Among 114 patients, 73 (64.0%) were aged ≥65 while 41 (36.0%) were aged <65 years. Further, 64 (56.1%) were males and 50 (43.9%) were females. Histological subtypes included adenocarcinoma, squamous cell carcinoma, large cell carcinoma, and small cell carcinoma in 76 (66.7%), 21 (18.4%), 2 (1.8%), and 15 (13.2%) cases, respectively. EGFR mutation, ALK, other molecular alterations, or no molecular alterations, as driver molecular alterations, were observed in 38 (33.3%), 5 (4.4%), 3 (2.6%), and 68 (59.6%) patients, respectively. PS at the onset of bone mets was 0, 1, 2, 3, and 4 in 10 (8.8%), 49 (43.0%), 31 (27.2%), 20 (17.5%), and 4 (3.5%) patients, respectively. Post-bone mets best PS was 0, 1, 2, 3, and 4 in 11 (9.6%), 72 (63.2%), 24 (21.1%), 6 (5.3%), and 1 (0.9%) patients, respectively. Post-bone mets first BOR was PR, SD, and PD in 43 (42.6%), 22 (21.8%), and 36 (35.6%) patients, respectively. The number of lines before post-bone mets first pharmacotherapy was 0, 1, and ≥2 in 87 (86.1%), 10 (9.9%), and 4 (4.0%) patients, respectively. The regimen of post-bone mets first pharmacotherapy included platinum-based drugs in 28 (27.7%), ICI in 11 (10.9%), molecular-targeted drugs in 38 (37.6%), platinum-based drugs combined with ICI in 15 (14.9%), and others in 9 (8.9%) patients. Surgery for bone metastases was performed in 8 (7.0%) patients whereas radiotherapy for bone metastases was done in 89 (78.1%). BMA was administered in 75 (65.8%) patients. Distant metastases other than bone metastases were observed in 86 (75.4%) patients: liver in 15 (13.2%), brain in 40 (35.1%), dissemination in 15 (13.2%), lung in 2 (1.8%), and others in 14 (12.3%). Further, 80 (70.2%) patients had bone metastases at the initial visit to our institute while 103 (90.4%) were dead at the final follow-up (Table 1).
Median post-bone metastasis survival was 13.6 months (1-year survival: 54%) (Figure 1). The relationship between post-bone metastasis survival and each parameter was next investigated using univariate analysis. First, age was not significantly involved in post-bone metastasis survival (age <65 years, n = 41), median post-bone metastasis survival of 15.3 (95% confidence intervals [CI]: 12.9–21.3) months; age of ≥65 (n = 73), median post-bone metastasis survival of 10.5 (95% CI: 7.0–18.7) months; p = 0.971) whereas gender significantly impacts post-bone metastasis survival (female, n = 50, median post-bone metastasis survival 22.0 (95% CI: 15.3–28.8) months; male, n = 64, median post-bone metastasis survival of 11.2 (95% CI: 7.1–13.6) months; p < 0.05) (Figure 2A and B). Small cell lung cancer (SCLC) provided significantly poorer survival than non-small cell lung cancer (NSCLC) in terms of histological subtype (NSCLC, n = 99, median post-bone metastasis survival of 15.4 [95% CI: 12.2–21.3] months; SCLC (n = 15), median post-bone metastasis survival of 7.1 (95% CI: 3.2–9.9) months; p < 0.001) (Figure 2C). Additionally, EGFR or ALK driver molecular alterations significantly influenced post-bone metastasis survival (driver molecular alterations negative, n = 71), median post-bone metastasis survival of 9.2 (95% CI: 6.1–12.9) months; positive driver molecular alterations (n = 43), median post-bone metastasis survival of 25.4 (95% CI: 17.6–31.4) months (p < 0.001) (Figure 2D). PS was significantly involved in post-bone metastasis survival not only at the onset of bone metastases (PS 0, n = 9, median post-bone metastasis survival of 25.0 [95% CI: 0.4–N/A] months; PS 1, n = 52, 13.8 [95% CI: 9.2–17.3] months; PS 2, n = 32, 17.6 [95% CI: 9.9–31.8] months; PS 3, n = 16, 6.5 [95% CI: 2.3–10.5] months; PS 4, n = 5, 5.4 [95% CI: 1.2–N/A]; p < 0.05) but also at its best after bone metastases (PS 0, n = 11, median post-bone metastasis survival of 28.8 [95% CI: 4.8–N/A] months; PS 1, n = 72, 15.3 [95% CI: 12.9–19.7] months; PS 2, n = 24, 7.0 [95% CI: 4.7–15.4] months; PS 3, n = 6, 2.3 [95% CI: 0.7–N/A] months; PS 4, n = 1, 1.2 [95% CI: N/A–N/A]; p < 0.001) (Figure 2E and F). Post-bone mets first BOR was significantly associated with post-bone metastasis survival when it comes to treatment (PR, n = 52, median post-bone metastasis survival of 17.3 (95% CI: 12.9–26.0) months; SD, n = 19, 17.6 (95% CI: 6.7–36.5) months; PD, n = 31, 7.9 (95% CI: 4.7–10.7) months; p < 0.001) (Figure 2G). Surgery and/or radiotherapy for bone metastases was not significantly related to post-bone metastasis survival (neither surgery nor radiotherapy, n = 17, median post-bone metastasis survival 20.7 [95% CI: 6.9–28.8] months; any surgery and/or radiotherapy, n = 97, 12.9 [95% CI: 9.9–16.0] months; p = 0.695) (Figure 2H). Additionally, BMA significantly improved post-bone metastasis survival (no BMA, n = 39, median post-bone metastasis survival of 7.8 (95% CI: 5.5–12.9) months; BMA, n = 75, 15.4 (95% CI: 12.9–20.7) months; p < 0.05) (Figure 2I). Lastly, molecular-targeted drugs significantly prolong post-bone metastasis survival as compared to other regimens (ICI, n = 11, median post-bone metastasis survival of 13.6 [95% CI: 3.8–26) months; molecular targeted, n = 38, 28.8 (95% CI: 20.7–35.3) months, platinum-based, n = 28, 12.9 (95% CI: 8.1–15.3) months; platinum-based + ICI, n = 15, 11.6 (95% CI: 4.8–14.7) months; p < 0.001) (Figure 2J).
Then, multivariate analysis was conducted to eliminate confounders, revealing that post-bone mets best PS, histological subtype, and BMA were independent significant factors to impact post-bone metastasis survival (Table 2). Next, we focused on patients with bone metastases at the initial visit. Post-bone mets best PS and driver molecular alterations were independent significant factors for post-bone metastasis survival (Table 3). Generally, SCLC is apparently associated with poor prognosis consistent with our result in Figure 2C. Hence, we next focused on NSCLC. Driver molecular alterations, which include EGFR mutation and ALK fusion or rearrangement, were a positive prognostic factor (no molecular alterations, n = 56, median post-bone metastasis survival of 11.6 [95% CI: 6.1–14.2] months; molecular alterations, n = 43, 25.4 [95% CI: 17.6–31.4] months; p<0.05) (Figure 3A). Additionally, molecular-targeted drugs were significantly associated with better prognoses than other pharmacotherapies, including ICI or platinum-based agents (ICI, n = 11, median post-bone metastasis survival of 13.6 [95% CI: 3.8–26] months; molecular targeted, n = 38, median post-bone metastasis survival 28.8 [95% CI: 20.7–35.3] months, platinum-based, n = 20, 14.7 [95% CI: 9.2–27.1] months; platinum-based + ICI, n = 11, 11.6 [95% CI: 4.8–15.3] months; p<0.005) (Figure 3B). Moreover, multivariate analysis shows that post-bone mets best PS and driver molecular alterations significantly affect post-bone metastasis survival (Table 4).
A subgroup analysis extracted a subpopulation with PS 2 or poorer at the initial visit to clarify whether PS improvement could improve post-bone metastasis survival and determine the involved factors. Post-bone metastasis survival was significantly prolonged in patients with improved PS after some treatment as compared to those without improvement (PS improvement, n = 30, median post-bone metastasis survival of 18.7 [95% CI: 10.5–26.4] months; no improvement, n = 23, 6.9 [95% CI: 2.3–10.7] months; p<0.05) (Figure 4). Our results indicate that post-bone mets best PS is strongly associated with prognosis. Thus, we aimed to identify associated factors with PS improvement. Univariate analysis was first performed, showing gender and post-bone mets first BOR as significant parameters for PS improvement, and multivariate analysis revealed post-bone mets first BOR as an independent significant factor (Tables 5 and 6). Next, we analyzed 101 patients who underwent pharmacotherapy to identify the factors related to post-bone mets first BOR, indicating that driver molecular alterations were likely associated with PR or SD of post-bone mets first BOR (p = 0.059) (Table 7). Moreover, multivariate analysis demonstrates molecularly targeted drug as an independent positive prognostic factor (Table 8).