A study of the histological basis of textural analysis.

In an attempt to quantitate histological structure by textural analysis, the characters of a stained microscopical field were recorded as a frequency polygon of the relative areas defined by successive increments of optical density: this is known as the optical density/area (OD/A) profile. In a study of lymph-nodes in Hodgkin's disease and of carcinoma of the female breast, each stained tissue section was found to have an OD/A profile of characteristic shape: moreover, in each of the lesions studied it was shown that the shape of the profile could be replicated by summing the profiles of each constituent cell type and intercellular matrix in proportions determined by differential cell counts and direct measurement of the degree of cell packing in the tissue. In an investigation of the influence of specimen preparation on OD/A profile, it was shown that density of staining and section thickness were important variables. Using the density of staining of tissue lymphocytes as an internal standard, a simple method was developed of compensating for differences in the density of Mayer's haemalum staining greater than those met with in practice. Attempts to compensate for intentionally large differences in section thickness were much less successful, but this was considered to be relatively unimportant since, in practice, the variation in thickness of sections cut on a modern microtome by a competent microtomist had been shown to be slight and to have a negligible effect on the OD/A profile.

Cytological basis of histological typing of diffuse Hodgkin's disease. Demonstration of an implied misnomer in the terminology of the Rye classification.

Differential cell counts were made on nine lymph nodes whose structure was replaced by diffuse Hodgkin's disease; two of these nodes had the classical histological appearance of the lymphocytic predominance subtype, four of the mixed cellularity subtype, and three of the lymphocytic depletion subtype. Our attempts to achieve valid sampling methods are recorded. The counts, in general, confirm the postulated histological basis of the Rye classification of the subtypes of the diffuse disease. The major discrepancy is that, contrary to the histological descriptions, our direct counts have shown that lymphocytes, are, in general, more numerous in the lymphocytic depletion than in the mixed cellularity subtypes. The cell counts also show that normal mononuclear cells (mainly fibroblasts and macrophage-type cells) are much more numerous in the mixed cellularity subtype than in the other forms of diffuse Hodgkin's disease; this feature has not been emphasised in the Rye classification. On the basis of our differential counts, a hypothesis is proposed that could explain the natural history of the different subtypes of diffuse Hodgkin's disease as the resultant of three processes: (a) tumour aggressiveness, (b) specific cell-mediated immunological reactions, and (c) non-immunological stromal responses.