A prognostic model for patients with primary extramedullary multiple myeloma

Abstract

Background: Extramedullary disease is a manifestation of multiple myeloma, the prognosis of which remains poor even in the era of novel drugs. Therefore, we aimed to develop a predictive model for patients with primary extramedullary multiple myeloma (EMM). Methods: Clinical and laboratory data of patients diagnosed with primary EMM between July 2007 and July 2021 were collected and analyzed. Univariate and least absolute shrinkage and selection operation Cox regression analyses (LASSO) were used to select prognostic factors for overall survival (OS) to establish a nomogram prognostic model. The performance of the model was evaluated using concordance index which was internally validated by bootstraps with 1,000 resample, area under the curve (AUCs), and calibration curves. Results: 217 patients were included in this retrospective study. Patients with EMM had a higher rate of belonging to the male sex, age >50 years, advanced Durie-Salmon stage III, hypercalcemia, and low hemoglobin level. Compared with patients with bone-related extramedullary disease, those with extraosseous-related extramedullary disease had a higher frequency of advanced Durie-Salmon stage III, lower rate of hypercalcemia, and elevated prothrombin time. The OS and progression-free survival (PFS) of patients with bone-related extramedullary disease were significantly higher than those of patients with extraosseous-related extramedullary disease. After the univariate and LASSO analyses, six prognostic factors, including performance status, number of extramedullary involved sites, β2-microglobulin, lactate dehydrogenase, monocyte-lymphocyte ratio, and prothrombin time, were integrated to establish a nomogram. The model showed robust discrimination with a concordance index (C-index) of 0.775 (95% confidence interval [CI], 0.713-0.836), internally validated with the corrected C-index of 0.756, and excellent performance in time-dependent AUCs compared with other staging systems. The AUCs for 1-, 3-, and 5-year OS were 0.814, 0.744, and 0.832, respectively. The calibration curves exhibited good consistency between the observed and nomogram-predicted OS. The 5-year OS of patients in the high-risk group (23.3%; 95% CI, 13.9%-39.3%) was much worse than that in the low-risk group (73.0%; 95% CI, 62.5%-85.4%; p < 0.001). Conclusion: The nomogram predictive model based on six clinical variables showed good prognostic performance and could better predict individual survival in patients with EMM.

Keywords: extramedullary multiple myeloma; overall survival; prognostic factors; prognostic nomogram; progression-free survival.

FIGURE 1

Flow chart of this study.

FIGURE 2

Overall survival (OS) and progression free disease (PFS) of 217 patients with primary EMM.

FIGURE 3

Development and evaluation of a predictive model (A) 1000 bootstrap replicates by Lasso Cox regression analysis for variable selection; (B) The nomogram based on data from 217 patients with primary EMM to predict individual prognosis; (C) Calibration curves for predicting OS at 1-, 3-, and 5-year; (D) Sensitivity and specificity at 1-, 3-, and 5-year of the predictive nomogram model were assessed in 217 patients by time-dependent ROC analysis; (E) Time-dependent receiver operating characteristic (ROC) curves (Nomogram [red], R-ISS [green], ISS [orange], DS [blue]).

FIGURE 4

OS of patients according to different stratification. (A) Survival curves of risk stratification based on our model (High risk vs Low risk). (B) Survival curves of different groups (EM-B vs EM-E). (C) Survival curves of different treatment (No radiotherapy vs Radiotherapy). (D) Survival curves of different treatment (No transplant vs Transplant).

FIGURE 5

PFS of patients according to different stratification. (A) Survival curves of risk stratification based on our model (High risk vs Low risk). (B) Survival curves of different groups (EM-B vs EM-E). (C) Survival curves of different treatment (No radiotherapy vs Radiotherapy). (D) Survival curves of different treatment (No transplant vs Transplant).