Academic Integrity Review: Mechanisms of Stroke After Intracranial Angioplasty and Stenting in the SAMMPRIS Trial

Academic fraud report · Geng Detector

Summary

This review applies the Geng six-style forensic framework to Derdeyn et al.'s 2013 Neurosurgery paper (DOI: 10.1227/NEU.0b013e318286fdc8) on stroke mechanisms in the SAMMPRIS trial. Overall verdict: questionable (yellow). One substantive issue was identified: a logical data inconsistency in Table 3 involving Patient 25, a 76-year-old female recorded with a 30-day mRS score of 5 in one column while another column states the patient died 3 days after the procedure. Since death corresponds to mRS 6, this is a clear internal contradiction in clinical trial record-keeping. The arithmetic and statistical checks (33/224 = 14.7% symptomatic stroke rate; 220 minus 6 aborted minus 1 failed crossing = 213) were internally consistent. The device timeline (Wingspan FDA approval 2005; trial 2011; submission 2012; publication 2013) was also consistent. Image-based forensics could not be performed because the provided PDF text did not include raw pixel data for Figures 1 and 2. Confidence is moderate; the single finding is more likely a data-entry error than deliberate manipulation, but it warrants verification against the original case report form.

Verdict

Questionable (yellow). One internal logical conflict in clinical data was identified; arithmetic and timeline checks passed; image forensics were not feasible with the supplied source.

Key findings

Internal data conflict in Table 3 (Patient 25): A 76-year-old female is assigned a 30-day mRS of 5 in one column, while a neighboring column records "Died 3 days after procedure." Per the modified Rankin Scale, death is scored as 6, making this a direct logical contradiction.
Arithmetic consistency: 33/224 = 0.14732, rounds to 14.7%, matching the reported 30-day symptomatic stroke rate. The flow 220 attempted − 6 aborted − 1 failed crossing = 213 is internally consistent.
Device timeline consistency: Gateway PTA Balloon Catheter and Wingspan Stent System were FDA-approved in 2005; SAMMPRIS preliminary results were announced in 2011; the manuscript was submitted in September 2012 and published in 2013. No anachronism detected.
Image-forensic limitations: Figures 1 and 2 could not be analyzed for duplication, splicing, or noise anomalies because the source PDF provided only textual descriptions and captions, not pixel data.

Evidence highlights

Patient 25, Table 3: 30-day mRS = 5; "Family withdrew care. Died 3 days after procedure." (DOI: 10.1227/NEU.0b013e318286fdc8)
Statistical check: 33 ÷ 224 = 0.14732 ≈ 14.7% (Results, pp. 777–778).
Patient flow arithmetic: 220 − 6 − 1 = 213 (Patients and Methods, p. 778).
Device chronology: Wingspan FDA approval 2005; trial result 2011; submission September 2012; publication 2013.

Notes

The most plausible explanation is a data-entry oversight in the case report form, not deliberate under-reporting of mortality. Nevertheless, the contradiction should be reconciled against the original CRF and, if confirmed as an error, corrected via an erratum to the journal.
No evidence of image manipulation could be established; absence of pixel-level analysis means duplication or splicing cannot be ruled out.
All arithmetic and chronological checks were internally self-consistent.
This report is AI-assisted and intended for academic discussion; final determination of misconduct requires institutional investigation.