Skip to main content
. Author manuscript; available in PMC: 2024 May 14.
Published in final edited form as: Annu Rev Anal Chem (Palo Alto Calif). 2023 Apr 17;16(1):205–230. doi: 10.1146/annurev-anchem-101422-090956

Table 3.

Proposed experimental and analytical specifications for spectroscopic imaging studies in histopathology

Category Essential specifications
Patient and clinical data Characteristics of patients selected (disease, control)
Source of material
Demographics of patients (e.g., age, sex, ethnicity)
Sample processing Sample history from acquisition to measurement
Processing steps used
Preservation and storage conditions Pathology reference slides (e.g., H&E or IHC stains)
Quality assessment processes (e.g., pathologist review)
Study parameters Goal of the study
Inclusion/exclusion criteria
Control samples
Case selection (e.g., prospective or retrospective, stratification or matching)
Target statistics (e.g., power)
Specimen preparation Substrate used
Protocol to cut and deposit
Chemical treatments (e.g., hexane wash for 24 h)
Measurement conditions Description of optical setup
Detailed measurement protocol
Quality reference and data (e.g., standard spectrum, 1951 USAF resolution test-equivalent sample)
Reproducibility checks for instrument
Data processing Description of all preprocessing steps (e.g., baseline correction, normalization derivatives, smoothing)
Spectral and other variables considered and selected (with selection algorithm)
SNR calculation method and SNR evaluation of data
Machine learning Model selection and alternatives considered or rejected during study
Algorithm
Full code and parameters
Sample sets for training, testing, and validating steps
Missing information/unbalanced data handling
Claims Relation to ground truth
Statistical measures of results including sample size
Quantitative measures (e.g., ROC curves, Kaplan-Meier plots, hazard ratios) with confidence intervals
Precise description of end points
Model verification
Study assessment Relationship between spectral data and diagnostic test performance
Sensitivity analysis
Limitations of the study
Comparison with clinical/research standard

Abbreviations: H&E, hematoxylin and eosin; IHC, immunohistochemical; ROC, receiver operating characteristic; SNR, signal-to-noise ratio; USAF, United States Air Force.