Prospective Evaluation of 18-F FDG PET/CT and Biopsies of Osteolytic Lesions and Random Bone Marrow Samples in Newly Diagnosed Multiple Myeloma Patients

Background

The degree of plasma cell (PC) infiltration in the bone marrow (BM) is an important diagnostic and prognostic marker in multiple myeloma. An infiltration of 60% or more has been included into the new criteria of the IMWG defining myeloma. PC infiltration can vary significantly within and among individual patients regarding growth patterns (focal, diffuse or mixed), bone destruction (best visible in CT), which may or may not be concomitantly present, and levels of PC metabolism (best detected by PET). Usually, BM examinations are performed by random biopsy and aspirate from the pelvis. It is up for debate whether the PC infiltration at this location is representative for the whole BM compartment or merely represents a local picture detail of the disease. In this prospective study we evaluated PC infiltration of osteolytic lesions (OL) and random BM biopsies and aspirates (RA) at the iliac crest with local parameters whole-body imaging with PET/CT.

Patients and Methods

64 transplant-eligible patients with newly diagnosed multiple myeloma (NDMM) were enrolled in this ongoing prospective study to investigate the genetic heterogeneity of malignant cells from OL in different parts of the BM compared with a RA of the pelvis. Target OLs were identified by low-dose whole-body CT scan. Sample pairs (n=64) were obtained by CT-guided biopsies of OLs as well as simultaneous RAs of the iliac crest at diagnosis and before maintenance therapy (n=19). To analyze differences between PC infiltration of the BM in RA compared to OL, we performed immunohistochemistry (IHC) on trephines of the iliac crest and on samples from OL. Whole-body ¹⁸F-FDG PET/CT was performed at diagnosis (n=53) and before initiation of maintenance therapy (n=42) assessing PET/CT characteristics like uptake patterns, number of focal lesions, maximal Standardized Uptake Value (SUVmax) of the respective lesion, SUVmax of normal BM as reference and delta SUVmax (SUVmax lesion-SUVmax reference) at diagnosis and before maintenance therapy.

Results and Discussion:

At baseline, samples from OLs were obtained in the pelvis (47 patients), in the spine (18) or in the extremities (4). PET/CT at diagnosis showed 3 different infiltration patterns: focal lesions in 11 patients, diffuse infiltration in 11 patients, and a mixed pattern in 31 patients. The median number of focal lesions per patient was 7 (range, 0 to >20). PET/CT-detectable lesions were most frequent in patients with a mixed pattern (median, 8 OL, 14/31 patients had >10 lesions). Patients with a focal pattern had a median number of 3 focal lesions; only one patient had >10 OLs. Interestingly, the number of PET/CT-detectable focal lesions at diagnosis neither correlates with ISS stage of the patients nor with their response to therapy.

At diagnosis, PC infiltration in OL was significantly higher in comparison to PC in random samples of the iliac crest (p=0.001). In 23 of 36 patients with a PC percentage in OL >=60%, the respective PC infiltration in RA of the iliac crest was <60%. The size of lesions (max. axial diameter measured in the accompanying CT scan) correlated with the extent of PC infiltration in IHC of OL (p=0.00014). However, comparing estimates of cellularity in CT and PET/CT, neither Hounsfield units (HU) nor SUV showed any correlation with PC infiltration of OL samples.

In a preliminary follow-up analysis of 19 patients, neither PC infiltration, size, HU nor SUV of OL showed any significant association with the outcome seen at the time of imaging analysis. However, our analysis showed that after induction therapy and ASCT, 9 of 10 patients with remaining PET-CT-detectable, ¹⁸F-FDG avid OLs would progress within 12 months (90%, 4 patients with focal, 6 patients with mixed patterns at baseline).

Conclusion

Our data suggests that the routine assessment of PC infiltration in RA of the iliac crest might underestimate the degree of PC infiltration in the whole skeleton of NDMM. PC infiltration correlated significantly with the size of the lesion in CT but neither with HU nor SUVmax of OL in PET-CT. This raises the question whether the imaging techniques being used will pick up signatures of non-viable tumor, such as necrotic tissue or inflammation, instead of or in addition to malignant plasma cells. Interestingly, patients with PET-detectable, ¹⁸F-FDG avid residual lesions after therapy were at high risk of progression within 12 months.