• MALT lymphoma
• gastric lymphoma
• histological grading

Gastric MALT lymphoma (GML) development is closely associated with Helicobacter pylori infection cases.¹ The majority of stage IE GML regress following H pylori eradication but assessing cure of the disease requires prolonged follow up. Residual lymphoid infiltrate in post-treatment gastric biopsies can be very difficult to interpret and histological criteria for the diagnosis of minimal residual disease or complete remission are not clearly defined. Molecular follow up by polymerase chain reaction (PCR) for the rearranged immunoglobulin heavy chain variable region shows that persistent monoclonal bands is observed in 44% of cases showing apparent complete histological remission.² The significance of ongoing PCR monoclonality in the absence of histological disease is still under investigation.

Thus histological evaluation of gastric biopsies remains the cornerstone to assess lymphoma response to therapy. In 1993, Wotherspoon et al proposed a histological scoring system that was initially designated to express the degree of confidence of a diagnosis of GML on gastric biopsies.³ This histological scoring has been used to evaluate the response to therapy in a number of subsequent trials but many investigators have found the system difficult to apply and of low interobserver reproducibility. Other studies have used the criteria of partial and complete remission defined by Neubauer and colleagues.⁴ Criteria of lymphoma response to therapy need to be standardised using a system that can be easily applied so that results of future clinical trials can be compared.

As part of multicentre clinical trials on GML, GELA (Groupe d’Etude des Lymphomes de l’Adulte) pathologists and one of the authors (ACW) established a post-treatment histological grading system based on evaluation on haematoxylin-eosin (H&E) stained sections of three essential diagnostic features: the lymphoid infiltrate, presence of lymphoepithelial lesions (LEL), and stromal changes. We classified the morphological features observed in post-treatment gastric biopsies as follows: “complete histological response” (CR), “probable minimal residual disease” (pMRD), “responding/residual disease” (rRD), and “no change” (NC) (table 1). These groups give clinically relevant information to the clinician. In particular, the category responding/residual disease (rRD) implies that overt lymphoma is present in association with features that suggest a degree of regression. This would imply to the clinician an ongoing response that does not require immediate use of alternative therapies.

To assess the reproducibility of this histological grading system, we selected at random 10 patients with GML enrolled in the GELA clinical trial (seven men and three women; median age 60 years (range 35–74)).⁵ A total of 45 sets of gastric biopsies stained with H&E were evaluated separately by each histopathologist blind to the clinical follow up data using the new follow up system. Three to six sequential gastric biopsies were analysed for all patients with a mean follow up of 19 months after H pylori eradication therapy. Interobserver agreement evaluated by the weighted kappa value gave excellent results, with values over 0.84, indicating very good agreement among the seven observers.

Assessing the lymphoma remission status is of great importance for clinical practice. Developing tools to evaluate residual disease are needed, not only for clinical practice but also to conduct clinical trials that aim to define therapeutic guidelines. We propose in this study a histological grading system for the evaluation of post-treatment gastric biopsies. Testing of this scheme in a small number of cases within the group developing this scheme has shown it to be highly reproducible. These results encourage further evaluation of this scheme on larger series, as well as investigation of its clinical significance and impact on clinical guidelines. In combination with molecular studies, this scheme could provide an interesting tool for the evaluation of residual disease in prospective studies on GML.