Apparatus and Methods for Automated Assessment of Tissue Pathology

Apparatus for automatically assessing tissue pathology, the apparatus comprising:

• a data processing unit connected to obtain images of cell nuclei in tissue samples, the data processing unit comprising:
• nucleus identification means for identifying the cell nuclei in the images;
• feature value computation means for computing values for a plurality of morphometric features of a cell nucleus identified by the nucleus identification means, the morphometric features including one or more texture features;
• and,
• index calculation means for computing a value for an index characterizing the tissue from the values of the morphometric features calculated by the feature value computation means;
• and,
• means for recording index values computed by the index calculation means.

A method for deriving an index indicative of a likelihood of an outcome for a tissue, the method comprising:

• obtaining images of a plurality of cell nuclei in sample of the tissue;
• computing values for a plurality of morphometric features of the plurality of cell nuclei;
• computing discriminant scores for the plurality of nuclei based upon the values of the morphometric features for the nuclei;
• classifying the nuclei into a bins of a series of bins based upon the discriminant scores for the nuclei;
• computing the index value based upon the relative numbers of the nuclei classified in the bins;
• and,
• recording the index value.

Apparatus for automatically assessing tissue pathology, the apparatus comprising:

• a data processing unit connected to obtain images of cell nuclei in tissue samples, the data processing unit comprising:
• nucleus identification means for identifying the cell nuclei in the images;
• feature value calculation means for computing values for a plurality of morphometric features of the plurality of cell nuclei;
• discriminant score computation means for computing discriminant scores for the plurality of nuclei based upon the values of the morphometric features for the nuclei;
• classification means for classifying the nuclei into a bins of a series of bins based upon the discriminant scores for the nuclei;
• index computation means for computing the index value based upon the relative numbers of the nuclei classified in the bins;
• and,
• means for recording the index value.

A method for assessing a likely outcome for an oral lesion, the method comprising:

• obtaining images of cell nuclei in a sample of tissue from the oral lesion;
• computing values for a plurality of morphometric features of the plurality of cell nuclei;
• and,
• from the values of the morphometric features, computing a value for an index indicative of a likely outcome for the lesion;
• and,
• recording the index value.