Howard's brick rule for partial objects: Pawley's Confocal Handbook discusses a fundamental problem in quantitative microscopy: what do you do with objects that cross the ROI boundary? A nucleus that's half inside and half outside the ROI could be counted in both, neither, or one. Counting in both overestimates; counting in neither underestimates. Howard's brick rule provides a consistent solution: include objects touching the top and right boundaries, exclude objects touching the bottom and left boundaries. This produces an unbiased count on average, though for small ROIs the variance can be significant.
ROI masking as spatial filtering: In signal processing terms, applying an ROI is multiplication by a spatial mask — a binary function that is 1 inside the ROI and 0 outside. This is a spatial filter that passes only the signal within the region of interest. The sharp edges of the mask can introduce boundary artifacts (analogous to the Gibbs phenomenon in Fourier analysis), which is why the boundary-handling rule matters for accurate quantification.
Biological rationale for spatial stratification: The tumor microenvironment is spatially organized. Immune cells at the invasive margin behave differently from those in the tumor center. PD-L1 expression varies by location. Spatial phenotyping results change dramatically depending on which tissue zone is analyzed. Whole-section averages obscure this spatial heterogeneity — ROI-based analysis preserves it. This is why modern pathology scoring systems (like the Immunoscore for colorectal cancer) require separate analysis of tumor center and invasive margin.
Objects at ROI boundaries create a counting problem — is a half-included cell counted or not? Howard's brick rule provides a consistent answer that avoids systematic bias. More fundamentally, ROIs exist because tissue biology is spatially organized — immune cells at the tumor border behave differently from those in the tumor center, and combining them would obscure the differences that matter clinically.
