Up-date on quantitative grading systems and prognosis

Several subjective grading systems have been proposed for different premalignant and malignant lesions with varying emphasis on the nuclear component. However, all subjective grading systems are prone to a significant degree of observer error because the subjective evaluation of nuclear features like size and staining density may be inaccurate and not reproducible. Therefore, quantitation-based methods are to be preferred to evaluate accurately the cell and its architectural features and to eliminate the subjective element of microscopy. The problems which have so far been considered and solved in some quantitation-based grading systems are as follows: 1. feature evaluation, selection and combination into indices and grading systems; 2. number of grades to be identified; 3. degree of certainty in the grading systems; 4. quality control and standardisation of the grading systems; 5. routine application of the grading systems and their efficacy in predicting tumour progression and/or regression (also in response to treatment); 6. current and future aims in researching new grading systems and improving the current ones. The solution of these problems has allowed the grading to be accurately evaluated; thus avoiding the poor reproducibility observed when features have been subjectively investigated.