What is the best first-line treatment for POEMS syndrome: autologous transplantation, melphalan and dexamethasone, or lenalidomide and dexamethasone?

POEMS syndrome is a rare plasma cell dyscrasia. This study compared the responses to and survival of 347 POEMS syndrome patients given three first-line treatment regimens: autologous stem cell transplantation (ASCT, N = 165) and melphalan + dexamethasone (MDex, N = 79), or lenalidomide + dexamethasone (LDex, N = 103). After a median 45-month follow-up, overall hematologic complete remission (CRH) was 46.4%, vascular endothelial growth factor complete remission (CRV) was 55.1%, and neurological remission (RN) was 93.8%. CRH was better with ASCT (49.7%) than with MDex (37.7%, p = 0.001). CRV was better with ASCT (66.2%) than with MDex (38.5%, p = 0.001) or LDex (47.7%, p = 0.008). Differences in RN achieved by three regimens (91.5% vs. 100% vs. 93.8%, p = 0.234) were not significant. Overall 3-year progression-free survival (PFS) was 80.5% and overall 3-year overall survival (OS) was 90.8%. PFS was 87.6% with ASCT and 64.9% with LDex (p = 0.003). OS in the three regimens did not differ (p = 0.079). In medium-high risk patients, ASCT had better CRH and CRV than MDex, and better PFS than LDex. Therefore, although all three treatments had reasonable responses and survivals, patients with higher risk may benefit more from ASCT treatment.


Introduction
POEMS syndrome is a rare plasma cell dyscrasia characterized by polyneuropathy, organomegaly, endocrinopathy, monoclonal gammopathy, and skin changes. There is currently no standard treatment [1]. Current systematic therapies include high-dose chemotherapy with autologous stem cell transplantation (ASCT), alkylator-based therapy, and therapy with novel agents. High-dose chemotherapy with ASCT can achieve excellent clinical responses, with a significant improvement of neuropathy, complete hematologic response (CR H ) in 60% of POEMS syndrome patients, and complete VEGF response (CR V ) in 50%. Five-year progression-free survival (PFS) of 75% and 5-year overall survival (OS) of 90% have been reported [2][3][4]. However, treatment-related complications are frequent, with peri-engraftment syndrome reported in 12-50% of patients [5,6]. Melphalan is an alkylation agent that is widely used for treatment of plasma disorders including myeloma and amyloidosis [7]. Melphalan plus dexamethasone (MDex) achieved 100% neurological response (R N ), 38.7% CR H , and 95.8% CR V in a prospective study of 31 POEMS syndrome patients, but long-term outcomes were not reported [8,9]. MDex is less toxic than ASCT, with no treatment-related mortality and a low incidence of severe adverse events. Lenalidomide is an immunomodulatory drug that is also cytotoxic to malignant plasma cells [10]. Lenalidomide plus dexamethasone (LDex) achieved an R N of over 95%, a CR H of 46.3%, a CR V of 42.5%, a 3-year PFS of 75%, and an OS of 90% in a phase II single-arm prospective study including 41 patients [11]. All three treatments can achieve acceptable remission rates and survival, but it is unclear which is the best first-line treatment for POEMS syndrome. The efficacy of these treatments has not previously been compared in large patient cohorts, but the rarity of the disease makes it difficult to conduct randomized controlled trials. This study retrospectively evaluated the treatment efficacy and survival of the three treatment regimens in cohort of newly diagnosed POEMS syndrome patients treated at a single center.

Patients
The medical records of POEMS syndrome patients admitted to our hospital between January 2000 and October 2017 and initially treated with ASCT, MDex, or LDex were evaluated. A total of 347 patients who satisfied the POEMS syndrome diagnosis criteria described by Dispenzieri et al. and had complete treatment and follow-up data were included [1]. All patients provided informed consent. The study was approved by the Institutional Review Board of Peking Union Medical College Hospital and conducted following the ethical guidelines of the Declaration of Helsinki.

Risk stratification
Patients were stratified to low, medium, or high risk groups using four baseline characteristics, age > 50 years, presence of pulmonary hypertension, presence of pleural effusion, and an estimated glomerular filtration rate < 30 ml/min/1.73 m 2 . The first three characteristics had a value of 1; the last had a value of 2. Patients with total scores 0, 1, and 2-5 were assigned to low, medium, and high risk groups, respectively [12].

Treatment regimens
As no standard algorithm is available to guide the choice of treatment, the choice of regimen was mainly based on the clinician's judgment. Patients eligible for ASCT were < 65 years of age, without serious systemic disease or organ dysfunction, severe pulmonary hypertension, severe capillary leak syndrome (hypotension and/or refractory ascites), and active infection, and with successful collection of adequate peripheral blood stem cells [13]. The protocol included mobilization with cyclophosphamide 3 g/m 2 and colony-stimulating factor and conditioning with high-dose melphalan 140-200 mg/m 2 . MDex therapy included melphalan 10 mg/m 2 + dexamethasone 40 mg on days 1-4, one cycle every 28 days for a total of 9 cycles. LDex therapy included lenalidomide 10-25 mg on days 1-21, dexamethasone 40 mg on days 1, 8, 15, 22, one cycle every 28 days for a total of 12 cycles. Prophylaxis of embolization with LDex therapy was by aspirin 100 mg qd.

Response criteria
CR H was assessed by immunofixation electrophoresis and free immunoglobulin light-chain negativity in both serum and urine. CR V was defined as a decrease in VEGF to a normal concentration (<600 pg/ml). Achieving either CR H or CR V was defined as CR H /CR V . Neurological dysfunction was one of the key clinical manifestations and affects patients' life quality the most. Therefore, we use R N to represent clinical remissions. R N was defined as a decrease in the Overall Neuropathy Limitations Scale (ONLS) score of at least 1 point [8].

Statistical analysis
PFS was the time from treatment to recurrence, the deterioration of clinical symptoms, or death. OS was defined as the time from treatment to death. The date of last follow-up was August 31, 2018. Statistical analysis was performed with SPSS 22 (SPSS, Inc., Chicago, IL, USA). The chisquare test, or Fisher's exact test when appropriate, was used to determine the significance of differences in the values of categorical variables. Time to responses were calculated from the date of treatment initiation. Survival curves were plotted by the Kaplan-Meier method and differences were compared with the log-rank test. p-Values < 0.05 were considered significant.

Patient characteristics
The patient baseline characteristics are shown in Table 1. The median age at the time of diagnosis was 48 (range, 21-74) years; 65.7% were men. The median ONLS score was 4 of a possible 0-12 points, and peripheral neuropathy was confirmed by electromyography in the 19 patients with ONLS scores of 0. A total of 61.1% of the patients were positive for IgA type heavy chain monoclonal immunoglobulin. The light chains were all λ type. Risk stratification resulted in 118 low-risk (34.0%), 150 medium-risk (43.2%), and 79 high-risk (22.8%) patients.
Of 347 included patients, 79 received MDex, 103 received LDex, and 165 were treated by ASCT. Compared with the MDex or LDex groups, patients treated with ASCT had a more favorable baseline clinical manifestations, with a lower percentage of patients > 50 years of age (p = 0.022), an extravascular water load (p = 0.001), or IgA type heavy chains (p = 0.048). Fewer high-risk patients were treated with ASCT than with MDex or LDex (p < 0.001).

Treatment response
Overall median follow-up was 45 months, total follow-up time was 1420 person-years. Median follow up was 90 months for MDex patients, 48 months for ASCT patients, and 44 months for LDex patients.
The total CR V rate was 55.1%. CR V was significantly better after ASCT than after MDex (38.5%, p = 0.001) or LDex (47.7%, p = 0.008). (Fig. 1a) Subgroup analysis found that the CR V benefit of ASCT compared with the other regimens was significant in medium-high risk (p = 0.001), but not in low-risk (p = 0.222). Moreover, CR V was prognostic for PFS in medium-high risk patients, but not for OS in these groups. (Supplement Fig. 1).
Combined CR H and CR V have been associated with PFS and OS, and can be used as a surrogate endpoint [14]. The overall CR H /CR V rate was 65.8% and the ASCT rate (75.3%) was higher than that with MDex (58.3%, p = 0.015) but not significantly different from the rate with LDex (p = 0.142). The difference in the CR H /CR V rates in the ASCT and MDex groups was observed in medium-high risk patients (p = 0.001), but not in low-risk patients (p = 0.748).
As for time to response, the overall median time to CR H was 10 months (1-179 months). The median time to CR H was 26.5 months (1-179 months) for MDex, 8.5 months (1-86 months) for ASCT and 7 months (1-46 months) for  (Fig. 1c). The differences in R N in the three treatment regimens, 100% with MDex, 91.5% with ASCT, and 93.8% with LDex were not significant (p = 0.234) ( Table 2).

Survival
Forty-nine of the 347 patients died during follow-up; two from lethal complications during ASCT. Fifty-one patients experienced disease progression and were alive at the end of the follow-up period. The overall 3-year PFS rate was 80.5% and was significantly higher in ASCT (87.6%) than in LDex (64.9%) patients (p = 0.003) but was not different from patients in the MDex group (84.6%, p = 0.568) (Fig. 2a). Subgroup analysis showed that PFS was longer with ASCT than with LDex in both low-risk (p = 0.016) and medium-high risk patients (p = 0.037) (Fig. 2c, d).   (Fig. 2b).

Risk of secondary malignancy
One secondary malignancy (hepatocellular carcinoma) was reported in the ASCT group. No secondary malignancies were reported in the LDex or MDex groups.

Discussion
There are currently no standard treatments for patients with POEMS syndrome. ASCT, LDex, and MDex are all effective, but the question "what is the best treatment for POEMS syndrome?" has not been answered [12]. The rarity of POEMS syndrome makes it difficult to conduct randomized double-blind-controlled trials. Consequently, most have been retrospective analyses of single regimens or single-arm prospective studies. This retrospective analysis of 347 patients compared the three most common treatment options in POEMS syndrome, each with a different therapeutic basis. In this patient series, ASCT had the highest response and PFS rates of the three regimens, especially for patients at medium-to-high risk. One explanation is that ASCT eradicates the underlying plasma cell clones more completely than LDex or MDex. However, there were fewer high-risk patients and more patients with mild clinical manifestations in the ASCT group than in the MDex and LDex groups, which included patients with severe organ dysfunction. OS was not significantly better in ASCT patients than in the other groups, possibly because of treatment-related mortality (1.2%), and long-term complications, such as secondary malignancy (0.6%). On the other hand, OS can be improved with second-line treatments for patients with suboptimal responses. Good responses have been achieved with a wide range of treatment choices for refractory POEMS syndrome patients, which partly compensates for deficits in maintaining long-term PFS with first-line treatment. Salvage treatment with lenalidomide or bortezomib-based regimens have both shown promising response rates [15][16][17]. Novel regimens including daratumumab may further broaden the treatment options for multidrug-resistant patients [18]. MDex and LDex regimens have both been shown effective in POEMS syndrome patients not eligible for ASCT, but it is not clear that they are less toxic alternatives to ASCT as first-line POEMS therapy [8,19]. In these patients, MDex regimens achieved response rates similar to ASCT in low-risk patients, but response was generally inferior in medium-high risk patients. However, long-term outcomes with MDex and ASCT were not significantly different. Melphalan therapy was generally not favored because of concerns of treatment-related stem cell damage, secondary myelodysplastic syndrome, and leukemia in myeloma patients, with a cumulative risk of 10-15% [20,21]. However, secondary malignancies or myelodysplastic syndrome were not reported in this patient series. The 3.2% occurrence of secondary malignancies reported in a Japanese cohort of POEMS patients was much lower than that seen in myeloma [2].
Lenalidomide was highly active in these newly diagnosed POEMS syndrome patients, with response rates and OS similar to those with ASCT. However, CR H and CR H / CR V or after LDex treatment were not translated into longterm PFS. PFS after LDex treatment was inferior to that achieved with ASCT, even in low-risk patients. This may result from lack of maintenance treatment. Therefore, maintenance therapy was suggested for LDex group because limited cycles of LDex seemed insufficient to maintain long-term PFS compared with ASCT, though it remains to be shown whether adding maintenance treatment can increase PFS as did in myeloma patients [22].
In conclusion, all three treatments had reasonable responses and survivals in newly diagnosed POEMS syndrome patients. However, ASCT might have better responses and survivals compared with the other two groups, especially for higher risk patients. The treatment should be based on a comprehensive evaluation of the physical and the economic status of the patient, the planned follow-up, and the experience of the physicians and treatment center. ASCT could be recommended for high-risk patients if they are eligible. Lenalidomide maintenance after LDex treatment may be beneficial in term of PFS.