PURPOSE: To investigate the patterns of local failure and the risk factors predictive of local failure and to establish the dose-response relationships influencing the probability of local control in patients with Stage I and II angiocentric T-cell or natural killer (NK)/T-cell lymphoma who were treated with radiotherapy (RT) alone. METHODS AND MATERIALS: We retrospectively reviewed the data from 102 patients with Ann Arbor Stage I and II angiocentric T-cell or NK/T-cell lymphoma who underwent RT alone to a median dose of 45 Gy (range, 20-70 Gy) between 1976 and 1998. The patterns of local failure, risk factors predictive of local failure, dose-response relationships, and survival data were analyzed. Because of the protean feature of local recurrences, the sites of local failure were allocated to one of three categories: true recurrence (TR), marginal recurrence (MR), and elsewhere recurrence (ER). RESULTS: Despite a higher complete remission rate (72%) after RT, 60 patients experienced treatment failure, including local failure in 48 (47%), regional failure in 3 (3%), and systemic failure in 28 (27%). The patterns of local failure were TR in 42, MR in 3, and ER in 5 patients. The median time to recurrence for TR/MR was shorter than that for ER (1 month for TR/MR vs. 12 months for ER). Patients with TR/MR had a more unfavorable prognosis than those experiencing ER (2-year survival rate after salvage treatment: 6% for TR/MR vs. 80% for ER; p <0.01). The dose-response curve was sigmoid in shape within the range of 20-54 Gy, which followed the plateau at doses in excess of about 54 Gy. A positive correlation was observed in the dose-response curve for the probability of local control (p = 0.017, logistic regression analysis). The overall 5-year actuarial survival and local recurrence-free survival rate for all patients was 42% and 53%, respectively. Achievement of complete remission was the most statistically significant risk factor predictive of TR/MR and the most important prognostic factor. CONCLUSION: Our data confirm that local failure remains the major obstacle for patients who receive RT alone and that achievement of complete remission is a particularly important determinant of treatment success. Although dose escalation up to >54 Gy cannot entirely reduce the incidence of TR/MR, we believe it is important to identify an appropriate subset of patients for whom an additional boost dose may be beneficial. Given the high rate of local failure, an investigational approach should be conducted to supplement RT using radiosensitizers or more effective chemotherapeutic agents in future trials.
PURPOSE: To investigate the patterns of local failure and the risk factors predictive of local failure and to establish the dose-response relationships influencing the probability of local control in patients with Stage I and II angiocentric T-cell or natural killer (NK)/T-cell lymphoma who were treated with radiotherapy (RT) alone. METHODS AND MATERIALS: We retrospectively reviewed the data from 102 patients with Ann Arbor Stage I and II angiocentric T-cell or NK/T-cell lymphoma who underwent RT alone to a median dose of 45 Gy (range, 20-70 Gy) between 1976 and 1998. The patterns of local failure, risk factors predictive of local failure, dose-response relationships, and survival data were analyzed. Because of the protean feature of local recurrences, the sites of local failure were allocated to one of three categories: true recurrence (TR), marginal recurrence (MR), and elsewhere recurrence (ER). RESULTS: Despite a higher complete remission rate (72%) after RT, 60 patients experienced treatment failure, including local failure in 48 (47%), regional failure in 3 (3%), and systemic failure in 28 (27%). The patterns of local failure were TR in 42, MR in 3, and ER in 5 patients. The median time to recurrence for TR/MR was shorter than that for ER (1 month for TR/MR vs. 12 months for ER). Patients with TR/MR had a more unfavorable prognosis than those experiencing ER (2-year survival rate after salvage treatment: 6% for TR/MR vs. 80% for ER; p <0.01). The dose-response curve was sigmoid in shape within the range of 20-54 Gy, which followed the plateau at doses in excess of about 54 Gy. A positive correlation was observed in the dose-response curve for the probability of local control (p = 0.017, logistic regression analysis). The overall 5-year actuarial survival and local recurrence-free survival rate for all patients was 42% and 53%, respectively. Achievement of complete remission was the most statistically significant risk factor predictive of TR/MR and the most important prognostic factor. CONCLUSION: Our data confirm that local failure remains the major obstacle for patients who receive RT alone and that achievement of complete remission is a particularly important determinant of treatment success. Although dose escalation up to >54 Gy cannot entirely reduce the incidence of TR/MR, we believe it is important to identify an appropriate subset of patients for whom an additional boost dose may be beneficial. Given the high rate of local failure, an investigational approach should be conducted to supplement RT using radiosensitizers or more effective chemotherapeutic agents in future trials.