Consolidative radiotherapy (RT) for aggressive NHL assumes that relapse is most likely in sites of initial disease involvement, however there is limited data examining this assumption. We examined patterns of failure in patients (pts) following systemic chemotherapy (CT), to determine factors predicting for site of recurrence.

Methods: In 2 ALLG intermediate-grade NHL studies, 234/519 pts (45%) achieved complete remission (CR) with anthracycline-based CT of whom 169 had full data available. Median age was 55 yr (range: 19–76); ECOG PS ≤2, 94%; bulky disease (≥10 cm), 42 (25%); elevated LDH, 79 (47%); stage I, 14 (8%); stage II, 49 (29%); stage III, 37 (22%); stage IV, 69 (41%). Regimens given were CHOP (45), MACOP-B (48), standard CEOP (epirubicin 75 mg/m2 substituted for doxorubicin) (38) and high dose (cyclophosphamide 1500 mg/m2, epirubicin 150 mg/m2) CEOP (38). Ten pts received RT to involved sites.

Results: Median follow-up time was 7.5 yr (range: 2.3–11.6). At close out, 69 pts (41%) had relapsed and 9 (5.3%) had died. Relapse sites: 24 pts local (14.2%), 21 remote (12.4%), 24 both (14.2%). Log-rank tests revealed earlier stage to be associated with increased time-to-first-failure (p=0.039) but not age, CT, histology (B or T cell), response (CR vs. CRu), bulk, RT or IPI. At 4 yr the estimated percentages of failure-free patients were 71%, 78%, 45% and 67% in stages I to IV, respectively. Early stage predicted local vs distant relapse (p=0.027) (Fisher’s exact test), with failures solely at original sites being 0, 2, 13.5 and 14.5% in stages I to IV respectively.

Conclusion: In this study, 70% of relapses included original sites, 35% solely so. Advanced stage was the strongest predictor of early and local relapse. No other criteria were associated with probability of relapse at original sites. RT potentially could prevent 1/3 of relapses and contribute to prevention of another 1/3, as systemic therapy becomes more effective.

Author notes

Corresponding author

Sign in via your Institution