The Response Evaluation Criteria in Solid Tumors Can be Substituted for the International Workshop Criteria?: Comparison Using Japan Clinical Oncology Group Study JCOG0203 for Untreated Indolent B-Cell Lymphoma

Introduction: The International Workshop Criteria (IWC) has been used for response evaluation in malignant lymphoma since 1999. In contrast, the Response Evaluation Criteria in Solid Tumors (RECIST) has been used for most solid tumors since 2000. The IWC is more complicated than RECIST, partly because it requires calculating the sum of the product of bidimensional diameters of the target lesions, whereas RECIST is based on unidimensional assessment of target lesions. If RECIST can be demonstrated to be more valid and precise than the IWC for malignant lymphoma, RECIST would be the substitute for IWC even for malignant lymphoma. The objective of this study was to evaluate whether RECIST could be substituted for the IWC using data from the JCOG0203 trial, a phase II/III study of R-CHOP-21 versus R-CHOP-14 in untreated, advanced-stage indolent B-cell lymphoma where the superiority of R-CHOP-14 in progression-free survival (PFS) was not shown (hazard ratio [HR], 0.92; 95% CI, 0.68-1.25; one-sided log-rank p = 0.30) (J Clin Oncol 2011;29:3990-8).

Methods: Three hundred patients aged 20-69 with stage III or IV indolent B-cell NHL were enrolled in the JCOG0203 trial between 2002 and 2007. Patients ineligible after the central pathological review (CPR) or patients with no target lesions for RECIST were excluded from this analysis. We calculated the kappa coefficient to evaluate the degree of agreement between IWC and RECIST. RECIST would be considered better in cases of a kappa coefficient of ≥0.7 because this indicates that both are in good agreement and RECIST is less complicated. If the kappa coefficient is <0.7, RECIST would be considered discordant with IWC. We evaluated which criteria were more predictive for PFS. In particular, criteria that can show a good separation of PFS, namely, the lower point estimate of the hazard ratio (HR) of a complete response (CR, including complete response and unconfirmed CR [CRu]) vs. a non-CR, is defined as more useful criteria. The landmark-time analysis was applied to PFS counted from day 168, when response status (CR vs. non-CR) for all patients was fixed.

Results: Overall, 269 patients were included in this analysis, 14 patients were ineligible by the CPR and 17 patients without target lesions for RECIST were excluded.The median age was 54 years, and 46.8% of the patients were men. Two-hundred fifty-six (95.2%) were diagnosed as having follicular lymphoma and 8 (3.0%) with extranodal marginal zone B-cell lymphoma. According to the Follicular Lymphoma International Prognostic Index, 85 (31.6%), 112 (41.6%), and 72 (26.8%) patients were classified as being at low, intermediate, and high risk, respectively.Based on the IWC, 207 patients (77.0%) had a CR/CRu, 54 (20.1%) a partial response (PR), 3 (1.1%) had stable disease (SD), 4 (1.5%) had progressive disease (PD), and 1 (0.4%) was not evaluable (NE). Based on RECIST, 120 (44.6%) had a CR, 137 (50.9%) had a PR, 6 (2.2%) SD, 4 (1.5%) had PD, and 2 (0.7%) were NE. The kappa coefficient between IWC and RECIST was 0.342 (95% CI, 0.261-0.424), indicating poor agreement. Subsequently, we calculated the HR (CR vs. non-CR) of PFS for each criterion with 264 patients in the landmark analysis. The HR of the IWC (0.473: 95% CI, 0.330-0.677, log-rank test p < 0.001) was lower than that of RECIST (0.635: 95% CI, 0.454-0.888, p = 0.0075). In terms of overall survival, the HR (CR vs. non-CR) was 0.500 (95% CI, 0.223-1.119, p = 0.0856) and 0.471 (95% CI, 0.197-1.130, p = 0.0843) according to IWC and RECIST, respectively.

Conclusion: RECIST cannot be substituted for the IWC in untreated, advanced-stage indolent B-cell NHL. Although fluorodeoxyglucose-positron emission tomography should be included in the standard response criteria currently used, introduction of unidimensional assessment is not recommended.