TY - JOUR T1 - A head-to-head comparison of the Lugano Classification and PERCIST for FDG-PET/CT response assessment in diffuse large B-cell lymphoma JF - Journal of Nuclear Medicine JO - J Nucl Med SP - 1688 LP - 1688 VL - 62 IS - supplement 1 AU - Nicklas Nielsen AU - Anne Nielsen AU - Oke Gerke AU - Karen Juul-Jensen AU - Thomas Larsen AU - Malene Hildebrandt Y1 - 2021/05/01 UR - http://jnm.snmjournals.org/content/62/supplement_1/1688.abstract N2 - 1688Purpose: Quantitative FDG-PET/CT response assessment has been investigated in patients with diffuse large B-cell lymphoma (DLBCL) for the potential to improve response prediction. Nevertheless, a comprehensive comparison of the current international guidelines of the Lugano Classification and the PET Response Criteria In Solid Tumors (PERCIST) has not previously been made. In a head-to-head investigation, we aimed to compare the Lugano Classification and PERCIST for the prediction of progression-free survival (PFS). Methods: In this retrospective cohort study, patients with DLBCL undergoing first-line R-CHOP(-like) therapy at Odense University Hospital, Denmark, were included from December 2013 - January 2020. Baseline, interim, and end of treatment (EOT) FDG-PET/CT scans were reevaluated using the Lugano Classification and PERCIST. Response was dichotomized into complete metabolic response (CMR) vs. non-CMR and responders vs. non-responders at interim. Non-response at interim was defined as Deauville score 5 (DS-5) with the Lugano Classification and PMR with ≤66% reduction in Standardized Uptake Value corrected for lean-body-mass (SULpeak), stable, or progressive metabolic disease (PERCIST-66%) for PERCIST. At EOT, response was dichotomized into complete metabolic response (CMR) vs. non-CMR. PFS was defined as the time from diagnosis until progression, relapse, or death, and patients who did not experience an event were censored at the date of data collection. The ability to predict PFS, the time-dependent predictive accuracy, and the proportion of agreement between the two sets of criteria were analyzed at interim and EOT. Results: A total of 170 patients were included for analysis. There were 35 events (progression, relapse, or death) after a median of 11.7 months (95% CI 6.7 - 16.7). In multivariable Cox regression adjusted for the International Prognostic Index (dichotomized into IPI 1-3 vs. 4-5), the following were predictive of PFS: DS-5 at interim (HR 13.9 (5.2 - 37.3), p < 0.001), PERCIST-66% at interim (HR 69.7 (13.3 - 365.4), p < 0.001), the Lugano Classification (non-CMR) at EOT (HR 4.6 (2 - 10.7), p < 0.001), and PERCIST (non-CMR) at EOT (HR 5 (2 - 12.7), p = 0.001) (Figure 1 A, B, C, and D, respectively). The PPV, NPV and accuracy for the prediction of events within 2 years was 100% (54.1 - 100), 83.3% (75.2 - 89.7), and 84.2% (76.4 - 90.2) for DS-5 at interim, 100% (39.8 - 100), 87.1% (78 - 93.4), and 87.6% (79 - 93.7) for PERCIST-66% at interim, 56.3% (29.9 - 80.2), 90.5% (83.2 - 95.3), and 86% (78.5 - 91.6) for the Lugano Classification (non-CMR) at EOT, 57.1% (28.9 - 82.3), 88.2% (78.7 - 94.4), and 83.3% (74 - 90.3) for PERCIST (non-CMR) at EOT (Table 1). The proportion of agreement was 98.4% (94.2 - 99.8) between DS-5 and PERCIST-66% at interim. There was perfect agreement between the Lugano Classification and PERCIST at both interim and EOT using the metabolic response categories. Conclusion: The Lugano Classification and PERCIST were equally predictive of PFS. Only non-response (as opposed to non-CMR) was predictive of poor PFS at interim, where a small subset of the events within 2 years was predicted. Non-CMR was predictive of unfavorable PFS at EOT where close to one-half of the events within 2 years were identified. There was perfect agreement using the metabolic response categories at both interim and EOT, and almost perfect agreement using non-response at interim between the Lugano Classification and PERCIST. View this table: ER -