Maintenance therapy and improved relapse-free survival in patients with monoclonal gammopathy of renal significance

Key Points

• •Maintenance therapy after induction and treatment duration of >6 months were associated with improved RFS in patients with plasma cell MGRS.
• •ASCT was not associated with improved RFS or OS.

Visual Abstract

Abstract

Monoclonal gammopathy of renal significance (MGRS) is a recognized clinical entity with multiple therapeutic approaches, often modeled after therapies used for multiple myeloma, including anti-CD38 antibodies, proteasome inhibitors, immunomodulatory agents, and autologous stem cell transplantation (ASCT). However, because of its rarity, the optimal use of newer agents, ASCT, and maintenance therapy remains poorly defined. We retrospectively analyzed 50 patients with biopsy-proven MGRS treated across 3 institutions. The most common histopathologic diagnosis was proliferative glomerulonephritis with monoclonal immunoglobulin deposits (72%). Median estimated glomerular filtration rate was 32 mL/min (interquartile range, 17-51), and median 24-hour proteinuria was 3.28 g. A total of 37 patients received plasma cell (PC)–directed therapy, and 8 patients received B-cell–directed therapy. Common PC-directed induction regimens included daratumumab-based therapy (32.4%), CyBorD (cyclophosphamide, bortezomib, and dexamethasone; 29.7%), and bortezomib-dexamethasone (29.7%). Maintenance therapy was administered to 13 of 37 treated patients in the PC group (35.1%). With a median follow-up of 6.15 years, the median relapse-free survival (RFS) was 3.74 years in the PC group and 19 years in the B-cell group. Among PC-treated patients, RFS and overall survival (OS) did not significantly differ by induction regimen or use of ASCT. In contrast, maintenance therapy and induction duration >6 months were associated with significantly improved RFS, which remained significant on multivariate analysis. These findings suggest that maintenance therapy and prolonged treatment duration may improve disease control in MGRS and support further evaluation of extended treatment strategies to prevent renal damage and improve long-term outcomes.

Introduction

Monoclonal gammopathy of renal significance (MGRS) is detected in ∼40% to 50% of patients with monoclonal gammopathy who undergo kidney biopsy.¹ Patients usually present with low levels of serum paraprotein and evidence of renal dysfunction. MGRS is caused by the clonal deposition of immunoglobulins in the kidneys,^2,3 remains a diagnosis of exclusion, and requires a kidney biopsy demonstrating well-defined MGRS-associated lesions, as outlined in the 2017 International Kidney and Monoclonal Gammopathy Research Group consensus on MGRS. Alternative biopsy-proven diagnoses include free light-chain deposition disease and immunoglobulin-related amyloidosis.⁴ Without treatment, patients with MGRS are at a high risk of progression to end-stage renal disease.

Treatment of MGRS focuses on the eradication of the underlying B-cell or plasma cell (PC) clone using directed therapies. Because of the rarity of the disease, evidence-based guidelines on optimal treatment strategies are limited. Current recommendations include limited-duration induction treatment with combinations of proteasome inhibitors, immunomodulatory agents, and anti-CD38 antibodies, followed by high-dose melphalan and stem cell rescue.^5,6 However, the use of high-dose chemotherapy in this patient population is challenging because melphalan requires dose adjustment in patients with kidney disease and is potentially toxic and costly. Moreover, the benefit of transplantation in the era of novel therapies has not been well studied. Daratumumab-based combination regimens have shown at least a 30% reduction in 24-hour proteinuria, with most of the patients achieving complete or very good partial hematologic responses.⁷ These outcomes are comparable to those obtained with high-dose melphalan. Finally, the role of maintenance strategies, such as lenalidomide, in MGRS remains undefined.

This study aimed to evaluate treatment outcomes in patients with MGRS treated with modern therapies, with particular emphasis on the impact of induction regimen type, autologous stem cell transplantation (ASCT), and maintenance strategies on relapse-free survival (RFS) and overall survival (OS).

Methods

Patient population

This was a retrospective study conducted across 3 institutions, including The Ohio State University Wexner Medical Center, the University of Illinois Chicago, and John Stroger Jr. Hospital of Cook County. Institutional review boards approved the study, and informed consent was waived because of its retrospective nature. Records of patients who had undergone a kidney biopsy for suspected MGRS between 2010 and 2025 were obtained from deidentified pathology biorepository databases at each of these institutions. Patients with multiple myeloma nephropathy and renal amyloidosis were excluded. Collected variables included age, gender, ethnicity, presence of a serum monoclonal protein, serum heavy- and light-chain levels, presence of a PC or B-cell lymphoproliferative clone, percentage of bone marrow (BM) PCs, kidney function, need for hemodialysis (HD) or kidney transplant, and 24-hour urine proteinuria measured at diagnosis, after induction, and at last follow-up.

Treatment and response definitions

Treatment was selected at the physician’s discretion, based on patient age, comorbidities, and the underlying clone. Patients were classified as having a PC clone based on whether a PC clone was detected in the BM by flow cytometry or immunohistochemistry, or when a non–immunoglobulin M monoclonal protein was identified in the serum or kidney biopsy.⁸ Patients who did not meet either of these criteria were categorized as receiving B-cell–directed therapy. The type and duration of induction, use of ASCT, number of lines of therapy, and use of maintenance therapy were recorded. Maintenance therapy was defined as administration of a single agent following induction in the absence of hematologic relapse.

Hematologic responses were assessed based on International Myeloma Working Group criteria.⁹ Renal responses were evaluated using criteria established for renal light chain amyloidosis (AL), with response defined as a >30% reduction of proteinuria without a >25% decline in estimated glomerular filtration rate (eGFR).¹⁰ In addition, the depth of proteinuria reduction over time was reported using the methodology described by Kastritis et al.¹¹ In patients with anuria, only hematologic response and improvement in eGFR were recorded.

Given the absence of validated criteria for disease progression in MGRS,⁵ a composite end point was used to define RFS. Events included an increase in monoclonal protein, an eGFR decline >25%, increased proteinuria, need for repeat kidney biopsy, or death. RFS was calculated from treatment initiation to the occurrence of any of these events.

Statistical analysis

Summary statistics were first used to describe the study cohort. The Kaplan-Meier curves with log-rank tests were used to visualize and assess differences in survival between groups at any time point. The Fisher exact test was performed to assess differences in the hematologic and renal responses across induction therapies and other variables. The Friedman test was used to compare medians among paired groups, followed by pairwise comparisons using the Wilcoxon signed-rank test with Bonferroni correction for multiple testing. Hodges-Lehmann median differences were calculated as estimates of effect size. Univariate and multivariate Cox proportional hazards regression analyses were performed. Firth-penalized Cox regression was applied when 0 events were present in 1 group. These models were used to evaluate the effects of demographic and clinical factors on RFS and OS, with and without adjustment for significant covariates through a backward-selection procedure. All statistical analyses were performed using SAS (SAS 9.4; SAS Institute, Inc, 2004), and all tests were 2-sided with a significance level of P value .05.

This study was approved by the University of Illinois Institutional Review Board (STUDY2024-0518), The Ohio State University Institutional Review Board (2024C0078) and the John Stroger Jr. Hospital of Cook County Institutional Review Board (24-037X).

Results

Baseline clinical characteristics

Patient characteristics are summarized in Table 1. A total of 50 patients with biopsy-proven MGRS were included in the study. Among them, 42 patients had a PC clone, and 8 patients had a B-cell clone that was CD19⁺ and CD20⁺. In addition, 2 individuals were identified in 2002 but were treated with modern agents, including 1 patient who had a B-cell clone and another who had a PC clone but did not initiate treatment until 2019 after re-engaging in care. The remaining diagnoses occurred between 2011 and 2025.

Table 1.

Baseline characteristics of the study cohort

	All (N = 50)	B-cell clone (n = 8)	PC clone (n = 42)
Age, median (range), y	61 (9-86)	48 (9-78)	61 (19-86)
Gender, male, n (%)	21 (42.0)	3 (37.5)	18 (43.0)
Ethnicity, n (%)
White	40 (80.0)	6 (75.0)	34 (80.9)
African American	6 (12.0)	1 (12.5)	5 (11.9)
Hispanic	3 (6.0)	0 (0.0)	3 (7.2)
Hawaiian	1 (2.0)	1 (12.5)	(0.0)
MGRS pathology, n (%)
PGNMID	36 (72.0)	7 (87.5)	29 (69.0)
MIDD	5 (10.0)	0 (0.0)	5 (11.9)
C3G-MIg	5 (10.0)	1 (12.5)	4 (9.5)
Fibrillary GN	1 (2.0)	0 (0.0)	1 (2.4)
Immunotactoid GN	2 (4.0)	0 (0.0)	2 (4.8)
Type 1 cryoglobulinemia	1 (2.0)	0 (0.0)	1 (2.4)
Detectable serum M-protein, n (%)
Yes	19 (38.0)	3 (37.5)	16 (38.0)
No	31 (62.0)	5 (62.5)	26 (62.0)
β2-Microglobulin ≥5.5 mg/L, n (%)		n = 5	n = 31
Yes	18 (50.0)	2 (40.0)	16 (51.6)
No	18 (50.0)	3 (60.0)	15 (48.4)
Abnormal serum FLC ratio, n (%)
Yes	18 (36.0)	0 (0.0)	18 (42.9)
No	32 (64.0)	8 (100.0)	24 (57.1)
Serum FLC, median (range), mg/dL
FLC κ	40.6 (7.9-395)	40.1 (17.9-125.6)	40.1 (7.9-395)
FLC λ	28.9 (8.4-525)	36.7 (15.3-134.6)	25 (8.4-525)
BM cells, median (range), %	2 (0-26)	2 (1-4)	2 (0-26)
Creatinine, median (range), mg/dL	2.1 (0.58-11)	2.1 (0.58-4.78)	2.07 (0.67-11)
Creatinine >1.77 mg/dL, n (%)
Yes	28 (56.0)	4 (50.0)	24 (57.1)
No	22 (44.0)	4 (50.0)	18 (42.9)
eGFR, n (%), mL/min per 1.73 m2
>60	8 (16.0)	2 (25.0)	6 (14.3)
30-59	20 (40.0)	2 (25.0)	18 (42.9)
15-30	11 (22.0)	5 (50.0)	6 (14.3)
<15	11 (22.0)	0 (0.0)	11 (28.5)
eGFR, median (range), mL/min per 1.73 m2	32 (3-100)	26.5 (1-90)	33 (3-100)
Albumin, median (range), g/dL	3.2 (1.9-5.2)	2.95 (2.6-3.4)	3.4 (1.9-5.2)
24-h urine proteinuria, median (range), g/24 h	3.28 (0.1-21.5)	7.4 (0.59-14.3)	2.95 (0.1-21.5)
Need for HD, n (%)
Yes	20 (40.0)	3 (37.5)	17 (40.5)
No	30 (60.0)	5 (62.5)	25 (59.5)
Kidney transplant, n (%)
Yes	12 (24.0)	2 (25.0)	10 (23.8)
No	38 (76.0)	6 (75.0)	32 (76.2)

C3G-Mig, C3 glomerulopathy with monoclonal gammopathy; FLC, free light chain; GN, glomerulonephritis; MIDD, monoclonal immunoglobulin deposition disease.

The median age at diagnosis was 61 years (range, 9-86). Patients with a B-cell clone were younger at the time of diagnosis. The median baseline eGFR was 26.5 mL/min per 1.73 m² (range, 1-90) for patients with a B-cell clone and 33 mL/min per 1.73 m² (range, 3-100) for patients with a PC clone. The median baseline proteinuria was 7.4 g/24 h (range, 0.59-14.3) in patients with a B-cell clone and 2.95 g/24 h (range, 0.1-21.5) for the PC group. Before starting treatment of MGRS, 11 patients with a PC clone were already on HD at initiation of induction therapy, and 4 patients (2 with a PC clone and 2 with a B-cell clone) received a kidney transplant.

A serum monoclonal spike was detected in 38% of all patients, and an abnormal free light-chain ratio was present in 42.9% of patients with a PC clone. A BM clone was present in 30% of all patients detected by flow cytometry or immunohistochemistry, with a median abnormal cell percentage of 2% (range, 0%-26%).

The most common renal histopathologic subtype was proliferative glomerulonephritis with monoclonal immunoglobulin deposits (PGNMID), observed in 36 patients, including 7 of 8 patients with a B-cell clone and 29 of 42 patients (72%) with a PC clone.

Treatment modalities by clone group

In total, 37 patients with a PC clone received treatment. A daratumumab-based regimen was used in 12 of 37 patients (32.4%). VD (bortezomib-dexamethasone) and CyBorD (cyclophosphamide, bortezomib, and dexamethasone) were each administered to 11 of 37 patients (29.7%). VRD (bortezomib, lenalidomide, dexamethasone) was used in 3 of 37 patients (8.2%; Table 2). The mean duration of induction therapy was 7 months (range, 0.47-32.9). In addition, 8 patients (21.6%) underwent ASCT, all as consolidation after first-line therapy. Maintenance therapy was administered to 13 of 37 treated patients (35.1%) and was based on the same backbone as the induction regimen (eg, bortezomib, daratumumab, or lenalidomide) but given less frequently. Among patients who received maintenance therapy, 5 of 13 patients (38%) underwent ASCT, whereas 3 of 37 patients (8.2%) underwent ASCT without maintenance (P = .094). Second-line treatment was administered to 13 of 37 patients (31%), most commonly with daratumumab. In addition, 4 patients with a PC clone (10.8%) received third-line therapy.

Table 2.

Treatment received by patients with a PC clone or a B-cell clone

	PC clone (n = 37)	B-cell clone (n = 8)
Type of induction, n (%)
Dara-CyBorD, Dara-RVd, Dara-VD	12 (32.4)
VRD	3 (8.2)
VD	11 (29.7)
CyBorD	11 (29.7)
Anti-CD20 antibodies (rituximab or obinutuzumab)		7 (87.5)
Prednisone-mycophenolate		1 (12.5)
Number of induction cycles, mean (IQR)	7 (1-16)	26 (1-169)
ASCT, n (%)	8 (21.6)	0
Maintenance, n (%)
Yes	13 (35.1)	1 (12.5)
No	24 (64.9)	7 (87.5)
Type of maintenance, n (%)
Dara	3 (8.1)
Bortezomib or carfilzomib	5 (13.5)
Lenalidomide/thalidomide/pomalidomide	4 (10.8)
Rituximab	1 (2.7)	2 (25)
None	24 (64.9)	6(75)
Number of maintenance cycles, mean (IQR)	20.8 (3-70)	2 (1-3)
Second-line treatment, n (%)	13 (35.1)	2 (25)
Type of second-line treatment
Dara single agent	9 (69.2)
Dara-cyclophosphamide or bortezomib	2 (15.4)
Bortezomib or carfilzomib	2 (15.4)
Ibrutinib		1 (50)
Eculizumab		1 (50)
Third-line treatment, n (%)	4 (10.8)	1 (12.5)
Type of third-line treatment, n (%)
Obinutuzumab	2 (50)
Rituximab	2 (50)
Ravulizumab		1 (100)

Dara, daratumumab; IQR, interquartile range; RVd, bortezomib, lenalidomide, and dexamethasone.

In contrast, in the B-cell clone group, anti-CD20 therapy was used in 7 of 8 patients (87.5%). The mean duration of induction was 26 months (range, 0.7-171.5). No patient proceeded to ASCT, and only 1 patient received maintenance therapy. Two of 8 patients received second-line treatment, and 1 of them received a third-line agent at relapse.

Evaluation of hematologic and renal responses

Among patients with a PC clone, 35 patients had a documented hematologic response, of whom 22 (62.8%) achieved a complete response (CR) or very good partial response (VGPR). In the B-cell clone group, only 4 patients had evaluable hematologic responses (Figure 1A). A renal response was observed in 75% of patients with either a B-cell or PC clone (Figure 1B).

Figure 1.

Hematologic and renal responses in patients with MGRS. (A) Hematologic responses in patients with a PC or a B-cell clone. (B) Renal responses in patients with a PC clone or a B-cell clone. (C) Paired proteinuria values (grams per 24 hours) in 15 patients with a PC clone who were not HD dependent. Proteinuria measurements were obtained at diagnosis, end of induction, and last follow-up. Diagnosis vs end of induction: ∗P = .05; diagnosis vs last follow-up: ∗∗∗P = .0008; end of induction vs last follow-up: P = .6. Statistical analysis was performed using the Friedman test (overall P = .0006), followed by Dunn multiple-comparisons test. (D) Hematologic responses in 35 evaluable patients with a PC clone treated with a Dara-based regimen (n = 12), VRD (n = 3), VD (n = 10), or CyBorD (n = 10). Dara, daratumumab; MR/SD, minimal response or stable disease; NA, not available; ns, not significant.

We then specifically focused on changes in proteinuria and eGFR in patients not receiving HD who had longitudinal paired samples.

Among evaluable patients with a PC clone, a significant reduction in proteinuria was observed at the end of induction (Figure 1C), with a median decrease of 1.58 g/24 h (95% confidence interval [CI], 0.39-4.77; P = .011) and 3.4 g/24 h at last follow-up (95% CI, 0.95-9.53; P = .005; supplemental Table 1). eGFR also moderately improved or remained stable (supplemental Figure 1A). Three patients experienced increased proteinuria, and 10 had worsening eGFR at the end of induction. Eight patients with a PC clone underwent repeat kidney biopsy after induction because of concern for inadequate renal response; early recurrence of proliferative glomerulonephritis was demonstrated in 4 cases. Seventeen patients required HD during the study period, including 11 who initiated HD before induction, and 6 who began HD afterward because of lack of response or disease progression. In addition, 8 patients underwent a kidney transplant after induction therapy.

Finally, there was a higher rate of renal response in patients who achieved a PR or greater hematologic response (86.4% vs 62.5%; supplemental Figure 1B). However, the difference was not statistically significant (P = .300).

In 4 evaluable patients with a B-cell clone, the median reduction in proteinuria after induction was 5.6 g/24 h, and all had stable or improved eGFR. One patient then experienced increased proteinuria, and 2 had worsening eGFR. Three patients started HD during the study period, all due to lack of response to therapy.

PC clone group response by type of treatment

In patients with a PC clone, hematologic responses were evaluated according to the induction regimen (supplemental Table 2). Treatments included 12 patients receiving a daratumumab-based regimen, 3 treated with VRD, 10 with VD, and 10 with CyBorD. No major differences were observed among regimens, although VD appeared inferior compared with other therapies (Figure 1D).

Renal responses were then assessed in the same patient population (supplemental Table 2). In total 33 patients were evaluable, including 11 patients who received a daratumumab-based regimen, 2 who received VRD, 10 who received VD, and 10 who received CyBorD. Notably, 10 patients (90%) treated with CyBorD achieved a renal response, whereas response rates were lower with other regimens; however, these differences were not statistically significant (supplemental Figure 1C).

No significant differences in hematologic or renal response were observed between daratumumab-based regimens and nondaratumumab regimens. Among 12 patients treated with daratumumab and 23 treated with other regimens, a VGPR or better was achieved in 8 of 12 patients (66.7%) receiving daratumumab and 14 of 23 patients (60.9%) receiving other therapies. Although the difference was not statistically significant, there was a trend toward improved hematologic responses in the daratumumab group (supplemental Figure 1D). Renal responses were observed in 7 of 11 patients (63.6%) treated with daratumumab and in 17 of 22 patients (77.3%) treated with other regimens (supplemental Figure 1E).

RFS and OS outcomes

We first analyzed RFS and OS in the entire cohort of 50 patients. With a median follow-up of 6.15 years (95% CI, 5.67-6.61), median RFS was 3.7 years (95% CI, 1.82 to not reached [NR]) in the PC group and 19.2 years (95% CI, 0.79 to NR) in the B-cell group. Although RFS appeared longer in the B-cell group, the difference was not statistically significant (P = .51; supplemental Figure 2A). Similarly, no significant difference in OS was observed between the 2 groups (Figure 2A).

Figure 2.

RFS and OS in patients with MGRS. (A) Kaplan-Meier curves for OS in patients with a PC clone (n = 42) or B-cell clone (n = 8). Start date: day of diagnosis. Log-rank P = .29. (B) Kaplan-Meier curves for RFS in 35 patients with a PC clone achieving CR or very good partial hematologic response (n = 22) vs PR or lower (n = 13). Start date: first day of induction therapy. Log-rank P = .01. (C) Kaplan-Meier curves for OS in patients with a PC clone requiring HD (n = 17) vs those not requiring HD (n = 25). Start date: day of diagnosis. Log-rank P = .05. (D). Kaplan-Meier curves for RFS in patients with a PC clone who received ≥6 months (n = 19) vs <6 months (n = 18) of induction therapy. Start date: first day of induction therapy. Log-rank P = .03. (E) Kaplan-Meier curves for RFS in patients with a PC clone who received maintenance therapy (n = 13) vs no maintenance therapy (n = 24). Start date: first day of induction. Log-rank P = .04. (F) Kaplan-Meier curves for OS in patients with a PC clone who received maintenance therapy (n = 13) vs no maintenance therapy (n = 24). Start date: day of diagnosis. Log-rank P = .11.

We then focused on patients with a PC clone and evaluated disease characteristics, hematologic response, induction duration, need for HD, use of ASCT, and maintenance therapy.

Within the PC group, RFS and OS were compared between patients with PGNMID (n = 29) and those with other histopathologic diagnoses (n = 13). Median RFS was 2.9 years (95% CI, 0.57 to NR) in the PGNMID group and 8.37 years (95% CI, 0.29 to NR) in the other group; however, this difference was not statistically significant (P = .61). Similarly, OS did not differ significantly between groups (P = .598), with median OS NR in either cohort.

Patients achieving CR or VGPR had longer RFS from the start of induction (median, 8.4 years [95% CI, 2.7 to NR] vs 1.19 years [95% CI, 0.44 to NR]; Figure 2B), and the difference was statistically significant (P = .017). Median OS was NR and did not differ significantly between groups (supplemental Figure 2B). No significant differences in RFS or OS were observed based on renal response (supplemental Figure 2C-D). RFS was not influenced by HD status (P = .6; supplemental Figure 2E), whereas OS was negatively affected by the need for HD (P = .05; Figure 2C).

Analysis by induction duration, comparing treatment longer or shorter than 6 months (∼6 cycles), demonstrated improved RFS (P = .03) and OS (P = .02) with longer therapy (Figure 2D; supplemental Figure 3A).

Among 42 patients with a PC clone, 8 (19%) underwent ASCT. No significant differences in RFS or OS were observed between patients who received ASCT and those who did not (supplemental Figure 3B-C). Similarly, survival outcomes did not differ between daratumumab-based and non–daratumumab-based induction regimens.

Maintenance therapy was associated with significant improvement in RFS (Figure 2E); median RFS was NR with maintenance vs 2.9 years (95% CI, 1.03-8.15) without maintenance (P = .04). Median OS was NR in the maintenance group and was 8.48 years in the nonmaintenance group. A trend toward improved OS with maintenance therapy was observed but was not statistically significant (P = .11; Figure 2F).

Among the 11 patients who underwent kidney transplantation, 8 had PGNMID. Five patients experienced posttransplant relapse, all of whom had PGNMID, and only 1 of these patients had received maintenance.

Univariate and multivariate analyses (supplemental Tables 3 and 4), including age, creatinine, pathologic diagnosis, amount of proteinuria, induction type and duration, and maintenance therapy, demonstrated that maintenance therapy and an induction duration of >6 months were independently associated with a reduced risk of relapse (hazard ratio, 0.13; 95% CI, 0.03-0.53; P = .043) and OS (hazard ratio, 0.21; 95% CI, 0.06-0.68; P = .03), respectively, after adjusting for relevant covariates.

Discussion

Multiple new therapies are available for the treatment of PC disorders; however, because of the rarity of MGRS, it has been difficult to define the best treatment strategy for this patient population. Here, we present the experience of 50 patients treated with modern therapies, including detailed information on staging, lines of therapy, use of ASCT, and maintenance strategies. Notably, this study comprises a well-defined cohort devoid of amyloid-associated MGRS, thereby enhancing the clarity and interpretability of the findings while highlighting disease-specific limitations.

High rates of hematologic and renal CRs and VGPRs were seen with modern induction therapy. Achieving a hematologic CR or VGPR has been shown to be the most significant predictor of preventing progression to HD and OS.⁴ Our findings align with previous studies demonstrating longer RFS in patients achieving a CR or VGPR. Similarly, sustained reduction in 24-hour proteinuria over time was also observed, highlighting the central role of renal response as a meaningful end point. Recent data from the Mayo Clinic demonstrated that complete or very good partial kidney response, defined as 24-hour proteinuria <200 mg/dL, or >60% proteinuria reduction, is associated with prevention of renal replacement therapy and improved OS.¹² Overall, these data support achievement of CR/VGPR coupled with sustained proteinuria reduction as the primary treatment goals for patients with MGRS.

Challenges in assessing treatment response and relapse are particularly evident in patients without measurable disease. Many patients, especially those with PGNMID, do not present with a detectable circulating monoclonal protein or have a very small clone size. Consequently, hematologic response cannot be reliably assessed using standard criteria and must be inferred from renal parameters.¹³ Similarly, traditional renal response criteria based exclusively on proteinuria reduction are inadequate in patients with oligoanuria. In the absence of validated criteria for progression, relapse in this study was defined using an end point incorporating both hematologic progression and kidney function (serial eGFR with proteinuria measurements, along with repeat kidney biopsy). We believe this approach represents a clinically meaningful outcome that better reflects disease activity in this patient population. Future validation of such a measure is needed, along with emerging tools such as serum mass spectrometry or RNA-based immunoglobulin sequencing in BM or kidney tissue, which may ultimately improve detection of early relapse.^14,15

Regarding optimal induction therapy, we did not observe statistical differences in renal or hematologic responses with daratumumab use, likely reflecting limited sample size rather than a true lack of efficacy. Previous studies have reported VGPR/CR rates >80% with daratumumab-based combinations.⁷ Daratumumab-CyBorD has demonstrated substantial improvement in hematologic and organ responses in amyloidosis and is likely one of the most active regimens in nonamyloid MGRS.¹⁶ Larger studies will clarify the role of daratumumab-based induction in this setting, especially in patients with PGNMID.¹⁷

A notable feature of this cohort is the predominance of PGNMID histology. Although the difference did not reach statistical significance, there was a trend toward inferior outcomes, consistent with previous reports describing higher relapse rates and frequent posttransplant recurrence.^18,19 Among the patients who relapsed after kidney transplantation in our cohort, PGNMID accounted for all cases. Although limited by the small sample size, these findings suggest that PGNMID represents a particularly difficult-to-treat MGRS subtype and may warrant more intensive or prolonged therapy.

ASCT showed no significant difference in RFS or OS in our cohort, which contrasts with previous reports suggesting improved outcomes in selected patients with AL-MGRS.²⁰ However, most patients in earlier studies received less intensive induction regimens, whereas 68% of patients in our cohort received triplet or quadruplet therapy upfront. Our findings therefore suggest that deeper responses achieved with modern induction may attenuate the incremental benefit of ASCT. Given the limited sample size, detecting a true difference between ASCT and no ASCT is challenging. Moreover, the independent contributions of prolonged induction, ASCT, and maintenance therapy are difficult to disentangle. Therefore, larger, multicenter studies are required to better define the individual impact of each treatment.

Maintenance therapy emerged as one of the most impactful factors in our cohort, being associated with significantly improved RFS and a trend toward improved OS. To our knowledge, this is the first study demonstrating the benefit of maintenance therapy in patients with MGRS. Notably, most patients receiving maintenance did not undergo ASCT, which may partly explain the observed benefit and aligns with previous AL amyloidosis data showing a maintenance benefit primarily in patients who did not undergo transplantation.¹³ However, these findings must be interpreted cautiously, as patients selected for maintenance may differ systematically from those who were not, and longer duration of therapy itself may act as a surrogate for treatment intensity.

Management of patients who underwent kidney transplantation represents another area of uncertainty highlighted by our study. All but 1 patient who underwent transplantation and did not receive maintenance experienced recurrence, raising the possibility that withholding maintenance, often because of concerns about immunosuppression and infection risk, may contribute to relapse. These observations suggest that patients undergoing kidney transplantation, particularly those with PGNMID, may still require some form of posttransplant PC-directed therapy, although the optimal agent, timing, and duration remain undefined.

Overall, our data suggest that greater intensity of therapy, whether delivered through prolonged induction, ASCT, maintenance therapy, or combinations thereof, appears to correlate with improved disease control, even in patients with small or difficult-to-measure clones.

This study has several limitations. Its retrospective design introduces potential selection bias; the relatively small sample size limits its statistical power, and the predominance of PGNMID may affect generalizability. Limited data on B-cell clones and post–kidney transplant maintenance therapy restricted our ability to draw firm conclusions regarding these subgroups.

Nonetheless, these findings help clarify several unresolved questions in MGRS management and generate hypotheses for future study, including the need for standardized response criteria, the role of prolonged or maintenance therapy in preventing relapse, the optimal management of kidney transplant recipients, and the particularly high-risk nature of PGNMID. Prospective multicenter studies incorporating pathology-specific end points by kidney pathology, molecular monitoring, and minimal residual disease assessment will be essential to refine risk-adapted treatment strategies in this rare disease.

Conflict-of-interest disclosure: F.C. reports consulting or advisory role fees from Sanofi; and research funding from Pfizer. M.S. reports consulting or advisory role fees from Janssen and Regeneron. The remaining authors declare no competing financial interests.