Contributions
Abstract: PB1707
Type: Publication Only
Background
Bone marrow infiltration (BMI) evaluation plays a key role in lymphoma staging, treatment and prognosis. The role of PET/CT in the assessment of BMI is still controversial, especially in non-Hodgkin’s lymphoma (NHL).
Aims
To evaluate the role of 18F-FDG PET/TC in bone marrow infiltration for the diagnosis of Non Hodgkin Lymphoma. We compared 18F-FDG PET/TC visual and quantitative analyses with bone marrow biopsy in NHL patients.
Methods
Fifty patients with newly diagnosed NHL from February 2011 to February 2016 were retrospectively analyzed. Of these, 26 (group A) patients had aggressive NHL and 24 (group B) indolent NHL. To detect BMI on the posterior iliac crest 3 different PET/CT evaluation methods were used: 1) visual analysis, 2) maximal standardized uptake values (SUVmax, cut-off >2.5), and 3) Deauville score (categorical). Each method was applied in the whole patients cohort, in group A and in group B. Images were blindly reviewed separately by 3 nuclear medicine physicians. PET/CT results were compared with the bone marrow biopsy performed after imaging in all patients. Decision-curve analysis was used to evaluate the increment in net benefit (NB) obtained considering the Deauville score over a biopsy-all strategy.
Results
The prevalence of a positive biopsy was 38% in whole cohort, 19% in group A and 58% in group B. In the whole cohort, sensitivity, specificity and accuracy were 21%, 84% and 60% for visual analysis; 58%, 55% and 50% for SUVmax; and 47%, 81% and 68% for Deauville score. In group A, sensitivity, specificity and accuracy were 0%, 76% and 62%, for visual analysis; 40%, 52% and 50% for SUVmax; and 20%, 71% and 62% for Deauville score. In group B, sensitivity, specificity and accuracy were 29%, 100% and 58% for visual analysis; 64%, 60% and 62% for SUVmax; and 57%, 100% and 75% for Deauville score. At probability threshold equal to the prevalence of a positive biopsy, the increase in NB by Deauville score was 0.11 in the whole cohort, 0.02 in group A and 0.33 in group B. In this latter group, biopsying patients on the basis of the Deauville score is a strategy that reduced the biopsy rate by 24%, without missing any BMI.
Conclusion
FDG-PET/TC visual analysis has a limited value for detecting BMI in patients with NHL, while quantitative analysis by Deauville score provides a higher diagnostic performance. Noteworthy, the high positive predictive value in patients with indolent NHL suggests a potential role of FDG-PET/TC in avoiding bone marrow biopsy in this subtype of lymphoma.
Session topic: 20. Aggressive Non-Hodgkin lymphoma - Clinical
Keyword(s): PET, NHL, Bone marrow involvement
Abstract: PB1707
Type: Publication Only
Background
Bone marrow infiltration (BMI) evaluation plays a key role in lymphoma staging, treatment and prognosis. The role of PET/CT in the assessment of BMI is still controversial, especially in non-Hodgkin’s lymphoma (NHL).
Aims
To evaluate the role of 18F-FDG PET/TC in bone marrow infiltration for the diagnosis of Non Hodgkin Lymphoma. We compared 18F-FDG PET/TC visual and quantitative analyses with bone marrow biopsy in NHL patients.
Methods
Fifty patients with newly diagnosed NHL from February 2011 to February 2016 were retrospectively analyzed. Of these, 26 (group A) patients had aggressive NHL and 24 (group B) indolent NHL. To detect BMI on the posterior iliac crest 3 different PET/CT evaluation methods were used: 1) visual analysis, 2) maximal standardized uptake values (SUVmax, cut-off >2.5), and 3) Deauville score (categorical). Each method was applied in the whole patients cohort, in group A and in group B. Images were blindly reviewed separately by 3 nuclear medicine physicians. PET/CT results were compared with the bone marrow biopsy performed after imaging in all patients. Decision-curve analysis was used to evaluate the increment in net benefit (NB) obtained considering the Deauville score over a biopsy-all strategy.
Results
The prevalence of a positive biopsy was 38% in whole cohort, 19% in group A and 58% in group B. In the whole cohort, sensitivity, specificity and accuracy were 21%, 84% and 60% for visual analysis; 58%, 55% and 50% for SUVmax; and 47%, 81% and 68% for Deauville score. In group A, sensitivity, specificity and accuracy were 0%, 76% and 62%, for visual analysis; 40%, 52% and 50% for SUVmax; and 20%, 71% and 62% for Deauville score. In group B, sensitivity, specificity and accuracy were 29%, 100% and 58% for visual analysis; 64%, 60% and 62% for SUVmax; and 57%, 100% and 75% for Deauville score. At probability threshold equal to the prevalence of a positive biopsy, the increase in NB by Deauville score was 0.11 in the whole cohort, 0.02 in group A and 0.33 in group B. In this latter group, biopsying patients on the basis of the Deauville score is a strategy that reduced the biopsy rate by 24%, without missing any BMI.
Conclusion
FDG-PET/TC visual analysis has a limited value for detecting BMI in patients with NHL, while quantitative analysis by Deauville score provides a higher diagnostic performance. Noteworthy, the high positive predictive value in patients with indolent NHL suggests a potential role of FDG-PET/TC in avoiding bone marrow biopsy in this subtype of lymphoma.
Session topic: 20. Aggressive Non-Hodgkin lymphoma - Clinical
Keyword(s): PET, NHL, Bone marrow involvement