The Role of 18F-FDG PET/CT in Multiple Myeloma Staging according to IMPeTUs: Comparison of the Durie–Salmon Plus and Other Staging Systems

We aimed at comparing the Durie–Salmon Plus (DS Plus) staging system based on Italian Myeloma criteria for PET USe (IMPeTUs) with other two staging systems in predicting prognosis of patients with all stages of newly diagnosed multiple myeloma (MM). A total of 33 MM patients were enrolled in this retrospective study. The variation between the DS Plus classification and Durie–Salmon staging system (DSS) or Revised International Staging System (RISS) classification was assessed. When staged by the DSS, patients in stage I and stage II did not reach the median overall survival (OS), and the median OS was 33 months for stage III (p=0.3621). When staged by the DS Plus, patients in stage I did not reach the median OS of stage I, and the median OS for stages II and III was 38 and nine months, respectively (p=0.0064). When staged by the RISS, patients in stage I did not reach the median OS, and the median OS was 33 and 16 months for stage II and stage III, respectively (p=0.0319). The concordances between two staging systems were 0.07 (DS Plus versus DSS) and 0.37 (DS Plus versus RISS), respectively. Multivariate analysis revealed that DS Plus stage III (HR: 11.539, p=0.021) and the Deauville score of bone marrow ≥4 (HR: 3.487, p=0.031) were independent prognostic factors associated with OS. Both the DS Plus based on IMPeTUs and RISS possessed a better potential in characterizing and stratifying MM patients compared with the DSS. Moreover, DS Plus stage III and the Deauville score of bone marrow ≥4 were reliable prognostic factors in newly diagnosed MM patients.


Introduction
As a clonal hematologic malignancy, multiple myeloma (MM) is characterized by bone marrow plasma cell infiltration and the presence of serum and urine monoclonal immunoglobulins. e prognosis of MM patients is highly variable; therefore, a reliable staging system is extremely important to optimize appropriate treatment as quickly as possible and avoid irreversible organ damage [1]. e most widely applied staging systems in MM patients include the International Staging System (ISS) and Durie-Salmon staging system (DSS) [2]. However, both of these two staging systems have some limitations. As a powerful and reproducible stage classification, the ISS simply segregates patients into three different groups based on the levels of β 2 -microglobulin and serum albumin. However, the ISS-derived outcome can be affected by serum albumin, which is a host factor and not disease-specific [3]. Moreover, the ISS relies solely on tumor biological parameters but does not integrate any medical imaging modalities. e DSS relies on a combination of clinical factors, such as the number of lytic bone lesions on a skeletal radiographic survey, serum calcium, level of hemoglobin, amount of M protein, and renal function [4]. However, the DSS has a poor reproducibility because its classification based on the extent and number of bone lesions found by X-ray is observer-dependent [5].
Recently, the Revised International Staging System (RISS) improves the prognostic value of the ISS by combining the variables in the ISS with the chromosomal abnormalities (CA) detected by interphase fluorescence in situ hybridization (t(14;16), t(4; 14), and del17p) and serum lactate dehydrogenase (LDH) in those patients with newly diagnosed MM [6]. According to one recent study, the RISS can also be used to stratify patients with relapsed/refractory MM [7].
In 2006, the Durie-Salmon Plus (DS Plus) staging system integrated new imaging techniques, such as 18 F-2-fluoro-2deoxyglucose ( 18 F-FDG) positron emission tomography/ computed tomography (PET/CT) and magnetic resonance imaging (MRI), into a new generation of MM staging and offered the opportunity to precisely stage patients by anatomic and functional techniques [8]. 18 F-FDG PET/CT provides prognostic information on symptomatic MM at the baseline and therapeutic follow-up, allowing the detection of extramedullary disease (EMD) [9]. However, no standard interpretation criteria have been proposed for the evaluation of 18 F-FDG PET/CT scans in MM, preventing reproducibility of data.
A group of Italian nuclear medicine experts, medical physicists, and hematologists have defined new visual interpretation criteria (Italian Myeloma criteria for PET USe; IMPeTUs) to standardize the 18 F-FDG PET/CT interpretation criteria and methods for extensive use in clinical practice for symptomatic MM patients. However, a larger series of patients should be adopted to define a visual cutoff for positivity [10].
It remains unclear whether the DS Plus based on IM-PeTUs can be used to appropriately interpret 18 F-FDG PET/CT imaging in MM patients. erefore, we aimed at comparing the DS Plus based on IMPeTUs with other staging systems in predicting prognosis of patients with all stages of newly diagnosed MM in this study.

Patients.
is retrospective study was approved by the Ethics Committee of the First Affiliated Hospital of Soochow University with waiver of informed consent.
Patients with newly diagnosed monoclonal plasma cell disease who had available data of RISS stage and DSS stage were enrolled in this analysis between May 2007 and December 2017. All patients underwent whole-body 18 F-FDG PET/CTfor initial diagnosis of disease.
Subjects were selected for inclusion if they met the following criteria: (1) the diagnosis of MM confirmed by the presence of an M-component in serum and/or urine plus clonal plasma cells in the bone marrow and/or a documented clonal plasmacytoma, (2) digital image data available for retrospective analysis, and (3) the time interval between assessment of hematological and immunologic parameters and 18 F-FDG PET/CT <3 weeks. e following patients were excluded: (1) patients who had insufficient information on the RISS, DSS, and DS Plus, (2) MM patients who had additional diseases, and (3) patients who had received treatment before 18 F-FDG PET/CT acquisition.

PET/CT Acquisition.
All patients underwent whole-body 18 F-FDG PET/CT according to the standard protocol in the same center. e patients were instructed to fast for at least 6 h. e blood glucose level of all the patients was lower than 11 mmol/L. Next, 60 min after the injection of 18 F-FDG (dose of 0.12 mCi/kg), imaging was started on a Discovery STE PET/CTscanner (General Electric Medical Systems, Milwaukee, WI, USA) with a CT of the whole body (140 kV, 120 mA, transaxial FOV 700 mm, pitch 1.75, rotation time 0.8 s, and slice thickness 3.75 mm). Subsequently, a whole-body emission scan in a 3D mode was performed from the base of the skull to 2.4. Image Analysis. 18 F-FDG PET images were blindly evaluated by two experienced nuclear medicine physicians in order to avoid any bias on image analysis. Image interpretation of 18 F-FDG PET/CT was based on IMPeTUs. To make sure IMPeTUs was suitable for DS Plus, bone lesions with a Deauville score ≥4 on PET scan and diffuse lytic lesions ranging between 0.5 and 1 cm in size with a Deauville score ≥3 were all identified as positive bone lesions. For EMD, positive lesions were also defined as the Deauville score ≥4 on PET scan. If there was discordance between the two independent physicians for image analysis, a third investigator was advised to make a decision.

Statistical Analysis.
To assess the variation between DS Plus classification and DSS or ISS classification, an exact weighted kappa value was calculated with STATA 11 software package (StataCorp, College Station, TX,  Contrast Media & Molecular Imaging 3 USA), which was expressed as a number between 0 and 1, with 0 representing complete nonconcordance and 1 representing complete agreement [11]. To better compare the three staging systems, DS Plus and DSS subgroups were combined within the main numerical grouping for analysis. OS was determined from the date of the 18 F-FDG PET/CT scan until the date of death or last follow-up. e OS was estimated by the Kaplan-Meier method. Comparisons among groups were made by a log-rank test. For multivariate analysis, variables that were considered clinically relevant or independently predictive of survival in univariate analysis were introduced into a Cox proportionalhazards model. Statistical analysis was performed by IBM SPSS 19.0 software (SPSS Inc., Chicago, IL, USA) and GraphPad Prism 5.0 software (GraphPad Software Inc., San Diego, CA, USA).

Characteristics of Patients.
A total of 59 MM patients met inclusion criteria for 18 F-FDG PET/CT. Among them, 20 patients had received treatment before 18 F-FDG PET/CT acquisition, one patient had myelodysplastic syndromes (MDS), one patient had gastric cancer, one patient had no β 2 -MG results, and three patients did not do the examination of cytogenetic abnormalities. According to our exclusion criteria, the abovementioned 26 patients were excluded from the present study. Finally, 33 consecutive patients (10 women, 23 men; mean age ± SD, 60 ± 10 years; range, 34-77 years) were retrospectively enrolled in this study. Table 2 lists detailed characteristics of patients.  Figure 2. e comparison of the DS Plus and RISS showed that 57.58% of patients had concordant stages across systems. Among four patients with RISS stage I MM, one, two, and one were classified as stages I, II, and III based on the DS Plus staging system, respectively. Among 15 patients with RISS stage II MM, six, seven, and two were classified as stages I, II, and III based on the DS Plus staging system, respectively. Among 14 patients with RISS stage III MM, two, one, and 11 were classified as stages I, II, and III based on the DS Plus staging system, respectively. Agreement between the DS Plus and RISS stages calculated using the weighted kappa statistic was 0.37 (95% CI, 0.12-0.62; p < 0.01), indicating fair concordance between DS Plus and RISS stages. Examples are given in Figure 3.   Contrast Media & Molecular Imaging

Survival according to the ree Staging Systems.
When staged by the DSS, patients in stage I and stage II did not reach the median overall survival (OS), and the median OS was 33 months for stage III (p � 0.3621; Figure 4(a)). When staged by the RISS, patients in stage I did not reach the median OS of RISS stage I (the median follow-up duration of the stage I patients was 59 months), and the median OS was 33 and 16 months for stage II and stage III, respectively (p � 0.0319; Figure 4(b)). When staged by the DS Plus, patients in stage I did not reach the median OS of stage I, and the median OS for stages II and III was 38 and nine months, respectively (p � 0.0064; Figure 4(

Prognostic Factors.
According to the univariate analysis, other factors, such as the presence of EMD (p � 0.015), subgroup B of DS Plus (p � 0.012), and the Deauville score of bone marrow ≥4 (p � 0.013), were significantly associated with shorter OS (Table 4).

Discussion
In recent years, new imaging techniques, such as PET or MRI, play an increasingly important role in staging of MM.  2), where BM3 indicates that bone marrow uptake is <liver but >mediastinum, F3 indicates 4 to 10 lesions (arrows), SP indicates the spine, ExtraSP indicates outside the spine with (4) indicating reference lesion uptake >liver uptake + 10%, L2 indicates that 1 to 3 lesions were also lytic, PM indicates a bone lesion in surrounding soft tissues with bone cortical interruption, EM indicates at least one extramedullary lesion, and N indicates that the extramedullary lesion was nodal disease with (2) indicating extramedullary lesion uptake ≤mediastinum. is patient was classified as IIA according to DS Plus. e impact of MRI on staging patients according to DS Plus has been analyzed in many studies [12][13][14]. Although 18 F-FDG PET/CT has been proved to be prognostically valuable in staging different groups [15][16][17], only few studies have been reported on DS Plus based on 18 F-FDG PET/CT [18]. One of the reasons is that no standard interpretation criteria have been proposed for the evaluation of 18 F-FDG PET/CT scans in MM. erefore, IMPeTUs has been proposed to standardize image interpretation criteria in order to make clinical trial results applicable and reproducible. Since IMPeTUs is only a descriptive criterion according to experience with 18 F-FDG PET/CT in lymphoma and MM [17,19,20], we decided to adopt the Deauville score ≥4 and diffuse lytic lesions with the Deauville score ≥3 as the cutoff values in this study. Our study confirmed that IMPeTUs was useful to define bone lesions and EMD, which were main criteria of the DS Plus. DS Plus based on IMPeTUs showed fair concordance with the classic RISS, indicating great capability to differentiate patients with good OS from those with a poorer prognosis.
To the best of our knowledge, the present study was the first study designed to compare the DSS, RISS, and DS Plus. In some studies, researchers have compared the DSS with the ISS and found that the ISS has a better reliability, simplicity, and predictability for OS compared with the DSS for MM patients [5,21,22]. e RISS, which combines the ISS together with the prognostic power of high-risk CA and LDH, has been proved to give a better differentiation of MM patients into three survival subgroups [6,7,23,24]. Our data also indicated that the RISS provided significant prognostic information.
Nowadays, it remains largely explored whether the DSS can be used to predict the outcome. Most studies have demonstrated that the DSS cannot show a significant difference in OS among stages I, II, and III [21,22,25], while several studies have reported that there are statistical differences in OS among the three groups in the DSS [2,26]. In the present study, we found that the DSS was not correlated with OS. In most of the studies comparing the frequency distribution of the same patients across the DSS and ISS, the number of patients classified as DSS stage III is larger than that of patients classified as ISS stage III [5,21,22,25,26]. e percentage of patients with DSS stage III in our study was also higher. e reason might be attributed to that the DSS includes hemoglobin and serum calcium, which can be affected by various factors and advanced lytic bone lesions in X-ray, leading to difficulty in appropriate interpretation due to the lack of standard criteria.
It is believed that DS Plus is a reliable method for both staging and prognostic classification. e prognostic significance of DS Plus has been reported in one recent study using MRI of the spine and pelvis in 85 patients [27]. However, Fechtner et al. have reported that the DS Plus is not better than the DSS for the prediction of OS [28]. Although only few studies have reported the DS Plus in combination with 18 F-FDG PET/CT, the prognostic implication of PET/CT has been documented in several articles [16,29,30]. Our study confirmed that DS Plus based on    PET/CT has been confirmed in several studies [9,16]. DS Plus could discriminate patients of different stages in this study, which might be attributed to accurate assessment of the number and site of focal PET positive bone lesions with or without osteolytic characteristics detected by 18 F-FDG PET/CT. In addition, criteria of IMPeTUs include the description of the metabolic state of the bone marrow. In a study of 18 MM patients, clinical course of the disease could be predicted by the associated 18 F-FDG bone marrow images [32]. However, in other studies, 18 F-FDG uptake of the bone marrow in PET/CT is not significantly associated with OS [29,33]. e reason might be attributed to that the intensity of accumulation was assessed for 18 F-FDG images using different methods. In our study, the univariate and multivariate analyses showed that the metabolic state of the bone marrow was an independent prognostic factor associated with OS.
ere were some limitations in the present study: First, the sample size was small. Stage I patients in the DS Plus, RISS, and DSS all survived due to limited patient number. Because PET/CT is not yet a standardized imaging tool in MM, only a few MM patients performed PET/CT based on physicians' advice. In addition, some patients were excluded because of incomplete data. Second, in the present study, positive bone lesions were defined as lesions with a Deauville score ≥4 on PET scan and diffuse lytic lesions ranging between 0.5 and 1 cm in size with a Deauville score ≥3, and such a definition might reduce the specificity and consequently increase the false-positive rate. However, lesions in 51.5% of patients were downstaged according to the DS Plus compared with the DSS in our study. erefore, we think that this definition is useful to guide therapy. However, a prospective study with a larger population should be adopted to define a cutoff for positivity.
Taken together, our findings indicated that DS Plus based on IMPeTUs was useful for the initial staging of MM. Both DS Plus and RISS possessed a better potential in predicting OS of MM patients compared with the DSS. DS Plus stage III and the Deauville score of bone marrow ≥4 were independent prognostic factors associated with OS.

Conflicts of Interest
e authors declare that they have no conflicts of interest.