Incidence of lung metastases
Of the 2714 test patients, 171 were identified to have distant metastases, of which 128 patients had lung metastases. Females were less likely to have lung metastases(15, 11.7%) compared to male patients (P =0.037). Patients with lung metastases had larger tumor sizes (49.96\(\pm\)90.25mm) than the patients without(35.06\(\pm\)23.16mm,P <0.001). Tumors of grade III/IV were more likely to have lung metastases(56, 43.8%) than those with grade I/II (P <0.001). Less follow-up time(8.92\(\pm\)10.31,P <0.001) and higher mortality due to the first tumor (74, 57.8%)( P <0.001) was seen in patients with lung metastases. Age, race and primary tumor site was not significantly associated with lung metastases. Interestingly, patients with lung metastases were quite likely to have more than two organs with distant metastases(31, 24.2%)(P <0.001) (Table 1). The detailed incidence rates of lung metastases showed in Table 2.
Upon multivariable logistic regression (Table 3) among the entire cohort, females versus males’ odds ratio (0.4, 95% CI: 0.17-0.93,P <0.001) was associated with significantly less odds of having lung metastases at diagnosis. Black and Hispanic people seem to incur a higher rate of lung metastases than other races (OR:1.91, 95% CI: 1.03-3.53, P =0.04; OR:2.31, 95% CI:1.16-4.62,P =0.018). Greater tumor sizes resulted in greater odds of lung metastases(OR :1.01, 95% CI: 1.00-1.01, P =0.01) among entire cohort. For each10 mm increase in tumor size, the odds of having lung metastases increases by 6.6%, controlling for other covariates. The predicted probability of lung metastases increased when tumor size increased (Figure 1).