1 INTRODUCTION

Leiomyosarcoma (LMS) is one of the most common soft tissue sarcomas, accounting for 12% of all soft tissue sarcomas.1 It is reported that LMS mainly occurs in 50–60 years old patients and is often involved in the uterus, retroperitoneal space, and soft tissue.2 About 25% of LMS patients will occur DM even through radical resection.3, 4 Okamoto et al. have reported that the common metastatic sites of LMS include lung, liver, and bone.5 Further studies have shown that lung is the most common metastatic site.3, 5 Some studies have demonstrated that the survival rate of LMS patients is improving after receiving chemotherapy and surgery. However, the prognosis of LMS patients with DM is still poor, and the 5-year survival rate was less than 20%.6-8 Leiomyosarcoma accounts for about 0.12% of all tumors, and because of the low incidence rate of LMS, the risk factors, and prognostic factors of DM are not yet clear.1 Therefore, identifying high-risk LMS patients who are at risk of developing DM is meaningful. Many studies have shown that malignant tumor prognosis can be more accurate, more effective, and more beneficial by using nomograms.9, 10 Takehara et al. have analyzed the clinical status and prognosis of uterine LMS patients.6 Xue et al. have investigated the prognosis of extremities LMS patients and established a prognostic nomogram.11 However, as far as we know, there is no research on building a web-based nomogram to estimate the DM risk in LMS patients. Besides that, there are no studies to predict the prognosis of LMS with DM. Therefore, we intend to use the Surveillance, Epidemiology, and End Results (SEER) database to evaluate DM incidence and risk factors in LMS and establish a visualized web-based nomogram.12 Furthermore, we intend to predict the prognostic factors of LMS patients with DM.

2 MATERIALS AND METHODS 2.1 Patients

The data included in the present study were downloaded from the SEER*Stat software version 8.3.6.12 Using the International Classification of Diseases for Oncology, 3rd edition (ICD-O-3), we identified all LMS patients (ICD-O-3 histologic type: 8890, 8891, 8893, 8896). We collected case diagnosis time between 2010 and 2016. The inclusive criteria were as follows: (1) patients with pathological diagnosis of LMS, (2) patients from the time of 2010–2016, according to the term “year of diagnosis,” (3) complete follow-up information, no data loss. The exclusion criteria were as follows: (1) patients missing essential details, including grade, stage, tumor size, surgery, radiotherapy, chemotherapy, survival time, and marital status. (2) follow-up status is missing. According to the ethics guidelines, neither informed consent nor approval of the ethics committee is required because we use public and anonymous data.

2.2 Data element

The following demographic and clinical characteristics were included: age, sex (Female and Male), race [white, black, and others (American Indian/AK Native, Asian/Pacific Islander)], marital status (married and unmarried), grade (I–II or III–IV), T stage (I–II or III–IV), N stage (N0 or N1), surgical treatment (No or Yes), radiation treatment (No or Yes), chemotherapy (No or Yes), tumor size, and the histologic type (LMS NOS, Epithelioid LMS, Bizare LMS, and Myxoid LMS). The primary site was divided into uterus, soft tissue, retroperitoneum, others (eye and orbit, bones and joints, other digestive organs, trachea mediastinum, and other respiratory organs), distant metastasis (DM; No metastasis, oligo metastasis, and multiple metastases). The survival analysis' primary outcome was the overall survival (OS), defined as the time from diagnosis to death due to any cause.

2.3 Statistical analysis

The eligible patients were randomly divided into training set (n = 2184, 70%) and testing set (n = 652, 30%). In this study, patients in the training set were used to establish a nomogram, and patients in the test group were used to verify the nomogram. In this study, p-value <.05 (bilateral) was considered statistically significant. Univariate and multivariate logistic regression models were used to analyze the risk factors of DM in LMS patients. Based on these independent risk factors, a nomogram was established by R software. The nomogram was then evaluated by receiver operating characteristic curve (ROC), calibration curve analysis, and decision curve analysis (DCA). A web-based nomogram was further prepared based on the nomogram by the “Dynnom” package. The survival time was measured by the Kaplan–Meier analyzes, and the difference between DM and without DM was tested by log-rank test. Cox proportional hazard regression model was used for univariate and multivariate analysis, and significant variables were obtained. All statistical analyses were performed using R software (http://www.Rproject.org, version 4.0.3).

3 RESULTS 3.1 Demographic and clinical characteristics

A detailed workflow was shown in Figure 1. According to the predetermined criteria, a total of 2184 LMS patients were included. There were 699 males (32.01%) and 1485 females (67.99%). As for race, most of the patients were White (n = 1667 [76.33%]), followed by Black (n = 324 [14.84%]) and Others (n = 193 [8.84%]). The most common site of primary tumor is soft tissue (n = 1176 [53.85%]), followed by uterus (n = 668 [30.59%]), retroperitoneal (n = 300 [13.74%]), and others (n = 40 [1.83%]). The most common histological type was LMS NOS (n = 2103 [96.29%]), the others (Epithelioid LMS, Bizare LMS, and Myxoid LMS) were 81 cases (3.71%). Differentiation in grades III–IV (n = 1434, 65.66%) was the most common among tumor classifications. T1–2 (n = 2068, 94.69%) and N0 (n = 2098, 6.68%) phases were common. Of all the patients, 1185 (86.31%) had no metastasis, 222 (10.16%) had oligo metastasis and 77 (3.53%) had multiple metastases. Treatment methods selected by LMS patients included surgery (n = 1995 [91.35%]), chemotherapy (n = 1489 [68.18%]), and radiotherapy (n = 1467 [67.17%]). More details are shown in Table 1.

The flow chart of the study design and analysis

TABLE 1. Demographic and clinical characteristics of patients diagnosed with LMS in SEER database from 2010 to 2016 Subject characteristics Total cohort Training cohort Validation cohort n % n % n % Age Median (range, years) 61 (1–101) 61 (2–97) 60 (6–101) Sex Male 699 32.01 489 31.92 210 32.21 Female 1485 67.99 1043 68.08 442 67.79 Race White 1667 76.33 1174 76.63 493 75.61 Black 324 14.84 227 14.82 97 14.88 Others 193 8.84 131 8.55 62 9.51 Marital status Married 1746 79.95 1217 79.44 529 81.13 Unmarried 438 20.05 315 20.56 123 18.87 Primary site Soft tissue 1176 53.85 835 54.50 341 52.30 Retroperitoneum 300 13.74 204 13.32 96 14.72 Uterus 668 30.59 467 30.48 201 30.83 Others 40 1.83 26 1.70 14 2.15 Histologic type LMS NOS 2103 96.29 1483 96.80 620 95.09 Epithelioid 51 2.34 28 1.83 23 3.53 Bizare 2 0.09 1 0.07 1 0.15 Myxoid 28 1.28 20 1.31 8 1.23 Grade I–II 750 34.34 534 34.86 216 33.13 III–IV 1434 65.66 998 65.14 436 66.87 T stage T1–T2 2068 94.69 1455 94.97 613 94.02 T3–T4 116 5.31 77 5.03 39 5.98 N stage N0 2098 96.06 1469 95.89 629 96.47 N1 86 3.94 63 4.11 23 3.53 Number of metastasis 0 1885 86.31 1309 85.44 576 88.34 1 222 10.16 162 10.57 60 9.20 >1 77 3.53 61 3.98 16 2.45 Tumor size Median (range, mm) 85 (1–989) 85 (1–989) 85 (1–989) Surgery Yes 189 8.65 142 9.27 47 7.21 No 1995 91.35 1390 90.73 605 92.79 Radiotherapy Yes 717 32.83 500 32.64 217 33.28 No 1467 67.17 1032 67.36 435 66.72 Chemotherapy Yes 695 31.82 478 31.20 217 33.28 No 1489 68.18 1054 68.80 435 66.72 Vital status Alive 1258 57.60 877 57.25 381 58.44 Dead 926 42.40 655 42.75 271 41.56 3.2 Risk factors for DM development in LMS patients

An odds ratio (OR) greater than 1 indicates that the exposure is a risk factor, a OR less than 1 indicates a protective factor, and a value equal to 1 indicates an unrelated factor.13 Age, sex, race, grade, T stage, N stage, site, size, and histologic type were related to LMS developing DM in univariate logistics analysis. In multivariate logistics analysis, the Black (OR = 1.445, 95% CI = 1.039–2.008, p-value = .028), grade III–IV (OR = 2.873, 95% CI = 2.030–4.067, p-value < .001), N1 stage (OR = 3.428, 95% CI = 2.125–5.532, p-value < .001), primary site in uterus (OR = 1.754, 95% CI = 1.239–2.483, p-value = .002) and tumor size (OR = 1.002, 95% CI = 1.001–1.004, p-value < .001) were risk factors for DM in LMS patients. More details are shown in Table 2 and Figure 2.

TABLE 2. Logistic regression model for analyzing the risk factors for developing distant metastases in patients diagnosed with LMS Univariate Multivariate OR 95%CI P-value OR 95%CI P-value Age Range (years) 0.991 0.983–0.999 .034 0.996 0.986–1.005 .349 Sex Male Reference Reference Female 1.449 1.098–1.912 .009 0.859 0.609–1.211 .386 Race White Reference Black 1.627 1.188–2.227 .002 1.445 1.039–2.008 .028 Others 1.291 0.852–1.956 .228 1.234 0.803–1.898 .338 Primary site Soft tissue Reference Reference Retroperitoneum 1.152 0.772–1.717 .489 1.033 0.683–1.562 .879 Uterus 2.312 1.774–3.013 <.001 1.754 1.239–2.483 .002 Others 0.707 0.215–2.327 .568 0.662 0.197–2.229 .506 Histologic type LMS NOS Reference Reference Epithelioid 1.971 1.020–3.809 .044 1.23 0.615–2.460 .559 Bizare 0 0 .999 0 0 .999 Myxoid 0.769 0.231–2.562 .668 0.661 0.191–2.287 .514 Grade I–II Reference Reference III–IV 3.573 2.553–5.001 <.001 2.873 2.030–4.067 <.001 T stage T1–2 Reference Reference T3–4 2.333 1.51.-3.604 <.001 1.001 0.614–1.634 .995 N stage N0 Reference Reference N1 4.064 2.576–6.410 <.001 3.428 2.125–5.532 <.001 Size Range (mm) 1.003 1.002–1.004 <.001 1.002 1.001–1.001 <.001