Baseline demographics and characteristics
Table 1 summarizes the baseline characteristics of the study patients. In the entire cohort of 174 patients, the median patient age was 63 years (range, 31–86 years), the ratio of male (n = 136) to female (n = 38) was 3.58:1, and the median follow-up duration was 20 months (range, 1–34 months).
Clinical outcomes of the whole cohort
For the entire cohort, the median PFS and median OS durations of the whole cohort were 8 months [95% CI, 6.5–9.5 months] and 14 months (95% CI, 11.2–16.8 months), respectively. The 12- and 24-month PFS rate was 36.3% and 22.1%, respectively. The 12- and 24-month OS rate was 53.4% and 31.2%, respectively.
Progression events occurred in 70.1% (122/174) patients in the whole cohort. The median PFS was 7 months (95% CI: 5.4–8.6 months) in patients who received camrelizumab as the second-line therapy versus 11 months (95% CI: 6.8–15.2 months) in patients who received camrelizumab as the first-line therapy (P = 0.021). The 12- and 24-month PFS rates were 29.0% and 14.7% in patients who received camrelizumab as the second-line therapy, respectively, and 46.1% and 29.5% in patients who received camrelizumab as the first-line therapy, respectively (Fig. 1a). The median PFS was 6 months (95% CI: 4.3–7.7 months) in patients who received ≤4 cycles of camrelizumab versus 11 months (95% CI: 6.7–15.3 months) in patients who received >4 cycles of camrelizumab (P = 0.007). The 12- and 24-month PFS rates were 28.3% and 23.8% in patients who received ≤4 cycles of camrelizumab, respectively, and 47.5% and 21.8% in patients who received >4 cycles of camrelizumab, respectively (Fig. 1b). The median PFS was 6 months (95% CI: 4.5–7.5 months) in the intermediate–poor LIPI group versus 9 months (95% CI: 6.4–11.6 months) in the good LIPI group (P = 0.01). The 12- and 24-month PFS rates were 27.1% and 11.6% in the intermediate–poor LIPI group, respectively, and 41.5% and 28.1% in the good LIPI group, respectively (Fig. 1c).
Death occurred in 56.9% patients (99/174) in the whole cohort. The median OS was 11 months (95% CI: 7.3–14.7 months) in patients who received camrelizumab as the second-line therapy versus 16 months (95% CI: 13.0–15.2 months) in patients who received camrelizumab as the first-line therapy (P = 0.036). The 12- and 24-month OS rates were 46.2% and 27.0% in patients who received camrelizumab as the second-line therapy, respectively, and 60.3% and 35.5% in patients who received camrelizumab as the first-line therapy, respectively (Fig. 2a). The median OS was 11 months (95% CI: 8.1–13.9 months) in patients who received ≤4 cycles of camrelizumab versus 17 months (95% CI: 14.3–19.7 months) in patients who received >4 cycles of camrelizumab (P = 0.018). The 12- and 24-month OS rates were 44.5% and 30.9% in patients who received ≤4 cycles of camrelizumab, respectively, and 65.8% and 33.4% in patients who received >4 cycles of camrelizumab, respectively (Fig. 2b). The median OS was 11 months (95% CI: 7.6–14.4 months) in the intermediate–poor LIPI group versus 17 months (95% CI: 13.3–20.7 months) in the good LIPI group (P = 0.03). The 12- and 24-month PSF rates were 44.3% and 17.7% in the intermediate–poor LIPI group, respectively, and 64.0% and 38.9% in the good LIPI group, respectively (Fig. 2c).
Univariate analysis showed that receiving camrelizumab as the first-line therapy, receiving >4 cycles of camrelizumab, and having a good LIPI were positively associated with prolonged PFS and OS. Multivariate analyses consistently revealed that receiving >4 cycles of camrelizumab and having a good LIPI were independent predictors of better PFS and OS (Table 2,3).
Clinical outcomes of different treatment patterns
To investigate the impact of different treatment methods on patients’ prognoses, we subdivided all patients into three groups according to whether they combined with chemotherapy or radiotherapy or neither. The camrelizumab monotherapy group included 17.2% (30/174) patients who received only camrelizumab. The camrelizumab + chemotherapy (CT) group included 52.3% (91/173) patients who received camrelizumab combined with chemotherapy but not radiotherapy. The camrelizumab + radiotherapy (RT) group included 30.5% (53/174) patients who received camrelizumab + RT but with or without chemotherapy. Among the patients in the camrelizumab + RT group, 28 patients received chemotherapy and 25 did not receive chemotherapy. In camrelizumab monotherapy, camrelizumab + CT, and camrelizumab + RT groups, the median PFS rates were 5 months (95% CI, 2.8–7.1 months), 9 months (95% CI, 6.6–11.4 months), and 9 months (95% CI, 5.2–12.8 months), respectively (P = 0.055). The Kaplan–Meier estimates for 12-month PFS rates of the three groups were 23.3%, 40.0%, and 35.6%, respectively, and for 24-month PFS rates were 11.7%, 28.8%, and 22.1%, respectively. Camrelizumab + CT led to a significantly prolonged PFS than camrelizumab monotherapy (P = 0.019). Moreover, camrelizumab + RT led to a prolonged PFS than camrelizumab monotherapy (P = 0.075); however, the difference was not statistically significant (Fig. 3a).
For patients receiving camrelizumab monotherapy, camrelizumab + CT, and camrelizumab + RT, the median OS was 10 months (95% CI, 2.5–17.5 months), 17 months (95% CI, 11.7–22.3 months), and 14 months (95% CI,11.7–22.3 months), respectively (P = 0.096). The Kaplan–Meier estimates for 12-month OS rates for the three treatment groups were 37.5%, 56.5%, and 46.2%, respectively, and for 24-months OS rates were 18.3%, 38.7%, and 20.5%, respectively. Camrelizumab + CT led to a significantly prolonged OS than camrelizumab monotherapy (P = 0.032). Conversely, the OS did not significantly differ between camrelizumab + RT and camrelizumab monotherapy groups (P = 0.228; Fig. 3b).
A PFS and OS benefit with LIPI was consistently observed across three different treatment methods. Notably, the exploratory subgroup analysis of PFS revealed that patients in the good LIPI group benefited from camrelizumab + RT (P = 0.008; Fig 4a). However, they did not benefit form camrelizumab monotherapy (P = 0.442; Fig. 4b) and camrelizumab + CT (P = 0.228; Fig 4c). In the subgroup analysis of OS, patients in the good LIPI group benefited from camrelizumab + RT (P = 0.011; Fig 5a). However, these patients did not benefit form camrelizumab monotherapy (P = 0.378; Fig 5b) or camrelizumab + CT (P = 0.111; Fig 5c).
Short-term efficacy outcomes of the whole cohort
Among the 174 patients included in the study, 12 patients did not consent to undergo radiographic examination after the treatment. Thus, the remaining 162 patients were assessed for short-term efficacy. Table 5 summarizes the objective tumor response to camrelizumab. In the 162 patients with measurable disease, CR, PR, and SD were achieved in 5 (3.1%), 66 (40.7%), and 79 (45.4%) patients, respectively. The ORR and DCR were 43.8% and 92.6%, respectively. The CR, PR, and SD rates in the three groups were as follows: camrelizumab monotherapy group [2 (7.7%), 10 (38.5%), and 12 (46.2%), respectively], camrelizumab + CT group [2 (2.4%), 38 (44.7%), and 37 (43.5%), respectively], and camrelizumab + RT group [1 (2.0%), 18 (35.3%), and 30 (58.8%), respectively]. The ORR and DCR rates in the three groups were as follows: camrelizumab monotherapy group (46.2% and 92.3%, respectively), camrelizumab + CT group (47.1% and 90.6%, respectively), and camrelizumab + RT group (37.3% and 96.1%, respectively; Table 4).
Safety and feasibility
During the camrelizumab treatment period, treatment-related adverse events (AEs) occurred in 167/174 (96.0%) patients (Table 4). Treatment-related pneumonitis (any grade) was reported in 22 (12.6%) patients. Immune-related reactive capillary endothelial proliferation (any grade) was reported in 94 (54.0%) patients. Treatment-related AEs of grade 3 or higher were reported in 67 (38.5%) patients, with the most common ones being decreased neutrophil count (23.0%), decreased WBC count (19.5%), anemia (7.5%), and pneumonitis (4.6%). Treatment-related AEs led to treatment interruption of any treatment component in 16 patients (9.2%), with treatment interruption being the most prominent in patients with pneumonitis (4.6%). Patients’ decision and disease progression led to treatment interruption of any treatment component in 13 (7.5%) and 46 (26.4%) patients. One patient (0.6%) died from a treatment-related AE of immune checkpoint inhibitor-induced myocarditis (Table 5).