Prognostic factors: seeking a shaft of light or getting lost in the woods? — A personal view from Sherwood forest

Quantitative analysis of colorectal tissue microarrays by immunofluorescence and in situ hybridization

The accuracy and reliability of in situ studies may be compromised by qualitative interpretations. Quantitation imposes a greater degree of objectivity, is more reproducible, and facilitates the clarity of definitions. The aim of this study was to validate the utility of laser imaging systems for the in situ quantitative analysis of gene expression in tissue microarrays. Immunofluorescence was employed to quantify the expression of the tumour suppressor p53, a marker of proliferation (Ki67), an endothelial cell marker (CD31), and the mismatch repair proteins human Mut L homologue 1 and human Mut S homologue 2 in an arrayed series of colorectal tissues (n = 110). Quantitative data on this panel of antigens were compared objectively with qualitative scoring of immunohistochemical chromogen deposition. In addition, the expression of vascular endothelial growth factor (VEGF)-A, placental growth factor, hepatocyte growth factor, and c-Met mRNA was quantified by phosphor image analysis of in situ hybridization reactions. The quantified data on p53, Ki67, and CD31 expression were significantly associated with the pathologist's score (p < or = 0.001). While hepatocyte growth factor and placental growth factor were not up-regulated, c-Met expression was increased up to 2.5-fold and the median VEGF-A expression was elevated 4-fold (p = 0.003) in this series of colorectal tumours. Laser imaging systems are therefore feasible for high-throughput, quantitative profiling of tissue microarrays.

Outcome after extended follow-up in a prospective study of operable breast cancer: key factors and a prognostic index

In 1990, 215 patients with operable breast cancer were entered into a prospective study of the prognostic significance of five biochemical markers and 15 other factors (pathological/chronological/patient). After a median follow-up of 6.6 years, there were 77 recurrences and 77 deaths (59 breast cancer-related). By univariate analysis, patient outcome related significantly to 13 factors. By multivariate analysis, the most important of nine independent factors were: number of nodes involved, steroid receptors (for oestrogen or progestogen), age, clinical or pathological tumour size and grade. Receptors and grade exerted their influence only in the first 3 years. Progestogen receptors (immunohistochemical) and oestrogen receptors (biochemical) were of similar prognostic significance. The two receptors were correlated (r=+0.50, P=0.001) and displaced each other from the analytical model but some evidence for the additivity of their prognostic values was seen when their levels were discordant.

Review of known prognostic variables

The prognosis of patients with primary breast cancer depends on prognostic factors per se and on treatment. In certain subgroups of patients, the overall treatment benefits of systemic adjuvant therapy are modest. Therefore, proper selection of patients and treatment modalities is absolutely necessary. Patients with primary breast cancer and a very good prognosis can be classified by known prognostic factors such as tumor size T < or = 1 cm, grade 1, hormone receptor positive tumors, and no other factors indicating poor prognosis. [table: see text] Patients with very poor prognosis are characterized by young age (< 35 years), large tumor size (T > 3 cm), grade 3, hormone receptor-negative tumors, and high S-phase fractions. Tumor size is the best prognostic factor for the decision to treat (breast conservation surgery, primary preoperative or adjuvant systemic therapy) or not to treat. Table 4 shows possible treatment decisions (type and timing of therapy) in primary breast cancer based on known prognostic and predictive variables. The decision to treat or not to treat outside clinical trials should be based on this kind of information. The type and aggressiveness of adjuvant systemic treatment should depend on the receptor status of the tumor and on the age and menopausal status of the patient, as classical predictive factors; however, prognosticators which can describe the metastatic potential of breast cancer cells have yet to be defined.