Patterns of Outcome Following Recurrence in Patients with Diffuse Large B-Cell Lymphoma (DLBCL): Long Follow-up From a Single Centre.

Despite improvements in outcome of patients (pts.) with DLBCL, the prognosis for those who relapse after first-line treatment remains very poor. This retrospective analysis reports outcome of pts. with DLBCL treated before the routine use of Rituximab, with a median (med.) follow-up of 12 years.

Between 1985 and 2003, 461 pts. were diagnosed with DLBCL at SBH. 384/461 (83%, male: 208, med. age: 60 years, range 17-95) were treated with curative intent (CI) and form the basis of this analysis. Those with primary CNS or primary mediastinal B-cell lymphoma and HIV +ve pts. were excluded. Histology at diagnosis included 31 pts. with ‘composite' lymphoma (29 follicular [FL], 1 lymphoplasmacytic, 1 ‘MALT') and 17 with ‘discordant' lymphoma (all FL).The majority, 142 (37%), had Stage IV disease, 99 (26%) stage I, 88 (23%) stage II and 55 (14%) stage III. The IPI score (ascribed retrospectively) was low-risk in 189 (49%), low-intermediate (int.) in 110 (29%), high-int. in 64 (17%) and high-risk in 21 (5%). 354/384 (92%) received an anthracycline-containing regimen, according to protocols in use at the time. 30 pts. (8%) with localised disease received radiotherapy alone. Prior to 2001, all pts. received prophylactic intra-thecal methotrexate; subsequently, only those at high risk for CNS relapse did so.

CR/CRu was achieved in 240/384 pts. (63%) and PR in 67 (17%). 53 (14%) did not respond (NR) and 24 (6%) died of causes related to initial therapy. Overall survival (OS) correlated with outcome to therapy (CR/CRu vs PR vs NR, p<0.0001). The med. overall (OS) and cause-specific survival were 6 and 12 years respectively. The med. remission duration (RD) has not been reached, after a med. follow-up for living pts. of 12 years (range 2-24). On univariate analysis, age, stage and IPI score at diagnosis correlated with both survival (p=0.0006, <0.0001, <0.0001) and RD (p=0.02, 0.02, 0.008) respectively. 92 of the 240 pts. (38%) in whom CR/CRu was achieved relapsed at a med. of 10 months; none in the CNS. 82/92 pts. (89%) had a repeat biopsy at recurrence which in 6 revealed FL and in 1 ‘MALT' lymphoma. In 14 pts. recurrence of DLBCL occurred after 5 years. The med. age at relapse was 65 years. Stage and IPI score were assessable in 80/85 pts. with ‘pure' DLBCL at relapse, namely: Stage I:26 pts. (31%), II:10 (12%), III:19 (22%), IV:25 (29%). 46% of pts. who relapsed with Stage I disease had Stage I at presentation and 72% who relapsed with Stage IV had the same stage at diagnosis. IPI scores at recurrence were: low-risk:42 (49%), low-int:25 (29%), high-int:9 (11%), and high-risk:4 (5%). 66/85 pts. (78%) received further treatment with CI, 2^nd CR/CRu being achieved in 36/66 (55%; 36/85 if all pts. with recurrent DLBCL are considered), and PR in 14 (21%). 16 pts. had NR (only 1 of them responded to subsequent alternative therapies). OS from recurrence (med. for all pts.= 22 months) correlated with outcome to ‘salvage' therapy (CR/CRu vs PR vs NR, p<0.0001). The med. 2^nd RD was 3 years. Amongst the 45 pts. aged '60 years at relapse and thus eligible for high-dose treatment (HDT) with autologous haematopoietic stem cell support, 19 (16 in CR/CRu and 3 in PR; med. age, 46 years) went on to receive it but only 5 of the latter are alive at last follow-up, 13 (including the 3 pts. treated in PR) having died of lymphoma (9) or treatment-related causes (4). The reasons for not proceeding to HDT (in 21 pts. achieving a 2^nd CR) were: age (8), comorbidity (4), early recurrence (2) failure to mobilise cells (2), toxicity of prior therapy (3) and pt. choice (2). On univariate analysis, survival from recurrence correlated only with the IPI score at recurrence (p=0.009). Overall, 65% of pts. who developed recurrent disease died of lymphoma (including 28/30 pts. in whom 2^nd CR/CRu was not achieved with the first ‘salvage' therapy). Thus, only 12/85 pts. with recurrent disease are currently alive, 3 have been lost to follow-up.

Despite the curative potential of HDT, these data confirm that recurrence of DLBCL, in an unselected patient population, carries an appalling prognosis. Improving initial therapy is therefore crucial. In pts. in whom conventional 2^nd-line therapy has failed, it is entirely justifiable to evaluate experimental therapies.

No relevant conflicts of interest to declare.