Contributions
Abstract: EP521
Type: E-Poster Presentation
Session title: Aggressive Non-Hodgkin lymphoma - Clinical
Background
18F‑fluorodeoxyglucose positron emission tomography/ computed tomography (PET-CT) is the standard imaging modality for disease staging and end of treatment assessment of diffuse large B-cell lymphoma (DLBCL). However, its use for response prediction has not been standardized, yet. Baseline total metabolic tumor volume (TMTV) ≥220 cm³ and change of maximum standardized uptake value (ΔSUVmax) after two cycles of R‑CHOP (rituximab, cyclophosphamide, doxorubicin, vincristine, prednisone) have been reported to be superior to the visual Deauville 5-point scale (DS).
Aims
Our aim was to assess the prognostic value of TMTV, ΔSUVmax and DS in contemporary “real-world” DLBCL patients uniformly treated with R‑CHOP given every 21 days (R-CHOP21).
Methods
We conducted a retrospective analysis of 144 patients with DLBCL who were diagnosed and treated with R‑CHOP21 at our institution during January 1st 2015 and December 31st 2019. This study was approved by the ethics committee of the Medical University of Graz. All patients underwent PET-CT at diagnosis and after four cycles of R‑CHOP21 (i‑PET4). SUVmax and TMTV were computed using the semiautomatic software Hermes Hybrid 3D Tumorfinder (Hermes Medical Solutions, Stockholm, Sweden). Statistical analyses were performed with Stata (Windows version 15.0, Stata Corp.). Median follow-up was estimated with a reverse Kaplan-Meier estimator. Primary endpoint for all time-to-event analyses was 5-year progression free survival (PFS). PFS was estimated with Kaplan-Meier estimators, and compared between groups using log-rank tests. Uni- and multivariable modeling of PFS functions was performed with Cox proportional hazards models. Harrell’s concordance index was calculated to evaluate model fit of different PET parameters.
Results
Results:
Thirty-seven patients were excluded (HIV, n=2; transformed low-grade lymphoma, n=23; primary testicular lymphoma, n=3; primary central nervous system lymphoma, n=4; no interim imaging, n=5). Patients had a median age of 67 years (interquartile range [IQR], 53-74) and a median R‑IPI score of 3 points (IQR, 2-4). During a median follow up of 2.4 years (IQR, 1.4-3.6), we observed 19 (18%) primary disease progressions during first line therapy, ten (9%) relapses after initial CR or PR, 16 (15%) deaths related to DLBCL and four deaths from other causes. The estimated 5-year PFS and OS of the whole cohort were 63% (95% CI, 51-73) and 75% (95% CI, 63-84), respectively.
Higher baseline TMTV predicted for higher risk of progression and death. The baseline TMTV threshold of >220cm³ could be confirmed (Table 1). At i-PET4, TMTV, SUVmax, and DS comparably predicted PFS with a good discriminatory performance (c‑indices between 0.76 and 0.80). Five-year cumulative incidences of primary disease progression or relapse were 32% and 76% in patients with DS of 1-3 points (n=67) and 4-5 points (n=34), respectively (Gray’s test p <0.0001). Complete metabolic response (CMR), a composite of interim TMTV of 0 and SUVmax of 0 (n=60 patients), showed 5-year PFS of 83% and 31% in patients with and without a CMR (log‑rank p<0.0001). In a multivariable model of PFS including CMR at i‑PET4, double expressor lymphoma biology, cell of origin, and NCCN‑IPI, only CMR prevailed as an independent predictor of PFS (HR 9.88, 95% CI 3.94-24.79, Wald test p<0.0001). ΔSUVmax was predictive of PFS but not superior to DS.
Conclusion
DLBCL patients with baseline TMTV >220cm³ and DS of ≥4 have an inferior PFS and are therefore prime candidates for new treatment strategies.
Keyword(s): DLBCL, PET, Prognosis
Abstract: EP521
Type: E-Poster Presentation
Session title: Aggressive Non-Hodgkin lymphoma - Clinical
Background
18F‑fluorodeoxyglucose positron emission tomography/ computed tomography (PET-CT) is the standard imaging modality for disease staging and end of treatment assessment of diffuse large B-cell lymphoma (DLBCL). However, its use for response prediction has not been standardized, yet. Baseline total metabolic tumor volume (TMTV) ≥220 cm³ and change of maximum standardized uptake value (ΔSUVmax) after two cycles of R‑CHOP (rituximab, cyclophosphamide, doxorubicin, vincristine, prednisone) have been reported to be superior to the visual Deauville 5-point scale (DS).
Aims
Our aim was to assess the prognostic value of TMTV, ΔSUVmax and DS in contemporary “real-world” DLBCL patients uniformly treated with R‑CHOP given every 21 days (R-CHOP21).
Methods
We conducted a retrospective analysis of 144 patients with DLBCL who were diagnosed and treated with R‑CHOP21 at our institution during January 1st 2015 and December 31st 2019. This study was approved by the ethics committee of the Medical University of Graz. All patients underwent PET-CT at diagnosis and after four cycles of R‑CHOP21 (i‑PET4). SUVmax and TMTV were computed using the semiautomatic software Hermes Hybrid 3D Tumorfinder (Hermes Medical Solutions, Stockholm, Sweden). Statistical analyses were performed with Stata (Windows version 15.0, Stata Corp.). Median follow-up was estimated with a reverse Kaplan-Meier estimator. Primary endpoint for all time-to-event analyses was 5-year progression free survival (PFS). PFS was estimated with Kaplan-Meier estimators, and compared between groups using log-rank tests. Uni- and multivariable modeling of PFS functions was performed with Cox proportional hazards models. Harrell’s concordance index was calculated to evaluate model fit of different PET parameters.
Results
Results:
Thirty-seven patients were excluded (HIV, n=2; transformed low-grade lymphoma, n=23; primary testicular lymphoma, n=3; primary central nervous system lymphoma, n=4; no interim imaging, n=5). Patients had a median age of 67 years (interquartile range [IQR], 53-74) and a median R‑IPI score of 3 points (IQR, 2-4). During a median follow up of 2.4 years (IQR, 1.4-3.6), we observed 19 (18%) primary disease progressions during first line therapy, ten (9%) relapses after initial CR or PR, 16 (15%) deaths related to DLBCL and four deaths from other causes. The estimated 5-year PFS and OS of the whole cohort were 63% (95% CI, 51-73) and 75% (95% CI, 63-84), respectively.
Higher baseline TMTV predicted for higher risk of progression and death. The baseline TMTV threshold of >220cm³ could be confirmed (Table 1). At i-PET4, TMTV, SUVmax, and DS comparably predicted PFS with a good discriminatory performance (c‑indices between 0.76 and 0.80). Five-year cumulative incidences of primary disease progression or relapse were 32% and 76% in patients with DS of 1-3 points (n=67) and 4-5 points (n=34), respectively (Gray’s test p <0.0001). Complete metabolic response (CMR), a composite of interim TMTV of 0 and SUVmax of 0 (n=60 patients), showed 5-year PFS of 83% and 31% in patients with and without a CMR (log‑rank p<0.0001). In a multivariable model of PFS including CMR at i‑PET4, double expressor lymphoma biology, cell of origin, and NCCN‑IPI, only CMR prevailed as an independent predictor of PFS (HR 9.88, 95% CI 3.94-24.79, Wald test p<0.0001). ΔSUVmax was predictive of PFS but not superior to DS.
Conclusion
DLBCL patients with baseline TMTV >220cm³ and DS of ≥4 have an inferior PFS and are therefore prime candidates for new treatment strategies.
Keyword(s): DLBCL, PET, Prognosis