Randomized and retrospective studies have shown that consolidation RT significantly decreases the risk of disease recurrence in DLBCL, though most were not powered to detect an improvement in overall survival [1–8]. Given that many patients with DLBCL are long-term survivors, optimizing disease control while minimizing acute and late side effects is critical. The risk of RT-induced late effects, including development of secondary malignancies and cardiovascular disease, is undoubtedly related to the total dose utilized and volume treated .
The optimal RT dose for patients who have a complete response to R-CHOP chemotherapy based on PET imaging remains unknown. A randomized trial from the United Kingdom found excellent in-field control rates with RT doses as low as 30 Gy . It is possible that RT doses less than 30 Gy are equally effective. Thus, we evaluated our own experience in patients with DLBCL treated predominantly with R-CHOP and all having negative post-chemotherapy imaging (88% assessed with PET). In this patient population, we observed excellent in-field control with lower doses of RT. Additionally, a relationship between RT dose and in-field control was not evident.
In the study from the UK, 640 patients with intermediate-to-high grade NHL (DLBCL: 85%; stage I-II: 86%) were randomized to 30 Gy versus 40–45 Gy . Eighty percent of patients received consolidation RT as part of a primary combined-modality regimen, though 13% had relapsed or refractory disease and 10% were treated with palliative intent. At a median follow-up of 5 years, in-field control was similar between the low and high dose groups (82% vs 83%; HR 1.09, p = 0.7). Overall, in-field control was lower than that seen in retrospective series where 5 year rates were greater than 90% [19, 20]. The inferior in-field control in the UK trial may be explained by inclusion of patients with refractory disease, 20% of patients receiving RT alone, and less rigorous radiographic assessment of chemotherapy response. It has been demonstrated that a positive post-chemotherapy PET scan compared to a negative scan predicts for worse 5-year in-field in patients who receive consolidation RT (71% vs 95%, p = 0.008) .
A retrospective series from the Netherlands suggests that RT doses less than 30 Gy may be equally effective in patients who have a complete response to chemotherapy . This study found excellent in-field control rates among 94 patients with stage I intermediate-to-high grade NHL who achieved a complete response on CT imaging after 4 cycles of CHOP. There was no difference in 5-year crude in-field control between patients who received 26 Gy versus 40 Gy (92% vs 95%).
Similar to the study from the Netherlands, we found excellent 5-year in-field control of 94% among patients who had a complete response to chemotherapy and underwent consolidation RT. There was no difference in IFC among sites that received ≤ 20 Gy versus those that received 26–30 Gy (94% vs 95%). Our study is unique in that the majority of patients received R-CHOP (63%) and underwent functional imaging (89%). In addition, lower doses of RT were utilized (median, 30 Gy; range, 12–40 Gy).
In contrast to our findings, there was evidence of a RT dose response in a retrospective series from MD Anderson of 162 patients with stage I-III DLBCL . Five-year in-field control was significantly worse for patients who received < 40 Gy compared with ≥ 40 Gy (83% vs 97%, p = 0.002). The main limitation of this study is that chemotherapy response assessment was based on plain X-ray imaging. Consistent with our study, there was a higher rate of in-field failure with tumors that were originally ≥ 10 cm.
A limitation of our study is that it is retrospective and non-randomized in nature. Further, there were few in-field failures, which may limit the power to detect a RT dose response. There was clearly a preference to treat more advanced stage patients with lower doses of RT. This may over- or underestimate a RT dose response because stage III-IV patients are at an overall higher risk of disease recurrence, but distant failures may eclipse the importance of in-field failures. We did not find the latter to be the case in that patterns of failure were similar between early and advanced stage patients. Further, the role of consolidation RT in advanced DLBCL is controversial and the benefit of radiation therapy, regardless of dose, has not been conclusively demonstrated. At DCI, we are currently treating DLBCL patients who have a complete response to 4 or more cycles of R-CHOP based on PET imaging with consolidation RT dose of 18–20 Gy as part of a phase II trial.