Background: Pleomorphic sarcoma of soft tissue1 (PSST) is a heterogeneous group of sarcomas with different prognoses. Survival patterns of this tumor group have not been thoroughly investigated. Prognostic models should be developed to support individualized evaluation of PSST patients.
Materials and methods: We collected cases of soft tissue tumors in the Surveillance, Epidemiology, and End Results2 (SEER) database, excluding tumors located in intra-abdominal and retroperitoneal regions. We included patients with histopathologic diagnoses of tumors in the PSST group (n = 14,685) and compared the survival patterns of the different histologic types. To develop a prognostic model of leiomyosarcoma, we divided leiomyosarcoma patients into two groups. The first group (n = 4136) was patients with missing values, which became our training cohort after imputation. The second group (n = 671) was the validation cohort.
Results: Leiomyosarcoma was the most common PSST (32.7 %). By pairwise comparison, we found that histologic types of PSST can be divided into 4 prognostic groups (p < 0.001). Regarding our model, the C-index, K statistic, and explained variation R2D were 0.84, 0.8, 0.52 (training cohort) and 0.76, 0.76, 0.4 (validation cohort), respectively.
Conclusions: We found that the survival patterns of PSST can be divided into four distinct groups of histology types. Our dynamic nomogram can individualize the management of leiomyosarcoma patients and, therefore, can be very helpful for clinician-patient communication due to its applicability of flexibility.
Keywords: Dynamic nomogram; Leiomyosarcoma; Pleomorphic sarcoma; Prognosis; SEER; Soft tissue sarcoma.
Copyright © 2022. Published by Elsevier GmbH.