The differeantial is that of an atypical marginal zone hperplasia versus marginal zone lymphoma, pediatric type. PCR clonality proof should distinguish between both; since in that case the population is clonal, I would dare to diagnose here the latter. How old is the patient?

18 yrs old. Thank you.

I agree with Alexander that reactive lymphoid hyperplasia should be excluded before the diagnosis of nodal marginal zone lymphoma is made and the presence of clonal IgH rearrangements the neoplastic nature of the process. Pediatric NMZL has an excellent prognosis and most of the time excisional biopsy is curative.  
In general, dealing with children thought me to be very careful in making the diagnosis of lymphoma, particularly if it is of a rare type. In a case like this one, I would consider florid PTGC in the differential diagnosis, particularly with these expended CD21+ follicular dendritic meshworks. Adolescent boys can have a very high number of PTGC. I find BCL2 not very helpful in distinguishing between neoplastic and reactive follicles in children since in PTGC there is a high number of mantle B-lymphocytes and T-cells that are BCL2 positive and some pediatric nodular lymphomas such as follicular, are usually BCL2 negative. CD3 is very helpful to determine the T-cell component. CD43 is not very helpful since both reactive nodal B-cell proliferations and lymphomas can be positive. The atypical marginal zone hyperplasia of MALT as mentioned is somewhat lower on my differential since it is usually seen in the tonsils and appendices of younger children but not so much in lymph nodes. Another important point in children is the presence of reactive monotypic proliferations, that some times can be even clonal by molecular techniques.  
Hope this helps.

Thank you. Really I have a clone and histologically prefer MzL. Fortunatelly only "watching and waiting" will be applied, so any differencies cannot happen between categories. But the question left: CAN WE (pathologists) DISCRIMINATE THESE LESIONS? With the best wishes.

The morphology has its limitations. We can do our best, but other studies may prove more helpful in the future. Some time the question is not how much we can do but how much what we do is clinically relevant. I admire people practicing hemepath with limited resource and relying mostly on their morphologic skills.