The effect of sampling on quantitative nuclear image features in histologic sections of lung carcinomas

A feasibility study showed that quantitative nuclear image (QNI) analysis, in which the morphology of the nucleus is described by a number of mathematical parameters, can be used to make the therapeutically and prognostically important distinction between small cell lung carcinoma (SCLC) and non-SCLC, which can be difficult to make with subjective histologic typing. In the present study, the effects of sample size and sample site on the QNI features were investigated. For all sample sites in a given tumor, comparison was made between the histologic classification, the ultrastructural findings and the classification based on the QNI features. Using a running mean, it was found that a sample size of 25 nuclei is sufficiently large. Histologic and quantitative classifications of samples from different sites of the same tumors were in agreement with regard to the separation of SCLC and non-SCLC in 19 of 20 sections. In the case in which disagreement occurred in one section, the ultrastructural findings supported the quantitative classification. These data indicate that sampling from different sites has no essential influence on the QNI classification of lung carcinomas.