The purpose was to identify features of malignant and non-malignant neoplastic breast disease on galactography and to estimate their predictive value. This is the largest reported study correlating galactographic morphological patterns with histopathology and the only blinded study. The study included 351 consecutive galactograms and 161 breast biopsies performed in patients with nipple discharge over a 10-year period. Three radiologists, blinded to clinical data and histological results, re-evaluated 158 previously performed galactograms of patients who had undergone excision biopsy. Extravasation or incomplete filling precluded reading in 9.5% of examinations. Among the remaining 143 examinations there were 11 cancers (7.7%), 56 papillomas (39.2%), 19 cases of intraductal papillomatous proliferation (13.3%), 55 cases of fibrocystic or secretory disease (38.5%) and two normals. A "filling defect/cut-off" pattern (n = 90) was found in 6 cancers (6.7%) and 58 cases of papilloma or papillomatous proliferation (64.4%). A "leafless tree" pattern was found only in benign cases (n = 12; 8.4%). In 32 of 143 cases (22.4%) a "ductal ectasia" pattern was present, in one case of which (3.1%) cancer was found. Cancer was identified in two of four cases with an "architectural distortion" pattern. Cancer is rare in patients with nipple discharge. A tendency towards a lower incidence of cancer associated with the "ductal ectasia" and "leafless tree" patterns was found. No statistical evidence was found to indicate that galactography provides an effective prospective diagnosis of malignancy. However, an abnormal galactogram strongly correlated (p < 0.001) with the presence of a breast neoplasm when both benign and malignant tumours were considered. The most important role played by galactography is in the localization of breast neoplasms and in the choice of appropriate surgical therapy.

Cytology is helpful when positive and galactography localizes the source of discharge. Biopsy is indicated when palpation, mammography, cytology, or galactography is suspicious.

INTRODUCTION5-Hydroxytryptamine-3 (5-HT 3 ) receptor antagonists have been found to be potent anti-emetic drugs for chemotherapy-or radiation-induced nausea and vomiting.1±4 So far, four selective 5-HT 3 receptor antagonists (ondansetron, formerly GR 38032F; tropisetron, formerly ICS 205-930; granisetron, formerly BRL 43694; and dolasetron, formerly MDL 73147EF), with comparable clinical ef®cacy, are commercially available for these indications. The mechanism of the anti-emetic effect is not fully understood. 4 It is ascribed in part to central effects on the 5-HT 3 receptors in the area postrema, where the chemoreceptor trigger zone is located, with neural connections to the vomiting centre, but also in part to the peripheral action on 5-HT 3 receptors of afferent vagal ®bres in the stomach and small bowel.4±8 It has been suggested that the anti-emetic effect of 5-HT 3 may not only be due to suppression of nausea and the vomiting re¯ex, but also to direct or indirect effects on gastric motility. Although evidence for this latter assumption has been found in several animal SUMMARY Background: In previous studies, tropisetron has been shown to accelerate gastric emptying of a solid meal. However, it is uncertain whether other speci®c 5-hydroxytryptamine-3 receptor antagonists, such as ondansetron, also have a gastroprokinetic effect in humans. Aim: To evaluate the effect of ondansetron on gastric half-emptying time (T 1/2 ) of a solid meal, gastric myoelectrical activity and hormone levels in 14 healthy volunteers. Methods: In a placebo-controlled, randomized, crossover study, we investigated the effects of ondansetron (8 mg intravenously) on the gastric emptying of solids (by scintigraphy), gastric myoelectrical activity (by

No abstract