Identifying CD38⁺ cells in patients with multiple myeloma: first-in-human imaging using copper-64–labeled daratumumab

Survival from multiple myeloma (MM) has improved over the past decade in large part because of therapeutic agents such as proteasome inhibitors, immunomodulatory agents, antibody therapy, and high-dose chemotherapy. These advances have led to high response rates and deep responses, including the achievement of minimal residual disease (MRD). Unfortunately, the majority of patients, including those who are MRD-negative, eventually relapse. It is postulated that relapse is due to residual myeloma in the patient, often not detected even with sensitive bone marrow (BM)–based techniques such as multicolor flow cytometry or polymerase chain reaction–based assays. The International Myeloma Working Group has recognized that marrow involvement in myeloma can be heterogeneous and that true MRD assessment in part depends on imaging.¹  Positron emission tomography/computed tomography (PET/CT) using ¹⁸F-fluorodeoxyglucose (¹⁸F-FDG) is a widely used imaging technique in MM as a noninvasive disease detection and monitoring strategy, and it is currently the only standardized PET tracer, in particular, with respect to response evaluation.^2 18F-FDG is a radioactive glucose analog and is taken up by cells with high metabolic activity, a hallmark of cancer cells. It is widely available and does not require local production, making it one of the most commonly used radiotracers. The importance of PET/CT imaging as an essential prognostic factor in MM is highlighted by the finding that in newly diagnosed patients the presence of more than 3 ¹⁸F-FDG-positive focal lesions are predictors of poor prognosis,³  whereas posttreatment PET-negativity alone or combined with a negative MRD signal (ie, flow cytometry or polymerase chain reaction–based sequencing) strongly correlates with better progression-free survival.^3-6  However, ¹⁸F-FDG PET/CT has limitations, in regard to both false positives due to infection and inflammation and false negatives associated with concomitant conditions, such as hyperglycemia, recent use of steroids, or very small lesions. Moreover, ¹⁸F-FDG uptake may be absent in patients expressing low levels of hexokinase 2 in myeloma cells.⁷ 

In addition, although myeloma dissemination in the BM, defined as the number of BM lesions, has a prognostic significance when assessed by magnetic resonance imaging,⁸  the pattern of marrow infiltration (focal or diffuse) does not necessarily equate with a need for immediate therapy and often is not associated with the presence of bone lesions. Given these limitations and the lack of a well-established alternative, the International Myeloma Working Group has stated that “identification of potentially more sensitive tracers that could overcome the limitations of FDG would be of great interest.”⁹ 

The monoclonal antibody daratumumab binds with high affinity to a unique epitope on CD38, a transmembrane glycoprotein that is highly and ubiquitously expressed on MM cells and at low levels on normal lymphoid and myeloid cells.¹⁰  Given its specificity for plasma cells, as well as our data showing that MM cells maintain higher CD38 levels compared with that in other immune subsets,^11,12  it is an attractive agent for use in imaging in myeloma. We chose ⁶⁴Cu as a radionuclide because of its suitable half-life (12.7 hours) for preparation and imaging along with the lack of tendency to accumulate in the bones. Our group showed that tracing MM dissemination in a preclinical model using daratumumab radio-conjugated with ⁶⁴Cu via the chelator DOTA (⁶⁴Cu-daratumumab) led to improved sensitivity and specificity over that of ¹⁸F-FDG PET/CT.¹³  Moreover, recently published clinical data using ⁸⁹Zr–labeled daratumumab showed visualization of the agent in the skeleton, liver, and spleen, but the specificity of the myeloma targeting has not been unequivocally verified.¹⁴  Here, for the first time, we are reporting the data of a ⁶⁴Cu-daratumumab dose-escalation phase 1 imaging trial, which was designed to establish the dose of unlabeled daratumumab that in combination with ⁶⁴Cu-daratumumab would be safe and optimal to increase MM targeting resolution. Also, as preliminary evidence, we used biopsies to detect the presence or absence of MM cells in regions where ¹⁸F-FDG and ⁶⁴Cu-daratumumab signaling were either concordant or discordant.

Between January 2018 and January 2020, a total of 12 patients were enrolled. Subject characteristics are described in Table 1. Our research was approved by our Institutional Review Board. All subjects gave written informed consent. Patients over 18 years of age with histologically confirmed myeloma (either newly diagnosed or relapsed disease), smoldering myeloma, or a solitary plasmacytoma were eligible. All participants had standard ¹⁸F-FDG PET/CT imaging within 60 days of registration but were not required to have measurable or detectable disease by standard PET. Adequate organ function defined as platelet count >50/mm³, absolute neutrophil acount >1000/mm³, and bilirubin <1.5 × upper limit of normal was required. The major exclusion criterion was daratumumab therapy within 3 months of enrollment. A total of 12 patients divided into 4 different cohorts (n = 3 for each cohort) were imaged.

Unlabeled daratumumab was IV administered, followed by injection of ⁶⁴Cu-daratumumab through either a peripheral or a central vein. Standard premedication for daratumumab was administered. Within 6 hours of unlabeled daratumumab administration, patients received 13.63 to 16.68 mCi of ⁶⁴Cu-daratumumab (∼5 mg). Prior to the injection of ⁶⁴Cu-daratumumab, patients were treated with 0 (cohort 1, n = 3), 10 (cohort 2, n = 3), 45 (cohort 3, n = 3), or 95 mg (cohort 4, n = 3) of unlabeled daratumumab to assess its effect on image quality. The schematic workflow of the trial design is reported in supplemental Figure 1.

Imaging was performed on a GE Discovery STE PET/CT scanner (axial field of view, 15.4 cm). PET images were performed in 3-dimensional mode (septa retracted) and corrected for tissue attenuation on the basis of coregistered CT acquired during the same examination. Attenuation and scatter-corrected PET images were reconstructed on a 128 × 128 matrix with a voxel resolution of 5.5 × 5.5 × 3.3 mm using Ordered Subset Expectation Maximization algorithm with 2 iterations and 20 subsets.

After injection of 15 mCi/5 mg of ⁶⁴Cu-daratumumab on day 0, scans were performed on days 1 and 2. Because of the limited amount of activity to be injected and the fact that only 18% of ⁶⁴Cu decays produce a positron, the count rates were expected to be low. To compensate, time per bed position was relatively long. Day 1 scans were acquired at 15 minutes per bed from the top of the head to thighs, while legs were scanned at 10 minutes per bed. On day 2, the time per bed was varied between 15 and 30 minutes within the area of interest, which was typically from the head to the pelvis.

Analysis of the imaged PET/CT data was performed on a GE AW Server. ⁶⁴Cu-daratumumab standardized uptake values (SUVs) were evaluated as SUV_mean and SUV_max in organs showing radiotracer uptake. Organ segmentation was performed using the threshold tool on the PET image data while matching the corresponding CT volume. For the BM uptake quantification, the lumbar vertebrae (L1-L5) were evaluated. A 50% SUV_max threshold was used to generate the volumes of interest (VOIs). Generated VOIs were manually edited to ensure that the VOI conformed to the vertebral marrow as seen on the CT. Because specific activity in the BM cells and the activity in blood contribute to vertebral uptake measurement, the background activity from blood was estimated to evaluate specific uptake. Briefly, SUV_mean of the blood (SUV_mean-blood) was measured by placing a VOI in the heart cavity. Using a red marrow to blood concentration ratio of 0.36,¹⁵  the blood contribution to the BM uptake was determined as $36 SUVmean−bloodSUVmean−BM %$⁠.

For subjects given >0 mg of unlabeled daratumumab, blood for pharmacokinetic analysis was drawn at ∼15 minutes, 30 minutes, 3 hours, 24 hours, and 48 hours after the end of infusion. Aliquots of whole blood and serum were counted on a gamma counter (Hidex AMG), and the percentage of injected activity in the circulation was calculated.

Imaging-based dosimetry was performed using the SUV_mean quantified in the liver, spleen, heart, and the lumbar vertebrae on day 1 and 2. Activity in each of these organs and the rest of the body was calculated using OLINDA-based standard phantom data. Radiotracer whole-body clearance was assumed to be negligible to maximize organ doses. Monoexponential decay fits to the time activity curves were used to calculate residence times and input to OLINDA, yielding the radiation dose estimates to the standard adult phantoms.

See supplemental Information for drug administration and protocol design details.

To evaluate the natural biodistribution of daratumumab, we administered ⁶⁴Cu-daratumumab without preinfusion of unlabeled daratumumab in the first cohort of patients (n = 3). We imaged 1 newly diagnosed (patient 1) and 2 relapsed MM patients (patients 2 and 3) (Figure 1A; Table 1). Although in all 3 patients we generally detected highly diffuse ⁶⁴Cu-daratumumab uptake in the spine, pelvis, liver, and spleen (Figure 1A-B), in patient 1 we observed a decrease in radiotracer uptake in the lower cervical spine (C7 SUV_max 5.9) compared with that in the lumbar spine vertebrae (L1 SUV_max 21.2) (Figure 1A-B top panel). Interestingly, this region of decreased uptake was the site of previously administered involved field radiation (Figure 1A-B). A relatively symmetric pelvic uptake was observed in all 3 patients (Figure 1C), with the exception of the right pelvic area (acetabulum) of patient 3, which was again identified as the site of involved field radiation (Figure 1C bottom panel).

We investigated whether normal organ uptake interferes with MM-specific biodistribution of ⁶⁴Cu-daratumumab. Hence, we proceeded with a dose escalation design using an increased amount (10, 45, and 95 mg) of unlabeled daratumumab before ⁶⁴Cu-daratumumab infusion. In the cohort of patients pretreated with 10 mg of unlabeled daratumumab and then given ⁶⁴Cu-daratumumab, we observed reduced radiotracer uptake in the liver (L-SUV_mean 11.7 ± 4.0) and spleen (S-SUV_mean 23.6 ± 3.7), and a concomitant increase in BM total uptake (BM-SUV_mean 19.2 ± 3.1) by day 2, compared with levels in the first cohort of patients in which the unlabeled antibody was not administered (day 2 SUV_mean: L, 16.1 ± 0.6; S, 47.2 ± 16.3; BM, 13.7 ± 1.1) (Figure 2A-B; Table 2).

Moreover, a further noticeable decrease in the liver and spleen SUV_mean (SUV_mean: L, 6.3 ± 1.9; S, 8.6 ± 2.4) was observed when the third cohort of patients, 2 with MM (patient 7 and patient 8) and one with a solitary plasmacytoma (patient 9), was pretreated with 45 mg of unlabeled antibody (Figure 2A-B; Table 2). Although the 45-mg cohort clearly showed a decrease in the liver and spleen (SUV_mean: L, 6.3 ± 1.9; S, 8.6 ± 2.4), an almost twofold increase in BM SUV_mean in patients 7 and 8 was observed (BM-SUV_mean 33.9 ± 11.0) over that in the patients given no unlabeled daratumumab (BM-SUV_mean: 13.7 ± 1.1) (Table 2). Of note, the ⁶⁴Cu-daratumumab distribution in the BM of the patient with solitary plasmacytoma showed a completely divergent pattern compared with that observed in the patients with systemic MM (Table 3) and is further discussed below. Finally, when the dose of unlabeled daratumumab was raised to 95 mg, minimal uptake was observed in any of the organs (Figure 2A-B).

The first cohort of patients exhibited rapid clearance of the radiotracer from the bloodstream, indicating that the antibody was preferentially retained by CD38⁺ cells in the BM, rather than circulating CD38⁺ cells. A general trend of gradual increase in circulating radiotracer antibody was observed with increased amount of unlabeled daratumumab at early time points (1 to 3 hours) (Figure 2C; supplemental Figure 2). However, this effect was not observed at the 24-hour time point, at which only the patient with a plasmacytoma and the fourth cohort of MM patients (95 mg) maintained high levels of circulating radiotracer (Figure 1C), supporting that the unlabeled daratumumab may had saturated CD38⁺ cells in the organs of these patients, as well as BM MM cells. As evidence of this saturation, a significant increase in circulating ⁶⁴Cu-daratumumab was observed in the 95-mg cohort of patients at both early (P = .013) and late (P = .005) time points over that in the other cohorts (Figure 2C).

For the first 3 cohorts, the background radioactivity from the blood was found to contribute <5% of the radioactivity measured in the BM (Table 3). This finding indicates that most of the radiotracer measured in the BM was specifically bound to cells within the BM. The saturation of the receptors in the BM in patients infused with 95-mg cold antibody was evident, with roughly 42% and 23% of the BM activity attributed to blood-based background signal by day 1 and day 2, respectively.

We further investigated the distribution of the radiotracer in the cohort given 45 mg unlabeled daratumumab by comparing uptake in the patients with systemic MM (n = 2) with that in the patient with a solitary plasmacytoma. In the patients with systemic MM, we observed a diffuse uptake of ⁶⁴Cu-daratumumab throughout the spine (Figure 3A) and a more heterogeneous uptake in the pelvis area, the ribs, and the extremities (Figure 3A-B). Asymmetric focal uptake was observed in the ribs of 1 patient and in the calvarium of both MM patients (Figure 3B). Interestingly, although the 45-mg dose of unlabeled daratumumab also reduced liver and spleen uptake in the patient with a plasmacytoma, as expected, the BM uptake was substantially reduced (day 2 BM-SUV_mean 8.4 ± 2.2 vs MM BM-SUV_mean 33.9 ± 11.0) (Figure 3B-C; Table 3), an effect we believe is aligned with the absence of MM BM involvement in this patient and with the possible saturation of all CD38 sites by the unlabeled daratumumab. Consistent with this hypothesis, we also did not observe ⁶⁴Cu-daratumumab uptake in the vertebral plasmacytoma, which underwent biopsy and was shown to be CD138⁺ and CD38⁺ by immunohistochemistry (supplemental Figure 3). Indeed, the level of radiotracer in the blood of the patient with a plasmacytoma was comparable to levels found in the cohort receiving a saturating dose of 95 mg unlabeled daratumumab and was higher than the levels in the other patients given the 45-mg dose (Figure 2C), further supporting that the 45-mg unlabeled dose, in the absence of MM BM involvement, may saturate many CD38⁺ sites. Consistent with these results, the radioactivity in the blood contributing as a background signal to the BM was only ∼1% (Table 3) in patients 7 and 8 with MM. Background contribution to the BM from the blood radioactivity was ∼12% on day 2 for the patient with plasmacytoma.

We compared ⁶⁴Cu-daratumumab vs ¹⁸F-FDG biodistributions generated by our data, and, when possible, performed biopsies in areas in which we found discordant results between these 2 tracers. In general, higher spinal, pelvic, and femoral uptake of ⁶⁴Cu-daratumumab compared with that of ¹⁸F-FDG was observed in all the patients analyzed.

Specifically, patient 1 (newly diagnosed) showed extensive uptake of both tracers (Figure 4A), and comparable focal uptake between the 2 was observed in the right and left humerus (Figure 4A). We also observed a diffuse uptake of ⁶⁴Cu-daratumumab in the pelvis and a more focal uptake in the femoral bones and legs (patient 7), but no specific ¹⁸F-FDG signals were observed in these areas (Figure 4A-B). A concordant positive signal (SUV_max of 14.7 on ⁶⁴Cu-daratumumab PET/CT vs 3.3 on ¹⁸F-FDG PET/CT) was also observed in the sternum of patient 1, a site that was confirmed positive for myeloma infiltration by biopsy (Figure 4C). In the iliac crest region of the same patient, we noted no ¹⁸F-FDG signal but observed ⁶⁴Cu-daratumumab uptake; this site was found to be positive for myeloma by biopsy (20% to 30% CD138⁺ MM plasma cells) (Figure 4C). Discordant uptake of the 2 tracers in the iliac crest was also observed in patient 8. The iliac crest of this patient was also positive by ⁶⁴Cu-daratumumab imaging and by biopsy (60% plasma cell infiltration), but negative by ¹⁸F-FDG (Figure 4D). Conversely, in another patient (patient 2), we observed positive uptake of ¹⁸F-FDG in the right clavicular region that instead was negative according to both ⁶⁴Cu-daratumumab signal and biopsy (Figure 4E). In patient 3, a positive signal by ¹⁸F-FDG PET/CT was seen in a pleural-based mass, with an SUV_max of 9.0. In contrast, this region did not show specific uptake of ⁶⁴Cu-daratumumab and was negative by biopsy, suggesting that the ¹⁸F-FDG uptake was a false positive (Figure 4F). Furthermore, focal ⁶⁴Cu-daratumumab uptake in the calvarium was observed in 8 out of 12 MM patients analyzed, and in patient 1, the focal uptake was mainly localized around calvarium bone lesions (Figure 4G), suggesting that antibody-based imaging may be superior to identify the presence of cancer cells at this site, which may be obscured by high uptake of ¹⁸F-FDG in the brain (Figure 4A,C,G).

Absorbed dose estimates calculated using the data from the third cohort of patients preinfused with 45 mg of cold antibody are shown in supplemental Table 1. Red marrow, spleen, and liver doses were 0.32 ± 0.22, 0.24 ± 0.07, and 0.16 ± 0.06 mSv/MBq, respectively. The whole-body effective dose was 0.07 ± 0.02 mSv/MBq.

We have validated our preclinical data that ⁶⁴Cu-daratumumab PET/CT imaging provides whole-body imaging of MM. In this first-in-human trial, we also demonstrated the safety and potential clinical applications of this immune-PET imaging strategy.

Daratumumab has become a widely used antibody in MM therapy.^16-18  Toxicities are moderate and manageable, the most common being infusion-related toxicities, which are frequently encountered with monoclonal antibodies. Both the amounts of ⁶⁴Cu-daratumumab (∼5 mg) and unlabeled daratumumab (≤95 mg) used in this study are considerably less than the typically prescribed 16-mg/kg dose.

Our data suggest that the 45-mg cold dose may be optimal for myeloma imaging, as liver and spleen uptake were greatly reduced while the marrow uptake was maximized, in contrast to the other cohorts. The higher dose of 95 mg appeared to saturate the CD38 receptors and limit uptake by known sites of disease. This concept of receptor saturation is further supported by results with the single patient with a solitary plasmacytoma, in whom a 45-mg cold dose may have exceeded the optimal amount by saturating all target sites.

In addition, this finding may have implications in patients pretreated with daratumumab. Although we have observed that patients who are no longer responding to daratumumab continue to exhibit expression of CD38 on their plasma cells,¹¹  it is possible that patients with high levels of circulating daratumumab from prior therapy may have receptor saturation at the time of administration of radiolabeled daratumumab. We will explore this aspect further in the phase 2 trial, where we will not exclude patients currently on daratumumab therapy. The study procedures were well tolerated; in 1 patient, we observed grade 3 toxicities of hypoxia and hypertension related to the infusion of unlabeled daratumumab, in keeping with the known side effects of the drug, which was successfully managed with additional diphenhydramine and hydrocortisone. Recently, 1 phase 3 noninferiority trial disclosed that a subcutaneous route of administration leads to significantly lower rates of these reactions (12.7% vs 35%).¹⁹  No serious adverse events were associated with ⁶⁴Cu-daratumumab.

The design of the immunological imaging agent needs to specifically consider the radionuclide, the chelator, and the antibody. A radionuclide with an appropriate half-life (order of hours to days) is essential. In this regard, both ⁶⁴Cu (t_1/2 = 12.7 hours) and ⁸⁹Zr (t_1/2 = 78.4 hours) are being explored for imaging MM disease burden, with their half-lives facilitating imaging times up to 2 days and a week, respectively. The advantages of ⁶⁴Cu-based imaging compared with ⁸⁹Zr-based imaging include the possibility of administering greater activity for better sensitivity or lowering the radiation dose to the patient for similar imaging results. The longer half-life of ⁸⁹Zr compared with ⁶⁴Cu results in an increased radiation dose from ⁸⁹Zr and typically limits the amount of injected radioactivity for comparable dosimetry. The improved dosimetry of ⁶⁴Cu over ⁸⁹Zr becomes especially important with respect to highly specific radiotracer uptake and retention in the BM. Since increased disease burden will result in increased radiotracer accumulation in the BM, the extended retention of radioactivity in the BM results in increased radiation dose. The delayed imaging with ⁸⁹Zr is advantageous for cases where the signal from disease is significantly hampered by background tissue uptake. In the ⁸⁹Zr study, imaging 5 to 6 days postradiotracer infusion has been suggested as a potential imaging timeframe on the basis of the femoral uptake and the clearance of activity from the blood. In our trial, we note that the background contribution of the blood radioactivity is a minor part of the activity measured in the BM. Of the 9 patients imaged in the first 3 cohorts, the only case of elevated levels of circulating radiotracer at the 24-hour time point, as could also be seen on the PET imaging scans, was of the patient with a solitary plasmacytoma, who had the highest blood SUV_mean compared with the other patients analyzed. By 48 hours, the presence of the radiotracer was not seen (Figure 3C), a result supported by the much lower activity measured in the blood at day 2 (supplemental Figure 2).

Quantitative evaluation indicated that >95% of the activity measured in the BM was specifically bound and <5% of the activity was attributed to background contribution from blood. This finding implies that ⁶⁴Cu-daratumumab is rapidly taken up by the BM of patients with MM, leaving little in the blood circulation by 48 hours, supporting rapid elimination of the radiotracer background and feasibility of imaging within the first 2 days of radiotracer infusion.

Although we do not hypothesize that the high BM SUV is an indication of widespread diffuse disease, one can postulate that there is likely more extensive bone and marrow involvement detected by ⁶⁴Cu-daratumumab imaging than in standard imaging, which will be investigated further.

Our trial is the first of radiolabeled daratumumab imaging to use biopsies to investigate discordances between ¹⁸F-FDG PET/CT and ⁶⁴Cu-daratumumab imaging. We showed that 2 lesions that were ¹⁸F-FDG positive but ⁶⁴Cu-daratumumab negative did not have evidence of MM by tissue biopsy and instead had evidence of an inflammatory reaction. Although we cannot exclude a heterogeneous distribution of myeloma being responsible for the failure to detect disease by biopsy, these patients did not progress in the regions where these biopsies occurred upon clinical follow-up. We also demonstrated that lesions that were ¹⁸F-FDG negative but ⁶⁴Cu-daratumumab positive were positive for myeloma involvement on biopsy. Also of note was the skull uptake: since there was no brain uptake of ⁶⁴Cu-daratumumab in contrast to standard ¹⁸F-FDG, we were able to demonstrate focal calvarial uptake, which allows better evaluation of skull plasmacytomas. These findings lay the groundwork for a phase 2 trial comparing the sensitivity and specificity of both imaging strategies. A key disadvantage in general of antibody-based imaging is the extended time for which the patient needs to be scanned compared with ¹⁸F-FDG PET. ¹⁸F-FDG PET scans are typically acquired at 2 to 3 minutes per bed position. ⁶⁴Cu-DOTA–daratumumab PET scans in this trial were acquired between 10 and 30 minutes per bed position on the Discovery STE PET/CT scanner to maximize the imaging signal. Given the high SUVs measured in patient BM, reduction in scan time can be explored in the future. Recent developments on whole-body PET scanners²⁰  would make antibody-based imaging much faster than the current protocols.

If successful, ⁶⁴Cu-daratumumab PET/CT may be applied in several settings. For smoldering MM, laboratory values, standard imaging methods, and clinical criteria are currently used to differentiate smoldering MM from active MM. However, this area remains in significant flux, as evidenced by the changing definitions of smoldering myeloma.^21,22  It would be of interest to investigate the utility of ⁶⁴Cu-daratumumab PET/CT in the early detection of disease as a potentially more sensitive imaging modality than standard ¹⁸F-FDG PET/CT, which in turn may lead to better selection of patients who could benefit from early therapy. We envision that evaluation of MRD in MM will eventually be supplemented or even replaced by such a molecular imaging modality. Furthermore, ⁶⁴Cu-DOTA–daratumumab PET/CT may be studied as a potential predictor of response to therapies. A key advantage of using ⁶⁴Cu-DOTA–daratumumab is the ease of chelation of numerous radionuclides to the DOTA chelator, potentially allowing imaging with ⁶⁴Cu-DOTA–daratumumab to guide daratumumab-based radiotherapies. Moreover, the possibility of performing imaging at earlier time points may favor a rapid initiation of therapy and facilitate improved patient management.

The authors acknowledge support from the Steven Gordon and Briskin Family Innovation Grant and are grateful to Young Kim for assistance with immunohistochemistry. The authors are also grateful for support from Steve Allen, Suzan King, and the Tom Klemens Family Foundation.

The Biostatistics and Mathematical Modeling Core and the GMP Manufacturing Core at City of Hope were supported by the National Institutes of Health, National Cancer Institute under award number P30CA033572. This research was also in part supported by National Institutes of Health, National Cancer Institute NIH-2-R01-CA238429-01 (F.P., J. Shively, and R.R.).

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Contribution: A.K., V.A., J.P., A. Chaudhry, E.C., G.M., R.R., A.M.W., J.W., S.J.F., D.C., P.Y., J. Shively, and F.P. designed research; A.K., V.A., E.K.P., A. Chaudhry, V.E.B.-a., N.B., N.N., M.R., F.S., C.K., and J. Simpson performed research; A.K., V.A., J.P., A. Chaudhry, J.F.S., D.Y., M.P., A. Chowdhury, and F.P. analyzed data; A.K., V.A., J.P., J. Simpson, J.F.S., and F.P. wrote the manuscript; and all authors reviewed the manuscript.

Conflict-of-interest disclosure: A.K. is a consultant for Bristol Myers Squibb, GlaxoSmithKline, Janssen Pharmaceuticals, Karyopharm Therapeutics, and Regeneron Pharmaceuticals; is on the scientific advisory board for Sutro Biopharma; serves on the speakers bureau for Amgen, Bristol Myers Squibb, and Takeda; and has stock ownership in Bristol Myers Squibb. J.P. is a consultant for Gilead. M.R. is on the speakers bureau for Janssen Pharmaceuticals. The remaining authors declare no competing financial interests.

Correspondence: Amrita Krishnan, The Judy and Bernard Briskin Center for Multiple Myeloma Research, City of Hope, 1500 E Duarte Rd, Duarte, CA 91010; e-mail: akrishnan@coh.org; and Flavia Pichiorri, The Judy and Bernard Briskin Center for Multiple Myeloma Research, Department of Hematological Malignancies Translational Science, City of Hope, 1500 E Duarte Rd, Duarte, CA 91010; e-mail: fpichiorri@coh.org.