Determination of chromosome 18q status. Loss of a chromosome 18q allele in a tumor was considered to be present when the relative intensity of the two alleles in the tumor DNA differed from the relative intensity in the normal tissue DNA by ±35%. MSI (microsatellite instability) tumors, in which abnormal size PCR fragments were obtained, were included among the tumors with no loss. The cut off was set at 35% after the verification of the procedure performed by mixing two homozygotes for two different alleles for each marker in varying concentrations (100%, 90%, 75%, 60%, 50%, 40%, 25%, 10%, 0%) (data not shown).

Evaluation of assays and assessment of 18q allelic loss frequency.The results of the 18q LOH analysis obtained by the conventional PCR/PAGE method and the multiplex fluorescent PCR are shown in Figures 1 and 2, respectively. The results obtained by the two methods were in complete agreement for all analyzed patients (data not shown). Due to its simplicity, speed, reproducibility and lesser manipulation requirements, the automated fluorescent multiplex PCR analysis is a method of choice for a routine diagnosis of allelic imbalances of clinical relevance.

Loss of heterozygosity at 18q was present in 31 of 92 (34%) analyzed cancers. This figure is slightly lower compared to the reports of other studies, ranging from 43% to 53% [18-20]. The discrepancy could be explained by the fact that our group consisted mainly of lower staged tumors (18 tumors-stages A and B vs. 13 tumors-stage C). Allelic loss in all the markers, i.e. total loss of chromosome 18q was detected in 29 out of 31 (93.5%) tumors. In one tumor (3.2%) partial loss of telomere markers D18S58 and D18S61 was detected and distal loss of the 18q tip (D18S58) was also present in one (3.2%) tumor. These two patients and their tumors did not differ in any way from the rest of the patients with 18q allelic imbalance of their tumors.

Association of 18q loss with some features of patients and tumors. The association of 18q allelic loss with several clinical and histopathological features in tumors of 92 colorectal cancer patients included in this study is shown in Table 3. Loss of heterozygosity at 18q was not associated neither with gender and age of the patients, nor with histotype and Dukes' stage of tumors. Association of 18q imbalance with left sided localization of tumors was with a border line significance. The prognostic significance of 18q imbalance on relapse of the disease, disese free survival and total survival is currently being evaluated in a prospective study of over 100 colorectal cancer patients and will be reported upon its completion.