I did not see the positive Tdt . But if you have changes of phenotype that you cannot explain, you should compare the clonal identy by JH -PCR.The phenotype in the second biopsy would correspond to a lymphoblastic B precursor- Could be secondary?

Tdt positivity is uneven and raises the possibility of false positivity. One can see many cytoplasmic positive cells; exclusive nuclear positivity is visible in a minority of cells.  
The morphology is typical of burkitt. A second clone with new cytogenetic changes is possible.  
t 8:14 can be present in lymphoblastic lymphoma, but i was not able to find tdt positivity in burkitt. Finally, many published cases of burkitt gastric lymphoma are associated with H.p.

This is very challenging case. We have recently seen similar case of a child who developed secondary acute lymphoblastic leukemia with 11 q23 abnormalities following therapy for Burkitt lymphoma (approximately one year after therapy for BL).  
 
Though we believe that this patient might also develop secondary acute lymphoblastic leukemia following therapy for his BL, expression of the CD10 by lymphoma would be unusual for pro-B ALL associated with 11 q23 abnormalities. TdT positivity (variable) is not an artifact, it looks real. Lack of bone marrow involvement is consistent with extramedullary presentation of ALL.  
 
Alternatively, there might be different mechanism explaining the unusual immature blastic phenotype of this patent lymphoma in 2008 biopsy. We have also recently seen another patient , who presented with precursor B lymphoblastic (blastoid) transformation of follicular lymphoma grade I demonstrating TdT+, Pax5+, CD79a+, CD10+, BCL6+,BCL2+, TdT+(variable) phenotype. Cytogenetic analysis showed in transformed cells the presence of c-myc rearrangement and t (14;18).  
 
Searching the literature we found an interesting paper by Hassan r et al. Europ J Haematol 2008; 80 (265-270), describing a case of Burkitt lymphoma/ leukemia transformed form a precursor B cell. In this paper the authors report a case of BL/ L with a FAB L3 morphology with phenotype and genotype characteristic of CD10+ pre- B-ALL associated with t(8;14) q 24;q32. In this child BL/L with pre-B phenotype was a primary presentation of lymphoma and there was no history of preceding BL with mature phenotype. However, this paper demonstrates that cells with Burkitt morphology may have phenotype of precursor B cell. Another similar case was described by R. Kamrokji et al. in Leukemia Research 2003; 27: 561-566.  
 
It would be very helpful if cytogenetic analysis of primary gastric tumor of this patient from 2007 and current biopsy of lymphoma from 2008 were available, at least FISH for c-myc rearrangement. It is possible that lymphoma cells in the biopsy from 2008 and 2007 may be related. To determine clonal relation we would also suggest performing PCR for Ig GR.  
 
We would also recommend careful reevaluation of BM biopsy for possible minimal involvement by lymphoma. Does the patient have circulating lymphoma cells in peripheral blood?  

Thank you for comments. Some points: 1. FALSE positivity: practically I never seen NUCLEAR and esp. TdT false positive. 2. TdT gain is not single change: we lost IgM and Bcl6 in the same laboratory IH environment (2 biopsies are stainde in the same time and conditions). 3. Material is scant- and I have not a possibility for c-myc now. So I would appreciate external help in this case. I can try to check IgH, but in this case material may be exhausted. 4. Nuclear infoldings/clefts esp. in IH stains are prominent. Is it fully compatible with Burkitts in the last biopsy? 5. Original 2007 BM changes are chalenging- THE DEFINITE CD20+ large cell infiltrate was present, but I have lost it in recutt slides, so expanded IH panel was not efective. repeated BM biopsy do not reveal any changes except slightly more numerous TdT+ intersticial precursors...

very interesting case and we get alot of knowledge. Thank you doctor for very informative.

After our clinicians, the tumor progression is seen DURING chemotherapy + immunotherapy (anti CD20) process (not responding tumor).

I have little to add to Dr. Hyjek's comments. I have never been faced to TdT positivity in relapsed or therapy-refractory Burkitt lymphomas, but I have seen follicular lymphomas with TdT+ high grade transformation. c-myc FISH could bring some clarity in this case. I consider the TdT-positivity in the 2008 biopsy specific. we have to consider that the given tretament regimens could have exherted a profound mutational pressure on the primary lymphoma and some of the changes in the second biopsy, e.g. lost CD20, probably lost IgM expression, could be ralted to the treatment... would a specific interparetion (transformation/de novo) change the treatment strategy?

It is indeed a very interesting case. The patient was admited to our department for the treatment of primary progression of his lymphoma (the primary disease was treated with intensive Burkitt-like chemotherapy and anti CD20 immunotherapy). He has a very bulky abdominal disease, hepatic and kidney lesions. The loss of CD20 is well described in case of rituximab treatment.  
He has received NHL type salvage therapy with some response. Lymphoblastic lymphoma may better respond to ALL type therapy however.

Sorry for the late comment. The TdT stain looks real to me. In fact there is little to add to the previous comments except that I would like to see images of both BMB's including the results of TdT on both BMB's.

The patient deceased due to basal meningitis. Autopsy was declined and cause of death was not evaluated (fungi? tumor? etc.?). B clonality was not done. There was no signs of BM involvement. So primary threpine biopsy was the single positive.

The last point: the patient was found to be HIV+.