Morphometry and Image Analysis

Visual analysis of microscopic images of cells and tissues has been in practice since long for achieving a diagnosis of pathological lesions. The pattern recognition may vary from person to person in subjective analysis. However, many definable quantitative parameters like diameter of cells & nuclei, circularity of cells, width or length of tissue components, area or diameters of certain special structures like glomeruli etc may be well analyzed through image analysis. Image is the basis of any kind of interpretation and diagnosis in anatomic pathology. Stereology and histometry had been in practice for >150 years. Stereology is based on the theorem forwarded by French Geologist Delesse in 1847. He demonstrated that the area properties of a mineral component in a rock as seen in a random cut surface, is equal to its volume proportion. Delesse principle was further propagated in histology by Weibel & Elias (1967). The term histometry refers to measurement of tissue components while morphometry has a wider meaning in the context of biology as it refers to measurement of size of cells, organisms, nuclei, subcellular components, area and volume fractions of tissue components. Stereology is the quantitative study of the three-dimensional properties of an object from the two-dimensional images using geometrical formulae and procedures. Volumetric parameters could be achieved from the images of accurately cut and well-stained microtome sections.

Older methods were based on the manual measurements on accurately enlarged micrographs and computation with respect to magnification, but the introduction of digital technology and availability of image analysis software(s) have revolutionized the morphometry at the light microscopic as well as electron microscopic (ultrastructural) level. Grey scale or multicolored microscopic images can be captured, archived and analyzed at convenience with utmost accuracy and precision. Image analysis can be divided into two broad categories; (1) morphometry related to spatial measurements and (2) cytometry/densitometry/fluorometry related to intensity measurements. The observer needs thorough knowledge of the peripheral settings, calibrations, utility tools, image analysis protocols, parameter's settings and data management. Inter and intra-observer variations could only be kept under check with perfect training in image analysis and specimen preparation. Though till date, applications of image analysis in routine diagnostic pathology are not considered seriously, but the quantitative parameters may be important for prognosis and or understanding the progression of pathological disorders by determination of nuclear irregularity and DNA content of tumor cells, quantitation of antigen expression, measurement of glomerular parameters and interstitial fibrosis in kidney biopsies and so on. Morphometry and image analysis is a valuable research tool in quantitative and analytical histology and cytology.

Image Analysis Laboratory Setup:

There is massive development in the field of image analysis in the last two decades. This is due to significant computational capability for data accumulation and storage. The other factor is rapid development of knowledge based systems for data analysis and diagnostic decision making. Presently there is widespread use of image analysis (IA) in pathology both in clinical and research areas. Reduction of cost of hardware, software and image capture devices is encouraging to setup image analysis systems and popularize this tool. An image analysis laboratory needs a work station comprised of a best quality binocular microscope equipped with high resolution digital camera and a P-IV computer with windows operating system and loaded with image analysis software.

Areas of Application of Image Analysis:

The principle areas of application of image analysis are: 1) Morphometric assessment: such as linear distance, area and perimeter and circularity measurements, object counting, size and shape etc., 2) Chromatin pattern recognition, 3) DNA quantification and ploidy estimation, 4) Quantification of immunocytochemistry, 5) Measurement of fractal dimension, 6) Cell to cell relation assessment, 7) Diagnostic decision making and 10) Chromosomal analysis.

Image Analysis Protocol:

1. The procedure adapted to process the specimens and to prepare the histological sections or smears should be same for all the specimens. Thickness of sections for morphometry and image analysis for evaluation of immunohistochemistry or DNA should always be kept constant.
2. Switch-on the system and verify the peripheral settings and calibration parameters.
3. Focus your slide and search the area of interest and acquire/grab the image.
4. Proceed for the interactive or density detection as per your requirement.
5. Record displayed results or save the data.
   

Comparison of Data from Different Centers:

Information technology has revolutionized the data management and analysis through ‘world wide web’ services and artificial intelligence. There are many knowledge-based expert systems available for data interpretation. Artificial neural network (ANN) is one of the efficient and popular expert systems for data interpretation. This is designed on the basis of the human brain. ANN has three layers 1) an input layer 2) single or multiple hidden layers and 3) an output layer. Input layer receives signal from external stimuli. The information is processed and finally results come through the output layer. ANN learns by example rather than the specific programming logic at the backend of the customized software, for a classification procedure.