The presentation and management of 24 patients with endometriosis (median age 34 (range 21-68)years) presenting to general surgeons over a period of 10 years (1985-1994) was reviewed. Patients presented with an abdominal wall swelling related to a previous Pfannenstiel incision (seven patients), umbilical swelling (four), inguinal canal swelling (two), incidentally following appendicectomy (five), terminal ileal obstruction (two), rectal bleeding (two) and urinary symptoms (two). Endometriosis was not suspected in most patients but was confirmed by surgical excision or resection with minimal morbidity. No recurrence occurred during a median follow-up of 53 (range 9-113) months. Endometriosis is a disease rarely seen by general surgeons and is often diagnosed incidentally or on histological examination. Cyclical symptoms associated with menstruation are present in 50 per cent of patients and should suggest the diagnosis in those presenting with scar-related and/or subcutaneous swellings. Simple excision or resection of the presenting lesion provides adequate treatment but, since pelvic endometriosis may be present, referral to a gynaecologist is recommended in every case.

SUMMARY Sections from 100 cervical biopsy specimens were studied by 12 consultant histopathologists to determine the robustness of the existing pathology terminology and classification. Analysis by K statistics showed good agreement in the diagnosis of CIN 3 and squamous carcinoma but an inability to distinguish accurately between the lesser grades of CIN.It is recommended that the classification be changed to low grade (present CIN 1 and 2) and high grade (present CIN 3) categories alone. There was very poor agreement in the identification ofcellular changes associated with human papilloma virus (HPV) infection.Several novel analytical methods of assessing the severity of uterine cervical intraepithelial neoplasia (CIN) have been proposed,'2 but histological assessment remains the basis for determination oftreatment, clinical management, and subsequent follow up of patients. Although clear criteria for the diagnosis and grading of CIN have been described,3 such assessments are subjective and prone to intra-and interobserver variation.45 The problems of histological assessment have been further complicated by the increasing recognition of human papilloma virus (HPV) infection"7 which may be an aetiological factor in the development of CIN.89 HPV infection may be indicated by koilocytosis and other changes that distort cellular appearances and so may apparently exaggerate the severity of the premalignant appearances ofthe cervical epithelium, particularly in the higher layers-making grading more difficult.It is reasonable that efforts should be made to establish the degree of confidence which can be given to the histological reporting of cervical biopsy lesions by pathologists and to determine the robustness of the existing terminology and classification. We describe the findings of a study of cervical biopsy specimens conducted by a group of 12 pathologists, all of consultant grade, but with varying degrees of experience.Accepted for publication 1 September 1988 Material and methods COMPOSITION OF PANELTwelve histopathologists were invited to join the study with a deliberate attempt by the organisers to obtain a composition representative of Scottish pathology as a whole. The members came from pathology laboratories in Aberdeen (n = 2), Dundee (n = 2), Edinburgh (n = 2), Airdrie (n = 1), Perth (n = 1), Stirling (n = 1) and Glasgow (n = 3) and varied in years of consultant experience (five to 25 years) and nature ofsubstantive post (university staffn = 5: NHS staff n = 7). All the members of the group had undertaken their postgraduate training in Scotland. CLASSIFICATION OF CERVICAL HISTOPATHOLOGYAt the initial meeting current cervical pathology terminology was reviewed and following discussion a proforma was designed for completion after examination of each slide in the circulation. This was modified in a minor way after the first circulation and the final form is shown in the figure. It was decided to keep the classification simple, but to relate it as closely as possible to everyday practice. The following ...

Summary The expression of mRNA for the epidermal growth factor (EGF) receptor, EGF and transforming growth factor a (TGF-a) was determined in 76 malignant, six borderline and 15 benign primary ovarian tumours using the reverse transcriptase-polymerase chain reaction and related to clinical and pathological parameters. Of the malignant tumours, 70% (53/76) expressed EGF receptor mRNA, 31% (23/75) expressed EGF mRNA and 35% (26/75) expressed TGF-a mRNA. For the borderline tumours, four of six (67%) expressed EGF receptor mRNA, 1/6 (17%) expressed TGF-a mRNA and none expressed EGF mRNA. Finally, 33% (5/15) of the benign tumours expressed EGF receptor mRNA, whereas 40% (6/15) expressed EGF mRNA and 7% (1/15) expressed TGF-a mRNA. The presence of the EGF receptor in malignant tumours was associated with that of TGF-a (P = 0.0015) but not with EGF (P = 1.00), whereas there was no relationship between the presence of EGF and TGF-a (P = 1.00). EGF receptor mRNA expression was significantly and positively associated with serous histology (P = 0.006) but not with stage or grade. Neither EGF nor TGF-a showed any link with histological subtype or stage. The survival of patients with malignant tumours possessing EGF receptor mRNA was significantly reduced compared with that of patients whose tumours were negative (P = 0.030 for all malignant tumours; P= 0.007 for malignant epithelial tumours only). In contrast, neither the expression of TGF-a nor EGF was related to survival. These data suggest that the presence of EGF receptor mRNA is associated with poor prognosis in primary ovarian cancer.

Sixteen cases of ductal (endometrioid) carcinoma of the prostate are presented. The tumour presents in elderly men (age range 65-87 years) with haematuria or obstructive symptoms. Serum prostate specific antigen may be normal or raised. On cytoscopy, there is often an exophytic lesion in the region of the verumontanum. Histologically, two variants are recognized: papillary and cribriform, of which there were eight cases each. Eight cases consisted of pure ductal carcinoma and seven were mixed, containing a variable proportion of micro-acinar carcinoma. The associated micro-acinar carcinoma had a Gleason score of at least 5. One case of carcinosarcoma with a ductal epithelial component was also included. All cases displayed positive immunohistochemical staining for prostate specific antigen and prostatic acid phosphatase and but were negative for the basal cell marker MA903. The tumour responds well to orthodox micro-acinar carcinoma therapy and appears notably sensitive to hormonal manipulation. Follow-up of the mixed group is restricted to a maximum of 3 years. Of the eight pure cases, five patients are still alive with survival periods of 11, 8, 7, 3 and 1 years. Three patients died of intercurrent disease of which one patient survived 12 years, having received no treatment. This tumour, therefore, can be regarded as having a good prognosis.