The Prognostic Role of Interim PET after First Chemotherapy Cycle and PET Sequential Evaluation of Response to ABVD in Hodgkin Lymphoma Patients - the Polish Lymphoma Research Group (PLRG) Observational Study

Background: Several studies confirmed the predictive role on treatment outcome of interim-PET after 2^nd ABVD cycle (iPET2) in Hodgkin lymphoma (HL). We hypothesized that interim PET after 1^st cycle (iPET1) might define chemosensitivity with a better accuracy than iPET-2. To test this hypothesis, PLRG launched in 2008 a prospective multicenter observational study aiming at assessing the prognostic role of iPET1 and the dynamic of sequential PET response to ABVD.

Methods: Adult pts with newly diagnosed early (stage I-IIA) and advanced (stage IIB-IV) consecutively enrolled in 11 Polish centers were risk-stratified by the EORTC/GELA criteria and treated according to the ESMO guidelines: ABVD x 3-4 cycles + IFRT in early stage, ABVD x 6 ± consolidation RT in advanced stage disease. Patients were scanned with iPET1 and iPET2 and no treatment change was permitted based solely on iPET results, with the exception of clinical or radiological evidence of overt HL progression. After the first interim analysis (52 pts enrolled, 2010), which demonstrated that all the iPET1 negative patients had also a negative iPET2, the protocol was amended, limiting the iPET2 scans only to pts with iPET1 Deauville score 5,4,3. Quality control for PET-CT was supervised by the Italian-Polish core lab using a standard methodology. PET scans were interpreted locally according to the Deauville 5-point scale: Score 1 to 3, was considered a negative (-), score 4 to 5 a positive (+) scan. Subsequently all PET scans were uploaded to the web platform WIDEN® for central review and Italian-Polish expert panel (EP) scored them afresh. Discorcondant cases were discussed in a joint review session with all the five EP members. Binary and overall concordance rates were calculated using k Cohen's and alpha Krippendorf's coefficients, respectively. Negative (NPV) and positive predictive values (PPV) of iPET1 were calculated using time to progression free survival (PFS) event.

Results: Between 2008 and 2014, 346 pts were registered. 35 pts were excluded from the analysis for absence/poor quality of images resulting in 108(35%) assessable pts with early and 203(65%) with advanced HL. Median age at diagnosis was 31(18-80) years. iPET1 was scored 1-3 in 87/108(81%), and 4-5 in 21/108(19%) of pts with early and in 133/203(65%), and 70/203(35%) with advanced stage, respectively. Out of 91pts with positive iPET1, 83 pts underwent iPET2, which remained (+) in 41/83(49,4%) pts. In 22 pts treatment was escalated. 11 of those pts, in whom the treatment escalation was decided solely on positive iPET were excluded from the analysis; the remaining had symptoms or CT evidence of progression. After a median follow-up of 40,2 (3,2-90,2) months 300 pts (103 "early" and 197 "advanced") were evaluable. 65(21,7%) of them (9 in early and 56 in advanced group) had a PFS event: in "early" group 9(9%) showed disease progression (4 with iPET1(-) and 5 with IPET1(+)) and 1 of them died. In advanced stage 49(25%) pts showed disease progression (16 with iPET1(-) and 33 withiPET1(+)) out of whom 13 died; 7 additional pts died without HL progression: 4 from toxicity and 3 from unrelated events. At 36 months NPV and PPV of iPET1 was 93% and 45% in "early" and 81% and 52% in "advanced" group, respectively. The dynamic of response to ABVD was assessed in 189 pts who underwent both iPETs. All 116 pts with iPET1(-) remained (-) in iPET2-(fast-responders). Out of 83pts with IPET1(+) 39 (47%) became iPET2(-)-(slow responders); the rest (34pts: 41%) remained iPET2(+)-(no responders). PFS for fast-responders @36 months was 85%, for slow-responders 80% (log rank p=0,36) and for no-responders 25% (log rank p=0,0000). The EP changed the local iPET1 score in 27 cases: in 14 from (+) to (-) in 13 from (-) to (+). The inter-observer-agreement among reviewers on evaluating a positive vs. negative interim PET scans was good, Fleiss' kappa = 0.73, comparable to that found in analogous studies (IVS, HD0607).

Conclusion: iPET1 fails to better identify chemosensitivity in ABVD-treated HL compared to iPET2. PPV of iPET1 is substantially inferior to the published results for iPET2. However NPV of iPET1 is comparable to iPET2 and therefore might guide early treatment de-escalation strategies.