Optimal cut-off values and diagnostic significance of clinical laboratory indicators in newly diagnosed multiple myeloma.

Introduction

Multiple myeloma (MM) is a malignancy originating from plasma cells, characterized by uncontrolled proliferation and accumulation of plasma cells in the bone marrow. This disease often leads to a spectrum of systemic manifestations, including renal failure, hypercalcemia, bone destruction, and anemia due to bone marrow failure [ 1 ]. In the United States, MM accounts for approximately 1.8% of all cancer cases and 18.7% of hematologic malignancies. Its prevalence is slightly higher among men than women, with a median age of onset at 69 years [ 2 , 3 ]. For the year 2023, the American Cancer Society estimates that 35,730 new cases of MM will be diagnosed, with an estimated 12,590 associated deaths [ 3 ].

The initial clinical presentation of MM is highly diverse, involving multiple organ systems. Literature reports indicate that anemia is observed in 73% of cases, osteolytic bone destruction in 79%, and acute kidney damage in 19% [ 4 ]. Additionally, patients may present with hypercalcemia (13%), lymphadenopathy (1%), leukopenia (20%), and thrombocytopenia (5%) [ 4 ]. Given that MM predominantly affects elderly individuals with various underlying health conditions, its characteristic symptoms can often go unnoticed, leading to a high risk of initial misdiagnosis [ 5 ]. According to estimates from available literature, the time elapsed from the onset of initial health changes to seeking medical attention ranges from 1 to 7 months, with the time from first medical consultation to final diagnosis spanning from 2 weeks to 17 months [ 6 ]. Initial misdiagnosis rates for MM patients can reach as high as 13.7–32.2%, and literature shows that due to the invasive and complex diagnosis process, less than half of patients receiving a correct diagnosis within the first 6 months [ 7 ].

A delay in the diagnosis of MM is often associated with more advanced disease progression, resulting in a cascade of adverse consequences, including a diminished quality of life and reduced long-term survival [ 8 ]. Timely recognition and management are paramount in improving patient outcomes. When MM is not identified early during the initial diagnosis by the primary healthcare provider and subsequently referred to specialists, such as oncologists and hematologists, the patient may experience a significant delay in diagnosis. This delay can lead to suboptimal control of disease symptoms [ 8 , 9 ]. Studies reveal that 40% of MM patients experience clinical symptoms for up to 46 months before diagnosis, with these patients often presenting multiple complications. Conversely, 40% of patients with symptoms lasting less than 3 months and timely diagnosis show no complications [ 10 ]. Furthermore, research indicates that prolonged diagnosis times do not significantly impact overall survival but do affect the onset of initial symptoms ( p < 0.043) and disease-free survival after diagnosis ( p < 0.003) [ 10 ]. Therefore, reducing the time to MM diagnosis can substantially enhance long-term patient prognosis and overall survival.

Current diagnostic criteria for MM, as outlined in guidelines from the National Comprehensive Cancer Network (NCCN) [ 11 ] and the International Myeloma Working Group (IMWG) [ 12 ], include the presence of 10% or more clonal plasma cells in bone marrow or evidence of biopsy-confirmed plasmacytoma, along with one or more myeloma defining events (MDEs). MDEs encompass established CRAB characteristics (hypercalcemia, renal failure, anemia, or lytic bone lesions) and three specific biomarkers: clonal bone marrow plasma cells ≥ 60%, serum free light chain (FLC) ratio ≥ 100 (with involved FLC levels ≥ 100 mg/L), and evidence of one or more bone destructions greater than 5 mm observable in MRI. Additionally, the revised international staging system incorporates data from serum biomarkers such as beta2 microglobulin, albumin, lactate dehydrogenase, fluorescence in situ hybridization [t(4;14), del(17p), t(14;16)] to assess progression-free survival and overall survival [ 13 ].

Hence, initial diagnostic workup for suspected MM typically includes a battery of tests: blood tests (complete blood count, comprehensive metabolic panel, calcium levels, coagulation profile, serum protein electrophoresis, immunofixation electrophoresis, β 2 microglobulin, etc.), urine analysis (urinalysis, urine protein electrophoresis, urine immunofixation electrophoresis, 24-hour urine light chain quantification, etc.), bone marrow examination (cytology, biopsy, immunohistochemistry, flow cytometry, fluorescence in situ hybridization, etc.), and radiological examination (local or whole-body MRI, whole-body positron emission tomography PET-CT), among others [ 14 , 15 ]. The complexity of the MM diagnostic process, coupled with the expense and invasiveness of some tests, restricts access to accurate diagnosis and treatment resources for many patients in underserved regions, leading to protracted and costly diagnostic journeys. Additionally, limited knowledge of MM among primary care practitioners and emergency physicians in certain areas exacerbates diagnostic challenges, with studies revealing that 50.6% (939/1854) of MM patients visited healthcare facilities multiple times before being referred to hematology departments [ 16 ]. If clinicians in initial diagnostic institutions conduct timely stratification and screening for suspected high-risk MM patients after evaluating their condition, it will increase the detection rate of MM in outpatient clinics. This will benefit the improvement of the quality of life and long-term prognosis of MM patients and reduce the incidence of related severe complications. Therefore, there is an urgent need in clinical practice for a non-invasive method that can quickly screen for MM through basic hematological tests. For example, refining diagnostic thresholds for indicators significantly related to MM diagnosis to initially screen high-risk MM patients and constructing an initial MM diagnosis model can assist general practitioners and hematologists. This approach will shorten the diagnosis time for MM and improve the overall prognosis of patients.

This study aims to utilize fundamental laboratory indicators such as complete routine hematological and biochemical parameters to identify optimal cutoff values for indicators associated with MM diagnosis. The refinement of threshold ranges will aid in clinical screening. Furthermore, traditional machine learning algorithms will be employed to construct an MM diagnostic model, enhancing the efficiency of MM diagnosis. This research aims to assist general practitioners and doctors in initial MM diagnosis, reducing the time taken for MM diagnosis, and improving the overall prognosis for patients.

Materials and methods

Participants

Based on the Medical Big Data Research Center of Zhejiang Provincial People’s Hospital, we conducted a retrospective analysis of 279 NDMM patients and 553 healthy subjects who visited the outpatient clinic of Zhejiang Provincial People’s Hospital from January 2008 to June 2023. Using January 2018 as the cutoff, the samples were divided into two groups, serving as the experimental group and the validation group in this study.

In the training group, we conducted a retrospective analysis of 184 patients diagnosed with MM. These patients received treatment at the outpatient department of Zhejiang Province People’s Hospital during the period from January 2008 to January 2018. Additionally, we included a control group consisting of 362 healthy individuals. The criteria for inclusion and exclusion of MM patients in the experimental group were as follows:

Inclusion Criteria: Adherence to the NCCN Guidelines for Diagnosis and Treatment (revised in 2022) [ 11 ]. Initial visit for MM diagnosis. Availability of complete clinical data. Exclusion Criteria: Receipt of specialized treatments such as radiotherapy and chemotherapy. Coexistence of other diseases, including malignant solid tumors, severe acute infections, severe autoimmune diseases, MM-related amyloidosis, complicating hematological disorders, and severe conditions affecting vital organs such as the heart, liver, kidney, and nervous system. Substantial gaps in clinical data.

To assess the diagnostic efficacy and generalizability of our model, we conducted a new retrospective analysis of 95 MM patients and 191 healthy subjects. These individuals sought medical care at our hospital between February 2018 and June 2023 and met the same inclusion and exclusion criteria mentioned above. This dataset, denoted as “dataset 2”, was used as the external validation cohort. The specific sample screening process is illustrated in Fig. 1 . Fig. 1 Flow chart of comprehensive sample screening

This study received approval from the Ethics Committee of Zhejiang Province People’s Hospital, and the research was conducted in accordance with the principles outlined in the Declaration of Helsinki (Ethical approval code: QT2023281). A retrospective study was conducted on electronic medical records in this experiment, and all individuals had provided informed consent before being included in the study.

Clinical and laboratory data collection

For this study, we retrospectively gathered biochemical and blood routine data from the electronic medical record database of Zhejiang Province People’s Hospital. In accordance with current clinical diagnostic criteria and routine diagnostic practices, we considered age and gender alongside a selection of pertinent laboratory indices, which include: Albumin (Alb), globulin (Glo), total protein (TP), albumin to globulin ratio (AGR), serum calcium (Ca), serum creatinine (Cr), lactate dehydrogenase (LDH), hemoglobin (Hb), platelet count (Plt), platelet distribution width (PDW). All these parameters were assessed at the patients’ initial clinic visit, ensuring simplicity, accessibility, and cost-effectiveness in data collection.

Statistical analysis

Statistical analysis was conducted using SPSS 26.0, SAS 9.4, and R 4.0.3 software. Categorical variables were described in terms of frequencies and percentages, and differences between the two groups were assessed using the chi-square (χ 2 ) test. Numerical variables were subjected to the Kolmogorov-Smirnov test to assess normality. If continuous variables were normally distributed, an independent T-test was employed and described as mean standard deviation. Conversely, non-normally distributed continuous variables were analyzed using the Wilcoxon rank-sum test and described as [M(P 25 , P 75 )]. Multivariate correlation analysis employed binary logistic regression to identify the most relevant features for model construction. Statistical significance was defined as p < 0.05.

By estimating the maximum Youden index for each clinical laboratory indicator after multifactorial selection, optimal cutoff points for each indicator in the diagnosis of MM were determined. In this experiment, traditional machine learning techniques were employed, and a model was established using the logistic regression algorithm. Logistic regression is an algorithm that assesses the probability of an event occurring through factors or features associated with the occurrence or impact of the event. Widely applied in medical research, it demonstrates high accuracy in disease diagnosis and risk prediction [ 17 , 18 ]. Dataset 1 was utilized for model development, while dataset 2 was employed for external validation. The logistic regression model’s performance was evaluated using the Receiver Operating Characteristic (ROC) curve and metrics such as sensitivity and specificity.

Results

Patient characteristics

The overall baseline characteristics of the total sample are shown in Table 1 . Chi-square test results indicate that there is no statistically significant difference in gender between the two groups. However, independent sample T-test results show a significant age difference between the groups. Among the clinical laboratory indicators, all but LDH show statistically significant differences ( p < 0.05).

Table 1 The clinical characteristics of all patients Characteristics MM group ( n = 279) Control group ( n = 362) Statistical values p Male 178 (63.80) 350 (63.29) χ 2 = 0.21 0.886 Age (years) 68.0 (59.0–74.0) 56.0 (34.0–69.0) Z = 9.385 < 0.001 Alb (g/L) 33.3 (28.7–8.2) 43.6 (41.5–45.4) Z= − 18.98 < 0.001 Glo (g/L) 42.5 (26.0–68.0) 28.4 (26.1–30.5) Z = 10.155 < 0.001 TP (g/L) 76.9 (65.5–98.6) 72.4 (69.3–74.6) Z = 4.896 < 0.001 AGR 0.79 (0.44–1.42) 1.55 (1.41–1.71) Z= − 14.54 < 0.001 Ca (mmol/L) 2.25 (2.11–2.41) 2.34 (2.28–2.39) Z= − 6.695 < 0.001 Cr (mg/dL) 1.060 (0.839–1.683) 0.900 (0.807–1.005) Z = 8.277 < 0.001 LDH (U/L) 174.0 (140.0–221.0) 176.0 (160.0-198.0) Z= − 0.905 0.365 Hb (g/dL) 9.20 (7.40–11.00) 14.50 (13.20–15.70) Z= − 21.04 < 0.001 Plt (/mm 3 ) 155,000 (113000–201000) 210,000 (173000–250000) Z= − 10.04 < 0.001 PDW (fL) 12.6 (10.7–14.9) 13.1 (11.6–14.9) Z= − 2.850 0.004 Normally distributed variables are expressed as “x ± s (mean ± standard deviation),” while non-normally distributed variables are expressed as “[M(P 25 , P 75 )]” Alb Albumin, Glo Globulin, TP Total Protein , AGR Albumin to Globulin Ratio, Cr Serum Calcium, Ca Serum Calcium, Hb Hemoglobin, LDH lactate dehydrogenase, Plt Platelet Count, PDW Platelet Distribution Width

The baseline characteristics of dataset 1 are presented in Table 2 . In dataset 1, gender and age distribution showed no statistically significant difference between the two groups based on the results of the chi-square test and Wilcoxon rank-sum test. However, among all clinical parameters, the MM group had higher levels of Glo, TP, Cr and LDH compared to the control group, with the difference in Glb being the most significant. On the contrary, the levels of Alb, AGR, Ca, Hb, PDW, and Plt in the MM group showed a decreasing trend compared to the control group, with Alb exhibiting the most significant difference. All clinical and laboratory parameters, except for LDH and TP, showed statistical significance ( p < 0.05).

Table 2 The characteristics of the development cohort (cohort 1) Characteristics Reserch group ( n = 184) Control group ( n = 362) Statistical values p Male 109 (59.23) 212 (58.56) χ 2 = 0.21 0.639 Age (years) 68.0 (58.0–75.0) 66.0 (57.0–74.0) Z = 0.46 0.651 Alb (g/L) 33.80 (29.80–39.30) 43.80 (41.6–45.5) Z= − 14.68 < 0.001 Glo (g/L) 37.10 (24.60–60.80) 28.80 (26.9–31.0) Z = 5.036 < 0.001 TP(g/L) 73.6 (63.4–89.8) 72.8 (70.3–75.0) Z = 1.310 0.190 AGR 0.91 (0.47–1.52) 1.54 (1.39–1.67) Z= − 9.74 < 0.001 Ca (mmol/L) 2.25 (2.12–2.42) 2.33 (2.27–2.38) Z= − 4.30 < 0.001 Cr (mg/dL) 1.049 (0.850–1.633) 0.906 (0.827–1.015) Z = 6.161 < 0.001 LDH (U/L) 176.47 (142.10–226.40) 174.50 (160.33–191.87) Z = 0.73 0.478 Hb (g/dL) 9.41 ± 2.59 14.16 ± 16.95 t = 25.73 < 0.001 Plt (/mm 3 ) 154,500 (110500–211500) 193,000 (161000–231000) Z= − 5.48 < 0.001 PDW (fL) 11.87 (10.43–15.51) 14.20 (11.30–15.47) Z= − 6.15 < 0.001 Normally distributed variables are expressed as “x ± s (mean ± standard deviation),” while non-normally distributed variables are expressed as “[M(P 25 , P 75 )]” Alb Albumin, Glo Globulin, TP Total Protein , AGR Albumin to Globulin Ratio, Cr Serum Calcium, Ca Serum Calcium, Hb Hemoglobin, LDH lactate dehydrogenase, Plt Platelet Count, PDW Platelet Distribution Width.

Table 3 provides an overview of the baseline characteristics of the samples in dataset 2. The chi-square test results revealed no significant difference between the two groups in terms of gender. However, independent sample T-tests indicated a notable age difference between the MM and control groups. In the MM group, TP, Cr and Glo levels were higher than those in the control group, with Glo displaying the most significant difference. Conversely, compared to the control group, the MM group showed lower levels of Alb, AGR, Ca, LDH, Hb, Plt, and PDW, with the most significant difference observed in the PLT group. All clinical and laboratory parameters, except for PDW and LDH, exhibited significant differences ( p < 0.05).

Table 3 The characteristics of the validation cohort (cohort 2) Characteristics Reserch group ( n = 95) Control group ( n = 191) Statistical values p Male 69 (72.63) 144 (75.39) χ 2 = 0.16 0.671 Age (years) 68.0 ± 10.7 33.0 ± 10.1 t = 32.05 < 0.001 Alb (g/L) 32.1 (27.5–35.7) 43.20 (41.3–45.0) Z= − 12.07 < 0.001 Glo (g/L) 51.7 (34.2–73.2) 27.3 (24.7–29.8) Z = 9.964 < 0.001 TP (g/L) 84.5 (70.0–101.9) 70.6 (67.3–73.9) Z = 6.505 < 0.001 AGR 0.63 (0.36–0.99) 1.61 (1.45–1.77) Z= − 10.93 < 0.001 Ca (mmol/L) 2.25 (2.11–2.37) 2.36 (2.28–2.43) Z= − 5.486 < 0.001 Cr (mg/dL) 1.077 (0.827–1.710) 0.878 (0.755–0.972) Z = 5.617 < 0.001 LDH (U/L) 171.0 (136.0–208.0) 182.0 (159.0-210.0) Z= − 2.245 0.024 Hb (g/dL) 8.70 (7.60–10.70) 14.90 (13.90–16.10) Z= − 12.84 < 0.001 Plt (/mm 3 ) 155,000 (116000–191000) 235,000 (195000–267000) Z= − 9.23 < 0.001 PDW (fL) 13.12 (11.07-15.00) 13.10 (11.90–15.60) Z= − 1.94 0.093 Normally distributed variables are expressed as “x ± s (mean ± standard deviation),” while non-normally distributed variables are expressed as “[M(P 25 , P 75 )]” Alb Albumin, Glo Globulin, TP Total Protein , AGR Albumin to Globulin Ratio, Cr Serum Calcium, Ca Serum Calcium, Hb Hemoglobin, LDH lactate dehydrogenase, Plt Platelet Count, PDW Platelet Distribution Width.

Multifactor analysis

Binary logistic regression analysis, with “yes” indicating the likelihood of developing MM as the dependent variable and statistically significant ( p < 0.05) variables identified through univariate analysis as independent variables, was performed using a stepwise sampling method with inclusion and exclusion criteria of 0.05 and 0.1, respectively. The results of the binary logistic regression analysis are shown in Table 4 , which revealed that Alb, Hb, and PDW were statistically significant ( p < 0.001).

Table 4 Logistic regression analysis of Alb, Hb, and pdw to predict the occurrence of MM Characteristics B S.E. Wald p OR (95%CI) Intercept 17.403 1.996 76.037 < 0.001 Alb − 0.243 0.037 38.089 < 0.001 0.87 (0.83–0.91) Hb − 0.121 0.018 69.672 < 0.001 0.90 (0.87–0.93) PDW − 0.054 0.021 11.074 < 0.001 0.94 (0.92–0.96) Alb Albumin, Hb Hemoglobin, PDW Platelet Distribution Width

Optimal thresholds for Hb, Alb, and PDW

Through the application of the logistic regression algorithm and ROC curve analysis, the maximum Youden index for each relevant indicator was determined following multifactor screening. The optimal cutoff point for each diagnostic test indicator in MM diagnosis was established by identifying the maximum Youden index (calculated as sensitivity + specificity − 1) as the optimal threshold. Specific results can be found in Table 5 . The optimal threshold value for Hb was identified as 11.6 g/dL, with an associated area under the curve (AUC) of 0.931 (95% CI 0.905–0.956). Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were observed at 0.799, 0.943, 0.877, and 0.902, respectively.

For Alb, the optimal threshold value was determined to be 39.30, with an AUC of 0.890 (95% CI: 0.857–0.923). Sensitivity, specificity, PPV, and NPV at this threshold were found to be 0.777, 0.898, 0.794, and 0.888, respectively. It is noteworthy that the AUC for PDW was the lowest, recorded at 0.663 (95% CI: 0.615–0.710). The optimal threshold value for PDW was identified as 14.1, resulting in a high sensitivity of 0.939 but lower specificity, PPV, and NPV at 0.327, 0.415, and 0.913, respectively, across the three groups. Detailed comparisons of these indicators can be found in Table 5 and Fig. 2 A.

Table 5 Optimal cutoff values for Alb, Hb and Pdw Characteristics AUC p Cut-off Value Sensitivity Specificity Youden Index PPV NPV Model 0.960(0.942–0.978) < 0.001 0.430 0.860 0.957 0.818 0.911 0.931 Alb 0.890(0.857–0.923) < 0.001 39.30 0.777 0.898 0.674 0.794 0.888 Hb 0.931(0.905–0.956) < 0.001 11.6 0.799 0.943 0.742 0.877 0.902 PDW 0.663(0.615–0.710) < 0.001 14.10 0.939 0.327 0.265 0.415 0.913 Alb Albumin, Hb Hemoglobin, PDW Platelet Distribution Width, AUC Area Under the Curve, PPV Positive Predictive Value, NPV Negative Predictive Value

Construction and validation of logistic regression diagnostic model

Considering the analysis results, a logistic regression model was developed using dataset 1. Figure 2 A illustrates the ROC curve for dataset 1, where the logistic regression model employing the three variables “Hb + Alb + PDW” achieved an AUC value of 0.960 (95% CI 0.942–0.978). This model exhibited strong sensitivity (0.860), specificity (0.957), PPV of 0.911, and NPV of 0.931. The formula for the model is as follows: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\text{l}\text{n}\left(\frac{\text{p}}{1-\text{p}}\right)\:=\:20.117\times\:\text{I}\text{n}\text{t}\:-\:2.27\times\:\text{x}1\:-\:0.089\times\:\text{x}2\:-\:0.039\times\:\text{x}3$$\end{document} ln p 1 - p = 20.117 × Int - 2.27 × x 1 - 0.089 × x 2 - 0.039 × x 3 x1 = Albumin, x2 = Hemoglobin, x3 = Platelet Distribution Width.

The calibration of the model was assessed using the Hosmer-Lemeshow (HL) test, a well-established method for evaluating the goodness of fit of a model and assessing the agreement between fitted values and observed value [ 19 ]. The result of the HL test indicated a p-value of 0.412, demonstrating that the model exhibited a good fit to the data. The model fitting results highlighted that Alb, Hb, and PDW were significant factors influencing the occurrence of multiple myeloma. Notably, lower levels of Hb, Alb, and PDW were associated with a higher likelihood of multiple myeloma. In addition to the HL test, a visual calibration curve was generated as part of the analysis. This calibration curve was constructed using the Bootstrap method. Its interpretation is as follows: If the calibration curve closely aligns with the 45°reference line, it indicates that the predicted values are in excellent agreement with the actual values. If the calibration curve lies above the reference line, it suggests that the predicted values tend to be greater than the actual values, indicating an overestimation of risk. Conversely, if the calibration curve is below the reference line, it implies an underestimation of risk [ 20 ]. For dataset 1, the calibration curve of the logistic regression model is visually represented in Fig. 2 B. This graphical representation provides further insights into the model’s performance and alignment with the observed data. Fig. 2 The ROC curve and calibration curve for development cohort. A The ROC curve for development cohort. B The calibration curve for development cohort

To further evaluate the model’s diagnostic efficacy, external validation was performed using Dataset 2. Figure 3 A illustrates the ROC curve for the logistic regression model applied to Validation Cohort, achieving an AUC of 0.979. The calibration curve for the logistic regression model applied to Validation Cohort is presented in Fig. 3 B. This finding demonstrates that the model constructed in this study exhibited high discriminative and generalization capabilities. Fig. 3 ROC curve and calibration curve for validation cohort. A The ROC curvefor validation cohort. B The calibration curve for validation cohort

Discussion

In clinical practice, the majority of MM cases originate from a precancerous condition known as monoclonal gammopathy of unknown significance (MGUS) [ 21 ]. MGUS can progress to a more advanced premalignant stage called smoldering multiple myeloma (SMM) [ 22 ]. In certain high-risk patients, MGUS and SMM may rapidly advance to symptomatic newly diagnosed multiple myeloma (NDMM) [ 23 ]. Timely identification and intervention in MGUS, SMM, and symptomatic MM patients can potentially alter their outcomes, preventing progression to relapsed and/or refractory multiple myeloma (RRMM) and its associated complications, ultimately extending overall survival [ 24 ]. The time to diagnosis thus represents a crucial factor in patient outcomes.

However, diagnosing NDMM remains a challenge, even for experienced clinicians, often necessitating collaboration among various medical disciplines (orthopedics, radiology, nuclear medicine, radiotherapy, hematology, oncology and etc.) before reaching a final diagnosis [ 25 ]. In this study, we identified a significant correlation of Alb, Hb, and PDW with the diagnosis of NDMM. Although these indicators are not mentioned in MM diagnostic guidelines, refining the optimal cutoff values for the above indicators revealed a substantial increase in the probability of diagnosing NDMM when Alb is below 39.3 g/L (3.93 g/dl), Hb is below 11.6 g/dL, and PDW is below 14.1fL. To further enhance the efficiency of early NDMM diagnosis based on these indicators, we constructed a logistic regression preliminary screening model using traditional machine learning algorithms. The results demonstrated good accuracy and generalization, effectively assisting in the initial diagnosis of NDMM.

Therefore, when MM patients present with symptoms such as back pain, pathological fractures, infections, or anemia syndromes in the outpatient setting, clinicians are recommended to perform preliminary selection of high-risk patients and utilize diagnostic models for further screening, which allows for timely intervention and specialist referrals. This approach will enhance the clinical diagnostic efficiency of MM, improve the long-term prognosis and quality of life for MM patients, and is of significant importance for the clinical diagnosis and early treatment of MM.

Anemia constitutes the initial presentation in 73% of NDMM patients, and a decrease in Hb levels exhibits notable specificity for NDMM diagnosis [ 4 ]. The underlying mechanisms contributing to low Hb levels in NDMM patients involve aberrant iron utilization, abnormally reduced serum erythropoietin levels, diminished bone marrow responsiveness to erythropoietin, hemolysis, and bone marrow infiltration [ 26 ]. In the current study, our findings suggest a substantial increase in the likelihood of MM when Hb levels fall below 11.6 g/dL, aligning with the clinical observation of anemia as an early symptom in MM patients. Research has demonstrated that reduced Hb levels serve as the sole independent cofactor associated with heightened fatigue symptoms and a decline in health-related quality of life among NDMM patients [ 27 ]. Furthermore, beyond its utility in MM diagnosis, Hb levels also emerge as an independent prognostic factor for survival in MM patients [ 28 – 30 ]. Additional studies have indicated that low Hb levels can independently elevate the risk of renal impairment in NDMM patients [ 31 , 32 ]. Consequently, clinical management often incorporates erythropoietin administration to achieve and maintain elevated Hb levels (between 11 and 12 g/dl) in a bid to enhance the prognosis and survival of NDMM patients [ 33 ]. In conclusion, the clinical screening for NDMM should vigilantly assess a patient’s Hb level. Building upon the results of this experimental study, clinicians should be attentive to individuals presenting with Hb levels below 11.6 g/dL and exhibiting symptoms of anemia.

Albumin is an important protein synthesized by human liver, which is abundant in human body, accounting for about 60% of total protein, and has special functions such as detoxification and maintaining osmotic pressure [ 34 ]. In clinical practice, Alb is often concerned by clinicians as an index reflecting the nutritional status of patients. Alb has outstanding prognostic value for MM. In some studies, Alb can be used as an independent predictor of Overall Survival (OS) and Progression-Free Survival (PFS) in MM patients [ 35 ]. Alb has been shown to be a major factor in enhancing the prognostic power of β 2 MG and is an absolutely necessary observation to identify truly low-risk MM patients [ 36 ]. Thus, β 2 -microglobulin (β 2 MG) levels and Alb levels in the International Staging System (ISS) are used to stratify and evaluate the prognosis of MM patients: Stage I, β 2 MG ≤ 3.5 mg/L, Alb ≥ 3.5 g/dL (median survival: 62 months); stage II, between stages I and III (median survival: 44 months); Stage III, β 2 MG ≥ 5.5 mg/L (median survival: 29 months) [ 13 ]. Inflammatory cytokines such as interleukin-6 (IL-6) are often secreted in the myeloma microenvironment. Studies have shown that Alb level is negatively correlated with IL-6 (the main growth and survival factor of myeloma cells) level, so MM patients often develop hypoalbuminemia [ 37 – 40 ]. In addition, Alb also has diagnostic and prognostic significance in other disciplines. For example, Alb level is closely related to renal function. It has been shown that Alb lower than 4.1 g/dL increases the risk of poor renal prognosis and decreased renal function [ 41 ]. Studies have also shown that the critical value of Alb for predicting adverse outcomes in patients with status epilepticus is 2.67 g/dL, and the concentration of Alb at admission is an independent predictor of short-term survival time in patients with status epilepticus [ 42 ]. The optimal cut-off value of Alb for the diagnosis of MM has not been indicated in related studies and guidelines. The results of our study showed that the newly diagnosed Alb level could be used as an independent diagnostic factor to predict the occurrence and development of MM, and when Alb was lower than 3.93 g/dL, the risk of MM diagnosis was greatly increased. Similar to our results, studies have also shown that Alb levels can be used to diagnose MM and fully demonstrate the clinical diagnostic value of Alb for NDMM [ 43 , 44 ].

Platelet distribution width(PDW) is an indicator reflecting changes in platelet size and is considered a marker of platelet activation, demonstrating prognostic value in coronary heart disease and certain solid organ tumors [ 45 – 47 ]. In our study, PDW emerged as an independent predictor of MM occurrence, with a significant increase in MM risk observed when PDW was below 14.1 fL. In our findings, although the PDW parameter exhibited low specificity and PPV for the diagnosis of MM patients, it demonstrated a high level of sensitivity and a relatively high NPV. In addition to MM, PDW has also been investigated in other hematological malignancies. A study from North India that reported significantly higher PDW levels in patients with chronic myeloproliferative neoplasms compared to controls [ 48 ]. In another study, PDW demonstrated excellent performance in differentiating chronic myeloproliferative neoplasms. The AUC for PDW was 0.68, indicating its diagnostic value for myeloproliferative neoplasms [ 49 ]. PDW’s diagnostic potential for myeloproliferative tumors has also been suggested in previous research. Our results, along with these findings, underscore PDW’s value as a diagnostic marker in NDMM.

Currently, apart from the three observational indicators, Alb, Hb, and PDW, identified in this study, other observational parameters and their optimal threshold values have also been proven to assist in the diagnosis and prognosis of MM. For instance, Red Cell Distribution Width (RDW) has demonstrated special diagnostic and prognostic value for MM, as elucidated by Xue et al. [ 50 ]. In their research, ROC curve analysis revealed that RDW > 13.5 exhibited a sensitivity of 94.5% and 63.7% for initial diagnosis of MM and a sensitivity of 92.3% and 63.7% for recurrent MM [ 50 ]. Additionally, studies have shown that RDW levels serve as an independent prognostic factor for MM patient survival and can predict treatment responses [ 51 , 52 ]. In a retrospective study involving 146 newly diagnosed MM patients with a median follow-up of 47 months, the normal RDW group displayed a better PFS compared to the high RDW group (24.2 months vs. 17.0 months, p = 0.029). Furthermore, multivariate analysis revealed that elevated RDW at diagnosis was an adverse prognostic factor for PFS in MM patients (HR 3.21, 95% CI 1.24–8.32) [ 53 ]. Similarly, other studies have demonstrated that pretreatment RDW levels are prognostic for MM, reflecting systemic inflammation and disease prognosis, thus making it a vital parameter for assessing treatment efficacy [ 54 ].

In addition to RDW, another diagnostic and prognostic indicator for MM is the Albumin to Globulin Ratio(AGR). A retrospective study from South Africa found that AGR can serve as a more cost-effective and readily accessible disease marker for assessing survival in MM patients [ 55 ]. In a study by Cai et al. involving MM patients with a median follow-up of 36 months, the high AGR group exhibited a significant association with improved OS and PFS. ROC curve analysis determined the optimal cutoff value for AGR to be 1.16, con-firming AGR as a potential prognostic biomarker for MM [ 56 ]. In this experiment, although we demonstrated the statistical significance of AGR for newly diagnosed MM patients through univariate analysis ( p < 0.05), the final results did not yield a conclusion that AGR can serve as a diagnostic marker for MM. The diagnostic value of AGR in MM requires further research and exploration.

Despite our study’s contributions, several limitations should be noted. Potential errors in data collection and handling of missing values may have influenced our regression analysis. However, secondary data correction and cleaning procedures were performed to minimize inaccuracies in model results. Moreover, external validation using data from other hospitals or regions is essential to confirm the model’s generalization beyond the single-center dataset employed in this study. The relatively small cohort size may have introduced bias in data collection and analysis, emphasizing the need for expanded cohorts to validate our findings.

In conclusion, our logistic regression model offers an effective tool for diagnosing suspected NDMM patients in outpatient settings, while also highlighting the diagnostic significance of Hb, Alb, and PDW. Further research is warranted to elucidate the underlying mechanisms of these markers in NDMM diagnosis.

Conclusions

This investigation substantiates that Hemoglobin (Hb), Albumin (Alb), and Platelet Distribution Width (PDW) stand as independent determinants in the diagnosis of Multiple Myeloma (MM). By ascertaining the optimal cutoff values for each diagnostic biomarker and refining the threshold ranges, this study furnishes a viable approach for non-specialist practitioners to conduct preliminary screenings of patients at their initial clinical encounters. Such an approach mitigates the risk of diagnostic oversights and inaccuracies among frontline healthcare providers. Furthermore, leveraging routine hematological and biochemical parameters, we constructed a diagnostic model for MM. This model demonstrates commendable sensitivity, accuracy, and generalization, thereby amplifying the efficacy of early MM detection in the context of initial screenings. Consequently, this research imparts valuable insights for the enhancement of MM diagnostic practices.

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Author contributions

Conceptualization, X.M.-T. and Y.R.-C.; methodology, Y.W.; software, M.N.-L.; validation, Y.P., Y.R.-C. and X.L.-S.; formal analysis, W.H. and C.Y.-S.; investigation, W.H. and C.Y.-S.; resources, Y.R.-C.; data curation, M.N.-L.; writing—original draft preparation, M.N.-L.; writing—review and editing, X.M.-T., Y.R.-C. and Y.W.; visualization, Y.W., Y.P. and X.L.-S.; supervision, X.M.-T.; project administration, Y.R.-C.; funding acquisition, X.M.-T., Y.R.-C. and X.L.-S. All authors have read and agreed to the published version of the manuscript.

Funding

This research was supported by Natural Science Foundation of Zhejiang Province (LY24H080004), Health Bureau Foundation of Zhejiang Province (2023KY479), and Chinese Traditional Medicine Administration Bureau Foundation of Zhejiang Province (2022ZQ012).

Data availability

The relevant supporting data are available from the corresponding author upon request.

Declarations

Ethics approval and consent to participate

This study received approval from the Ethics Committee of Zhejiang Province People’s Hospital, and the research was conducted in accordance with the principles outlined in the Declaration of Helsinki (Ethical approval code: QT2023281). A retrospective study was conducted on electronic medical records in this experiment, and all individuals had provided informed consent before being included in the study.

Competing interests

The authors declare no competing interests.

References

1. Monteith BE Sandhu I Lee AS Management of multiple myeloma: a review for general practitioners in oncology Curr Oncol 2023 30 4382 401 10.3390/curroncol30050334 37232792 PMC10217377 Monteith BE, Sandhu I, Lee AS. Management of multiple myeloma: a review for general practitioners in oncology. Curr Oncol. 2023;30:4382–401. 37232792 10.3390/curroncol30050334 PMC10217377

2. Siegel RL Miller KD Fuchs HE Jemal A Cancer statistics, 2021 CA Cancer J Clin 2021 71 7 33 10.3322/caac.21654 33433946 Siegel RL, Miller KD, Fuchs HE, Jemal A. Cancer statistics, 2021. CA Cancer J Clin. 2021;71:7–33. 33433946 10.3322/caac.21654

3. Siegel RL Miller KD Fuchs HE Jemal A Cancer statistics, 2022 CA Cancer J Clin 2022 72 7 33 10.3322/caac.21708 35020204 Siegel RL, Miller KD, Fuchs HE, Jemal A. Cancer statistics, 2022. CA Cancer J Clin. 2022;72:7–33. 35020204 10.3322/caac.21708

4. Kyle RA Gertz MA Witzig TE Lust JA Lacy MQ Dispenzieri A Review of 1027 patients with newly diagnosed multiple myeloma Mayo Clin Proc 2003 78 21 33 10.4065/78.1.21 12528874 Kyle RA, Gertz MA, Witzig TE, Lust JA, Lacy MQ, Dispenzieri A, et al. Review of 1027 patients with newly diagnosed multiple myeloma. Mayo Clin Proc. 2003;78:21–33. 12528874 10.4065/78.1.21

5. Smith A Howell D Patmore R Jack A Roman E Incidence of haematological malignancy by sub-type: a report from the haematological malignancy research network Br J Cancer 2011 105 1684 92 10.1038/bjc.2011.450 22045184 PMC3242607 Smith A, Howell D, Patmore R, Jack A, Roman E. Incidence of haematological malignancy by sub-type: a report from the haematological malignancy research network. Br J Cancer. 2011;105:1684–92. 22045184 10.1038/bjc.2011.450 PMC3242607

6. Howell DA Hart RI Smith AG Macleod U Patmore R Cook G Myeloma: patient accounts of their pathways to diagnosis PLoS ONE 2018 13 e0194788 10.1371/journal.pone.0194788 29617390 PMC5884527 Howell DA, Hart RI, Smith AG, Macleod U, Patmore R, Cook G, et al. Myeloma: patient accounts of their pathways to diagnosis. PLoS ONE. 2018;13:e0194788. 29617390 10.1371/journal.pone.0194788 PMC5884527

7. Dvorak C Common complaints, difficult diagnosis: multiple myeloma J Am Acad Nurse Pract 2006 18 190 4 10.1111/j.1745-7599.2006.00122.x 16681705 Dvorak C. Common complaints, difficult diagnosis: multiple myeloma. J Am Acad Nurse Pract. 2006;18:190–4. 16681705 10.1111/j.1745-7599.2006.00122.x

8. Howell D Smith A Appleton S Bagguley T Macleod U Cook G Multiple myeloma: routes to diagnosis, clinical characteristics and survival—findings from a UK population-based study Br J Haematol 2017 177 67 71 10.1111/bjh.14513 28146275 PMC5396308 Howell D, Smith A, Appleton S, Bagguley T, Macleod U, Cook G, et al. Multiple myeloma: routes to diagnosis, clinical characteristics and survival—findings from a UK population-based study. Br J Haematol. 2017;177:67–71. 28146275 10.1111/bjh.14513 PMC5396308

9. Neal RD Tharmanathan P France B Din NU Cotton S Fallon-Ferguson J Is increased time to diagnosis and treatment in symptomatic cancer associated with poorer outcomes? Systematic review Br J Cancer 2015 112 Suppl 1 S92 107 10.1038/bjc.2015.48 25734382 PMC4385982 Neal RD, Tharmanathan P, France B, Din NU, Cotton S, Fallon-Ferguson J, et al. Is increased time to diagnosis and treatment in symptomatic cancer associated with poorer outcomes? Systematic review. Br J Cancer. 2015;112(Suppl 1):S92–107. 25734382 10.1038/bjc.2015.48 PMC4385982

10. Kariyawasan CC Hughes DA Jayatillake MM Mehta AB Multiple myeloma: causes and consequences of delay in diagnosis QJM Mon J Assoc Phys 2007 100 635 40 10.1093/qjmed/hcm077 17846059 Kariyawasan CC, Hughes DA, Jayatillake MM, Mehta AB. Multiple myeloma: causes and consequences of delay in diagnosis. QJM Mon J Assoc Phys. 2007;100:635–40. 10.1093/qjmed/hcm077 17846059

11. Callander NS Baljevic M Adekola K Anderson LD Campagnaro E Castillo JJ NCCN Guidelines ® insights: multiple myeloma, Version 3.2022 J Natl Compr Canc Netw 2022 20 8 19 10.6004/jnccn.2022.0002 34991075 Callander NS, Baljevic M, Adekola K, Anderson LD, Campagnaro E, Castillo JJ, et al. NCCN Guidelines ® insights: multiple myeloma, Version 3.2022. J Natl Compr Canc Netw. 2022;20:8–19. 34991075 10.6004/jnccn.2022.0002

12. Chng WJ Dispenzieri A Chim CS Fonseca R Goldschmidt H Lentzsch S IMWG consensus on risk stratification in multiple myeloma Leukemia 2014 28 269 77 10.1038/leu.2013.247 23974982 Chng WJ, Dispenzieri A, Chim CS, Fonseca R, Goldschmidt H, Lentzsch S, et al. IMWG consensus on risk stratification in multiple myeloma. Leukemia. 2014;28:269–77. 23974982 10.1038/leu.2013.247

13. D’Agostino M Cairns DA Lahuerta JJ Wester R Bertsch U Waage A Second revision of the International Staging System (R2-ISS) for overall survival in multiple myeloma: a European Myeloma Network (EMN) Report within the HARMONY Project J Clin Oncol 2022 40 3406 18 10.1200/JCO.21.02614 35605179 D’Agostino M, Cairns DA, Lahuerta JJ, Wester R, Bertsch U, Waage A, et al. Second revision of the International Staging System (R2-ISS) for overall survival in multiple myeloma: a European Myeloma Network (EMN) Report within the HARMONY Project. J Clin Oncol. 2022;40:3406–18. 35605179 10.1200/JCO.21.02614

14. Rajkumar SV Multiple myeloma: 2020 update on diagnosis, risk-stratification and management Am J Hematol 2020 95 548 67 10.1002/ajh.25791 32212178 Rajkumar SV. Multiple myeloma: 2020 update on diagnosis, risk-stratification and management. Am J Hematol. 2020;95:548–67. 32212178 10.1002/ajh.25791

15. Cowan AJ Green DJ Kwok M Lee S Coffey DG Holmberg LA Diagnosis and management of multiple myeloma: a review JAMA 2022 327 464 77 10.1001/jama.2022.0003 35103762 Cowan AJ, Green DJ, Kwok M, Lee S, Coffey DG, Holmberg LA, et al. Diagnosis and management of multiple myeloma: a review. JAMA. 2022;327:464–77. 35103762 10.1001/jama.2022.0003

16. Lyratzopoulos G Neal RD Barbiere JM Rubin GP Abel GA Variation in number of general practitioner consultations before hospital referral for cancer: findings from the 2010 National Cancer Patient Experience Survey in England Lancet Oncol 2012 13 353 65 10.1016/S1470-2045(12)70041-4 22365494 Lyratzopoulos G, Neal RD, Barbiere JM, Rubin GP, Abel GA. Variation in number of general practitioner consultations before hospital referral for cancer: findings from the 2010 National Cancer Patient Experience Survey in England. Lancet Oncol. 2012;13:353–65. 22365494 10.1016/S1470-2045(12)70041-4

17. Tolles J Meurer WJ Logistic regression: relating patient characteristics to outcomes JAMA 2016 316 533 4 10.1001/jama.2016.7653 27483067 Tolles J, Meurer WJ. Logistic regression: relating patient characteristics to outcomes. JAMA. 2016;316:533–4. 27483067 10.1001/jama.2016.7653

18. Meurer WJ Tolles J Logistic regression diagnostics: understanding how well a model predicts outcomes JAMA 2017 317 1068 9 10.1001/jama.2016.20441 28291878 Meurer WJ, Tolles J. Logistic regression diagnostics: understanding how well a model predicts outcomes. JAMA. 2017;317:1068–9. 28291878 10.1001/jama.2016.20441

19. Paul P Pennell ML Lemeshow S Standardizing the power of the Hosmer-Lemeshow goodness of fit test in large data sets Stat Med 2013 32 67 80 10.1002/sim.5525 22833304 Paul P, Pennell ML, Lemeshow S. Standardizing the power of the Hosmer-Lemeshow goodness of fit test in large data sets. Stat Med. 2013;32:67–80. 22833304 10.1002/sim.5525

20. Van Calster B Vickers AJ Calibration of risk prediction models: impact on decision-analytic performance Med Decis Mak 2015 35 162 9 10.1177/0272989X14547233 25155798 Van Calster B, Vickers AJ. Calibration of risk prediction models: impact on decision-analytic performance. Med Decis Mak. 2015;35:162–9. 10.1177/0272989X14547233 25155798

21. Dhodapkar MV MGUS to myeloma: a mysterious gammopathy of underexplored significance Blood 2016 128 2599 606 10.1182/blood-2016-09-692954 27737890 PMC5146746 Dhodapkar MV. MGUS to myeloma: a mysterious gammopathy of underexplored significance. Blood. 2016;128:2599–606. 27737890 10.1182/blood-2016-09-692954 PMC5146746

22. Ho M Patel A Goh CY Moscvin M Zhang L Bianchi G Changing paradigms in diagnosis and treatment of monoclonal gammopathy of undetermined significance (MGUS) and smoldering multiple myeloma (SMM) Leukemia 2020 34 3111 25 10.1038/s41375-020-01051-x 33046818 Ho M, Patel A, Goh CY, Moscvin M, Zhang L, Bianchi G. Changing paradigms in diagnosis and treatment of monoclonal gammopathy of undetermined significance (MGUS) and smoldering multiple myeloma (SMM). Leukemia. 2020;34:3111–25. 33046818 10.1038/s41375-020-01051-x

23. Kyle RA Remstein ED Therneau TM Dispenzieri A Kurtin PJ Hodnefield JM Clinical course and prognosis of smoldering (asymptomatic) multiple myeloma N Engl J Med 2007 356 2582 90 10.1056/NEJMoa070389 17582068 Kyle RA, Remstein ED, Therneau TM, Dispenzieri A, Kurtin PJ, Hodnefield JM, et al. Clinical course and prognosis of smoldering (asymptomatic) multiple myeloma. N Engl J Med. 2007;356:2582–90. 17582068 10.1056/NEJMoa070389

24. Goldman-Mazur S Kumar SK Current approaches to management of high-risk multiple myeloma Am J Hematol 2021 96 854 71 10.1002/ajh.26161 33725367 Goldman-Mazur S, Kumar SK. Current approaches to management of high-risk multiple myeloma. Am J Hematol. 2021;96:854–71. 33725367 10.1002/ajh.26161

25. Caers J Garderet L Kortüm KM O’Dwyer ME van de Donk N Binder M European myeloma network recommendations on tools for the diagnosis and monitoring of multiple myeloma: what to use and when Haematologica 2018 103 1772 84 10.3324/haematol.2018.189159 30171031 PMC6278986 Caers J, Garderet L, Kortüm KM, O’Dwyer ME, van de Donk N, Binder M, et al. European myeloma network recommendations on tools for the diagnosis and monitoring of multiple myeloma: what to use and when. Haematologica. 2018;103:1772–84. 30171031 10.3324/haematol.2018.189159 PMC6278986

26. San Miguel JF García-Sanz R Recombinant human erythropoietin in the anaemia of multiple myeloma and Non-hodgkin’s lymphoma Med Oncol 1998 15 Suppl 1 S29 34 9785334 San Miguel JF, García-Sanz R. Recombinant human erythropoietin in the anaemia of multiple myeloma and Non-hodgkin’s lymphoma. Med Oncol. 1998;15(Suppl 1):S29-34. 9785334

27. Gascón P Arranz R Bargay J Ramos F Fatigue- and health-related quality-of-life in anemic patients with lymphoma or multiple myeloma Support Care Cancer 2018 26 1253 64 10.1007/s00520-017-3948-5 29116407 Gascón P, Arranz R, Bargay J, Ramos F. Fatigue- and health-related quality-of-life in anemic patients with lymphoma or multiple myeloma. Support Care Cancer. 2018;26:1253–64. 29116407 10.1007/s00520-017-3948-5

28. Caro JJ Salas M Ward A Goss G Anemia as an independent prognostic factor for survival in patients with cancer: a systemic, quantitative review Cancer 2001 91 2214 21 10.1002/1097-0142(20010615)91:12<2214::AID-CNCR1251>3.0.CO;2-P 11413508 Caro JJ, Salas M, Ward A, Goss G. Anemia as an independent prognostic factor for survival in patients with cancer: a systemic, quantitative review. Cancer. 2001;91:2214–21. 11413508

29. Tao ZF Fu WJ Yuan ZG Wang DX Chen YB Hou J Prognostic factors and staging systems of multiple myeloma Chin Med J 2007 120 1655 8 10.1097/00029330-200710010-00002 17935664 Tao ZF, Fu WJ, Yuan ZG, Wang DX, Chen YB, Hou J. Prognostic factors and staging systems of multiple myeloma. Chin Med J. 2007;120:1655–8. 17935664

30. Lin C Shen H Zhou S Liu M Xu A Huang S Assessment of infection in newly diagnosed multiple myeloma patients: risk factors and main characteristics BMC Infect Dis 2020 20 699 10.1186/s12879-020-05412-w 32972385 PMC7517606 Lin C, Shen H, Zhou S, Liu M, Xu A, Huang S, et al. Assessment of infection in newly diagnosed multiple myeloma patients: risk factors and main characteristics. BMC Infect Dis. 2020;20:699. 32972385 10.1186/s12879-020-05412-w PMC7517606

31. Ling W Yu W Sun H Lv M Analysis of related risk factors and reversal value of renal injury in elderly patients with multiple myeloma J BUON Off J Balkan Union Oncol 2017 22 735 40 28730783 Ling W, Yu W, Sun H, Lv M. Analysis of related risk factors and reversal value of renal injury in elderly patients with multiple myeloma. J BUON Off J Balkan Union Oncol. 2017;22:735–40. 28730783

32. Chen J Liu H Li L Liu Z Song J Wang G Clinical features and treatment outcome of elderly multiple myeloma patients with impaired renal function J Clin Lab Anal 2019 33 e22888 10.1002/jcla.22888 31004381 PMC6595344 Chen J, Liu H, Li L, Liu Z, Song J, Wang G, et al. Clinical features and treatment outcome of elderly multiple myeloma patients with impaired renal function. J Clin Lab Anal. 2019;33:e22888. 31004381 10.1002/jcla.22888 PMC6595344

33. George LL Malik MN Miller EJ Hicks K Khanam R Saterehaseman A Special Considerations for supportive care and management of complications in elderly patients with multiple myeloma Clin Lymphoma Myeloma Leukemia 2021 21 812 22 10.1016/j.clml.2021.07.013 34384735 George LL, Malik MN, Miller EJ, Hicks K, Khanam R, Saterehaseman A, et al. Special Considerations for supportive care and management of complications in elderly patients with multiple myeloma. Clin Lymphoma Myeloma Leukemia. 2021;21:812–22. 10.1016/j.clml.2021.07.013 34384735

34. Mazzaferro EM Edwards T Update on albumin therapy in critical illness Vet Clin N Am Small Anim Pract 2020 50 1289 305 10.1016/j.cvsm.2020.07.005 32839002 Mazzaferro EM, Edwards T. Update on albumin therapy in critical illness. Vet Clin N Am Small Anim Pract. 2020;50:1289–305. 10.1016/j.cvsm.2020.07.005 32839002

35. Solmaz S Uzun O Sevindik OG Demirkan F Ozcan MA Ozsan GH The effect of haemoglobin, albumin, lymphocyte and platelet score on the prognosis in patients with multiple myeloma Int J Lab Hematol 2023 45 13 9 10.1111/ijlh.13958 36053932 Solmaz S, Uzun O, Sevindik OG, Demirkan F, Ozcan MA, Ozsan GH, et al. The effect of haemoglobin, albumin, lymphocyte and platelet score on the prognosis in patients with multiple myeloma. Int J Lab Hematol. 2023;45:13–9. 36053932 10.1111/ijlh.13958

36. Cheng T Wang X Han Y Hao J Hu H Hao L The level of serum albumin is associated with renal prognosis and renal function decline in patients with chronic kidney disease BMC Nephrol 2023 24 57 10.1186/s12882-023-03110-8 36922779 PMC10018824 Cheng T, Wang X, Han Y, Hao J, Hu H, Hao L. The level of serum albumin is associated with renal prognosis and renal function decline in patients with chronic kidney disease. BMC Nephrol. 2023;24:57. 36922779 10.1186/s12882-023-03110-8 PMC10018824

37. Palumbo A Avet-Loiseau H Oliva S Lokhorst HM Goldschmidt H Rosinol L Revised International staging system for multiple myeloma: a report from International Myeloma Working Group J Clin Oncol 2015 33 2863 9 10.1200/JCO.2015.61.2267 26240224 PMC4846284 Palumbo A, Avet-Loiseau H, Oliva S, Lokhorst HM, Goldschmidt H, Rosinol L, et al. Revised International staging system for multiple myeloma: a report from International Myeloma Working Group. J Clin Oncol. 2015;33:2863–9. 26240224 10.1200/JCO.2015.61.2267 PMC4846284

38. Schwarze MM Hawley RG Prevention of myeloma cell apoptosis by ectopic bcl-2 expression or interleukin 6-mediated up-regulation of bcl-xL Cancer Res 1995 55 2262 5 7757973 Schwarze MM, Hawley RG. Prevention of myeloma cell apoptosis by ectopic bcl-2 expression or interleukin 6-mediated up-regulation of bcl-xL. Cancer Res. 1995;55:2262–5. 7757973

39. Bataille R Jourdan M Zhang XG Klein B Serum levels of interleukin 6, a potent myeloma cell growth factor, as a reflect of disease severity in plasma cell dyscrasias J Clin Investig 1989 84 2008 11 10.1172/JCI114392 2592570 PMC304085 Bataille R, Jourdan M, Zhang XG, Klein B. Serum levels of interleukin 6, a potent myeloma cell growth factor, as a reflect of disease severity in plasma cell dyscrasias. J Clin Investig. 1989;84:2008–11. 2592570 10.1172/JCI114392 PMC304085

40. Lichtenstein A Tu Y Fady C Vescio R Berenson J Interleukin-6 inhibits apoptosis of malignant plasma cells Cell Immunol 1995 162 248 55 10.1006/cimm.1995.1076 7743552 Lichtenstein A, Tu Y, Fady C, Vescio R, Berenson J. Interleukin-6 inhibits apoptosis of malignant plasma cells. Cell Immunol. 1995;162:248–55. 7743552 10.1006/cimm.1995.1076

41. Mihou D Katodritou E Zervas K Multiple myeloma staging based on the combination of beta-2-microglobulin and albumin: the role of albumin in the model Hematology 2007 12 527 31 10.1080/10245330701384161 17852450 Mihou D, Katodritou E, Zervas K. Multiple myeloma staging based on the combination of beta-2-microglobulin and albumin: the role of albumin in the model. Hematology. 2007;12:527–31. 17852450 10.1080/10245330701384161

42. Liao Q Li SZ Zeng QQ Zhou JX Huang K Bi FF The value of serum albumin concentration in predicting functional outcome of status epilepticus: an observational study Epileptic Disord 2023 25 150 9 10.1002/epd2.20001 37358922 Liao Q, Li SZ, Zeng QQ, Zhou JX, Huang K, Bi FF. The value of serum albumin concentration in predicting functional outcome of status epilepticus: an observational study. Epileptic Disord. 2023;25:150–9. 37358922 10.1002/epd2.20001

43. Cui R Zhang S Wang M Jia T Zhai Y Yue Y Establishment of multiple Myeloma diagnostic model based on logistic regression in clinical laboratory Clin Lab 2020 10.7754/Clin.Lab.2019.190832 32776728 Cui R, Zhang S, Wang M, Jia T, Zhai Y, Yue Y, et al. Establishment of multiple Myeloma diagnostic model based on logistic regression in clinical laboratory. Clin Lab. 2020. 10.7754/Clin.Lab.2019.190832. 32776728 10.7754/Clin.Lab.2019.190832

44. Chen S Zhou M Yang J Cheng YR Li Y Lu K Significance of common blood test indexes in the diagnosis and prognosis of multiple myeloma Clin Lab 2022 10.7754/Clin.Lab.2021.210716 35443602 Chen S, Zhou M, Yang J, Cheng YR, Li Y, Lu K, et al. Significance of common blood test indexes in the diagnosis and prognosis of multiple myeloma. Clin Lab. 2022. 10.7754/Clin.Lab.2021.210716. 35443602 10.7754/Clin.Lab.2021.210716

45. Gunaldi M Erdem D Goksu S Gunduz S Okuturlar Y Tiken E Platelet distribution width as a predictor of metastasis in gastric cancer patients J Gastrointest cancer 2017 48 341 6 10.1007/s12029-016-9886-5 27826813 Gunaldi M, Erdem D, Goksu S, Gunduz S, Okuturlar Y, Tiken E, et al. Platelet distribution width as a predictor of metastasis in gastric cancer patients. J Gastrointest cancer. 2017;48:341–6. 27826813 10.1007/s12029-016-9886-5

46. Takeuchi H Noda D Abe M Anami K Miyawaki M Osoegawa A Evaluating the platelet distribution width-to-plateletcrit ratio as a prognostic marker for patients with breast cancer Anticancer Res 2020 40 3947 52 10.21873/anticanres.14386 32620636 Takeuchi H, Noda D, Abe M, Anami K, Miyawaki M, Osoegawa A, et al. Evaluating the platelet distribution width-to-plateletcrit ratio as a prognostic marker for patients with breast cancer. Anticancer Res. 2020;40:3947–52. 32620636 10.21873/anticanres.14386

47. Zheng YY Wang L Shi Q Mean platelet volume (MPV) and platelet distribution width (PDW) predict clinical outcome of acute ischemic stroke: a systematic review and meta-analysis J Clin Neurosci 2022 101 221 7 10.1016/j.jocn.2022.05.019 35636058 Zheng YY, Wang L, Shi Q. Mean platelet volume (MPV) and platelet distribution width (PDW) predict clinical outcome of acute ischemic stroke: a systematic review and meta-analysis. J Clin Neurosci. 2022;101:221–7. 35636058 10.1016/j.jocn.2022.05.019

48. Parashar Y Kushwaha R Kumar A Agarwal K Singh US Jain M Haemostatic profile in patients of myeloproliferative neoplasms—a tertiary care centre experience J Clin Diagn Res 2016 10 Ec01 4 28050371 10.7860/JCDR/2016/19985.8840 PMC5198324 Parashar Y, Kushwaha R, Kumar A, Agarwal K, Singh US, Jain M, et al. Haemostatic profile in patients of myeloproliferative neoplasms—a tertiary care centre experience. J Clin Diagn Res. 2016;10:Ec01-4. 28050371 10.7860/JCDR/2016/19985.8840 PMC5198324

49. Chandrashekar V Tarigopula A Can platelet distribution width be used to predict the possibility of chronic myeloproliferative neoplasms? Lab Med 2019 50 174 9 10.1093/labmed/lmy061 30329092 Chandrashekar V, Tarigopula A. Can platelet distribution width be used to predict the possibility of chronic myeloproliferative neoplasms? Lab Med. 2019;50:174–9. 30329092 10.1093/labmed/lmy061

50. Huang C Wei H Lan F Lu Y Li S Qin X Assessment of red blood cell distribution width and multiple myeloma in a Guangxi population: a retrospective study Clin Lab 2019 10.7754/Clin.Lab.2018.180738 30868861 Huang C, Wei H, Lan F, Lu Y, Li S, Qin X. Assessment of red blood cell distribution width and multiple myeloma in a Guangxi population: a retrospective study. Clin Lab. 2019. 10.7754/Clin.Lab.2018.180738. 30868861 10.7754/Clin.Lab.2018.180738

51. Wang J Xie X Cheng F Zhou X Xia J Qian X Evaluation of pretreatment red cell distribution width in patients with multiple myeloma Cancer Biomark A 2017 20 267 72 10.3233/CBM-170032 28826172 Wang J, Xie X, Cheng F, Zhou X, Xia J, Qian X, et al. Evaluation of pretreatment red cell distribution width in patients with multiple myeloma. Cancer Biomark A. 2017;20:267–72. 10.3233/CBM-170032 28826172

52. Chen X Liu J Duan J Xiong H Liu Y Zhang X Is RDW a clinically relevant prognostic factor for newly diagnosed multiple myeloma? A systematic review and meta-analysis BMC Cancer 2022 22 796 10.1186/s12885-022-09902-9 35854269 PMC9297629 Chen X, Liu J, Duan J, Xiong H, Liu Y, Zhang X, et al. Is RDW a clinically relevant prognostic factor for newly diagnosed multiple myeloma? A systematic review and meta-analysis. BMC Cancer. 2022;22:796. 35854269 10.1186/s12885-022-09902-9 PMC9297629

53. Lee H Kong SY Sohn JY Shim H Youn HS Lee S Elevated red blood cell distribution width as a simple prognostic factor in patients with symptomatic multiple myeloma BioMed Res Int 2014 2014 145619 24963470 10.1155/2014/145619 PMC4055253 Lee H, Kong SY, Sohn JY, Shim H, Youn HS, Lee S, et al. Elevated red blood cell distribution width as a simple prognostic factor in patients with symptomatic multiple myeloma. BioMed Res Int. 2014;2014:145619. 24963470 10.1155/2014/145619 PMC4055253

54. Zhou D Xu P Peng M Shao X Wang M Ouyang J Pre-treatment red blood cell distribution width provides prognostic information in multiple myeloma Clin Chim Acta 2018 481 34 41 10.1016/j.cca.2018.02.009 29452082 Zhou D, Xu P, Peng M, Shao X, Wang M, Ouyang J, et al. Pre-treatment red blood cell distribution width provides prognostic information in multiple myeloma. Clin Chim Acta. 2018;481:34–41. 29452082 10.1016/j.cca.2018.02.009

55. Laudin GE Levay PF Coetzer B Globulin fraction and albumin: globulin ratio as a predictor of mortality in a South African multiple myeloma cohort Int J Hematol Oncol 2020 9 Ijh27 10.2217/ijh-2020-0003 33014331 PMC7521188 Laudin GE, Levay PF, Coetzer B. Globulin fraction and albumin: globulin ratio as a predictor of mortality in a South African multiple myeloma cohort. Int J Hematol Oncol. 2020;9:Ijh27. 33014331 10.2217/ijh-2020-0003 PMC7521188

56. Cai Y Zhao Y Dai Q Xu M Xu X Xia W Prognostic value of the albumin-globulin ratio and albumin-globulin score in patients with multiple myeloma J Int Med Res 2021 49 300060521997736 10.1177/0300060521997736 33682516 PMC7944530 Cai Y, Zhao Y, Dai Q, Xu M, Xu X, Xia W. Prognostic value of the albumin-globulin ratio and albumin-globulin score in patients with multiple myeloma. J Int Med Res. 2021;49:300060521997736. 33682516 10.1177/0300060521997736 PMC7944530