1. ¹Head and Neck Surgery Department, Institut Gustave-Roussy, 39 rue Camille Desmoulins, Villejuif 94805, France
2. ²Pathology Department, Centre René Huguenin, St Cloud 92210, France
3. ³Pathology Department, Institut Gustave-Roussy, Villejuif 94805, France
4. ⁴Genetics Department, Institut Gustave-Roussy, Villejuif 94805, France

Correspondence: Dr F Janot, E-mail: janot@igr.fr

Received 15 November 2002; Revised 30 January 2003; Accepted 10 February 2003.

Abstract

Oesopharyngeal brush (OPB) sampling with cytological analysis can yield exfoliated cells from asymptomatic tumours of the upper aero-digestive tract and the oesophagus. In this study, we compared cytological evaluation and molecular analysis for the detection of exfoliated cancer cells sampled with an OPB. A total of 56 patients with a known unique head and neck squamous cell carcinoma (HNSCC) and five healthy controls were enrolled prospectively. Exfoliated cells from these 61 patients were collected with an OPB before initial endoscopy. p53 mutations and UT 5085 microsatellite instability (MI) were analysed in the HNSCC tumour, lymphocytes and the corresponding OPB DNA samples. p53 mutations and UT5085 MI were detected in 31 out of 56 and 14 out of 56 HNSCC, respectively, but not in any of the five controls. Direct sequencing of p53 was able to detect mutations in OPB DNA in only two out of 29 patients harbouring a p53-mutated primary tumour. Microsatellite instability was detected in OPB DNA of 11 out of 13 informative (bandshift detected in tumour) patients, whereas cytological analysis detected abnormal cells in only six of the same 13 patients (P=0.03). In informative patients, all positive OPB samples at cytological analysis were also positive at molecular analysis of UT5085, and both analyses confirmed the two negative samples. Molecular analysis of OPB from eight uninformative patients and from five healthy controls were all negative. OPB sampling with MI-based molecular analysis could be efficient for early detection of recurrent HNSCC. This result prompts us to use other microsatellite markers in order to maximise the percentage of informative patients.