The stage-modified IPI (MIPI), histology and a combined treatment influence the clinical outcome of 401 patients with primary extranodal head and neck B-cell lymphomas (PHNBCL) (IELSG23)

The purpose of this study was to evaluate the clinical outcome, prognostic factors and the rate of CNS recurrence in PHNBCL patients. From December 1990 to June 2004, 401 patients, with a median age of 60 years (range, 19–85), were diagnosed as having PHNBCL in 9 IELSG centers. The tumor site was: Waldeyer’s ring (69%), followed by parotid and salivary glands (12%), nasal cavity and paranasal sinuses (8%), thyroid (6%) and palate and oral cavity (5%). Regarding histology 77% of cases had DLBCL,18% indolent lymphomas (MALT=42; FL=29), while only 5% had Burkitt lymphoma or MCL. DLBCL and indolent lymphomas (IL) were selected for the present analysis. Stage II was slightly higher in DLBCL (67% vs 54%; P=0.03), while the rate of other adverse features such as advanced age (>60 yrs), bulky disease (>10 cm), B symptoms, ECOG-PS>1, No of extranodal sites>1, elevated LDH, beta 2-microglobulin or ESR, reduced albumin and >1 MIPI risk factors (RF) (50 vs 40%) was not statistically different into 2 groups. One hundred ninety four patients (51%) were treated with single therapy, mostly consinsting in anthracyclin containing chemotherapy (CHT), while 183 patients (49%) were given 3–4 courses of CHT, followed by IF RT. Only 26/250 DLBCL (10%) and 2/35 IL (6%) patients received CNS prophylaxis (Methotrexate 12 mg i.t.; median of cycles 3, range1–6). Two hundred sixty two DLBCL (85%) and 63 IL patients (91%) achieved a complete remission. Sixty DLBCL and 25 IL remitters, eventually relapsed (23 vs 40 %; P=0.007), 40 and 35% in the same site, 53 and 60% in other sites (1/234 patients with DLBCL of Waildeyer’s ring or paranasal sinuses, in CNS, 0.4%) and 7 and 5 in both. After a median follow-up of 42 and 35 months the DFS varied according to histology (DLBCL, 73% vs IL 52%; P=0.005), while OS and EFS were comparable into two groups (OS, 72% vs 74%; EFS, 56% vs 45%). In DLBL patients, the 5-year EFS varied according to MIPI (0–1 RF, 44% vs >1 RF, 69%; P=0.0001), and treatment (CHT, 42% vs CHT+IFRT, 60%; P=0.0001), while in IL patients MIPI and treatment did not influence EFS. Cox multivariate analysis showed that MIPI and treatment were independent prognostic factors for OS in DLBCL. Also in IL patients the MIPI score maitained its independent value for OS. In conclusion, this retrospective analysis shows that MIPI, histology and a combined treatment influence the outcome of patients with PHNBCL. Moreover, in the present series a very low rate of CNS recurrence occurred in high risk patients, who did not receive adequate prophylaxis, suggesting that CNS prophylaxis could not be mandatory in PHNBCL patients. This should be confirmed by prospective studies of clinical outcome.