Interpretation of biopsies of "normal" urothelium in patients with superficial bladder cancer. MRC Superficial Bladder Cancer Sub Group.

In the course of a Medical Research Council trial of intravesical chemotherapy, biopsies were taken from apparently normal bladder urothelium near to newly diagnosed superficial bladder cancers in 417 patients. Differences were noted in the rates at which histological features were described in different centres. To gain more information about the reproducibility of the pathological findings, a group of 6 pathologists (5 from the UK and 1 from the USA), all having a special interest in urological pathology, were asked to examine a representative sample of 92 slides. They were then asked to re-examine 30 of them after an interval of at least 6 months. At first examination and at re-examination the slides were assessed using a standard proforma. However, the definitions of the categories were left unspecified for the pathologists to use their own criteria. The 5 UK pathologists then met to establish a consensus view of each slide. The results indicated that: (1) The reporting of non-dysplastic changes varied so much between pathologists as to render it of little value to clinical practice. (2) There were wide variations between different pathologists in the reported incidence of dysplastic change. (3) On a second review the pathologists reproduced their own assessment on only 62% of occasions. (4) Even after discussion between pathologists there was no consensus on the diagnosis of mild as opposed to moderate dysplasia. Consensus was reached on all biopsies which showed either severe dysplasia or carcinoma in situ. (5) In adopting a policy of taking urothelial biopsies, urologists should be aware of the imprecision and lack of reproducibility in the interpretation of such biopsies. (6) Biopsies of cystoscopically normal urothelium may not be a useful guide in defining therapy.