Method for automated processing of digital images of tissue micro-arrays (TMA)
A method and system for automated quantitation of tissue micro-array image (TMA) digital analysis. The method and system automatically analyze a digital image of a TMA with plural TMA cores created using a needle to biopsy or other techniques to create standard histologic sections and placing the resulting needle cores into TMA. The automated analysis allows a medical conclusion such as a medical diagnosis or medical prognosis (e.g., for a human cancer) to be automatically determined. The method and system provides reliable automatic TMA core gridding and automated TMA core boundary detection including detection of overlapping or touching TMA cores on a grid.
1. A method for automated processing of digital images of tissue micro-arrays (TMA), comprising: differentiating a plurality of objects of interest from a background portion of a digital image of a tissue micro-array (TMA) of a tissue sample to which a chemical compound has been applied by adjusting a contrast of the digital image to create a contrast adjusted digital image; identifying a plurality of boundaries for a plurality of TMA cores in the differentiated plurality of objects of interest in the contrast adjusted digital image based on pixel intensity of potential TMA cores; displaying the contrast adjusted digital image with the plurality of TMA core boundaries graphically displayed on a graphical user interface; and creating a medical conclusion using the plurality of TMA cores having the identified boundaries in the contrast adjusted digital image;
wherein the identifying step includes: locating a plurality of centers of the plurality of differentiated objects of interest in the contrast adjusted digital image; applying a digital filter to the located plurality of centers of the plurality of differentiated objects in the contrasted adjusted digital image to remove unwanted objects; expanding a plurality of areas of interest around the filtered plurality of centers of the plurality objects in the contrasted adjusted digital image; determining overlapping objects, if any, from the expanded plurality of areas of interest in the contrast adjusted digital image; and determining a plurality of boundaries around the expanded plurality areas of interest in the contrast adjusted digital image to determine a plurality of boundaries of TMA cores; and wherein pixels with an intensity value less than a combined value of a mean intensity value and a standard deviation value of an area of interest are selected as pixels belonging to the area of interest.
2. The method of claim 1 further comprising providing a computer readable medium having stored therein instructions for causing one or more processors to execute the steps of the method.
3. The method of claim 1 wherein the differentiating step includes making the plurality objects of interest darker and the background portion lighter in the digital image.
4. The method of claim 1 wherein the chemical compound includes a Haematoxylin and Eosin (H/E) stain.
5. The method of claim 1 wherein the predetermined factors further include size, shape, length, width, core boundary characteristics, overlapping core areas, and core grid position.
6. The method of claim 1 wherein the locating step includes locating the plurality centers of the plurality of differentiated objects of interest using a Gaussian kernel.
7. The method of claim 6 wherein the Gaussian kernel includes a Gaussian kernel of sigma three.
8. The method of claim 1 wherein the applying step includes applying a digital filter based on a pre-determined range of sizes of a TMA core in the contrasted adjusted digital image.
9. The method of claim 8 wherein the pre-determined range of sizes is less than 0.6 millimeters and greater than 2.0 millimeters.
10. The method of claim 1 wherein the medical conclusion includes a medical diagnosis or medical prognosis for a human cancer including a human breast cancer or a human prostrate cancer.