Characteristics of patients
A total of 103 newly diagnosed ALCL patients with a median age of 31 years (range: 5-68 years) were eligible to be evaluated. Ninety-three (90.3%) patients were ≤ 60 years, the gender ratio was 1.5:1 (male: female), and ALK expression ratio was 68:35 (positive: negative). The median follow-up time was 33 months (range: 1-84 months). Twenty-eight patients (27.2%) experienced recurrence or disease progression, and 14 patients (13.6%) died. The median for ALCs was 1.5 ×109/L (range: 0.54-4.20), and the median for AMCs was 0.5 ×109/L (range: 0.04-1.29). The patient characteristics are shown in Table 1.
Optimal cut-off values for the LMR and LMR/LDH
The optimal cut-off value for the baseline LMR was 2.4, and the area under the curve (AUC) was 0.687 (95% CI: 0.532-0.843; P = 0.025; Fig. 1A). A total of 75 patients (72.8%) with LMR > 2.4 were defined as the high baseline LMR group, and 28 patients (27.2%) with LMR ≤ 2.4 were defined as the low baseline LMR group. The optimal cut-off value for LMR/LDH was 1.4, and the AUC was 0.738 (95% CI: 0.625-0.851; P = 0.004; Fig. 1B). A total of 57 patients (55.3%) had an LMR/LDH > 1.4, and 46 patients (44.7%) had an LMR/LDH ≤ 1.4.
The low baseline LMR Is significantly associated with high IPI score, B symptom, increased β2-MG, and Ann-Arbor stage Ⅲ-Ⅳ.
Several clinical factors were associated with the low baseline LMR in ALCL patients (Table 1). Specifically, low baseline LMR was common in high-risk patients with an IPI score 3-5 (56.5%), while it was rare in patients with IPI score 0-2 (18.8%) (P < 0.001). Similarly, low baseline LMR was more frequent in patients with B symptoms (35.8%) than in patients without B symptoms (18.0%) (P = 0.042). And low baseline LMR was significantly related to increased serum β2-MG (P = 0.036). Additionally, compared with 12.9% of patients with Ann-Arbor stageⅠ-Ⅱ, 33.3% of advanced patients with Ann-Arbor stage Ⅲ-Ⅳ had an LMR ≤ 2.4 (P = 0.033).
The baseline LMR can predict early response to treatment and prognosis for ALCL patients
The CCR rate of the low baseline LMR group (57.1%) was remarkably lower than that for the high baseline LMR group (80.0%, P = 0.019). When a good response (CR+PR) was obtained in the high LMR group, the LMR was decreased significantly (P = 0.030) but not for patients with a poor response (P = 0.086). No significant difference was found in the low LMR group regardless of whether a response was achieved or not (P values were 0.137 and 0.358).
The Low baseline LMR was significantly related to poor OS (Fig. 2A, P = 0.006) and PFS (Fig. 2B, P = 0.013), and the 5-year OS and PFS were 51.3% and 32.6%, respectively, in the low LMR group, and 90.4% and 70.5%, respectively, in the high LMR group. Meanwhile, low baseline LMR/LDH was significantly associated with shorter OS (Fig. 2C, P = 0.002) and PFS (Fig. 2D, P = 0.029) when compared with that of the high baseline LMR group, and the 5-year OS and PFS were 57.4% and 52.3%, respectively, in the low LMR/LDH group, and 92.8% and 74.9%, respectively, in the high LMR/LDH group.
The baseline LMR is highly consistent with ALK expression in predicting prognosis of patients with ALCL
In this study, we found that the CCR rate in the ALK-negative group (57.1%) was significantly lower than that in the ALK-positive group (82.4%) (P = 0.006). The OS (Fig. 3A, P = 0.005) and PFS (Fig. 3B, P = 0.012) in the ALK-negative group were significantly poorer than that in the ALK-positive group, and the 5-year OS and PFS were 60.4% and 47.1%, respectively, in the ALK-negative group, and 92.8% and 76.2%, respectively, in the ALK-positive group. Although the baseline LMR and ALK expression are highly consistent in terms of predicting efficacy and prognosis, there was no significant correlation between the two (R = 0.10).
To compare the prognostic power of the baseline LMR, ALK expression, and their combination for ALCL patients, we divided the 103 cases into four subgroups: Group a: patients who were ALK positive and had a baseline LMR > 2.4, Group b: patients who were ALK positive and had a baseline LMR ≤ 2.4, Group c: patients who were ALK negative with a baseline LMR > 2.4, and Group d: patients who were ALK negative with a baseline LMR ≤ 2.4. We found that a baseline LMR > 2.4 combined with ALK positive were significantly associated with better OS (Fig. 3C, P < 0.001) and PFS (Fig. 3D, P = 0.002) when compared with the other three groups. In contrast, patients who were ALK negative with a baseline LMR ≤ 2.4 had the worst prognosis compared to patients who were ALK positive and/or those with a baseline LMR > 2.4.
The survival of patients with an increased LMR after treatment is greater than that for patients with a decreased LMR after treatment.
To observe changes in LMR during treatment and their effect on prognosis, the LMR at various time points during treatment and the follow-up was dynamically monitored, and the patients were divided into two subgroups: Group 1, those with an increased LMR at the end of therapy (greater than or equal to the baseline LMR), and Group 2, those with a decreased LMR after treatment (less than the baseline LMR). It was found that patients in Group 1 had a longer OS (Fig. 4A, P = 0.014) and PFS (Fig. 4B, P = 0.018) than those in Group 2, and the 5-year OS and PFS were 92.05% and 75.88%, respectively, in Group 1, and 69.48% and 57.68%, respectively, in Group 2. Additionally, we found that different chemotherapy regimens had no effect on the LMR (Suppl Fig. 1); nevertheless, at 6-9 months of the follow-up, the LMR of patients treated with chidamine was significantly lower than that for patients not treated with chidamine until 12 months later (Fig. 5A, P = 0.046 at 6 months and 0.037 at 9 months, respectively). As expected, compared with chemotherapy, chidamine combined with chemotherapy demonstrated a mild advantage in PFS and OS for both newly diagnosed (Fig. 5B and Fig. 5C) and R/R patients (Fig. 5D and 5E), but there was no statistically significant difference (All P values > 0.05).
In addition, the LMR at recurrence was lower than that at last follow-up in the high and low LMR groups, but there was no statistically significant difference (Suppl Fig. 2, P = 0.082 and 0.317, respectively).
Survival analysis and prognosis factors
Univariate analysis is shown in Table 2. ECOG PS ≥ 2, Ann-Arbor stage ≥ Ⅲ, ALK negative, IPI Score ≥ 3, ALC ≤ 1.5 ×109/L, AMC > 0.5 ×109/L, elevated serum LDH level, elevated serum β2-MG, extranodal site involvement, LMR ≤ 2.4, and LMR/LDH ≤ 1.4 could be significantly associated with shortened OS and PFS. Clinical factors with statistical significance for OS and PFS (P < 0.05) were included in the multivariate analysis (Table 3). Multivariate analysis demonstrated that LMR was a poor prognostic factor for OS (P = 0.02) and PFS (P = 0.013).