About
Independent clinical evidence and validation methodology for AI medical devices (SaMD).
Serteser Consulting is run by Burak Serteser (BME MSc), a biomedical engineer who has developed a medical-AI medical device and published it in an international peer-reviewed journal. For medical-AI / SaMD companies: independent clinical validation studies, the CER statistical core, and regulatory evidence (TITCK / EU-MDR / FDA); and for academic teams: systematic reviews and biostatistics.
Principal
Burak Serteser
Burak Serteser
Biomedical Engineer (BME MSc) · SaMD Clinical-Evidence Methodologist · Founder of Salnus
At Salnus I developed a medical-AI medical device and published its clinical study in an international peer-reviewed journal. I turn that experience into independent methodology support under the Serteser Consulting umbrella: clinical validation study design for SaMD / medical-AI companies, standalone performance, the CER statistical core, and reporting-guideline (TRIPOD-AI / STARD-AI) compliance. As a methodologist independent from the team that develops the model, I build the part of the regulatory file that is the most heavily scrutinized yet the least negotiable.
The academic side is the foundation of this methodology: an international peer-reviewed clinical study (2026), PROSPERO-registered systematic reviews and meta-analyses, deep-learning-based segmentation studies, and clinical data pipeline projects. Biostatistics, systematic review, and meta-analysis consulting for academic teams is ongoing.
Publications & Research
Active portfolio
- · Clinical study published in an international peer-reviewed journal (2026)
- · nnU-Net four-bone knee CT segmentation (n=167, manuscript in preparation)
- · Hip osteoarthritis AI SR + meta-analysis (under peer review)
- · ONFH MRI AI SR + meta-analysis (in writing)
- · Knee osteoarthritis AI SR (in writing)
Method and Standards
Working framework
- · IMDRF SaMD N41 + MDCG 2020-1 clinical evaluation
- · EU-MDR Rule 11 classification + CER (TITCK transposes it)
- · TRIPOD-AI / STARD-AI / DECIDE-AI reporting standards
- · Pre-locked SAP + standalone / external validation
- · PROSPERO + PRISMA 2020 + Cochrane RoB 2 (systematic review)
- · KVKK- and GDPR-compliant data management
Sample Work
Anonymized project examples
For client confidentiality, the specialty, institution, and names are not shared; the format and scope are as follows.
Clinical-Evidence
SaMD standalone validation + CER statistics
- Scope
- For a medical-AI / SaMD company, a pivotal clinical validation study independent of the development team, plus the CER statistical core.
- Process
- Clarifying the intended-use statement, pre-locked SAP, standalone performance (external validation), reference standard protocol, TRIPOD-AI / STARD-AI compliant reporting, CER statistics section (MDCG 2020-1).
- Output
- Locked protocol + SAP, standalone performance report, CER statistics and clinical evaluation section, named methodologist signature in the regulatory file.
Duration: 6-10 weeks
Academic
Systematic review + meta-analysis
- Scope
- A full-process systematic review and meta-analysis for a team of 1 researcher + 1 supervisor in a medical specialty.
- Process
- PROSPERO registration, PRISMA 2020 flow diagram, dual independent screening (Rayyan), RoB 2 assessment, random-effects pooling with R metafor, GRADE evidence.
- Output
- Publication-ready draft (Vancouver), PRISMA 2020 supplementary files, forest plot and funnel plot figures, raw extraction CSV.
Duration: 4-5 months
Enterprise
Clinical data pipeline
- Scope
- For a health technology company, end-to-end pipeline setup from raw DICOM series to a segmentation training set.
- Process
- Anonymization, format normalization (NIfTI), quality control script, nnU-Net training configuration, KVKK-compliant storage architecture.
- Output
- Working pipeline (Python + Docker), README, file structure documentation, training log + initial validation metrics.
Duration: 6-8 weeks
Individual
AI mentoring
- Scope
- For an academic, training on AI tools across literature review, drafting, and revision processes.
- Process
- One-on-one online sessions; prompt engineering, ICMJE Vancouver limits, plagiarism and hallucination control checklist, personalized workflow.
- Output
- Personal prompt library, workflow document, follow-up messaging support.
Duration: 6 hours (3 sessions)
Contact
First, a short introduction.
A free 15-minute call to clarify the scope. For written communication, burak@serteser.com.tr.