In an attempt to rectify this, expensively advertised government sponsored screening campaigns have been launched in the past decade, sometimes coinciding with election campaigns and often without the support of a medical consensus. Recent high profile mistakes have highlighted the inadequacies of the current service, where human error has led to false negative and positive reporting of breast and cervical cancer. These errors have occurred throughout Britain, most recently for cervical cancer in Kent and for breast cancer in Devon.
These errors have arisen because screening is currently a subjective rather than an objective process depending on an individual’s interpretation of microcalcification or architectural distortion on a mammogram or the appearances of nucleus and cytoplasm on a smear. In order to limit these errors computerised screening processes need to be introduced that provide an objective analysis of the risk of malignancy including an assessment of quantifiable molecular changes that occur in cancer. Is this possible? We believe that it may be, if there is an investment in nascent technology, and we recommend that a portion of the enormous amounts of money that will be lost in future medical negligence settlements should be directed from the public purse to the development of molecular screening technologies.
Our understanding of tumorigenesis has progressed and for some tumours the genetic events leading to malignancy are well understood. Critical and specific genetic events involving oncogenes, tumour suppressor genes, or DNA repair enzymes have been characterised in many tumours, and these molecular flags of malignancy could be exploited for screening cancers. Genetic changes can be identified by mutation analysis or by looking for microsatellite instability, which is a feature of many tumours.
There is already evidence that these molecular techniques can be used for the early identification of various common cancers. K-ras mutations occur as one of the first in a cascade of genetic events in the development of colonic cancers. These changes are detectable in the stools of patients with preinvasive lesions as well as in patients with a first degree relative with colorectal cancer, and in patients with recurrent adenomas following surgery for a carcinoma. K-ras mutations are not diagnostic of malignancy because they are also present in patients with inflammatory bowel disease. Caldas and his colleagues have found K-ras mutations in the stools of patients both with pancreatic carcinoma and the precursor lesion, pancreatic mucinous ductal cell hyperplasia. This is a particularly exciting finding in the context of a tumour that typically presents late when the prognosis is poor.
K-ras and p53 mutations occur as early changes in other tumours. Examination of archival sputum specimens has led to the observation of mutations in patients who went on to develop adenocarcinoma of the lung, which in one patient had a lead time of one year.
“Modern screening processes must be developed” |
Microsatellite alterations were found in histopathologically normal bronchial specimens and cytological samples considered to have only minimal atypia in patients with a range of histological subtypes of lung cancer. It may be that these field changes could be detectable before the cancer could be picked up by conventional diagnostic means. Microsatellite instability had also been found in the urine of patients with cystoscopically confirmed bladder tumours. Microsatellite instability has a higher specificity than conventional cytology, predicating for malignancy in 95% as compared with 50% of patients.
Currently these techniques involve polymerase chain reaction amplification of the sample DNA, followed by a hybridisation technique to detect the mutation or separation of the products by electrophoresis to identify microsatellite alterations. The latter technique is simpler, with options for mutation screening limited by the size of the gene and the number of possible mutations which can occur. The advent of DNA microchip technology, however, offers the prospect of a real breakthrough for automating this analysis. Oligonucleotide sequences containing specific mutations are impregnated on to a glass slide and the sample DNA is then hybridised on to it. With this technology it has already been shown to be possible to screen a selected number of mutations in the BRCA1 gene. DNA microchip technology will allow us to screen large populations quickly and cheaply by automated means with the advantages of high throughput, cost efficient analysis.
Another possible approach to developing a new generation of screening technologies may be from capitalising on recent advances in imaging. The first digital x ray system is currently working efficiently and effectively in Britain. It should therefore be possible, using image analysis, to assess mammograms and avoid observer error. Similar computerisation should be possible for the digitalised image from a cervical smear.
The time has come for the community, through the Medical Research Council and the NHS research and development programme, to invest in preventive medicine and the identification of the molecular markers of malignancy. Modern screening processes must be developed to allow populations to be examined through automated procedures. The potential benefit is enormous: a real reduction in cancer mortality.