Period 1 included 54 patients; period 2 included 61 patients. The difference in survival was significant (P = .01).
A, Survival of patients according to the most intensive therapy received. The therapies consisted of a combination of rituximab and polychemotherapy (PCT) with anthracycline, rituximab-PCT without anthracycline, PCT with anthracycline (without rituximab), PCT only (without anthracycline and/or rituximab), and less-intensive therapies only (including single-drug chemotherapy and radiotherapy). The global difference between the curves was significant (P = .002). B, Sixty-three patients received treatment at any time with rituximab-PCT; 52 patients only received other treatments (including PCT without rituximab). The difference between rituximab-PCT regimens and other treatments was significant (P < .001).
Grange F, Joly P, Barbe C, Bagot M, Dalle S, Ingen-Housz-Oro S, Maubec E, D’Incan M, Ram-Wolff C, Dalac S, Templier I, Esteve E, Quereux G, Machet L, Leduc M, Dereure O, Laroche L, Saiag P, Vergier B, Beylot-Barry M. Improvement of Survival in Patients With Primary Cutaneous Diffuse Large B-Cell Lymphoma, Leg Type, in France. JAMA Dermatol. 2014;150(5):535-541. doi:10.1001/jamadermatol.2013.7452
Primary cutaneous diffuse large B-cell lymphoma, leg type (PCDLBCL-LT), occurs in elderly patients and has been considered as a lymphoma with a poor prognosis, with estimated 5-year specific survival rates of approximately 50%. The hypothesis of an improvement in prognosis over time has not been studied.
To evaluate this hypothesis in a large series of patients and investigate factors associated with prognosis as well as improvement in the prognosis.
Design, Setting, and Participants
A retrospective multicenter study was conducted including dermatology departments belonging to the French Study Group on Cutaneous Lymphoma. Participants were 115 patients with PCDLBCL-LT diagnosed between 1988 and 2003 (period 1) or between 2004 and 2010 (period 2).
Main Outcomes and Measures
Age, sex, period of diagnosis, number of skin lesions, tumor stage, tumor location (leg vs nonleg), lactate dehydrogenase level, type of therapy (with or without a combination of rituximab and polychemotherapy [PCT]), and outcome were recorded. Baseline characteristics and outcome were compared according to period of diagnosis and type of therapy. Prognosis factors were identified by univariate and multivariate survival analyses.
The mean age of the patients was 76.9 years, and 47% of the patients were older than 80 years. The 3- and 5-year specific survival rates improved between period 1 and period 2, from 55% to 74% and from 46% to 66%, respectively (P = .01). Patients had similar baseline characteristics during both periods, but rituximab-PCT regimens were administered to 88.5% of the patients in period 2 vs 16.7% in period 1 (P < .001). The 3- and 5-year specific survival rates were 80% and 74%, respectively, in patients who received a rituximab-PCT regimen compared with 48% and 38% in those who received less-intensive therapies. No significant difference was observed between both groups in age and baseline prognostic factors. In multivariate analysis, treatment without rituximab-PCT was the only adverse prognostic factor (odds ratio, 4.6 [95% CI, 2.4-9.1]; P < .001), whereas the number of skin lesions (P = .06) and location on the leg (P = .07) had only borderline significance.
Conclusions and Relevance
A major improvement in the survival of patients with PCDLBCL-LT has occurred over time in France, mainly as a result of the use of intensive rituximab-PCT regimens in most patients, including very elderly ones. Until further prospective clinical trials are conducted, such regimens should be considered as the standard of care in these patients.
Primary cutaneous diffuse large B-cell lymphoma, leg type (PCDLBCL-LT), was recognized as a unique entity in the World Health Organization/European Organization for Research and Treatment of Cancer classification of cutaneous lymphomas in 2005.1 This lymphoma is characterized clinically by its occurrence in elderly persons with a preferential (but not exclusive) leg localization and histologically by confluent sheets of medium to large cells with round nuclei typically positive for CD20, MUM-1, and BCL2 in the absence of t(14;18) translocation, which resembles the activated B-cell type of nodal diffuse large B-cell lymphomas (DLBCLs).1- 7 In contrast with the more frequent other subtypes of cutaneous B-cell lymphomas (ie, primary cutaneous marginal zone B-cell lymphoma and primary cutaneous follicle center lymphoma), PCDLBCL-LT has a poor prognosis, with 5-year survival rates ranging from 41.0% to 73.1% in studies4,6,8- 11 and approximately 50% to 55% in most reviews.1,7,12,13
A combination of rituximab and the standard cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP) regimen has been recognized as the standard therapy for patients with DLBCLs.14,15 Although PCDLBCL-LT is a specific disease entity occurring in older patients, this combination regimen was also used more frequently in this rare cutaneous lymphoma or replaced by less-intensive combinations of rituximab and polychemotherapy (PCT) in the oldest and/or frailest patients.
In previous studies,16 we provided preliminary data suggesting an improvement of prognosis in patients with PCDLBCL-LT as a result of a combination of rituximab and PCT and/or other improvements in treatment management. At the outset of the present study, we hypothesized that specific survival of patients with PCDLBCL-LT had increased over time, and then investigated for factors associated with such an evolution, focusing on period of diagnosis and type of treatment.
The study was approved by the institutional review board of Reims University Hospital, Reims, France. A retrospective review was carried out on cases of PCDLBCL-LT included in the registry of the French Study Group on Cutaneous Lymphoma (FSGCL) between January 1, 1988, and December 31, 2010. Inclusion criteria were (1) a histologic diagnosis of PCDLBCL-LT according to standard criteria1 as assessed by an expert panel of pathologists, (2) the absence of extracutaneous localization at diagnosis, and (3) available clinical data in the patient’s medical records. All medical records were reviewed.
The following clinical characteristics at diagnosis were recorded: date of diagnosis, age, sex, performance status (ie, 0, asymptomatic; 1, symptomatic but completely ambulatory; 2, symptomatic, with <50% of the time spent in bed during the day; 3, symptomatic, with >50% of the time spent in bed during the day; and 4, bedridden), anatomic site (head or neck, arm, anterior aspect of the trunk, posterior aspect of the trunk, or leg), number of skin lesions, TNM clinical category at diagnosis according to the International Society for Cutaneous Lymphomas/European Organization of Research and Treatment of Cancer classification,17 lactate dehydrogenase level, and B symptoms (ie, systemic symptoms including fever, sweats, and/or weight loss). According to date, 2 diagnosis periods were defined: from January 1, 1988, to December 31, 2003 (period 1), and from January 1, 2004, to December 31, 2010 (period 2). These 2 periods were a priori defined for comparison because we hypothesized that significant changes in management of the care of patients with PCDLBCL-LT progressively occurred from 2003 to 2004, with a possible effect on clinical outcome. Data on therapy included initial and subsequent therapies and whether the patient received rituximab-PCT at any time. For this purpose, the most intensive therapy received at any time was recorded, including rituximab-PCT with anthracycline (ie, standard or age-adapted rituximab-CHOP), rituximab-PCT without anthracycline, PCT (without rituximab and with or without anthracycline), and less-intensive therapies (including radiotherapy, corticosteroids, and/or single-agent chemotherapy, as well as rituximab alone). Follow-up data were recorded until December 31, 2011, including achievement of a complete response, cutaneous relapse, extracutaneous progression of the disease, final status, and date and cause of death. Causes of death were ascertained by physician members of the FSGCL who monitored patients in most cases and in some cases by questioning their general practitioners. The follow-up time ranged from 8 days to 159 months (mean, 39 months). Eighty-five patients (73.9%) were monitored until death, until the end point of the study, or for longer than 5 years; 30 patients (26.1%) were lost to follow-up after less than 5 years (range, 4.2-55.5 months).
Data are described using mean (SD) for quantitative variables and number and percentages for qualitative variables. Comparisons between subgroups of patients according to diagnosis period and type of therapy were performed using a χ2 test; Fisher exact test; 2-tailed, unpaired t test; or Wilcoxon rank-sum test, as appropriate.
Specific survival duration was calculated from diagnosis to date of disease-related death or censoring. Patients whose deaths were unrelated to lymphoma were considered censored. The survival curves according to period of diagnosis and type of treatment were established by the Kaplan-Meier method. Prognostic factors were identified by univariate analysis using the log-rank test and by multivariate analysis using a Cox proportional hazards regression model. Factors significant at the 0.20 level in univariate analysis were included in a stepwise regression multivariate analysis with entry and removal limits set at 0.20. P < .05 was considered significant. Statistical analysis was performed using SAS, version 9.3 (SAS Institute, Inc).
One hundred fifteen patients met the inclusion criteria. Sixty of these patients had been included in a previous study.4 Fifty-four received the diagnosis of PCDLBCL-LT in period 1 and 61 received the diagnosis in period 2. The main characteristics of patients and baseline characteristics in the entire study group and according to the diagnosis period are reported in Table 1. No significant difference was observed between the periods for these variables. The female to male ratio in the entire study group was 1.47. Age ranged from 44 to 97 years (mean, 76.9; median, 78.6 years). The performance status was 0, 1, 2, 3, and 4 in 43, 25, 13, 5, and 2 patients, respectively, and was not available in 27 cases. Ninety-one patients (79.1%) had leg lesions, whereas 24 patients (20.9%) only had lesions located at other sites. A total of 27.4%, 51.3%, and 22.3% of the patients had T1, T2, and T3 category disease, respectively. According to the inclusion criteria, no patient had nodal or visceral involvement at diagnosis. High lactate dehydrogenase level and B symptoms were observed in 30.8% and 8.0% of patients, respectively.
First-line and subsequent therapies and clinical outcome in the entire series and according to the period of diagnosis are given in Table 1. Radiotherapy was the first-line therapy in 44.4% of the cases in period 1, but the frequency dramatically decreased in period 2 (16.4%; P = .001). In contrast, rituximab-PCT was first administered in 9.3% of the patients in period 1 vs 68.9% in period 2 (P < .001). Overall, 88.5% of the patients in period 2, compared with 16.7% in period 1, received rituximab-PCT (with or without anthracycline) at any time (P < .001). Conversely, the proportion of patients who received PCT without rituximab (with or without anthracycline) or less-intensive therapies significantly decreased between the 2 periods (Table 1). Major outcome differences were observed between the periods. The 3- and 5-year specific survival rates improved over time from 55% to 74% and from 46% to 66%, respectively (P = .01) (Table 1 and Figure 1).
Survival curves according to level of therapy (as reported in Table 1) are shown in Figure 2. No significant difference was observed between patients who received rituximab-PCT with vs without anthracycline (P = .82), and between patients given PCT containing anthracycline, PCT not containing anthracycline, and less-intensive therapies (P = .87) (Figure 2A). Patients were therefore categorized as those given rituximab-PCT (with or without anthracycline) and those who were given other treatment (Figure 2B).The 3- and 5-year specific survival rates were 80% and 74% in the first group vs 48% and 38% in the second group, respectively (P < .001) (Figure 2B and Table 2).
In view of this major difference in survival, these 2 groups were compared for baseline characteristics (Table 2). No significant difference was observed for age and previously recognized prognostic factors, including number and location of skin lesions and stage at diagnosis,4 contrasting with major differences in the period of diagnosis and outcome (Table 2).
In univariate analysis, disease-related death was significantly associated with location on the leg (P = .03), period 1 (P = .01), and treatment without rituximab-PCT (P < .001). Age (P = .23), sex (P = .55), performance status (P = .21), number of skin lesions (ie, multiple, corresponding to T3-T2 categories, vs single, corresponding to T1 category) (P = .18), and advanced (T3 vs T1-T2 category) (P = .12) did not reach significance. Multivariate analysis (Table 3) identified treatment without rituximab-PCT as the major prognostic factor (odds ratio, 4.6 [95% CI, 2.4-9.1]; P < .001), whereas number of skin lesions (P = .06) and location on the leg (P = .07) had only borderline significance. The period of diagnosis was highly related to type of therapy (Table 2) and had no effect on survival after taking this variable into account. When the period of diagnosis and the type of therapy were excluded from analysis and only the intrinsic characteristics of patients and their tumors were considered, the location on the leg (P = .03) was the only factor associated with lower disease-related survival in univariate and multivariate analyses.
To our knowledge, the present study is the largest to evaluate survival and prognostic factors in PCDLBCL-LT. When comparing patients who received the diagnosis between 1988 and 2003 (period 1) with those who received the diagnosis between 2004 and 2010 (period 2), we observed similar characteristics at diagnosis, contrasting with an absolute augmentation of approximately 20% in 3- and 5-year specific survival rates. This improvement was even more dramatic—nearly a doubling of the 5-year survival rate—when patients given rituximab-PCT regimens (most during period 2) were compared with those who received only radiotherapy and/or single-agent or combination chemotherapy without rituximab (most during period 1). No significant baseline differences were observed between these groups except for the period of diagnosis. In particular, age did not differ significantly between patients who received the standard rituximab-CHOP regimen or an age-adapted rituximab-PCT regimen and those who received other treatments. These results strongly suggest that the prognosis of this life-threatening lymphoma has dramatically improved over time, mainly by the use of rituximab-PCT. In contrast, our data do not suggest any improvement with anthracycline-based PCT administered without rituximab.
Among both cutaneous and systemic lymphomas, PCDLBCL-LT is characterized by its occurrence in elderly patients and its poor prognosis.1,4 Indeed, PCDLBCL-LT generally occurs much later in life in comparison not only with other subtypes of primary cutaneous B-cell lymphomas but also with its nodal and systemic counterpart: DLBCL. In a study18 of 448 patients with DLBCL, 32% were older than 70 years. In contrast, we found in the present study that approximately 3 of 4 cases of PCDLBCL-LT occurred in patients older than 70 years, with nearly half older than 80. Although PCDLBCL-LT may represent an intrinsically aggressive subtype of lymphoma, this specific age at onset could explain in large part the poor outcome reported to date in most series. Before 2003, older patients with PCDLBCL-LT were systematically considered too frail to receive a standard rituximab-PCT regimen. Most of them were given radiotherapy, palliative care, and/or low-dose chemotherapy, which resulted in an overall 5-year specific survival rate of approximately 50%.1,4,7 During 2002-2005, it was demonstrated that the addition of rituximab to the standard CHOP regimen improved response rates, short-term survival,14,19 and long-term outcome18,20 without increasing toxic effects in patients aged 60 to 80 years with a DLBCL. By this time, rituximab-CHOP was increasingly considered to be a satisfactory option or even a possible standard of care in patients with PCDLBCL-LT monitored by clinicians from the FSGCL.12,21 However, results reported on DLBCL were not directly transferable to patients with PCDLBCL-LT, who were, on average, 15 years older and included a large proportion of “very elderly patients.” Few studies18,22- 24 reported data on chemotherapy in patients older than 80 years. Many difficulties often encountered with these patients are related to the presence of other diseases, diminished organ function, and altered drug metabolism.24 In day-to-day practice, this resulted in frequent replacement of the standard rituximab-CHOP regimen with rituximab regimens including lower doses (rituximab–mini-CHOP) or no anthracycline. Contrary to reports18 that the use of less-intensive regimens reduces the response rate and shortens survival in patients with DLBCL, we did not observe a poor outcome in the present study when doses and types of rituximab-PCT regimens were adapted to the age and general condition of patients with PCLDBCL-LT.
The main clinical adverse prognostic factors identified in previous series2,4 of patients with PCDLBCL-LT were older age, numerous skin lesions and/or an advanced T stage at diagnosis, and the location on the leg. Interestingly, only the leg location had a significant adverse prognostic value in univariate analysis in the present series including numerous patients given rituximab-PCT, and neither this factor nor the number of skin lesions reached significance in a model including the period and type of treatment. These findings suggest that the use of rituximab-PCT could annul the negative effect of previous adverse prognostic factors.
Furthermore, response and survival rates in patients who received rituximab-PCT in the present series were higher than those reported in elderly patients with DLBCL who were given rituximab-CHOP.19,20 These data suggest that PCDLBCL-LT could be less aggressive than its systemic counterparts, and that the poor prognosis reported to date may result in large part from insufficiently intensive treatments and/or the absence of addition of rituximab to traditional chemotherapy regimens.
Although not systematically recorded and analyzed in the present study, other improvements in the management of cancer in older patients could have contributed to a better prognosis over time. In particular, the treatment and prevention of chemotherapy-induced sepsis was perhaps improved by the use of standardized antibiotic regimens and granulocyte colony-stimulating factor. Previous studies16,23 have shown that severe neutropenia and consecutive infections were the major causes of morbidity and mortality in patients with PCDLBCL-LT treated with rituximab-PCT regimens. Similar observations18,25,26 in patients with nodal DLBCL led to the recommendation that granulocyte colony-stimulating factor should be systematically used when the risk of neutropenia exceeds 20%. In view of the reduced hematopoietic reserve capacity in elderly patients, this recommendation has been widely adopted by clinicians of the FSGCL in period 2, probably contributing to more favorable outcomes over time.
The prognosis of PCDLBCL-LT has dramatically improved in the past decade in France as a result of therapeutic improvements. Patients with PCDLBCL-LT and their families should now be informed of the approximate 65% to 75% five-year specific survival rate (instead of the 50% previously widely accepted) and, in the absence of contraindications, should be encouraged to accept age-adapted rituximab-PCT regimens regardless of their age.24 Rare cases of PCDLBCL-LT refractory to rituximab-PCT or more frequent cases with rapid recurrence after initial response remain a challenging issue and should be addressed by additional studies. A phase 2 clinical trial evaluating the efficacy of lenalidomide in these cases is under way in France.
Accepted for Publication: August 13, 2013.
Corresponding Author: Florent Grange, MD, PhD, Department of Dermatology, Hôpital Robert Debré, avenue du général Koenig, 51092 Reims CEDEX, France (firstname.lastname@example.org).
Published Online: March 19, 2014. doi:10.1001/jamadermatol.2013.7452.
Author Contributions: Dr Grange had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Grange.
Acquisition of data: Grange, Joly, Barbe, Dalle, Ingen-Housz-Oro, Maubec, D’Incan, Ram-Wolff, Dalac, Templier, Esteve, Quereux, Machet, Leduc, Dereure, Laroche, Saiag, Vergier, Beylot-Barry.
Analysis and interpretation of data: Grange, Barbe, Bagot, Leduc, Dereure, Beylot-Barry.
Drafting of the manuscript: Grange.
Critical revision of the manuscript for important intellectual content: Joly, Barbe, Bagot, Dalle, Ingen-Housz-Oro, Maubec, D’Incan, Ram-Wolff, Dalac, Templier, Esteve, Quereux, Machet, Leduc, Dereure, Laroche, Saiag, Vergier, Beylot-Barry.
Statistical analysis: Barbe.
Obtained funding: Grange.
Administrative, technical, or material support: Joly, Barbe, Leduc, Dereure, Beylot-Barry.
Study supervision: Grange, Joly, Bagot, Laroche.
Conflict of Interest Disclosures: None reported.
Funding/Support: This study was supported by a grant from the Société Française de Dermatologie.
Role of the Sponsor: The Société Française de Dermatologie had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.