Construction and validation of prognostic nomogram model for primary malignant cardiac tumours based on SEER database: A retrospective study

J Pak Med Assoc. 2025 Jul;75(Suppl 2)(7):S72-S78. doi: 10.47391/JPMA.SRPH-12.

ABSTRACT

OBJECTIVE: To build a prognostic nomogram for predicting primary malignant cardiac tumour patients' overall survival.

METHODS: The retrospective study was conducted in January 2023, and comprised data from January 2020 to December 2022 of primary malignant cardiac tumour patients obtained from the Surveilance, Epidemiology, and End Results database. The data was divided into training cohort A and validation cohort B. A prognostic nomogram for predicting overall survival was generated by means of independent prognostic factors. The performance nomogram was validated using the receiver operating characteristic curve, calibration curve, decision curve analysis and risk stratification. Data was analysed using SPSS 24.

RESULTS: Of the 528 patients, 371(70.3%) were in cohort A; 275(52.1%) males and 253(47.9%) females with 404(76.5%) aged 24-76 years. There were 157(29.7%) patients in cohort B; 81(51.6%) males and 76(48.4%) females with 124(79%) aged 24 76 years (p>0.05). Age, American Joint Cancer Staging Committee stage, histology and chemotherapy were independent risk elements for overa l survival, and were used for generating a nomogram (p<0.001). The area under the curve of traditional staging systems was surpassed by nomogram in both cohorts (p<0.001). The calibration curves suggested a strong concordance between predicted and practical survival (p<0.001). Decision curve analysis revealed that the nomogram had satisfactory clinical application value (p<0.001). The risk stratification system provided evidence of the nomogram's capacity to precisely identify high-risk patients (p<0.001).

CONCLUSIONS: A useful and reliable prognostic nomogram for predicting overa l survival in primary malignant cardiac tumour patients was developed and validated, improving oncologists' ability to accurately assess patient outcomes.

PMID:41248641 | DOI:10.47391/JPMA.SRPH-12