Medical-quality and evidence standards
CER-AI separates source quality, relevance, implementation and product validation so that one type of evidence is not presented as another.
Source selection
Primary peer-reviewed publications are preferred for clinical concepts. Textbooks and manufacturer documentation are identified separately and used only for their appropriate educational or device-definition roles.
Claim-to-source traceability
Selected publications are mapped to the specific pathway or safety concept they support. The complete citation remains in the single medical reference registry.
Limitations retained
Associations, retrospective studies, case reports and device documentation are not presented as equivalent forms of evidence, and published thresholds are not generalized beyond their stated context.
Independent product validation
No publication listed here is described as external validation of the complete CER-AI software unless a dedicated CER-AI validation study is performed and published.
Core evidence mapped to CER-AI pathways
Original ectasia risk framework
Supports the classical multivariable ectasia-risk framework involving topographic pattern, residual stromal bed, age, corneal thickness and refractive magnitude.
Independent validation
Supports use of the ERSS as an independently evaluated preoperative LASIK screening system. CER-AI retains ERSS as a distinct pathway rather than mathematically blending it with tomography systems.
Ectasia risk after PRK
Provides PRK-specific context for ectasia risk and demonstrates why a LASIK-derived risk score should not be treated as an absolute individual risk calculator for PRK.
Post-LASIK ectasia risk factors
This is the explicit literature source used in the current CER-AI NICE implementation documentation. CER-AI displays the NICE pathway separately from ERSS and BAD-D.
Practical subjective scoring system
Provides the identified publication for the PS3 pathway. CER-AI keeps PS3 findings visible as their own risk channel and documents triggered factors rather than presenting PS3 as a replacement for tomography or tissue-safety review.
Enhanced ectasia susceptibility detection
Supports the broader role of multivariate tomographic analysis and artificial intelligence in detecting ectasia susceptibility. CER-AI does not claim to reproduce proprietary indices from this study unless their required inputs and implementation are explicitly available.
Belin/Ambrósio ectasia screening
Provides contemporary background on the Belin/Ambrósio Display and the role of tomographic parameters in ectasia screening. In CER-AI, Final BAD-D remains an independent Pentacam-derived risk channel.
Percent Tissue Altered
Supports the clinical relevance of procedure-related tissue alteration in post-LASIK ectasia risk. CER-AI keeps structural/tissue-safety calculations conceptually separate from intrinsic corneal susceptibility pathways.
PTA in suspicious topography
Extends the tissue-alteration literature to eyes with suspicious preoperative topography and is relevant to CER-AI's separation of structural impact from topographic/tomographic susceptibility.
Manufacturer field definitions
Supports device-specific screen and field terminology only. Manufacturer documentation is identified separately from peer-reviewed clinical outcome evidence and does not validate CER-AI's risk classification.
Evidence interpretation principles
- Independent channels: ERSS, Final BAD-D, NICE and other supported pathways are interpreted separately rather than collapsed into an opaque single score.
- Source provenance: Pentacam-derived variables should retain their defined screen/source context.
- Procedure safety is separate: residual stromal bed, pachymetry, ablation planning and postoperative keratometry constraints are not overridden by a reassuring ectasia score.
- No unsupported extrapolation: CER-AI does not present proprietary indices, unpublished formulas, or unavailable device outputs as though they had been reproduced.
- Clinical decision support: the software is intended for qualified ophthalmic professionals and is not an autonomous diagnosis or treatment system.
Medical References
The complete CER-AI literature registry is maintained as a single consolidated reference list, including ERSS, NICE, PS3, BAD-D, Pentacam tomography, PRFI, PTA, RTA, SCORE, biomechanics, PRK/LASIK ectasia and related background evidence.