a (+/- denotes positive/negative status)

Table 2: Patients categorised according to CTC status

controls. These were an age matched population and their consistent negativity demonstrates the reliability and reproducibility of the assay, as did the analysis in duplicate of all samples.

Each patient was placed into 1 of 4 categories (Table 2): category 1 patients were CTC negative at both samplings (n=13), category 2 patients were CTC negative pre-operatively but CTC positive post-operatively (n=11), category 3 patients were positive at both samplings (n=8) and category 4 patients were CTC positive pre-operatively but negative post-operatively (n=3).

CTC status was not found to correlate with Dukes' stage, recurrence, tumour location or survival. Gender was identified to be a significant factor in CTC status (Table 3). In comparing each CTC category between males and females, males were significantly more likely than females to be in category 2 (p=0.03) and females were significantly more likely to be in category 3 than males (p=0.03). Further analysis grouping categories 1 and 2 together, i.e. negative pre-operative categories) and

Table 3: CTC status and gender

The distribution of patients according to gender for each category is shown.

categories 3 and 4 (positive pre-operative categories) revealed that females were significantly more likely to be CTC positive pre-operatively than males (p=0.007). However, when the distribution in females alone was compared between each category there was no significant difference. In contrast when the distribution of males alone were compared between categories they were significantly more likely to be in category 1 or 2, i.e. CTC negative pre-operatively, than category 3 or 4 (p=0.001).

Further analysis to investigate whether the gender of this cohort of patients was associated with age, Dukes stage, recurrence, tumour location or survival did not identify any correlation.

This study has described the highly specific detection of viable tumour cells in the peripheral blood of colorectal cancer patients in a technique demonstrated to be both reproducible and reliable. CTC were detected pre-operatively in 11/35 (31%) of the patients. For 8 of these patients the CTC remain detectable at 7 days, however the remaining 3 patients were CTC negative indicating that live CTC were no longer present. 13/35 (37%) of the patients fell into Category 1 with CTC not being detected either pre or post-operatively. It may be expected that lower Dukes stage tumours would be less likely to have CTC. It is a reasonable assumption that the presence of CTC would be an indicative factor for the presence of metastases and consequently CTC would be more common in higher Dukes stage tumours. However, the present study demonstrates that the metastatic process is not this simple as the category 1 patients consisted of 1 Dukes A, 5 Dukes B, 3 Dukes C and 4 Dukes D patients. The cancer biology of this category of tumours may intrinsically differ from tumours that are disseminated into the circulation. It is important to remember that this study only addresses the phenomenon of blood borne metastases as other studies have shown that lymphogenic tumour cell dissemination may be equally as important. Weitz et al, (1998) using Cytokeratin 20 mRNA RT-PCR studied 279 lymph node, blood and bone marrow samples from 20 colorectal cancer patients. A high proportion of patients with histopathologically tumour-free lymph nodes were found to have tumour cells in these nodes and/or the mesentery vessels by this technique but only 2 patients had tumour cells in the blood. These authors suggest that lymphogenic tumour cell dissemination is a very common and early event in colorectal cancer preceding hematogenous tumour cell dissemination, however it must be noted that RT-PCR is not specific for viable cells.

An interesting recent study by Nozawa et al (2003) assessed telomerase activity in blood samples from mesenteric (tumor-drainage) vein and peripheral vessels of 41 colon cancer patients in relation to liver metastases. The authors identified high telomerase activity of mesenteric samples reflecting the existence of liver metastasis of colorectal cancer.

Many studies have shown that surgical manipulation can provoke cell dissemination (Van der Pompe et al, 1998; Weitz et al, 1998; Crisan et al, 2000) and this is supported by the 11 Category 2 patients patients who were CTC negative pre-operatively and CTC positive post-operatively. However, as statistical analysis of the results did not identify any significance between the presence of CTC and clinicopathological factors it indicates that CTC may not be an important prognostic factor in overall patient survival. This conclusion is supported by Bessa et al, (2003) who, using RT-PCR based methodology, concluded that postoperative detection of CTC had no prognostic significance in patients with colorectal cancer undergoing surgical resection with curative intent.

The observation that gender was a highly significant factor in relation to the presence of CTC with females generally being more likely to have CTC pre-operatively than males is an extremely interesting finding. This supports the idea that there may be differences in the biology of the same cancer between genders. Such an observation has been made by a number of other studies in colorectal and other cancers, and a number of hormones and other factors implicated to behave differently between genders. Aberrant hypermethylation of promoter CpG islands is an important mechanism for the inactivation of tumour suppressor genes and in gastric cancer, Kang et al, (2003) identified that male patients showed higher numbers of methylated genes than females. In colorectal cancer, a large study (n=867) revealed loss of hMLH1 expression is more likely (p<0.0001) to occur in females than males (Kakar et al, 2003). Similarly, studies of Apolipoprotein E gene polymorphism in 206 colorectal cancer patients and 353 healthy controls revealeda strong association with both colorectal cancer risk and prognosis in a gender dependent manner (Watson et al, 2002).

In patient treatment, gender has also been implicated to be a predictive factor in response to treatment with 5-Fluorouracil (Yamashita et al, 2002). The authors studied the expression of Dihydropyrimidine dehydrogenase (DPD), the initial rate limiting enzyme in the catabolism of 5-fluorouracil. DPD expression levels are believed to correlate with the 5-FU sensitivity of malignant tumours. DPD expression was quantitated in 97 tumour specimens and 92 adjacent normal tissue specimens from 97 patients. The DPD expression in the tumour tissues was significantly lower in females than males although in the normal tissues there was no significant difference between the genders. The authors conclude that CRC patients who will benefit most, because of lowered DPD expression, must be given priority and female gender is a predictive factor for a better response to chemotherapy with 5-FU.

From this study it would appear that there is a factor(s) specific to males which means they are less likely to have CTC pre-operatively than females. Not only is the incidence of pre-operative CTC significantly less for males than females, but also when considering the male group alone, males are significantly more likely to be in a CTC negative pre-operative group. However it is important to remember that the group sizes are small and the study therefore needs to be expanded to fully elucidate such differences.

It is a very interesting phenomenon that the biology of cancer differs between genders and as evidence for such a difference accumulates it is becoming increasingly important to consider gender in future cancer studies. This is supported by a recent study by McArdle et al, (2003) who showed that following apparent curative resection for colorectal cancer and after adjusting for case-mix, male gender adversely affected 5-year survival. The data presented in our study supports the fact that gender differences must be considered when designing an individual patient's treatment.

In conclusion, this study has demonstrated methodology to determine the presence of viable CTC in colorectal cancer patients. The ability to study the presence of viable CTC is of paramount importance in further understanding the metastatic process and this study has described a highly reproducible and reliable assay which may be applied to a wide range of cancer studies. This methodology has previously been found valuable in studies of metastatic breast cancer patients (Soria et al, 1999), hepatocellular carcinoma patients (Tatsuma et al, 2000), non small cell lung carcinoma patients and Dukes stage C or D colon cancer patients (Gauthier et al, 2001). For colorectal cancer patients the technique could have an important role to play, in conjunction with faecal occult blood testing and lower GI endoscopy, in the diagnosis of symptomatic patients. The technique may also be very valuable in monitoring patients during chemo- or radiotherapy.

The study has indicated that although the presence of CTC is not clinically significant in this cohort size they are strongly associated with gender. This association, implicating differences in the biology of colorectal cancer between gender, should be assessed in other cancers and indicates an important new avenue for cancer research.