Screening for Early Esophageal Cancer: Whom and How?
Sautereau
DOI: https://doi.org/10.1055/s-1999-19
1999-05-01
Endoscopy
Abstract:The majority of esophageal cancers are epithelial in nature. The two most frequent microscopic types are squamous cell carcinoma and adenocarcinoma. Until recently, squamous cell carcinoma was much more prevalent (20 to 30 fold) than adenocarcinoma in Western countries. Over the past few decades the frequency of adenocarcinoma has increased to such an extent that it has, statistically, practically caught up with squamous cell carcinoma. The reason for this growth remains obscure. From an epidemiological point of view, tobacco, alcohol, various irritants such as heat, spices and chronic esophageal diseases like caustic stenoses, Plummer-Vinson syndrome or achalasia are major risk factors in squamous cell carcinoma. The major factor in adenocarcinoma is Barrett's oesophagus (BE). The diagnosis of advanced squamous cell carcinoma is of little practical value for the patient's future, given the catastrophic prognosis, regardless of the treatment prescribed. The five-year survival rate does not surpass 20 % in the best of cases. On the other hand, the prognosis for squamous cell carcinoma in situ or carcinoma limited to the submucosa is excellent. The rates of survival after surgical treatment, in certain Chinese studies, have been as high as 85 % after five years [1]. Unfortunately these results cannot be compared to occidental statistics for the following reasons [2]. First, the epidemiology of oesophageal carcinoma is different in China and Europe. In certain regions of northern China, nutritional factors are responsible for the younger age of onset of the illness as compared with that in Europe, thus accounting for effectiveness of surgical treatment even if it is aggressive. In Europe, cancer of the oesophagus affects principally alcoholic smokers whose general state of health and nutrition, being mediocre, are not suited to an aggressive surgical approach or radiotherapy without a high mortality rate. Secondly, the incidence of oesophageal carcinoma in Europe is relatively low compared with certain regions of the world. In Central Asia, certain Chinese provinces and South Africa, the prevalence of esophageal cancer in people over 75 years old varies from 20 to 25 % and represents the principal cause of death by cancer. In Europe the incidence is on average ten times lower. It does not rise above 10 per 100 000 population except in certain regions (Brittany and Normandy) where it can reach 30 per 100 000. Lastly, there is no systematic screening in Europe and its diagnostic yield would in any case be much too low. However screening targeted an the principal high-risk group should help diagnose a greater number of early-stage carcinomas. These different data explain, on the one hand, the difficulties in defining a universal strategy in all populations for screening for esophageal cancer, and on the other hand, that the diagnosis of micro-invasive cancer is rare in Europe and the paucity of statistical studies in the West. Yet, acceptable rates of cure can only be attained through early diagnosis. In Western countries, early diagnosis is possible when two conditions are satisfied. The first is to identify high-risk patients [3 4]. Ninety percent of the cases of squamous cell cancer in Western Europe and North America occur in alcoholic smokers. The etiological fraction linked to excess alcohol consumption is around 50 % and that to tobacco 40 %. The role of these two factors also appears to be important in Latin America and Japan, it is less significant in South Africa and India. In all these countries, a policy of primary prevention taking these two factors into account, would seem to be a priority. In high-risk patients, minor symptoms should instigate appropriate investigation. Contrary to conventional thinking, esophageal carcinoma limited to the submucosa causes symptoms in more than 2/3 of patients. Only in-situ or intramucosal carcinomas remain mainly asymptomatic [2]. Patients with cancers of the upper aero-digestive tracts have a high frequency of synchronous or metachronous tumours (15 % have carcinoma at multiples sites). Even if the efficiency of screening is excellent in these patients, there is still no impact on survival rates because of the gravity, in many cases, of the initial cancer. Head and neck cancers are associated with cancer of the upper aera-digestive tracts (including the esophagus) in 10 to 40 % of cases [5]. Systematic panendoscopic studies have demonstrated a synchronous association of two carcinomas in 9 to 14 % of patients. Metachronous cancers are localised mainly in the esophagus and lung; laryngeal cancers are more often associated with lung cancers (45 to 76 %), and pharyngeal cancers with those of the oesophagus (19 to 27 %). Particular attention should thus be paid to patients with an early-stage head and neck cancer. A panendoscopic examination (bronchoscopy, pharyngo-esophagoscopy and laryngoscopy) is recommended at the time of the initial diagnosis and then every six months for five years [2]. Other etiologies are more rare. The prevalence of cancer associated with achalasia varies from 2 to 7 % in most studies. In general, it is localised to the middle third of the esophagus. The average delay between the onset of the first symptoms of achalasia and those of cancer is 17 to 20 years. This complication occurs most frequently in patients who have not been treated or treated insufficiently for achalasia, suggesting that stasis and chronic irritation of the esophageal mucosa are factors that favour malignant transformation. The cancers are usually squamous cell carcinoma (90 %) and more rarely adenocarcinomas (10 %). Regular surveillance of patients with a history of achalasia thus appears to be justified, particularly where there has been therapeutic failure. However, studies are still necessary to evaluate the benefits of endoscopy and determine the optimal time interval between follow-up visits. The development of cancer on a benign esophageal stenosis is much more rare and is quite often related to a caustic stenosis. The risk of the appearance of an oesophageal cancer 24 years after the ingestion of caustic substances is 1000-fold. The prevalence of antecedents of caustic burns in cancers of the esophagus varies from 1 to 4 %. The surveillance of patients having suffered such burns is therefore, in theory, justified. The other etiologies (coeliac sprue, endoscopic sclerotherapy, gastrectomies, oesophageal membranes and diverticula, chronic oesophagitis) do not justify systematic screening. Certain genetic predispositions must be recognised. Tylosis is a rare genetic disease characterised by a hyperkeratosis of the palms and soles which is transmitted as an autosomal dominant trait. It is associated with a thousands fold risk of esophageal squamous cell carcinoma as compared with the general population. It has been suggested that this syndrome could be linked to an anomaly in the metabolism of vitamin A. Regular surveillance of patients with tylosis is thus justified. It is worth noting that epidemiological studies have reported familial aggregations of cases of esophageal cancer in several regions of the world. The frequency of adenocarcinomas, which represent, on average, 10 % of cancers of the esophagus, has considerably increased over the past few decades in Northern Europe and the U.S.A., the majority of them avising from Barrett's oesophagus (BE) [6]. The presence of BE increases the risk of cancer by a factor of 30 to 50 as compared with the general population. However, the prevalence of BE in the general population is greatly underestimated in endoscopic series and a certain number of arguments seem to indicate, that surveillance of BE has a minor impact on global mortality rates due to adenocarcinoma of the esophagus or the gastro-esophageal junction. The absence of any impact on death rates for the general population does not, however, signify that it is of no interest on an individual level. Several studies have demonstrated that not only the percentage of cancers diagnosed at an early stage was higher in patients who had undergone periodic endoscopic surveillance than in those who had not, but also the survival rate was significantly better when the tumours were discovered during follow-up rather than fortuitously. The inclusion of a patient into a surveillance protocol must take into account age and state of health before considering surgical treatment. Only BE, in the presence of a specialised columnar epithelium, regardless of whether segmental involvement is long or short justifies endoscopic follow-up. Certain patients such as obese men and excessive smokers and drinkers, with a long segment of BE and antecedents of Barret's ulcer or peptic stenosis would be more at risk of cancer and could justify closer surveillance. The follow-up of patients with BE is based, at the moment, on endoscopy with staged biopsies (cartography) in the search for dysplasia, the only recognised marker of a high risk of cancer. This marker poses a certain number of problems, namely poor diagnostic reproducibility and that the disease's natural course is not well known on an individual level. Even if no surveillance protocol has been perfectly codified, a few recommendations can be deduced from present knowledge of the natural evolution of dysplasia. Bi-annual follow-up can be recommended for patients with no dysplasia, however, if low-grade lesions are confirmed by two successive examinations, closer surveillance is necessary, every six months or yearly according to some authors. There is controversy surrounding the attitude to be adopted for high-grade dysplasias: same authors recommend esophagectomy but others perform this surgery only if there is also histological proof of cancer. The development of techniques for physical photodestruction of glandular metaplasia (electrocoagulation with argon plasma or YAG laser) could modify these strategies. The second condition necessary for the diagnosis of early-stage esophageal carcinoma is to carry out appropriate investigations in high-risk patients [2]. Abrasive balloon cytology consists of introducing a balloon probe into the gastric cavity of patients, inflating and then withdrawing it through the esophagus. Exfoliated esophageal cells are then collected by the abrasive balloon surface (silk netting, silicate crystals etc…). In endemic areas, such as Henan province in China or certain regions of Iran, this technique appears to be justified for mass screening. In Western countries, of low or medium risk, screening should be individualised and reserved for high-risk patients. This technique is generally considered to be inappropriate due to the risks linked to esophageal varices or gastric ulcers, which are frequent in these patients. Even though endoscopy [7 8] remains the method of choice for the confirmation or elimination of a suspicion of esophageal carcinoma in situ, it is far from being infallible. The arrival of fibroscopy then videoendoscopy did not modify detection statistics for early stage esophageal carcinoma in Western countries. According to Savary [9], the failure of endoscopy to provide an early diagnosis stems from our insufficient knowledge of the mucosal patterns of early cancer in squamous epithelium, a cursory examination of the esophagus, and the absence of recourse to staining. Endoscopic morphologic descriptions are protean. After an exhaustive study, the Chinese have described four types: occult, erosive, nodular and rare papillary. A French study has proposed adding to the former, a verrucous type which consists of a lesion that is both nodular and depressed with an irregular, mamillated surface and an infiltrating type which appears as an ulcerated zone that is thickened and fungated. Mixed features may also be observed. Erythematous flat forms are difficult to see even if they are very often multicentric and multifocal. The occult type is neither visible to the naked eye nor by endoscopy and can only be detected after staining. The classification of superficial cancers of the glandular type (adenocarcinoma) is identical to that of squamous cell carcinoma [2 3 4 5 6 7 8 9 10]. Detection of lesions is improved by the use of staining techniques. Toluidine blue [8 9 10 11], a metachromatic stain from the thiazine group, has a particular affinity for RNA and DNA and therefore stains suspect areas in the mucosa, which are richer in nuclei than normal mucosa. It reveals pre-invasive lesions, demonstrates their multicentricity, and locates satellite centres and flat or occult forms of dysplasia or carcinoma in situ. False negatives can be attributed to submucosal infiltration of the tumour or hyperkeratosis of a lesion which is more easily seen macroscopically since there is often leucoplakia or an exophytic appearance. The sensitivity is 82 % and the specificity 75 %. Lugol's solution [12 13] reacts specifically with glycogen in the epithelium, staining normal mucosa brown, whereas inflamed areas, gastric heterotopia and cancer are not stained. It is easier to use than toluidine blue and produces fewer false negatives. In the study by Fagundes et al. [14], Lugol was found to have a sensitivity of 46 %, a specificity of 90 %, a positive predictive value of 26 % and a negative predictive value of 96 %. In this study the chances of discovering oesophageal dysplasia were eight times higher in unstained areas when compared with areas that stained uniformly. The dyeing capabilities of methylene blue, which stains absorbent cells, can be used to identify foci of intestinal metaplasia in Barret's mucosa during endoscopy. These staining techniques should be used in high-risk patients if the performance of endoscopy is to be improved. New diagnostic methods such as autofluorescence [15] are under evaluation. Neoplastic tissues emit a different fluorescence spectrum from tumorous tissue when excited by a laser beam. The results, to date, seem to be of value in the detection of endoscopically normal lesions in Barret's epithelium. Likewise, potential markers have been studied such as cellular proliferation modifications and the DNA content of epithelial cells, using, in particular, flux cytometry, and above all the search for genetic alterations in Barret's mucosa, the most important of which is mutation of the p53 gene. Detection of this mutation would allow physicians to identify high-risk patients very early, confirming the presence of a genuine neoplastic process in Barret's epithelium.
gastroenterology & hepatology,surgery