Addition of Brentuximab Vedotin to Gemcitabine in Relapsed or Refractory T-Cell Lymphoma: Final Analysis of a Lysa Multicenter, Phase II Study. “the TOTAL Trial”

Background : Outcome of relapse/refractory (R/R) peripheral T-cell lymphoma (PTCL) patients (pts) is usually poor. [Mak, 2013]. Brentuximab vedotin (Bv) monotherapy is approved for R/R anaplastic large cell-lymphoma (ALCL) with long term responders [Pro, 2017] and for 1^st line in association with polychemotherapy [Horwitz, 2019] in all CD30+ PTCL (with no restriction related to expression cut-off). We designed a phase 2 study for R/R CD30+ PTCL combining gemcitabine (G) and Bv with the aim to increase the overall response rate (ORR) by 15%, compared to gemcitabine monotherapy (35%) [O’Connor, 2019] and among secondary endpoints to evaluate BV maintenance and the impact of both CD30 expression and soluble CD30 (sCD30) on response and survival.

Patients and Methods: : Patients with confirmed CD30+ (≥5%) PTCL with measurable disease who failed or were refractory to 1-3 systemic therapy (excluding G and Bv) and ECOG performance status < 3 were eligible.

Pts received 4 (28-days) cycles of G and Bv induction phase (GBvIP) (G : 1000mg/m² at D1 and D15 plus Bv: 1.8 mg/kg at D8) followed, in pts with complete (CR) or partial remission (PR) and non-eligible for SCT, by up to 12 (21-days) cycles of Bv (1.8 mg/kg) maintenance phase (BvMP).

The primary endpoint was ORR (CR + PR) according to Lugano criteria (CT-scan-based). CD30 expression was semi-quantitatively assessed by immunohistochemistry (IHC) and serum sCD30 levels were determined by ELISA. (NCT03496779).

Results: From April 2018 to October 2019, 71 pts with PTCL (47 males, 24 females) were enrolled. Central pathology review according to 2017 WHO criteria found nodal TFH-PTCL (34 ; 47.9%) [including AITL (27 ; 38%) and other nodal PTCL-TFH (7 ; 9.9%)] ; ALK- ALCL (14 ; 27%) ; PTCL-NOS (9 ; 13%) ; ALK+ ALCL (5 ; 7%), EATL (2 ; 2.8%) and other entities (7 ; 9.9%). Median age was 66 years (20-79), 65 pts (91.6%) were stage III-IV. ECOG status was ≥ 2 in 45 pts (63%). The number of prior lines were 1 (57), 2 (11) or 3 (3) and 28 pts (39.4%) had a refractory disease.

In intention to treat analysis, the ORR after GBvIP was 46.5% including CR (14;19.7%), PR (19;26.8%). Among the 33 responders, 27 initiated BvMP as well as 1 non responding pt (with SD but a complete metabolic response). The median number of BvMP cycles was 9 (range 1-12). At least one G≥3 AE were recorded in 58/71 pts during GBvIP and 11/28 pts during BvMP. AE led to treatment discontinuation in 9/71 pts (GBvIP) and 14/28 pts (BvMP). Overall a polyneuropathy of G≥1 was recorded in 9/71 pts (GBvIP) and 14/28 pts (BvMP).

After a median follow-up of 26.5 (0.5-42.1) months (mo), the median progression free survival (PFS) and overall survival (OS) were respectively 4.5 mo (95%CI [3.5 - 10]) (Figure 1A) and 12.9 mo (95%CI [8.6 – 25.5]). Among the 33 pts in PR/CR after GBvIP, the duration of response since inclusion (DOR) was 15.8 mo (95%CI [10.4 – (-)])

By univariate analysis, ORR was negatively influenced by high LDH (p=0.008) with borderline impact of ECOG (p=0.058), and both PFS and OS were reduced by refractory disease status at screening (p=0.031 and p=0.007) and non-ALCL histology (p=0.049 and p=0.041).

CD30 expression was investigated on tumor samples collected either at screening (55.5%) or (if not available) at diagnosis (39,7%) or at previous relapse (4.8%). In pts with ALCL, CD30 expression was constant and high (>75% of tumor cells) along with a significant higher level of sCD30 compared to non ALCL pts. In non ALCL pts, the median % of CD30-positive tumor cells was 12.5% [0 – 12.5]. Focusing on non-ALCL pts we found a high impact of sCD30 level >120 ng/mL on both ORR, DOR, PFS and OS while CD30 expression on tumor cells did not appear to influence the results. (Figure 1B ; Table 1)

Conclusion : According to our initial hypothesis, the addition of Bv to G is active in R/R CD30+ PTCLs both in terms of ORR, CR. In responding pts, the DOR exceeding 15 mo with very long responders suggests the value of Bv maintenance. Not surprisingly ALCL were associated with better results, but this treatment is also effective in non ALCL pts including a significant proportion of TFH lymphoma pts.

Besides ALCL pts, biomarkers are needed to predict which pts could benefit the most of Bv based therapy. In line with a previous report [Horwitz, 2014], we did not find a clear impact of CD30 expression on tumor cells, however we demonstrate that baseline serum sCD30 at treatment initiation was strongly correlated with both response and outcome.