Counting particles in tissue sections: choices of methods and importance of calibration to minimize biases.

Investigators must choose between counting methods to quantify microscopic particles in tissues. The conventional profile-based ("model-based" or "2D-") counting methods have been criticized for their potential biases due to assumptions about shapes, sizes, and orientation of particles when converting profile counts into cell numbers. New stereological methods ("design-based" or "3D-") methods such as the optical disector or physical disector were initially introduced as being inherently unbiased. Recent calibration analyses and comparisons of results from different investigators have revealed the potential for significant biases in the most efficient and most frequently used design-based method, the optical disector. This review aims to objectively assess the strengths and limitations of current profile- and disector-based cell counting methods by examination of studies in which these methods have been calibrated against the "gold-standard", counts obtained by 3-dimensional reconstruction of serial sections. Advantages and disadvantages of each counting method and the associated embedding and sectioning techniques are compared and frequent mistakes and pitfalls of each technique are discussed. The importance of a calibration step for each technique is emphasized, and a protocol is provided for a quick and simple calibration by a "sampling" 3-D reconstruction of limited serial sections. Trends in the usage of counting methods are analyzed in four major journals. It is hoped that this review will be helpful, for both investigators and manuscript reviewers, in clarifying some of the contentious issues in the choice and implementation of appropriate methods for particle counting in tissue sections.