Redefining PET response criteria for complete metabolic response in patients with newly diagnosed multiple myeloma

18F-fluorodeoxyglucose (FDG) positron emission tomography (PET)/computed tomography (CT) is currently used to determine the therapeutic response of patients with multiple myeloma (MM). It is used to delineate myeloma bone or extramedullary (EMD) lesions, evaluate tumor metabolic activity, and monitor treatment response. Following the International Myeloma Working Group (IMWG) criteria, it is essential to measure the measurable residual disease, but also the complete metabolic response (CMR), inside and outside the bone marrow (BM).¹ In the absence of standard imaging criteria for PET CMR, a combined analysis of two prospective imaging substudies from large phase III trials (IFM/DFCI 2009² and EMN02/HO95³) was conducted. Both trials enrolled transplant-eligible patients with newly diagnosed MM (NDMM) and evaluated PET/CT response after therapy applying the Deauville scale (DS) without predefined cutoffs.⁴

The findings of this joint analysis and the suggested redefined response criteria (included at the end of this article) were published recently in the Journal of Clinical Oncology by Elena Zamagni and colleagues and are summarized here.⁴

Study design

The imaging subanalysis involved two trials, IFM/DFCI 2009 and EMN02/HO95, enrolling a total of 228 transplant-eligible patients with NDMM.

• IFM/DFCI2009 prospectively evaluated 700 patients treated with a combination of eight cycles of bortezomib, lenalidomide, and dexamethasone (VRD) vs VRD plus autologous stem cell transplantation (ASCT), followed by lenalidomide maintenance.
  • The Imaging Young Myeloma (IMAJEM, NCT01309334) substudy compared axial magnetic resonance imaging and FDG-PET/CT in 134 patients at diagnosis, after three cycles of induction therapy, and before maintenance therapy or premaintenance (PM).
  • The primary endpoint was the bone lesion baseline detection rate. The secondary endpoint was the prognostic effect of the imaging techniques on progression-free survival (PFS) and overall survival (OS) after three cycles and at PM.
• EMN02/HO95 evaluated 1,503 patients prospectively comparing single vs double ASCT vs proteasome inhibitor-based therapy after three to four cycles of bortezomib-based induction and consolidation vs no consolidation, followed by lenalidomide maintenance. Patients were monitored for a median duration of 62.9 months.
  • The EMN02/HO95 imaging substudy prospectively evaluated 94 patients with FDG-PET/CT at diagnosis, after four cycles of induction therapy and PM.
  • The two primary endpoints included assessing the prognostic significance of PET/CT at diagnosis and after therapy and standardization of PET/CT evaluation via centralized imaging revision and definition of interpretation criteria.

PET/CT scans analyzed jointly were only at baseline and PM in an effort to standardize PET/CT evaluation and define criteria for PET CMR after therapy by determining the prognostic impact of such assessment on PFS. The five-point DS was used to describe BM score (BMS) and focal lesion (FL; FL score [FS]) uptake.

Parameters of physiological interest related to reference organs, liver, and mediastinal blood pool (MBP) were measured using a spherical volume of interest with a radius > 3 cm in the central portion of the liver and a volume of interest within the aorta lumen.

• Liver and MBP mean and maximum standardized uptake value (SUVmax) and BM SUVmax of prominent lesions were annotated and visually analyzed.
• Median BM SUVmax (but not FLs) of the lumbar vertebrae L3 to L5 was defined.
• The uptake in FLs, BM, or EMD > DS1 (DS1 = no uptake) was based on PET positivity, either at baseline or PM.
• The impact of PET/CT scans on clinical outcomes was analyzed.

Results

• FLs were found in 78% of all 228 patients at baseline, and 80% had an FS ≥ 4.
• Median FL and BM SUVmax values were slightly higher in IFM/DFCI 2009 than in EMN02/HO95 (5.7 vs 4.2 and 3.7 vs 2.68, respectively). By contrast, similar reference MBP and liver mean SUVs and distribution of FS, BMS, and EMD (11% of patients) were found.
• A significantly higher number of patients in the EMN02/HO95 trial scored an FS of 3 (36.9% vs 4.2% in IFM/DFCI 2009). By contrast, an equal number of patients in both trials had a BMS of 2 and 3.
• All patients showed BM diffuse uptake, including 35.5% with BMS ≥ 4.
• FLs at PM were visually detectable in 31% of patients, including 2% extramedullary cases, and an FS of 3 and ≥ 4 were detected in 24% and 67.7% of them, respectively.
• Diffuse uptake by residual BM was significantly lower at PM than baseline in 98% of patients (Table 1). Indeed, in 85% of patients, the BMS decreased from ≥ 4 at baseline to < 4 at PM.
• None of the patients with an FS of 2–3 had a BMS > 4.
• The FS and BMS at PM were < 3 in 53.5% and 71.2% of patients, respectively, and < 4 in 79% and 91.4% of patients, respectively.

Table 1. PET/CT assessment before maintenance therapy (PM)⁴

PET/CT results	Overall	IFM/DFCI 2009	EMN02/H095
BM, bone marrow; CT, computed tomography; DS, Deauville scale; FL, focal lesion; IQR, interquartile range; PET, positron emission tomography; PM, premaintenance; SUVmax, maximum standardized uptake value. *Only a DS score > 1 was considered as positive for presence of FLs and/or BM uptake.
Number of patients	199	119	80
Presence of FL uptake, DS score*	62 (31.2)	24 (20.2)	38 (47.5)
2	5 (8.1)	0 (0.0)	5 (13.2)
3	15 (24.2)	1 (4.2)	14 (36.9)
4	36 (58.0)	20 (83.3)	16 (42.1)
5	6 (9.7)	3 (12.5)	3 (7.9)
Presence of BM uptake, DS score*	195 (98.5)	118 (100.0)	77 (96.3)
2	103 (52.8)	63 (53.5)	40 (51.9)
3	75 (38.5)	43 (36.4)	32 (41.6)
4	16 (8.2)	11 (9.3)	5 (6.5)
5	1 (0.5)	1 (0.8)	0 (0.0)
Median BM SUVmax (IQR)	2.30 (1.80–3.08)	2.60 (2.10–3.40)	1.85 (1.54–2.51)
Median FL SUVmax (IQR)	3.67 (2.71–5.02)	5.37 (4.20–6.93)	3.07 (2.30–3.85)

Prognostic value of pre-maintenance PET/CT assessment⁴

• A DS score < 3 was associated with prolonged PFS and OS in FLs but not in BM.
• A DS score of 4 was associated with significant changes in PFS and OS for FLs and BM and, therefore, represents the optimal cutoff for PET/CT CMR assessment at PM.
• Both FS and BMS < 4 before maintenance were strong predictors of prolonged PFS and OS:

• • Median PFS in FS < 4 was 44.9 vs 26.6 months in FS ≥ 4 (HR, 0.6; 95% CI, 0.38–0.95; p = 0.03). For BMS, patients with BMS < 4 vs ≥ 4 achieved a median PFS of 41.9 vs 26.6 months, respectively (HR, 0.48; 95% CI, 0.25–0.92; p = 0.028).
  • Estimated 5-year OS rates for patients with FS < 4 and ≥ 4 were 77.7% and 64.1%, respectively (HR, 0.48; 95% CI, 0.25–0.92; p = 0.028), and for BMS < 4 vs ≥ 4 they were 76.7% vs 52.1%, respectively (HR, 0.29; 95% CI, 0.13–0.65; p = 0.003).

Conclusion

FDG-PET/CT is a reliable technique to predict outcomes in patients with NDMM, especially post-ASCT. Reduced FL and BM FDG uptake compared with the liver represent the standardized definition of PET CMR before maintenance under the newly proposed PET response criteria (Table 2), indicating sustained disease control. In both BM and FLs, a DS score of < 4 (with the liver uptake as reference) is the strongest predictor of improved PFS and OS.

However, clinical studies with independent prospective series of patients extensively applying measurable residual disease techniques at the BM level and imaging are still needed to validate and redefine the role of PET CMR and PET response criteria in treating patients with MM inside and outside clinical trials.

Table 2. Proposed refinement of PET response criteria before maintenance⁴                                                                                      

PET response before maintenance	Response criteria
BM, bone marrow; DS, Deauville scale; FL, focal lesion; PET, positron emission tomography.
Complete metabolic response	Uptake ≤ liver activity in BM sites and FLs previously involved, including extramedullary and paramedullary disease (DS score 1–3)
Partial metabolic response	Decrease in number and/or activity of BM/FLs present at baseline, but persistence of lesion(s) with uptake > liver activity (DS score 4 or 5)
Stable metabolic disease	No significant change in BM/FLs compared with baseline
Progressive metabolic disease	New FLs compared with baseline consistent with myeloma