BCL2 Rearrangement in Follicular Lymphoma
In this study, we included 53 patients with FCL presenting in the skin and further evaluated if clinical and morphologic features may be of diagnostic value between PCFCL and SSIFCL in view of the data obtained by further staging procedures, as in the real-life setting. We also evaluated the diagnostic and prognostic value of BCL2 gene rearrangement detected by FISH in seven (13.2%) of 53 patients, including four (8.5%) of 47 with PCFCL and three (50%) of six with SSIFCL. Then, the presence of a BCL2 rearrangement was found significantly associated with positivity of initial staging procedures (P = .02) but cannot serve as a diagnostic marker between the two entities since it may be observed in some PCFCL cases. The long follow-up period (median, 39 months) also showed that BCL2 rearrangement was associated with a higher risk of secondary extracutaneous spreading in patients with PCFCL (P = .05).
Given these results, our study suggests that FISH could be included as an ancillary procedure in FCL presenting in the skin, especially in cases with CD10/BCL2 coexpression. Such coexpression was observed in all SSIFCL cases and in 35% of PCFCL cases, including three of five PCFCL cases with secondary extracutaneous involvement. The single expression of BCL2 observed in more than 50% of PCFCL cases was not found correlated with such clinical evolutive features. In different studies, BCL2 expression in PCFCL has varied considerably, ranging from 0% to 86%. Moreover, the presence of T cells strongly immunoreactive for the BCL2 protein requires a careful comparison of BCL2, CD3, and CD20 immunostaining as well as the use of a high cutoff (at 50%) to assess the BCL2 positivity of an FCL case. By comparison with indolent cases, the architectural pattern of the 11 cases with extracutaneous involvement was not found specific with either a follicular growth pattern with dense infiltration of the dermis (type A pattern) or a focal periadnexal infiltrate (type B pattern), even if the type A pattern was observed more frequently in SSIFCL. Five of these 11 patients, however, were harboring a BCL2 rearrangement. Therefore, only combined BCL2 and CD10 expression could be relevant to select patients for FISH testing among any other histologic criteria. Analyzing a surgical biopsy specimen including the reticular dermis and fat rather than a punch biopsy specimen, as recommended by others, permitted us to increase the reliability of histopathologic analysis of the above criteria and FISH testing on a representative tumor sample.
Moreover, bone marrow biopsy has been considered an "optional" staging procedure in PCFCL according to the International Society for Cutaneous Lymphoma/European Organization for Research and Treatment of Cancer recommendations. Indeed, in patients with PCFCL and otherwise negative staging, bone marrow biopsy is often not performed in our national practice. Only one of our patients (of 43 patients with bone marrow examination) displayed a bone marrow infiltration as the only evidence of extracutaneous disease. Interestingly, this case displayed a BCL2/CD10 phenotype associated with BCL2 rearrangement. In the largest series reported so far by Senff et al, 22 (11%) of 193 patients with FCL first presenting in the skin showed a bone marrow infiltration. However, it was the only evidence of extracutaneous disease in nine of these 22 patients (nine of 193 [4.7%]), in accordance with the rate (2.3%) observed here. Moreover, the 5-year overall and disease-specific survival of these nine patients was 44% and 63%, respectively, compared with 84% and 95%, respectively, in 157 patients without extracutaneous disease. These data may support that bone marrow examination should be considered a systematic staging procedure in patients with FCL presenting in the skin. Nevertheless, isolated bone marrow involvement would not directly change the therapeutic management, since hematologists recommend a "watch-and-wait strategy" for FCL without clinical symptoms or cytopenias. Our data therefore suggest evaluating the presence of BCL2-CD10 coexpression and BCL2 rearrangement to determine if these features are recurrently observed by other groups in FCL cases with extracutaneous spreading, primarily or secondarily. The small subset of PCFCL cases exhibiting a BCL2 rearrangement seems to harbor a disease closer to nodal FCL, since they seem to be at risk for developing secondary extracutaneous spreading and should require adequate clinical monitoring.
FISH-positive cases have been further analyzed by PCR amplification of BCL2-JH rearrangement using the BIOMED-2 protocol. Here, molecular detection provided concordant data on skin specimens with FISH in all cases but one. Moreover, the absence of BCL2 rearrangement in peripheral blood samples of patients with PCFCL permitted excluding skin contamination with BCL2-rearranged circulating B cells, as described in healthy individuals but also in patients with FISH-negative PCFCL. For such reason, we would recommend performing FISH on skin sections rather than PCR to assess the presence of BCL2 rearrangement within a significant proportion of tumoral cells to avoid false positivity due to bystander cells, as previously reported.
Besides BCL2 rearrangement detection, we also investigated other potential prognostic factors for cutaneous relapse and extracutaneous spreading. In this study, 20 (42.6%) of 47 patients had cutaneous relapse, a rate observed in other studies. As reported by others, clinical features such as age or sex did not influence the DFS in PCFCL. We observed that predominant small/medium centrocytic cell morphology was statistically associated with a higher probability of cutaneous relapse(s) in the univariate and multivariate analysis. We assessed this prognosis factor using both the usual nonparametric log-rank test that considers only the first cutaneous relapse for each patient and the shared γ-frailty model that considers all cutaneous relapses for each patient. We also believe that such a model could be particularly relevant to the study of disease progression in different types of cutaneous lymphomas in which the number of relapses and their time interval may represent an indicator of the disease aggressiveness, not otherwise modeled so far in this field.
Finally, our study identified that BCL2 rearrangement is a rare event in PCFCL (8.5%) but may identify patients at risk of extracutaneous involvement together with other associated features, such as type A architectural pattern and BCL2-CD10 coexpression. Alternatively, we suggest searching a BCL2 rearrangement by FISH using separation probes in cases harboring the above features. Moreover, small/medium centrocyte morphology, and potentially BCL2/CD10 coexpression, seems to be predictive for a higher risk of cutaneous relapse.
Discussion
In this study, we included 53 patients with FCL presenting in the skin and further evaluated if clinical and morphologic features may be of diagnostic value between PCFCL and SSIFCL in view of the data obtained by further staging procedures, as in the real-life setting. We also evaluated the diagnostic and prognostic value of BCL2 gene rearrangement detected by FISH in seven (13.2%) of 53 patients, including four (8.5%) of 47 with PCFCL and three (50%) of six with SSIFCL. Then, the presence of a BCL2 rearrangement was found significantly associated with positivity of initial staging procedures (P = .02) but cannot serve as a diagnostic marker between the two entities since it may be observed in some PCFCL cases. The long follow-up period (median, 39 months) also showed that BCL2 rearrangement was associated with a higher risk of secondary extracutaneous spreading in patients with PCFCL (P = .05).
Given these results, our study suggests that FISH could be included as an ancillary procedure in FCL presenting in the skin, especially in cases with CD10/BCL2 coexpression. Such coexpression was observed in all SSIFCL cases and in 35% of PCFCL cases, including three of five PCFCL cases with secondary extracutaneous involvement. The single expression of BCL2 observed in more than 50% of PCFCL cases was not found correlated with such clinical evolutive features. In different studies, BCL2 expression in PCFCL has varied considerably, ranging from 0% to 86%. Moreover, the presence of T cells strongly immunoreactive for the BCL2 protein requires a careful comparison of BCL2, CD3, and CD20 immunostaining as well as the use of a high cutoff (at 50%) to assess the BCL2 positivity of an FCL case. By comparison with indolent cases, the architectural pattern of the 11 cases with extracutaneous involvement was not found specific with either a follicular growth pattern with dense infiltration of the dermis (type A pattern) or a focal periadnexal infiltrate (type B pattern), even if the type A pattern was observed more frequently in SSIFCL. Five of these 11 patients, however, were harboring a BCL2 rearrangement. Therefore, only combined BCL2 and CD10 expression could be relevant to select patients for FISH testing among any other histologic criteria. Analyzing a surgical biopsy specimen including the reticular dermis and fat rather than a punch biopsy specimen, as recommended by others, permitted us to increase the reliability of histopathologic analysis of the above criteria and FISH testing on a representative tumor sample.
Moreover, bone marrow biopsy has been considered an "optional" staging procedure in PCFCL according to the International Society for Cutaneous Lymphoma/European Organization for Research and Treatment of Cancer recommendations. Indeed, in patients with PCFCL and otherwise negative staging, bone marrow biopsy is often not performed in our national practice. Only one of our patients (of 43 patients with bone marrow examination) displayed a bone marrow infiltration as the only evidence of extracutaneous disease. Interestingly, this case displayed a BCL2/CD10 phenotype associated with BCL2 rearrangement. In the largest series reported so far by Senff et al, 22 (11%) of 193 patients with FCL first presenting in the skin showed a bone marrow infiltration. However, it was the only evidence of extracutaneous disease in nine of these 22 patients (nine of 193 [4.7%]), in accordance with the rate (2.3%) observed here. Moreover, the 5-year overall and disease-specific survival of these nine patients was 44% and 63%, respectively, compared with 84% and 95%, respectively, in 157 patients without extracutaneous disease. These data may support that bone marrow examination should be considered a systematic staging procedure in patients with FCL presenting in the skin. Nevertheless, isolated bone marrow involvement would not directly change the therapeutic management, since hematologists recommend a "watch-and-wait strategy" for FCL without clinical symptoms or cytopenias. Our data therefore suggest evaluating the presence of BCL2-CD10 coexpression and BCL2 rearrangement to determine if these features are recurrently observed by other groups in FCL cases with extracutaneous spreading, primarily or secondarily. The small subset of PCFCL cases exhibiting a BCL2 rearrangement seems to harbor a disease closer to nodal FCL, since they seem to be at risk for developing secondary extracutaneous spreading and should require adequate clinical monitoring.
FISH-positive cases have been further analyzed by PCR amplification of BCL2-JH rearrangement using the BIOMED-2 protocol. Here, molecular detection provided concordant data on skin specimens with FISH in all cases but one. Moreover, the absence of BCL2 rearrangement in peripheral blood samples of patients with PCFCL permitted excluding skin contamination with BCL2-rearranged circulating B cells, as described in healthy individuals but also in patients with FISH-negative PCFCL. For such reason, we would recommend performing FISH on skin sections rather than PCR to assess the presence of BCL2 rearrangement within a significant proportion of tumoral cells to avoid false positivity due to bystander cells, as previously reported.
Besides BCL2 rearrangement detection, we also investigated other potential prognostic factors for cutaneous relapse and extracutaneous spreading. In this study, 20 (42.6%) of 47 patients had cutaneous relapse, a rate observed in other studies. As reported by others, clinical features such as age or sex did not influence the DFS in PCFCL. We observed that predominant small/medium centrocytic cell morphology was statistically associated with a higher probability of cutaneous relapse(s) in the univariate and multivariate analysis. We assessed this prognosis factor using both the usual nonparametric log-rank test that considers only the first cutaneous relapse for each patient and the shared γ-frailty model that considers all cutaneous relapses for each patient. We also believe that such a model could be particularly relevant to the study of disease progression in different types of cutaneous lymphomas in which the number of relapses and their time interval may represent an indicator of the disease aggressiveness, not otherwise modeled so far in this field.
Finally, our study identified that BCL2 rearrangement is a rare event in PCFCL (8.5%) but may identify patients at risk of extracutaneous involvement together with other associated features, such as type A architectural pattern and BCL2-CD10 coexpression. Alternatively, we suggest searching a BCL2 rearrangement by FISH using separation probes in cases harboring the above features. Moreover, small/medium centrocyte morphology, and potentially BCL2/CD10 coexpression, seems to be predictive for a higher risk of cutaneous relapse.