Statistics from Altmetric.com
We read with interest the article by Dinis-Ribeiro et al addressing the follow up of “atrophic chronic gastritis and intestinal metaplasia (IM)”.1 The authors conclude that: (1) “in patients with atrophic chronic gastritis or with type I IM, a three yearly follow up could be suitable”; and (2) “patients with type III IM may benefit from 6–12 monthly (follow up?)”.
How precancerous lesions are histologically assessed and followed up are fields of our interest and we would raise some methodological concerns about the published study.
In assessing atrophy, it would be advisable to adopt the classification(s) proposed by the current international literature. The original Sydney system was recently revised by a group of specialists in gastrointestinal pathology (including the authors of the original classification), prompting important changes in the previous diagnostic criteria.2,3 The new version was also validated by testing its interobserver consistency. The adoption of such internationally shared criteria facilitates comparisons between studies.
As for the histological classification of dysplasia, the Dinis-Ribeiro study applied the Vienna criteria, which include category 4.3 (suspicious for invasive carcinoma) among the non-invasive neoplasia categories (NiN). From a biological standpoint at least, this category is quite distinct from the NiN categories. Recently, two classifications have been proposed for gastric NiN arising in the stomach; here again, adopting the World Health Organisation criteria would enable an easier comparison between this and other studies.4,5
Finally, the authors report that the two pathologists assessing the slides agreed in 85% of cases; it would have been better to express interobserver consistency properly, in terms of K statistics.
To define the “entry biopsy” as “first or intermediate” is a contradiction in terms, which may introduce a bias in the calculation of the follow up time and which influences the validity of the results. The authors state that 144 patients were included in the study and, a few lines later, that 239 pairs of endoscopy biopsies were considered. In view of the fact that they also say that no less than two biopsy samples were taken at each endoscopy, the numbers become bewildering.
In dealing with precancerous lesions, extensive sampling protocols (always including the angular mucosa) are mandatory.5,6 To say that “more than 15% of patients had more than four biopsies for each endoscopy” is not satisfactory, either for the patient’s safety or for any speculation vis-à-vis the “follow up model”—particularly because the follow up ranged from 3.2 to 36.2 months in 41 of the 144 patients.
An important outcome of the study would be the demonstration that low grade NiN can progress to more severe lesions (invasive or non-invasive?), but the clinical value of this observation is considerably reduced by the short follow up and the difficulty in correlating the number of biopsy samples (239) with the number of patients (144).
On the whole, we found the message emerging from the Denis-Riberio study a valuable contribution to our understanding of the natural history of gastric carcinogenesis. Our critical comments are intended simply as a reminder that caution is needed in recommending follow up protocols unless all the essential conditions can be met to support such recommendations.