3.1 Clinical characteristics
The baseline characteristics of the enrolled patients, 85 males (60.7%) and 55 females (39.3%), are shown in Table 1. The median age was 62 years (range 17–83 years), 55.0% of the patients being over 60 years old and 15.0% being over 70 years old. Almost all patients, 95.0%, presented with stage III or IV disease at the time of initial diagnosis. Bone marrow involvement was observed in 19.3% of the patients, while extranodal involvement at two or more sites was noted in 20.7%. The most frequently affected extranodal sites were the bone marrow (19.3%), nasal cavity (15.0%), and lung (10.0%). Laboratory data were obtained concurrently with the initial diagnosis. Upon initial diagnosis, anemia was identified in 56.0% of the patients. Elevated LDH was found in 61.6% of the patients. Hypoalbuminemia was present in 56.9%, high β2-microglobulin (β2-MG) (≥ 2.4 mg/L) was observed in 81.0%, and elevated IgA > 400 mg/dL were noted in 77.8% of the patients.
Table 1
AITL clinicodemographic characteristics(n = 140)
Variable | | | No. (%) |
Age years | | | |
Median(range) | | | 62.0(17–83) |
> 60 years | | | 77/140(55) |
> 70 years | | | 21/140(15) |
Sex | | | |
male | | | 85/140(60.7) |
female | | | 55/140(39.3) |
Ann Arbor stage | | | |
I | | | 1/140(0.7) |
II | | | 6/140(4.3) |
III | | | 53/140(37.9) |
IV | | | 80/140(57.1) |
LDH>240 U/L | | | 85/138(61.6) |
Aalbumin level<35g/L | | | 78/137(56.9) |
IgA level < 400mg/dL | | | 35/45(77.8) |
CRP ≥ 10mg/L | | | 51/80(63.7) |
β2-MG ≥ 2.4mh/L | | | 81/100(81) |
Anemia (Hb < 100g/L) | | | 101/139(72.7) |
Ki-67 > 60% | | | 42/125(33.6) |
Positive HBsAg | | | 14/124(11.3) |
Positive HCV | | | 2/120(1.7) |
Positive EBER | | | 64/101(63.7) |
Positive EBV-DNA | | | 27/105 (25.7) |
Median | | | 400 |
Average | | | 747.7 |
NO.of extranodal sites ≥ 2 | | | 29/140(20.7) |
Extranodal sites | | | |
bone marrow | | | 27/140(19.3) |
Nose | | | 21/140(15.0) |
Lungs | | | 14/140(10.0) |
Parotid gland | | | 11/140(7.9) |
Bone | | | 8/140(5.7) |
Liver | | | 6/140(4.3) |
Skin | | | 5/140(3.6) |
Note: LDH, lactate dehydrogenase; CRP, C-reactive protein; β2-MG, β2-macroglobulin; HBsAg, Hepatitis B surface antigen; HCV, Hepatitis C virus. |
3.2 Pathological characteristics
AITL is distinguished by a distinctive follicular architecture marked by clusters of TFH-associated markers, including CD4, PD-1, CXCL13, BCL6, and CD10[28]. In this study, immunohistochemistry showed that the percentage of CD10-positive cells was 63.3%, the percentage of CXCL13-positive cells was 89.0%, the percentage of CD30-positive cells was 78.2%, the percentage of BCL6-positive cells was 77.9%, the percentage of PD-1-positive cells was 89.0%, the percentage of PD-L1-positive cells was 99.2%, and the positivity rate for Ki-67 (exceeding 60%) was 33.6% (Suppl. Table 1).
3.3 First-line treatment
Among the first-line treatments, 116 patients received CHOP-like regimens, while 24 patients received non-CHOP-like regimens. The non-CHOP-like regimens included 13 of chidamide monotherapy, one of gemcitabine monotherapy, and seven of combination therapies involving gemcitabine. Specifically, the latter included three patients treated with the GDP regimen (gemcitabine, cisplatin, and dexamethasone) combined with chidamide, three patients treated with P-GemOx, and one patient treated with the PD1 antibody in combination with GDP. A comparative analysis of baseline characteristics between the group treated with CHOP-like regimens and the group treated with non-CHOP-like regimens indicated a significant difference in the proportion of patients aged 70 years and above (P < 0.05). No statistically significant differences were observed in sex, age, white blood cell count, hemoglobin, platelet count, LDH, C-reactive protein (CRP), Ki-67 index greater than 60%, β2-MG, albumin, ECOG PS, bone marrow involvement, Ann Arbor stage, or the extent of extranodal disease (Supplementary Table 2). In the first-line treatment, chidamide was administered to 51 patients, representing 36.4% of the total patients, while BV was given to 6 patients, accounting for 4.3% of the total. A total of 140 patients were evaluated for treatment efficacy, 36 (25.7%) of whom achieved CR, 34 (24.3%) achieved PR, 18 (12.9%) achieved SD, and 52 (37.1%) achieved PD, resulting in an ORR of 62.9%. In the first-line treatment, patients who received CHOP-like regimens exhibited CR and PR rates of 28.4% and 23.3%, respectively, with an ORR of 65.5%. For those treated with non-CHOP-like regimens, the respective rates were 12.5% and 29.2%, yielding an ORR of 50.0%. Although the CR rate and ORR for the CHOP-like regimens were modestly higher than those for the non-CHOP-like regimens, the difference was not statistically significant (P = 0.261) (Fig. 1A). Among the responding patients, the median DOR was 495.4 ± 406.8 days for the CHOP-like group and 383.2 ± 384.8 days for the non-CHOP-like group (P = 0.353) (Fig. 1C).
To further analyze the value of combining chidamide with first-line treatment, patients were categorized into two groups for efficacy evaluation: one receiving chidamide-containing regimens and the other receiving chidamide-free regimens. The baseline characteristics of these groups were similar (Supplementary Table 3). The efficacy evaluation yielded the following outcomes: the CR rates for the chidamide-containing and chidamide-free groups were 21.6% and 28.1%, respectively; the PR rates were 23.5% and 24.7%; and the ORRs were 60.8% and 66.7%. The addition of chidamide to first-line treatment failed to improve the remission rate (Fig. 1B). Among patients who achieved a treatment response, the DOR of the group administered chidamide was 421.3 ± 372.8 days, whereas the DOR of the group not receiving chidamide was 520.9 ± 422.1 days (P = 0.230). When chidamide was combined with first-line CHOP-like regimens, the DOR was 461.3 ± 389.2 days, compared to 515.6 ± 419.2 days for the CHOP-like regimens without chidamide (P = 0.553). The DOR for the first-line combination of chidamide with non-CHOP-like regimens was 261.3 ± 258.6 days, and for the non-CHOP-like regimens without chidamide, it was 578.2 ± 499.8 days (P = 0.156) (Fig. 1D-F).
In the first line, maintenance therapy was administered to 19 patients (13.6%). Of them, 11 (57.9%) received chidamide, 3 (15.8%) received interferon alpha, 1 received BV, and 4 received thalidomide or lenalidomide. Overall, maintenance therapy did not significantly prolong the DOR (maintenance group: 537.1 ± 464.4 days; nonmaintenance group: 484.9 ± 414.4 days, P = 0.676). Chidamide as a maintenance therapy did not significantly prolong DOR (chidamide maintenance group: 394.5 ± 164.5 days; without chidamide maintenance group: 488.3 ± 418.1 days, P = 0.215). Six patients who received chidamide in the first-line treatment were then given chidamide maintenance, all of whom had initially been treated with a CHOP-like regimen in combination with chidamide. The DOR for these 6 patients was 416.8 ± 182.9 days, while for the other patients it was 484.8 ± 414.3 days (P = 0.451), indicating that the use of long-term chidamide in first-line treatment also failed to prolong the DOR.
Eight patients (5.7%) underwent ASCT as a first-line treatment, including 6 at CR1 and 2 at PR1. The DOR of patients who received first-line ASCT was 478.0 ± 191.0 days, and for the other patients it was 469.4 ± 404.4 days (P = 0.927). The failure of first-line ASCT consolidation to prolong the DOR may have been due to the small sample size.
3.4 Second-line treatment
A total of 48 patients received second-line treatment. Twenty-five of them (52.1%) entered second-line treatment due to PD after first-line therapy; these patients composed the refractory group. The other 23 patients (47.9%) entered second-line treatment after recurrence and progression after a response to the first-line treatment, comprising 8 patients who achieved CR, 7 who achieved PR, and 8 with SD; these were defined as the relapsed group. As the second-line treatment regimen, 40 patients (83.3%) utilized non-CHOP-like regimens, which included 9 with chidamide combinations, 12 with gemcitabine-based treatments, and 2 with the DHAP regimen (dexamethasone, cisplatin, and cytarabine). In contrast, only 8 patients (16.7%) received CHOP-like regimens. In the second-line treatment, chidamide was combined with various agents, including azacitidine, gemcitabine, the CHOP regimen, BV, lenalidomide and temozolomide. In the evaluation of the efficacy of second-line treatment, 6/48 (17.1%) achieved CR, 5/48 (14.3%) achieved PR, 3/48 (8.6%) experienced SD, and 21/48 (60.0%) exhibited PD. Within the refractory group, 3/23 (23.1%) patients achieved CR, 1/23 (7.7%) achieved PR, 1/23 (7.7%) achieved SD, and 8/23 (61.5%) achieved PD. In the relapsed group, 3/25 (13.6%) patients achieved CR, 4/25 (18.2%) achieved PR, 2/25 (9.1%) achieved SD, and 13/25 (59.1%) achieved PD.
After second-line treatment, 15 patients were evaluated for DOR. The overall DOR for the second-line treatment was 366.7 ± 431.2 days. The DOR for the refractory group was 444.7 ± 525.2 days, and the DOR for the relapsed group was 298.4 ± 352.1 days (P = 0.532). Based on the above results, the efficacy of first-line treatment does not seem to affect the efficacy of second-line treatment, in terms of either the response rate or the duration of response.
3.5 Third-line treatment
A total of 15 patients received third-line treatment, with 11 (73.3%) advancing to this stage after PD during the second-line treatment. The remaining 4 (26.7%) patients entered third-line treatment after achieving either PR or SD as second-line treatment. The third-line treatment regimens predominantly consisted of the following: combinations with chidamide in 4 patients, which included azacitidine, lenalidomide, VP-16, and ICE; single-agent chidamide in 2 patients; regimens incorporating etoposide in 3 patients; and regimens containing gemcitabine in 2 patients. Thirteen patients were evaluable for their response to treatment. Of them, 1/13 (7.7%) achieved CR, 2/13 (15.4%) achieved PR, 3/13 (23.1%) experienced SD, and 7/13 (53.8%) showed PD. Notably, the only patient who achieved CR did so after receiving salvage ASCT from a fully matched sibling donor, following a PR in the second-line treatment.
3.6 Fourth/fifth-line treatment
Seven patients received fourth-line treatment, 5/7 (71.4%) of whom underwent PD as a third-line treatment and 2/7 (28.6%) of whom underwent PD after SD as a third-line treatment. The fourth-line treatment regimens included gemcitabine + everolimus in 3 patients, chidamide combined with lenalidomide and bortezomib in 1 patient, VP-16 + dexamethasone in 1 patient, PD-1 + GDP in 1 patient, and BV + chidamide in 1 patient. After treatment, 6 patients were evaluable for response, with 2/6 (33.3%) achieving SD and 4/6 (66.7%) experiencing PD. A total of 3 patients received fifth-line treatment, all of whom underwent PD as a fourth-line treatment; none of these patients achieved a treatment response, and they all ultimately passed away.
3.7 Outcome
The median duration of follow-up was 27.8 months (95% CI, 18.7 to 36.7). For the entire cohort of patients, the 2-year PFS rate was 36.1%, and the median PFS was 9.1 months (95% CI, 6.2 to 12.0 months). The 2-year OS rate was 60.3%, and the median OS was 32.8 months (95% CI, 16.4 to 49.2 months) (Fig. 2A, B). Eighty-five patients (60.7%) experienced PD within 24 months (POD24). The median PFS for the POD24 group was significantly shorter, at 5.1 months (95% CI, 4.2 to 6.0 months), than that for the non-POD24 group, who had a median PFS of 49.2 months (95% CI, 42.8 to 55.6 months) (P < 0.001) (Fig. 2C). The median OS for the POD24 group was 13.8 months (95% CI, 10.8 to 16.8 months), while the median OS for the non-POD24 group was not reached (P < 0.001) (Fig. 2D). The POD24 group exhibited a significantly poorer prognosis than the non-POD24 group.
The results of the univariate analysis (Table 2) showed that age, bone marrow involvement, elevated LDH levels, β2-microglobulin content, albumin levels, and the initial treatment regimen were factors negatively affecting PFS(P < 0.05). Age over 70 years, bone marrow involvement, elevated LDH levels, elevated β2-microglobulin level, and hypoalbuminemia were factors negatively affecting OS(P < 0.05). Age over 70 years old, bone marrow involvement and not receiving a CHOP-like regimen in the first-line are three risk factors for both PFS and OS. Multivariate analysis showed that bone marrow involvement, higher LDH, and higher β2-MG were factors that negatively affected PFS (P < 0.05). Higher β2-MG was the only factor negatively affecting OS (P < 0.05) (Table 2).
Table 2
univariate analysis for OS and PFS and multivariate analysis for OS
| PFS | | OS | |
| univariate analysis | | multivariate analysis | | univariate analysis | | multivariate analysis | |
Variable | HR | 95%CI | P | | HR | 95%CI | P | | HR | 95%CI | P | | HR | 95%CI | P | |
Sex, male | | 1.056 | 0.697–1.598 | 0.797 | | | | | | | 1.156 | 0.678–1.969 | 0.595 | | | | | | |
Age > 70years | | 2.271 | 1.331–3.879 | 0.003 | | 1.890 | 0.964–3.706 | 0.064 | | 2.329 | 1.227–4.421 | 0.010 | | 1.907 | 0.856–4.249 | 0.114 |
Ann Arbor stage ≥ III | | 1.217 | 0.494–3.001 | 0.669 | | | | | | | 1.160 | 0.362–3.723 | 0.803 | | | | | | |
CRP ≥ 10mg/l | | 1.679 | 0.948–2.973 | 0.075 | | | | | | | 1.623 | 0.781–3.375 | 0.194 | | | | | | |
Extranodal site > 1 | | 1.310 | 0.796–2.155 | 0.288 | | | | | | | 1,171 | 0.587–2.333 | 0.654 | | | | | | |
Bone marrow involvement | | 1.896 | 1.160–3.098 | 0.011 | | 2.160 | 1.174–3.972 | 0.013 | | 1.712 | 0.919–3.210 | 0.090 | | 1.231 | 0.578–2.625 | 0.588 | |
LDH > 240 u/L | | 1.820 | 1.169–2.832 | 0.008 | | 1.834 | 1.067–3.154 | 0.028 | | 2.032 | 1.138–3.628 | 0.017 | | 2.015 | 0.993–4.088 | 0.052 | |
β2-MG > 2.4mg/L | | 4.081 | 1.749–9.524 | 0.001 | | 2.995 | 1.248–7.190 | 0.014 | | 7.382 | 1.782–30.587 | 0.006 | | 4.554 | 1.067–19.361 | 0.041 | |
Albumin < 35g/L | | 1.967 | 1.280–3.024 | 0.002 | | 1.253 | 0.731–2.149 | 0.412 | | 2.710 | 1.493–4.921 | 0.001 | | 1.742 | 0.844–3.596 | 0.113 | |
IgA > 400mg/dL | | 0.940 | 0.355–2.487 | 0.900 | | | | | | | 1.065 | 0.298–3.804 | 0.922 | | | | | | |
Ki-67index > 60% | | 0.867 | 0.543–1.383 | 0.548 | | | | | | | 0.877 | 0.482–1.596 | 0.668 | | | | | | |
1st line with chop-like | | 1.694 | 1.020–2.815 | 0.042 | | 1.558 | 0.855–2.840 | 0.148 | | 1.565 | 0.806–3.038 | 0.186 | | 1.229 | 0.606–2.786 | 0.501 | |
1stline with chidamide | | 1.027 | 0.673–1.565 | 0.903 | | | | | | | 1.332 | 0.783–2.264 | 0.290 | | | | | | |
1stline with BV | | | 1.657 | 0.670–4.100 | 0.274 | | | | | | | 1.996 | 0.620–6.428 | 0.247 | | | | | | |
1stline with maintenance | | 0.909 | 0.473–1.750 | 0.776 | | | | | | | 0.801 | 0.325–1.972 | 0.629 | | | | | | |
LDH, lactate dehydrogenase; CRP, C-reactive protein; β2-MG, β2-microglobulin; OS, overall survival; PFS, progression free survival; 95% CI, 95% confidence interval. |
3.8 Validation of different prognostic models
We tested seven prognostic models that were developed for risk stratification of lymphoma patients. Our patients were scored based on the original model criteria and stratified into high-risk and low-risk groups. As shown in Figs. 3 and 4, all seven prognostic models showed good predictive power for both PFS and OS. The predictive efficacy of each scoring system was calculated as the C-index. The top three models for predicting PFS were the AITL score (0.670 ± 0.036), the PIAI (0.655 ± 0.048), and the Chinese AITL score (0.631 ± 0.029). For predicting OS, the top three models were the PIAI (0.698 ± 0.057), the Chinese AITL score (0.655 ± 0.034), and the AITL score (0.644 ± 0.046). We present the 5-year OS and PFS estimates from the original models and calculated the 1- to 4-year OS and PFS estimate ranges for all models, as shown in Table 3. Given that the Chinese AITL score demonstrated exceptional predictive efficacy for both OS and PFS and considering the significant differences in survival between the high-risk and low-risk groups defined by the Chinese AITL score (4-year PFS: 37.9% vs. 78.6%, 4-year OS: 10.1% vs. 69.3%, p < 0.001), we suggest that the Chinese AITL score may be a more appropriate prognostic model for Chinese AITL patients.
Table 3
Model (N of patients) | | Original study | | | Current study N = 140 | | |
| Model variables | Risk groups(point) | years -OS% | years -PFS (%) | Risk groups (point) | years -OS% | years -PFS (%) |
IPI N = 2031 | Age > 60 years | | | | | | |
| ECOG PS > 1 | Low risk (0–1) | 5Y-73.0 | 5Y-70.0 | Low risk (≤ 2) | 4Y-49.8 | 4Y-29.0 |
| Stage III/IV | Low-intermediate (2–3) | 5Y-51.0 | 5Y-50.0 | High risk (> 2) | 4Y-39.7 | 4Y-20.5 |
| EN sites > 1 | High-intermediate (4–5) | 5Y-43.0 | 5Y-49.0 | | | |
| LDH > ULN | High risk (≥ 6) | 5Y-26.0 | 5Y-40.0 | | | |
PIAI N = 243 | Age > 60 | | | | | | |
| ECOG PS > 1 | Low risk (0–1) | 5Y-44.0 | 5Y-28.0 | Low risk (≤ 2) | 3Y-43.9 | 3Y-21.5 |
| EN Sites ≥ 1 | High risk (2–4) | 5Y-24.0 | 5Y-15.0 | High risk (> 2) | 3Y-14.8 | 3Y-6.7 |
| Presence of B symptoms | | | | | | |
| PLT < 15×10*9 | | | | | | |
ATPI N = 207 | Age > 60 | | | | | | |
| EN Sites > 1 | Low risk (0–1) | 3Y-85.0 | 5Y-73.0 | Low risk (≤ 2) | 2Y-80.7 | 1Y-52.6 |
| PLT < 15×10*9/L | Low-intermediate (2) | 3Y-62.0 | 5Y-37.0 | High risk (> 2) | 2Y-32.4 | 1Y-16.7 |
| WBC > 1×10*12/L | High-intermediate (3) | 5Y-51.0 | 5Y-36.0 | | | |
| Hb < 100 g/L | High risk (4–6) | 5Y-12.0 | 5Y-13.0 | | | |
| IgA > 400mg/dl | | | | | | |
AITL score N = 282 | Age ≥ 60 years | | | | | | |
| ECOG PS > 2 | Low risk (0–1) | 5Y-63.0 | 5Y-41.0 | Low risk (≤ 2) | 3Y-45.6 | 1Y-36.9 |
| CRP > ULN | Intermediate- risk (2) | 5Y-54.0 | 5Y-37.0 | High risk (> 2) | 3Y-8.6 | 1Y-5.9 |
| Presence of B symptoms | High risk (3–4) | 5Y-21.0 | 5Y-13.0 | | | |
Chinese AITL score N = 164 | Age > 70 years | | | | | | |
| LDH > ULN | Low risk (0) | 5Y-69.0 | NA | Low risk (≤ 0) | 4Y-78.6 | 4Y-69.3 |
| Albumin < 35g/L | Intermediate risk (1) | 5Y-41.5 | NA | High risk (> 0) | 4Y-37.9 | 4Y-10.1 |
| | High risk (2–3) | 5Y-23.7 | NA | | | |
NCCN-IPI N = 1650 | Age, years | | | | | | |
| > 40, < 60 | Low risk (0–1) | 5Y-96.0 | 5Y-91.0 | Low risk (≤ 3) | 4Y-58.9 | 4Y-45.4 |
| > 60, ≤ 75 | | | | High risk (> 3) | 4Y-38.4 | 4Y-14.3 |
| > 75 | Low-intermediate (2–3) | 5Y-82.0 | 5Y-74.0 | | | |
| ECOG PS ≥ 2 | | | | | | |
| LDH | High-intermediate (4–5) | 5Y-64.0 | 5Y-51.0 | | | |
| > 1, ≤3 | High risk (≥ 6) | 5Y-33.0 | 5Y-50.0 | | | |
| > 3 | | | | | | |
| Stage III/IV | | | | | | |
| EN sites ≥ 1 | | | | | | |
PIT N = 385 | Age > 60 years | | | | | | |
| ECOG PS ≥ 2 | Low risk (0) | 5Y-58.9 | NA | Low risk (≤ 2) | 4Y-48.3 | 4Y-19.4 |
| LDH > ULN | Low-intermediate (1) | 5Y-45.6 | NA | High risk (> 2) | 4Y-40.8 | 4Y-24.0 |
| Bone marrow invasion | High-intermediate (2) | 5Y-39.7 | NA | | | |
| | High risk (3–4) | 5Y-18.3 | NA | | | |
Note: OS, Overall survival; PFS, progression-free survival; AITL, angioimmunoblastic T-cell lymphoma; IPI, International prognostic index; NCCN-IPI, National Comprehensive Cancer Network; PIAI, the alternative Prognostic Index for AITL; PIT, Prognostic Index for T-cell lymphoma Prognostic index for AITL; ATPI, AITL Prognostic Index; AITL score, angioimmunoblastic T-cell lymphoma prognostic index; Chinese AITL score, prognostic. index of angioimmunoblastic T-cell lymphoma in China; ECOG PS, Eastern Cooperative Oncology Group performance score; LDH, lactic dehydrogenase; WBC, white blood cell; Hb, hemoglobin; PLT, platelet; Ig, immunoglobulin; ULN, upper limit of normal; EN, extranodal site involvement; extranodal site involvement of bone marrow, central nervous system, lung, liver or gastrointestinal tract.; NA, not assessed. |