A: ELCA versus manual aspiration thrombectomyFinal TIMI-3 flow4165/193 (85.5)173/218 (79.4)1.58 (0.67–3.75)0.18900.0000.566 Shishikura 2013, 28.0%Myocardial blush grade 3268/98 (69.4)39/98 (39.8)3.57 (0.07–185.10)0.15300.0000.358 Arai 2023, 50.1%Slow-flow/no-reflow428/193 (14.5)41/218 (18.8)0.78 (0.22–2.78)0.579310.2000.235 Shimojo 2023, 46.1%In-hospital/≤ 30-day mortality33/130 (2.3)12/138 (8.7)0.31 (0.02–4.04)0.18700.0000.824 Arai 2023, 71.3%In-hospital MACCE43/193 (1.6)19/218 (8.7)0.22 (0.04–1.26)0.07000.0000.905 Arai 2023, 62.1%Stratum B: ELCA versus PCI alone (propensity-matched data)Final TIMI-3 flow4560/642 (87.2)542/605 (89.6)0.88 (0.38–2.03)0.662290.0810.246 Matsuda 2025, 52.7%Myocardial blush grade 33410/546 (75.1)370/498 (74.3)1.06 (0.13–8.43)0.915800.3910.009 Matsuda 2025, 47.0%Slow-flow/no-reflow4 listed, 3 estimable147/715 (20.6)135/671 (20.1)0.93 (0.29–3.02)0.817740.1670.018 Matsuda 2025, 34.8%Short-term mortality318/619 (2.9)33/619 (5.3)0.44 (0.05–3.80)0.246460.3550.167 Matsuda 2025, 50.7%Composite endpoints2 not pooledSTELA 46/368 (12.5); Iiya 2025 3/180 (1.7)STELA 45/368 (12.2); Iiya 2025 20/180 (11.1)STELA 1.03 (0.66–1.59); Iiya 2025 0.14 (0.04–0.46)N\A89N\AN\AN\ARandom-effects model throughout: REML between-study variance, Hartung–Knapp confidence intervals with the ad hoc variance correction, so no interval is narrower than the conventional random-effects interval. No fixed-effect estimate is reported (2.6.1). I2 and τ2 are taken from the same random-effects fit; Cochran’s Q is the homogeneity test statistic. All heterogeneity statistics refer to the data shown in this tableStratum B uses each study’s propensity-matched dataset: Masuda 2022 (71 pairs), Matsuda 2025/STELA (368 pairs), Iiya 2025 (180 pairs); the unadjusted total-cohort analysis is in Supplementary Table S3. A 0.5 continuity correction was applied to all four cells of any study with a zero cell. Dörr 2007 recorded no slow-flow/no-reflow events in either arm (0/14 vs 0/13) and is non-estimableNo outcome in either stratum is statistically significant. Leave-one-out analyses are given in Fig. S1CI, confidence interval; ELCA, excimer laser coronary angioplasty; MACE, major adverse cardiovascular events; MACCE, major adverse cardiovascular and cerebrovascular events; OR, odds ratio; PCI, percutaneous coronary intervention; TIMI, Thrombolysis in Myocardial Infarction Summary of pooled effect estimates by comparator stratum Random-effects model throughout: REML between-study variance, Hartung–Knapp confidence intervals with the ad hoc variance correction, so no interval is narrower than the conventional random-effects interval. No fixed-effect estimate is reported (2.6.1). I2 and τ2 are taken from the same random-effects fit; Cochran’s Q is the homogeneity test statistic. All heterogeneity statistics refer to the data shown in this table Stratum B uses each study’s propensity-matched dataset: Masuda 2022 (71 pairs), Matsuda 2025/STELA (368 pairs), Iiya 2025 (180 pairs); the unadjusted total-cohort analysis is in Supplementary Table S3. A 0.5 continuity correction was applied to all four cells of any study with a zero cell. Dörr 2007 recorded no slow-flow/no-reflow events in either arm (0/14 vs 0/13) and is non-estimable No outcome in either stratum is statistically significant. Leave-one-out analyses are given in Fig. S1 CI, confidence interval; ELCA, excimer laser coronary angioplasty; MACE, major adverse cardiovascular events; MACCE, major adverse cardiovascular and cerebrovascular events; OR, odds ratio; PCI, percutaneous coronary intervention; TIMI, Thrombolysis in Myocardial Infarction Forest plots of final Thrombolysis in Myocardial Infarction grade 3 flow (a), myocardial blush grade 3 (b), slow-flow/no-reflow (c), short-term mortality (d) and short-term major adverse cardiac and cerebrovascular events (e) in excimer laser coronary angioplasty versus percutaneous coronary intervention with mechanical aspiration thrombectomy (Stratum A). All estimates are from a random-effects model with restricted maximum likelihood between-study variance and Hartung–Knapp confidence intervals Generalized linear mixed model analyses. Panels (a) short-term mortality and (b) major adverse cardiac and cerebrovascular events are Stratum A, excimer laser coronary angioplasty versus percutaneous coronary intervention with mechanical aspiration thrombectomy. Panel (c) slow-flow/no-reflow is Stratum B, excimer laser coronary angioplasty versus percutaneous coronary intervention alone Forest plots of final Thrombolysis in Myocardial Infarction grade 3 flow (a), myocardial blush grade 3 (b), slow-flow/no-reflow (c) and short-term mortality (d) in excimer laser coronary angioplasty versus percutaneous coronary intervention alone (Stratum B), using each study’s propensity-matched dataset. Panel (e) shows the composite endpoint, which was not pooled because the two contributing studies ascertained events over incompatible follow-up horizons; the individual study estimates are displayed without a summary diamond. All pooled estimates are from a random-effects model with restricted maximum likelihood between-study variance and Hartung–Knapp confidence intervals Stratum A: ELCA vs. mechanical aspiration thrombectomy Final TIMI 3 flow Four studies reported final TIMI 3 flow, which was numerically more frequent with ELCA (165/193 [85.5%] vs. 173/218 [79.4%]; OR 1.58, 95% CI 0.67–3.75; P = 0.19; I2 = 0%), though the difference was not significant Fig. 3A. MBG 3 Two studies reported MBG 3, which was numerically higher with ELCA (68/98 [69.4%] vs. 39/98 [39.8%]; OR 3.57, 95% CI 0.07–185.10; P = 0.15; I2 = 0%). With only two contributing studies, the Hartung–Knapp interval is very wide, and the difference did not reach statistical significance Fig. 3B. Slow-flow/No-reflow With four contributing studies, slow-flow or no-reflow was lower with ELCA (28/193 [14.5%] vs. 41/218 [18.8%]; OR 0.78, 95% CI 0.22–2.78; I2 = 31%; τ2 = 0.20) Fig. 3C. Short-term mortality Three studies reported mortality, which was lower with ELCA (3/130 [2.3%] vs. 12/138 [8.7%]; OR 0.31, 95% CI 0.02–4.04; P = 0.19; I2 = 0%) [GLMM: OR 0.26, 95% CI 0.07–0.92]. The estimate did not reach significance under small-sample inference, and the events are sparse; the GLMM was directionally concordant but nominally significant, a discrepancy reflecting the sparse-data setting rather than a robust effect Fig. 3D and 4A. Short-term MACCE Four studies reported MACCE, which was lower with ELCA (3/193 [1.6%] vs. 19/218 [8.7%]; OR 0.22, 95% CI 0.04–1.26; P = 0.07; I2 = 0%) [GLMM: OR 0.17, 95% CI 0.05–0.56]. As for mortality, the difference was not significant once small-sample inference was applied Figs. 3E and 4B. Stratum B: ELCA vs. PCI alone Final TIMI 3 Flow Four studies contributed after propensity matching. Final TIMI 3 flow was achieved in 560/642 (87.2%) versus 542/605 (89.6%) (OR 0.88, 95% CI 0.38–2.03; I2 = 29%; τ2 = 0.081). The analysis was dominated by Matsuda 2025/STELA, whose matched cohort showed a lower rate with ELCA (310/368 vs. 328/367; OR 0.64, 95% CI 0.41–0.98) Fig. 5A. MBG 3 Three studies reported MBG 3, with no significant difference (410/546 [75.1%] vs. 370/498 [74.3%]; OR 1.06, 95% CI 0.13–8.43; P = 0.92). Heterogeneity was considerable (I2 = 80%; τ2 = 0.391; Q P = 0.009), reflecting opposite directions of effect in Kujiraoka 2023 and the STELA matched cohort Fig. 5B. Slow-flow/No-reflow Four studies contributed; Dörr 2007 was not estimable (zero events in both arms). Across the three estimable studies the event rates were 147/715 (20.6%) versus 135/671 (20.1%) (OR 0.93, 95% CI 0.29–3.02) [GLMM: OR 0.97, 95% CI 0.69–1.36]. Heterogeneity was substantial (I2 = 74%; τ2 = 0.167; Q P = 0.018), because Kujiraoka 2023 favored ELCA while the STELA matched cohort favored PCI alone Figs. 5C and 4C. Short-term mortality Three studies reported mortality, which was numerically lower with ELCA (18/619 [2.9%] vs. 33/619 [5.3%]; OR 0.44, 95% CI 0.05–3.80; I2 = 46%). In the unadjusted cohorts the counts were 22/678 (3.2%) versus 185/2,417 (7.7%) (OR 0.43, 95% CI 0.12–1.53), so propensity adjustment does not change the direction of effect, and neither analysis is precise enough to establish a mortality benefit Supplementary Table S3. Figure 5D. Short-term MACE Two studies reported a composite endpoint over incompatible ascertainment horizons (in-hospital MACCE in STELA versus approximately two-year MACE in Iiya 2025), with considerable heterogeneity (I2 = 89%), and were therefore not pooled: 46/368 (12.5%) versus 45/368 (12.2%) in the STELA matched cohort (OR 1.03, 95% CI 0.66–1.59) and 3/180 (1.7%) versus 20/180 (11.1%) in Iiya 2025 (OR 0.14, 95% CI 0.04–0.46). The discordance reflects the difference in follow-up horizon and endpoint definition as much as any device effect, and the two estimates should be considered independently Fig. 5E. Leave-one-out influence analysis Because the STELA registry contributed most of the weight to every Stratum B outcome, the leave-one-out analysis is reported in full for that stratum Fig. S1. Omitting STELA moved every estimate away from the null and eliminated the between-study heterogeneity: TIMI 3 flow OR 1.22 (95% CI 0.32–4.71; I2 = 0%), MBG 3 OR 0.92, slow-flow/no-reflow OR 0.71 (0.07–6.76; I2 = 0%), and mortality OR 0.24. No estimate became significant, and no omission reversed a conclusion. The registry therefore attenuates the pooled effect and is the source of the observed inconsistency rather than the driver of any apparent benefit. In Stratum A, leave-one-out omission produced no material change. Excluding the two Stratum B studies at serious risk of bias, or the era-based study, moved the angiographic estimates towards PCI alone without producing a significant result Supplementary Table S3. Discussion This systematic review and meta-analysis of ten comparative studies and over 3,500 patients with ACS finds that adjunctive ELCA-containing strategies show no statistically significant advantage over either comparator once small-sample inference and adjusted estimates are used. The largest point estimate, MBG 3 against manual aspiration thrombectomy, remains directionally favorable but rests on two studies with a very wide confidence interval. All pooled outcomes are of very low certainty. The dissociation between equivalent epicardial patency and tissue-level perfusion aligns with the proposed mechanism of ELCA: unlike aspiration thrombectomy, it vaporizes thrombus in situ, generating inert debris too small to obstruct the microcirculation [10, 12]. In vitro, ELCA also reduced platelet reactivity (the “stunned platelet” phenomenon), an experimental observation of unproven clinical relevance [31]. In this analysis, short-term mortality did not favor ELCA. Against PCI alone, the earlier association rested on unadjusted counts from a single registry contributing more than 80% of the weight, which itself reported a propensity-matched analysis showing no association (adjusted HR 0.951, 95% CI 0.529–1.712); the matched estimates yield an OR of 0.44 (0.05–3.80). Against aspiration thrombectomy, mortality rested on only 15 events. Biological plausibility is not absent, since each 1% increase in microvascular obstruction raises mortality by 14% before infarct-size adjustment (HR 1.14, 95% CI 1.09–1.19) [5], but CMR data indicating that microvascular obstruction without intramyocardial hemorrhage may not independently predict mortality weaken the causal chain from improved blush to survival [32]. As noted, routine aspiration thrombectomy is no longer recommended (Class 3, No Benefit) due to lack of mortality benefit and a potential stroke risk [6, 7, 9]. Contemporary guidance for high thrombus burden and no-reflow favors selective bailout rather than routine mechanical thrombus removal, and this is the context in which adjunctive ELCA is positioned [33]. ELCA, with its different photoablation mechanism, may represent an alternative, but the guidelines do not mention it, reflecting the paucity of randomized evidence. The CARMEL trial (n = 151) reported high device and angiographic success with greater benefit in thrombus-rich lesions [34], and the Cardiovascular Laser Society endorsed ELCA for thrombotic lesions [11]; that endorsement predates the largest registry analysis, which found no reduction in adverse cardiac events. No stroke signal has been reported, likely because of in-situ vaporization or possibly because of small samples and underreporting. The MBG-3 result against aspiration thrombectomy (OR 3.57, 95% CI 0.07–185.10) is the most mechanistically coherent finding. The 30 percentage-point absolute difference would be clinically meaningful: MBG 3 is associated with 3% mortality versus 23% for MBG 0/1, and MBG 2/3 independently predicts lower 3-year mortality (adjusted HR 0.57, 95% CI 0.39–0.82) [35, 36]. However, this finding rests on 196 patients from two studies, both at serious risk of bias, neither of which enrolled STEMI exclusively, and both with unblinded endpoint ascertainment; the interval spans more than three orders of magnitude. Estimates should therefore be read in light of their GRADE ratings and their behavior under sensitivity analyses Supplementary Table S3, not point estimates alone. Against PCI alone, MBG 3 showed no difference (OR 1.06, 0.13–8.43). Final TIMI-3 flow did not differ in either stratum (Stratum A: OR 1.58, 95% CI 0.67–3.75; Stratum B: OR 0.88, 95% CI 0.38–2.03), with near-ceiling rates (> 80% in both arms) limiting discriminative capacity; TIMI-3 flow is necessary but insufficient for myocardial reperfusion [37]. Slow-flow/no-reflow was non-significant in both strata, and its inconsistency across studies likely reflects differing definitions and era-related antiplatelet imbalances rather than a device effect. Clinical heterogeneity Beyond comparator stratification, the evidence base is heterogeneous in patient selection, thrombus burden, procedural protocol, follow-up (index hospitalization to approximately two years), and endpoint definition (in-hospital MACCE to multi-year target-vessel MACE, with two incompatible slow-flow definitions). These sources do not all bias in the same direction: enrichment of the ELCA arms for high thrombus burden would bias against ELCA on angiographic endpoints, whereas preferential selection of younger and more stable patients would bias toward apparent mortality benefit. A STEMI-restricted analysis could not be meaningfully performed in the aspiration stratum, so those results apply to predominantly STEMI ACS populations, not to STEMI specifically. Strengths and limitations This review has several strengths: PRISMA 2020/Cochrane methodology with prospective PROSPERO registration; pre-specified comparator stratification; dual risk-of-bias and GRADE assessment; independent reproduction of every pooled estimate from primary event counts; use of adjusted in preference to crude estimates; and a pre-specified sensitivity-analysis programme comprising leave-one-out, restricted and non-overlapping-cohort analyses, and a GLMM sparse-event analysis. Several limitations constrain interpretation. The evidence is overwhelmingly observational (9/10 studies), with the only included RCT underpowered and part-funded by the device manufacturer [27]. ELCA was preferentially used in lesions with higher thrombus burden Table 2, and unmeasured confounders such as operator expertise and patient frailty may still influence results [38]. All four Stratum A studies are at serious risk of bias, so no risk-of-bias-restricted sensitivity analysis was possible in that stratum and its estimates cannot be separated from confounding by indication. Because ELCA was almost always part of a broader procedural strategy, pooled estimates compare ELCA-containing strategies versus comparators, not isolated laser photoablation. Co-interventions were inconsistently reported and unevenly distributed Table 3, a primary threat to validity. Two era-based studies had marked antiplatelet imbalances (prasugrel loading 82.1% vs 0% in Kujiraoka 2023; 100% vs 40% in Shibata 2019) alongside concurrent changes in stent technology, conflating device, and era effects, which the authors of both studies acknowledge likely favored the ELCA arm [26, 28]. Four studies came from institutions that also contributed to the multicenter STELA registry (Kujiraoka 2023, Masuda 2022, Arai 2023, and Iiya 2025), so some patients may have been counted more than once, violating the independence assumption; excluding those single-center studies did not change any Stratum B conclusion Supplementary Table S3. Endpoint ascertainment was unblinded and non-core-laboratory in eight of ten studies, introducing detection bias for subjective angiographic endpoints [39], and geographic concentration limits generalizability. Few studies per outcome precluded funnel-plot analysis; the expected direction of publication bias would inflate the benefit. ELCA also has a steep learning curve, with complication rates falling with operator experience [40, 41], and operator volume was unreported. Lack of individual patient data prevented subgroup analyses by catheter size, fluence, thrombus burden grade or onset-to-balloon time [42]. Clinical implications Current evidence does not support the routine use of ELCA during primary PCI. Although ELCA is used mainly as bailout after conventional measures have failed in large-thrombus-burden lesions, any mortality benefit remains unproven: in the largest registry, the association disappeared after propensity matching, and the same matched cohort showed worse corrected TIMI frame count and higher peak CK-MB with ELCA. Rather than adoption, these data support a randomized trial in high-thrombus-burden lesions at experienced, high-volume centers. Future directions Future multicenter RCTs should include blinded core-lab adjudication, CMR quantification of MVO and infarct size, standardized MACE definitions with ≥ 12-month follow-up, stratification by TIMI thrombus grade and ELCA catheter size, pre-specified recording of all co-interventions, and geographically diverse enrolment [32, 39, 40]. The forthcoming trial UMIN000032954 may provide the first adequately controlled data. Conclusion This systematic review and meta-analysis found that adjunctive ELCA-containing strategies were not associated with improved epicardial flow, myocardial blush, or short-term mortality once small-sample inference and adjusted estimates from the dominant contributing registry were used, and all pooled outcomes were graded as very low certainty. Given the scarcity of randomized data, the predominance of observational designs, substantial clinical and definitional heterogeneity, and loss of the mortality signal on adjusted analysis, these estimates are hypothesis-generating and should not guide routine practice. Their main value is in establishing effect-size priors for multicenter RCTs with blinded core-laboratory adjudication in patients with high thrombus burden. Supplementary Information Below is the link to the electronic supplementary material. Supplementary file1 (PNG 441 KB) Supplementary file2 (DOCX 30 KB) Abbreviations American college of cardiology American college of emergency physicians Acute coronary syndrome American heart association Confidence interval Cardiac magnetic resonance Excimer laser coronary angioplasty Generalized linear mixed model Grading of recommendations, assessment, development and evaluations Harmonizing outcomes with revascularization and stents in acute myocardial infarction I-squared (heterogeneity statistic) Intra-aortic balloon pump Intramyocardial hemorrhage Major adverse cardiovascular and cerebrovascular events Major adverse cardiovascular events Myocardial blush grade Microvascular obstruction National association of EMS physicians Non-ST-elevation myocardial infarction Odds ratio Percutaneous coronary intervention Percutaneous excimer laser coronary angioplasty Preferred reporting items for systematic reviews and meta-analyses International prospective register of systematic reviews Propensity-score matching Randomized controlled trial Risk of Bias 2 (tool for randomized trials) Risk of bias in non-randomized studies of interventions Society for cardiovascular angiography and interventions ST-elevation myocardial infarction Thrombus aspiration during percutaneous coronary intervention in acute myocardial infarction study Thrombus aspiration in ST-elevation myocardial infarction Thrombolysis in myocardial infarction Trial of routine aspiration thrombectomy with PCI versus PCI Alone Unstable angina pectoris Author contributions All authors contributed to the study. Fouad Hanna and Ali Dway were responsible for the conceptualization and design of the study. All authors were responsible for Screening and Data extraction. Fouad Hanna, Abdelrahman El-Helbawy, and Sara T. Zaki were responsible for the creation of the figures and tables and data analysis. Fouad Hanna, Habib Rehman Khan, Luiz F. Ybarra, and Ali Dway shared the responsibility for writing the original draft of the manuscript. All authors commented on previous versions of the manuscript and approved the final submission. Revision and final corrections of the manuscript were performed by Fouad Hanna, Habib Rehman Khan, Ali Dway, and Mazen Negmeldin Yassin. Luiz F. Ybarra and Habib Rehman Khan served as the supervisors. Data availability Data available upon reviewer request and derived from already published studies. Declarations In accordance with the journal’s policy on Artificial Intelligence Generated Content (AIGC) and the COPE positi on statement, the authors declare the use of Gemini tool for the purpose of developing the graphical abstract. This tool was utilized to assist in the visual formation of the figure that summarizes the primary findings of this research. The authors have transparently described this use here and confirm that they are fully responsible for the accuracy of the information presented in the figure, for the correct referencing of supporting work, and for the final editorial decisions regarding the image. Gemini tool was not used as an author, nor was it involved in research design or data interpretation. Consent to participate Not applicable. Clinical trial number Not applicable (this research does not report the results of a clinical trial). Competing interests All authors have no competing interests or financial ties to disclose. Footnotes Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Ali Dway and Fouad Hanna contributed equally to this work. References Associated Data Supplementary Materials Supplementary file1 (PNG 441 KB) Supplementary file2 (DOCX 30 KB) Data Availability Statement Data available upon reviewer request and derived from already published studies.