TAS-102 Followed By Regorafenib Identified As Optimal Treatment Sequence in Pretreated mCRC

Treatment with trifluridine/tipiracil (TAS-102; Lonsurf) prior to regorafenib (Stivarga) appears to provide greater benefit than treatment with regorafenib followed by TAS-102 in patients with metastatic colorectal cancer (mCRC) following progressed on 2 or more chemotherapy regimens, according updated findings from the phase 2 SOREGATT/PRODIGE 68 study (NCT04450836) presented at the at the2024 ESMO Gastrointestinal Cancers Congress.¹

Results showed that 47.9% of patients in the study analysis population (n = 234) were treated with 2 or more cycles of each treatment sequence; this comprised 40% of patients in arm A (TAS-102 followed by regorafenib; n = 115) and 55.5% in arm B (regorafenib followed by TAS-102; n = 119). Sixty percent of patients in arm A and 44.5% of those in arm B did not complete at least 2 cycles of each regimen.

“Despite the fact that there was a premature termination of this study due to the publication of data from [the phase 3] SUNLIGHT trial [NCT0437187] ...which was very good news for our patients…this study continued to provide important information,” presenting author Michel P. Ducreux, MD, PhD, stated in an oral presentation. “...There is indeed a sequence effect that is proven, and starting with TAS-102 was better for the patients.”

Ducreux is head of the Gastrointestinal Oncology Unit and Gastrointestinal Oncology Tumor Board at Gustave Roussy, as well as a professor of oncology at Paris-Saclay University in France.

Both regorafenib andtrifluridine/tipiracil are FDA approved for patients with mCRC who have progressed on 2 prior lines of chemotherapy, Ducreux explained. However, there is a lack of randomized data investigating the optimal sequencing of these agents. Accordingly, the phase 2 study evaluated the number of patients who can receive both treatment sequences and compare survival outcomes to determine the best regimen in this setting.

This multicenter, international, comparative, open-label trial enrolled patients aged 18 years and older with mCRC. Patients must have had an ECOG performance status (PS) of 0 or 1, and previously progressed on fluoropyrimidine-, oxaliplatin-, and irinotecan-based chemotherapy; a VEGF inhibitor; and an EGFR inhibitor, if they have RAS wild-type disease. Upon enrollment, patients were stratified by ECOG PS, age (65 years of age vs older than 65 years), and country.

Notably, investigators originally planned to screen 340 patients. However, screening was prematurely halted in light of results from the SUNLIGHT trial of third-line TAS-102 plus bevacizumab (Avastin).

Following screening, 234 patients were randomly assigned 1:1 to a parallel study group. In arm A, patients received regorafenib followed by TAS-102; this sequence was reversed for patients assigned to arm B. During cycle 1, regorafenib was administered at a 80 mg daily dose at week 1, 120 mg per day at week 2, and 160 mg at week 3, followed by 1 week off. For all subsequent treatment cycles, regorafenib was administered at the 160 mg dose. TAS-102 was administered orally at 35 mg/m2 twice daily on days 1 to 5, and 8 to 12 during each 4-week treatment cycle. Treatment with either regimen continued until disease progression or unacceptable toxicity.

The primary end point of the study was the feasibility of administering regorafenib followed by TAS-102, and vice versa, defined as the percentage of patients treated with 2 or more cycles of each treatment regimen. Secondary efficacy end points included overall survival (OS), progression-free survival (PFS) per RECIST v1.1, disease control rate (DCR), objective response rate (ORR), time-to-treatment failure (TTF), and time to deterioration (TTD). Safety and quality of life were additional secondary end points.

Of the patients randomly assigned to arms A and B, 112 and 119 were treated, respectively. Three patients did not complete treatment due to general state alterations, disease progression, and physician decision.

The median number of cycles received across arms was 4 (range, 1-17; interquartile range [IQR], 2-6). A median of 2 cycles were received by patients in phase 1 (range, 1-13; IQR, 2-3) and in phase 2 (range 1-14; IQR, 2-3), respectively. Reasons for ceasing treatment included progressive disease (85.7% in phase 1; 78.7% in phase 2), toxicities (10.8%; 16.8%), death (2.2%; 2.6%), investigator decision (1.3%; 0.6%), and patient decision (0.0%; 1.3%).

Reasons not to switch included impairment of patients’ general condition (37.3% overall; 22.7% in arm A; 58.1% in arm B), death (17.3%; 22.7%; 9.7%), progression (10.7%; 13.6%; 6.5%), patient decision (6.7%; 11.4%; 0.0%), toxicity (9.3%; 9.1%; 9.7%), expiration of 4-week delay (5.3%; 9.1%; 0.0%), investigator decision (6.7%; 6.8%; 6.5%), and other reasons (6.7%; 4.5%; 9.7%). Notably, this information was missing for 1 patient in arm B.

In the overall analysis population (n = 234), 58.5% of patients were male and 65.8% of patients had an ECOG PS of 1. The median age of patients was 67 years (range, 32, 86; IQR, 59-73), the median body mass index was 24 (range, 14-39; IQR, 21-28), and the median time of initial tumor diagnosis was 29.9 months (range 7-137.3; IQR, 18.6-49.8). The majority of patients had tumors in their left colon (40.6%), followed by the right colon (32.5%) and the rectum (26.9%).

Additional assessment of secondary end points showed that there was no statistically significant difference in median OS between the 2 arms. The median OS was 6 months in arm A vs 6.9 months in arm B (HR, 0.87; 95% CI, 0.67-1.14; P = .3). Benefit was shown with the treatment sequence in patients 65 years of age or older (HR, 0.87; 95% CI, 0.66-1.14; P = .3066) and in arm B (HR, 0.87; 95% CI, 0.66-1.14; P = 0.2954). Conversely, the HR was 1.00 for patients younger than 65, as well as for those in arm A. HRs were 1.00 and 1.69 (95% CI, 1.26-2.27; P = .0005) in patients with an ECOG PS of 0 vs 1, respectively.

“In terms of multivariate analysis for OS, the most important clinical factor, and we know this in our clinical practice, was the PS of the patients. The fact that they [had an] ECOG [PS of] 0 was a better prognostic factor,” Ducreux emphasized.

The median PFS was 1.87 months (95% CI, 1.84-1.94) in arm A and 1.97 months (95% CI, 1.84-2.10) in arm B, translating to an HR of 0.81 (95% CI, 0.62-1.05; P = .10). The 6- and 12- months PFS rates in arm A were 8.70% (95% CI, 4.79%-15.27%) and 1.74% (95% CI, 0.48%-6.12%); these respective rates were 10.08% (95% CI, 5.86%-16.80%) and 2.52% (95% CI, 0.86%-7.15%) in arm B.

During the presentation, Ducreux emphasized that “It is important to consider that these are true-life patients with a quite good PS. [However], this is, in terms of prognosis, a group of patients with a poor prognosis… [they have a] median OS of 6 to 7 months, and the median PFS was very limited.”

The time from randomization to progression on the 2nd sequence of treatment (PFS2) was 3.7 months (95% CI, 2.86-3.84) in arm A and 3.9 months (95% CI, 3.65-4.17) in arm B (HR, 0.75; 95% CI, 0.57-0.97; P = .03). The 6- and 12- months PFS rates in arm A were 16.07% (95% CI, 10.41%-23.98%) and 3.57% (95% CI, 1.40-8.82); these respective rates were 26.89% (95% CI, 19.74%-35.49%) and 5.88% (95% CI, 2.88%-11.65%) in arm B.

The median TTF in arms A and B were 3.19 months (95% CI, 3.27-3.45) and 3.71 months (95% CI, 3.52-4.01 months), respectively. The 6- and 12-month TTF rates were 12.17% (95% CI, 7.39%-19.40%) and 3.48% (95% CI, 1.36%-8.60%) in arm A; in arm B, these were 22.69% (95% CI, 16.09%-31.00%) and 5.04% (95% CI, 2.33%-10.56%). The median duration of treatment was 102 days (range, 1-648; IQR, 52-157) in the overall population, 94 (range, 7-516; IQR, 48.5-125) days in arm A and 111 days (range, 1-648; IQR, 74-178) in arm B.

Ducreux added that, “In the multivariate analysis, there was a clear difference in the independent value of [according to the treatment] arm and also…in terms of the ECOG PS of the patient.”

Regarding safety, all patients across both arms experienced an adverse effect (AE). In arms A and B, 68% and 64% of patients, respectively, experienced grade 3 or higher AEs. One or more grade 3 or higher treatment-related AEs (TRAEs) occurred in 38% and 37% of patients in arms A and B, respectively. At least 1 serious AE was reported in 5% and 7% of patients in these respective arms.Grade 3 or higher AEs observed included blood and lymphatic system disorders (11.71% in arm A; 24.37% in arm B), cardiac disorders (0.90%; 0.0%), gastrointestinal disorders (1.80%; 5.04%), general disorders (9.01; 6.72%), hepatobiliary disorders (0.90%; 0.84%), infections (0.90%; 1.68%), metabolism and nutrition disorders (6.31%; 0.0%), musculoskeletal and connective tissue disorders (0.90%; 0.0%), renal and urinary disorders (0.0%; 1.68%), respiratory, thoracic and mediastinal disorders (0.90%; 0.0%), skin and subcutaneous tissue disorders (11.71%;4.20%), and vascular disorders (2.70%; 1.68%).

“A translational analysis is underway to determine which patients responded to regorafenib,” Ducreux concluded.

Disclosures: Dr Ducreux discloses receiving institutional funding from Roche, Merck Serono, and Keocyt; participation on advisory boards for Roche, Merck Serono, Amgen, Abbvie, Sanofi, Bayer, Zymeworks, Servier, Beigene, MSD, Arcus, Astellas, Scandion, and Abcely; and serving as a speaker in symposiums for Roche, Merck Serono, Ipsen, Beigene, Amgen, Bayer, MSD, Servier, and Pierre Fabre.

References

1. Ducreux M, Ben Abdelghani M, Tougeron D, et al. PRODIGE 68 - UCGI 38 - SOREGATT: A randomized phase II study comparing the sequences of regorafenib (reg) and trifluridine/tipiracil (t/t) after failure of standard therapies in patients (pts) with metastatic colorectal cancer (mCRC). Presented at: 2024 ESMO Gastrointestinal Cancers Congress; June 26-29, 2024; Munich, Germany. Abstract 3O.
2. Sequences of REGorafenib and trifluridine/​tipiracil in patients with metastatic colorectal cancer (SOREGATT). ClinicalTrials.gov. Updated March 1, 2024. Accessed June 27, 2024. https://clinicaltrials.gov/study/NCT04450836