Routine Bone Marrow Biopsy Adds Little Diagnostic Information in Patients with Newly Diagnosed Hodgkin Lymphoma Undergoing PET/CT Staging

Abstract 2627

Accurate staging is essential for the optimal choice of first-line treatment in Hodgkin lymphoma (HL). Bone marrow biopsy (BMB) is still widely recommended as part of the routine staging workup. However, this procedure may not be necessary for patients staged with [18]F-fluorodeoxyglucose positron emission tomography/computed tomography (PET/CT). The aims of this study were; 1) to determine if bone marrow infiltration can be excluded based on staging PET/CT result, and 2) to investigate whether BMB still adds useful information in the PET/CT era.

Patients with newly diagnosed HL undergoing pre-therapeutic staging at three Danish university centers were eligible for inclusion in this retrospective study, provided both PET/CT and BMB were performed at staging. Patients were identified in the Danish Lymphoma Registry (LYFO). BMB results were first obtained from LYFO and validated by a review of local pathology records. PET/CT records were retrieved and reviewed. The patterns of skeletal FDG uptake were categorized as follows: 1) uni-focal, 2) bi-focal, 3) multi-focal or 4) diffuse homogeneous FDG uptake. Clinical stage and risk assessment were performed with and without the account of BMB results, according to the Ann Arbor classification, the International Prognostic Score (IPS), and the German Hodgkin Study Group risk criteria for limited disease.

A total of 392 HL patients (Aalborg = 59, Aarhus = 67, Rigshospitalet = 266) were included, of whom 372 patients had classical HL and 20 had nodular lymphocyte predominant HL. According to staging PET/CT, 203 patients had limited stage disease and 189 patients had advanced stage disease. The median age at diagnosis was 39 years (range 15–87) and the female:male ratio was 3/4. BMB was positive for HL in 24 patients (6%), all of whom had been independently assessed by PET/CT as stage III (n=5) or IV (n=19) disease. Thus, no patients with PET/CT stage I-II disease, irrespective of the presence of B-symptoms, had HL infiltration in the BMB (p<0.0001, Fisher's exact test).

Skeletal FDG accumulation was seen in 90 patients and characterized as uni-focal (n=15), bi-focal (n=8), multi-focal (n=45) or diffuse (n=22). Among the 68 patients with focal skeletal FDG accumulation, 20 patients had positive BMB as compared to none of the patients with diffuse FDG uptake. Sixty out of 68 patients with focal skeletal FDG uptake had PET/CT response assessment during therapy. All patients who responded to therapy showed resolution of the pathological skeletal FDG uptake. The presence of focal skeletal FDG accumulation identified HL infiltration in the BMB with a sensitivity and specificity of 83% and 87%, respectively. The positive and negative predictive values of focal skeletal FDG accumulation for HL infiltration in the BMB were 29% and 99%, respectively. With both positive BMB and focal skeletal FDG accumulation considered to reflect true bone/bone marrow involvement, the sensitivity for bone/bone marrow disease was 94% for PET/CT and 33% for BMB. In terms of prognostic information, BMB upstaged 1% of the patients (n=5) from stage III to stage IV, with a one-point change of the IPS as an additional consequence. However, none of the 392 patients were allocated to another risk group or treatment group on the basis of BMB results.

At present there is no optimal gold standard for diagnosing bone/bone marrow disease in HL. Unilateral iliac crest BMB examines a very limited part of the total bone marrow volume, while PET/CT visualizes abnormal FDG accumulation throughout large parts of the skeletal system. Whether abnormal skeletal FDG uptake represents HL in all cases remains to be clarified. The most consistent finding of the present study was the absence of positive BMBs in PET/CT assessed stage I-II disease, where routine BMB seems to have no relevance. Furthermore, the omission of staging BMB would have had no impact on the treatment strategy in this cohort of 392 newly diagnosed HL patients.