
==== Front
ESMO Open
ESMO Open
ESMO Open
2059-7029
Elsevier

S2059-7029(24)01472-8
10.1016/j.esmoop.2024.103703
103703
Original Research
mFOLFOX6 versus mFOLFOX6 + aflibercept as neoadjuvant treatment in MRI-defined T3-rectal cancer: a randomized phase-II-trial of the German Rectal Cancer Study Group (CAO/ARO/AIO 0214)
Hofheinz R.-D. ralf.hofheinz@umm.de
1∗
Herrle F. 23
Dechow T. 4
von Weikersthal L.F. 5
Welslau M. 6
Lettmaier S. 7
Burkart C. 8
Kubicka S. 9
Kochen L. 10
Merx K. 1
Krause K. 11
Ebert M. 121314
Rödel C. 151617
Fokas E. 18
Ghadimi M. 19
Reissfelder C. 2
Gaiser T. 20
1 Universitätsmedizin Mannheim, Mannheim Cancer Center, Mannheim
2 Universitätsmedizin Mannheim, Chirurgische Klinik, Mannheim
3 RoMed Klinik, Allgemein- und Viszeralchirurgie, Prien am Chiemsee
4 Hämatologisch-Onkologische Praxis, Ravensburg
5 Gesundheitszentrum St. Marien GmbH, Amberg
6 Hämato-Onkologischer Studienkreis am Klinikum Aschaffenburg, Aschaffenburg
7 Universitätsklinikum Erlangen, Strahlenklinik, Erlangen
8 MVZ Villingen, Villingen-Schwenningen
9 Kreiskliniken Reutlingen GmbH, Klinikum am Steinenberg, Reutlingen
10 Frankfurter Institut für Klinische Krebsforschung IKF GmbH, Frankfurt/Main
11 AIO Studien gGmbH, Berlin
12 Universitätskmedizin Mannheim, II. Medizinische Klinik, Mannheim
13 DKFZ-Hector Krebsinstitut an der Universitätsmedizin Mannheim, Mannheim
14 Molecular Medicine Partnership Unit, EMBL, Heidelberg
15 Universitätsklinikum Frankfurt, Klinik für Strahlentherapie und Onkologie, Frankfurt am Main
16 German Cancer Consortium (DKTK), Frankfurt am Main
17 Goethe Universität Frankfurt, Frankfurt Cancer Institute (FCI), Frankfurt am Main
18 Universitätsklinikum Köln, Klinik und Poliklinik für Radioonkologie, Cyberknife und Strahlentherapie, Center for Integrated Oncology Aachen Bonn Cologne Duesseldorf (CIO ABCD), Cologne
19 Universitätsklinikum Göttingen, Klinik für Allgemein-, Viszeral- und Kinderchirurgie, Göttingen
20 Institut für Pathologie, Speyer, Germany
∗ Correspondence to: Prof. Ralf-Dieter Hofheinz, TagesTherapieZentrum, Mannheim Cancer Center, Universitätsmedizin Mannheim, Universität Heidelberg, Theodor-Kutzer Ufer 1-3, 68167 Mannheim, Germany. Tel: +49-621-383-2855 ralf.hofheinz@umm.de
10 9 2024
9 2024
10 9 2024
9 9 103703© 2024 The Author(s)
2024
https://creativecommons.org/licenses/by-nc-nd/4.0/ This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Background

Neoadjuvant chemotherapy is an option for patients with locally advanced rectal cancer at low risk for local recurrence. This randomized phase II trial investigated whether the addition of aflibercept to modified FOLFOX6 (mFOLFOX6) could improve the rates of centrally confirmed pathological complete remissions (pCR) and (disease-free) survival in magnetic resonance imaging (MRI)-staged cT3 rectal cancer.

Patients and methods

Patients with rectal cancer fulfilling the following criteria were included: lower border of tumor >5 cm and <16 cm from anal verge; circumferential resection margin >2 mm and T3-tumor with a maximum infiltration of 10 mm, as determined by MRI. Patients were randomized 1 : 2 to six cycles mFOLFOX6 ± aflibercept. Surgery was scheduled 4 weeks after chemotherapy. Primary endpoint was the rate of centrally confirmed pCR. The study was designed to detect an improvement of pCR from 10% to 27% (power 80%, type I error 20%).

Results

A total of 119 randomized patients started treatment (39 patients mFOLFOX6, arm A, and 80 mFOLFOX + aflibercept, arm B). The incidence of all grade adverse events was similar in both arms, however, adverse events grade ≥3 were more than twice as high in the experimental arm due to hypertension. Surgical complications were comparable. Aflibercept did not improve the pCR rate (arm A 26% versus arm B 19%, P = 0.47) and more patients in arm B had node positivity. With a median follow-up of 40.1 months, the 4-year disease-free survival was 83% in arm A and 85% in arm B (P = 0.82). Only two patients in arm A and one patient in arm B developed local recurrence.

Conclusions

In patients with locally advanced rectal cancer and MRI-defined low risk of local recurrence, neoadjuvant mFOLFOX6 + aflibercept was feasible and did not compromise surgery. Survival data were favorable in both arms, but pCR rates were not increased by the addition of aflibercept.

Highlights

• Neoadjuvant chemotherapy was non-inferior to chemoradiotherapy in selected patients with locally advanced rectal cancer.

• First randomized trial to investigate the addition of angiogenesis inhibitor to neoadjuvant mFOLFOX in locally advanced rectal cancer.

• FOLFOX6 + aflibercept was feasible and safe also in terms of surgery. pCR rates and (disease-free) survival were not improved.

Key words

aflibercept
neoadjuvant chemotherapy
rectal cancer
==== Body
pmcIntroduction

Treatment strategies for locally advanced rectal cancer (LARC) are becoming increasingly complex. The generally recommended treatment of locally advanced, non-metastatic tumors (defined as cT3-4 and/or cN+ tumors) has been preoperative short-course radiotherapy or chemoradiotherapy (CRT) followed by total mesorectal excision (TME) for many years.1 Aims of preoperative treatments are reduction of risk of local recurrence and distant metastases, improvement of resectability to enable R0 resection, preservation of sphincter function in tumors located in the lower third, and avoidance of permanent stoma. The treatment strategy is determined based on staging using rigid rectoscopy, endosonography, whole-body computed tomography (CT), and quality-assured magnetic resonance imaging (MRI) of the pelvis.

Given the long-term toxicity of radiotherapy and the ability to define patients at low risk for local relapse using MRI radiotherapy-free neoadjuvant combination chemotherapy has been under investigation. The PROSPECT phase-III study demonstrated non-inferiority of neoadjuvant chemotherapy (nCT) using six cycles of FOLFOX compared with neoadjuvant 5-fluorouracil (FU)-based CRT in patients with cT2N1 and T3 N0/1 tumors with free circumferential resection margin (CRM) and ability for sphincter-sparing surgery with respect to the primary endpoint, disease-free survival (DFS).2 Thus, nCT represents a new treatment option for these types of tumors.

The addition of monoclonal antibodies to nCT for LARC patients may be a strategy to further optimize treatment results. In a pilot phase II study, Schrag and coworkers3 investigated the addition of bevacizumab to FOLFOX without CRT. A pathological complete remission (pCR) rate of 25% and an R0 resection rate of 100% was achieved. Thus, data of this trial indicated that adding an anti-angiogenic drug to FOLFOX chemotherapy may improve efficacy (as measured by the rate of pCR).

Aflibercept is an anti-angiogenic agent with a broader mode of action, acting as a soluble receptor that binds to human vascular endothelial growth factor A (VEGF-A), VEGF-B, and the placental growth factor. It is licensed for the treatment of metastatic colorectal cancer.

In this randomized phase-II trial, we sought to determine whether the addition of aflibercept improves the efficacy (regarding the pCR rate) of neoadjuvant FOLFOX-based chemotherapy in patients with LARC staged cT3 CRM-negative with MRI.

Patients and methods

Study design

The trial (CAO/ARO/AIO 0214) was a phase II, open label, randomized, investigator-initiated multicenter study of neoadjuvant modified FOLFOX6 (mFOLFOX6) ± aflibercept, conducted by the German Rectal Cancer Study Group in Germany and Austria. Trial registration: NCT03043729 (date of initial registration: 6 February 2017); EudraCT-No.: 2015-002773-38 (date of initial registration: 2 August 2016). Eligible patients were adults with previously untreated locally advanced rectal or rectosigmoid adenocarcinoma staged T3 with negative MRI-predicted CRMs who were candidates for sphincter-sparing surgical resection.

All participants had a baseline staging that included a CT scan of the chest and abdomen to rule out metastatic disease. The colorectal surgeon carried out a baseline rigid rectoscopy and identified that the tumor was amenable to sphincter-preserving TME and had a distal edge located between 5 and 16 cm of the anal verge. Patients had staging with endoscopic ultrasound, as well as a pelvic MRI to estimate tumor size and the extent of nodal involvement. Only T3-tumors not causing symptomatic bowel obstruction were included irrespective of nodal status (N0 or N+). Tumor infiltration into perirectal fat had to be <10 mm, provided CRM was >2 mm. Other key inclusion criteria were: Eastern Cooperative Oncology Group (ECOG) status 0-1, as well as adequate bone marrow, hepatic, renal, and metabolic function. Patients were excluded if they had distant metastasis/CNS metastasis, hypercalcemia, a pre-existing permanent neuropathy ≥grade 2 [according to National Cancer Institute Common Terminology Criteria for Adverse Events (NCI-CTCAE) v.4.03], as well as an uncontrolled hypertension (defined as systolic blood pressure >150 mmHg and/or diastolic blood pressure >100 mmHg), or history of hypertensive crisis, or hypertensive encephalopathy.

Patients were randomized 1 : 2 in arm A or arm B. The actual analyzed sample size was 119 (39 patients in arm A; 80 patients in arm B). All patients gave written informed consent, and the study was conducted in accordance with the International Conference on Harmonization Good Clinical Practice (ICH-GCP) standards and with the Declaration of Helsinki, as well as the German drug law and the German GCP regulation.

Objectives and outcome measures

Primary objective of the study was to investigate the efficacy of the mFOLFOX6/aflibercept combination compared with mFOLFOX6 alone as neoadjuvant treatment in terms of the centrally evaluated objective pathological tumor. The pCR (ypT0N0) was assessed in a standardized manner by a central pathology using Dworak regression grading.4

Secondary objectives were to assess efficacy and safety of the mFOLFOX6/aflibercept combination compared with mFOLFOX6 alone using the endpoints DFS, overall survival (OS), and relapse-free survival (RFS), surgical morbidity, and 28-day-mortality, rate of R0-wide, R0-narrow R1 and locoregional R2 resection, as well as tumor downstaging and downsizing.

Randomization and treatment

Eligible patients were randomized 1 : 2 in arm A or arm B by using permuted block randomization. During the treatment phase, patients in arm A received six cycles of mFOLFOX6 alone [oxaliplatin (85 mg/m2, 2 h i.v.), leucovorin (400 mg/m2, 2 h i.v.), 5-FU (400 mg/m2 as bolus, followed by 2400 mg/m2 i.v. over 46 h)]. Patients in arm B received five cycles of aflibercept (4 mg/kg, 1 h i.v.) before mFOLFOX6 and in the sixth preoperative cycle mFOLFOX6 alone. All drugs were given on day 1 of a 14-day cycle.

TME surgery was scheduled 4 weeks (21 to 42 days) after the sixth cycle of chemotherapy. Tumor assessment with MRI was carried out at baseline and 3-4 weeks after the last administration of chemotherapy to rule out progression of the primary tumor and to enable salvage treatment (generally CRT) and to plan for surgery. The necessity of salvage treatment was judged within an interdisciplinary tumor board.

Toxicity was assessed continuously during the study from date of informed consent until date of end of treatment 4 weeks after surgery (or until earlier termination) according to NCI-CTCAE version 4.03. Dose reductions and treatment interruptions were allowed for the management of adverse events (AEs) deemed to be related to study treatment. Aflibercept could be reduced to 2 mg/kg in case of an adverse reaction and was discontinued in patients requiring more than one dose reduction of aflibercept or more than two dose delays due to toxicity.

Patients were followed for survival and subsequent therapies for at least 12 but no more than 36 months after last patient last treatment or until death, loss to follow-up or withdrawal of consent, whichever occurred first.

Sample size and analysis population

It was assumed that the proportion for pCR in arm A (mFOLFOX6) is 10%. The sample size is calculated such that a difference of absolute 17% (therefore pCR in arm B 27%) could be detected with a type I error rate of 20% and a power of 80% using a two-sided Fisher’s exact test. This pCR rate of 27% was considered as reasonable according to the results of the phase II study conducted by Schrag et al.3 Based on these assumptions and using a randomization ratio of 1 : 2, and assuming a dropout rate of 5%, the sample size was calculated to be 119 (arm A: 40; arm B: 79). At the time of the statistical design of our study, we were not aware of any larger prospective studies for FOLFOX as nCT. We therefore set the expected rate of pCR at 10%, which corresponds approximately to the percentage of pCR expected for standard radiochemotherapy for rectal cancer.

Analysis of the primary endpoint was based on the full analysis set (FAS), consisting of all randomized patients who received at least one complete cycle of chemotherapy, and on the per-protocol population, comprising all patients of the FAS who additionally underwent surgery and for whom a central pathologic evaluation is available. Patients were analyzed according to the study arm they were assigned to at randomization.

Statistical analysis

The analysis based on the FAS population was regarded as the primary hypothesis test; all other analyses were considered as exploratory. The primary endpoint, the percentage of patients with pCR, was analyzed using a two-sided Fisher’s exact test at a 20% significance level.

DFS was measured from the date of arm assignment until the date of local recurrence or distant metastasis or death. Similarly, OS was measured from the date of arm assignment until death of any cause. RFS in resected patients was measured as the length of time after completion of primary treatment (nCT + surgery) until documented relapse (i.e. local relapse, liver metastasis, systemic metastases). Incomplete time-to-event observations were handled as censored measurements. OS, DFS, and RFS were estimated by using the Kaplan–Meier method and were compared between groups using a two-sided log-rank test, for which P < 0.05 was considered statistically significant.

Safety analysis was carried out for all patients who received at least one dose of study medication and according to the treatment received.

Data cut-off for analysis was 7 November 2023. Statistical calculations were carried out using SAS software, version 9.4 or higher (SAS Institute, Cary, NC), or R, version 3.6.1 or higher (The R Foundation for Statistical Computing, Vienna, Austria).

Results

Patients

Between July 2017 and January 2021, a total of 120 patients at 23 study sites were randomly assigned to arm A (mFOLFOX6; n = 39) or arm B (mFOLFOX6 plus aflibercept; n = 81). After randomization, one patient in arm B was excluded due to ineligibility and therefore did not receive any study treatment. The remaining patients in arm A (n = 39) and arm B (n = 80) were included in the FAS (Figure 1). Their baseline characteristics were well-balanced (Table 1).Figure 1 CONSORT diagram. mFOLFOX6, infusional modified 5-fluorouracil, leucovorin, and oxaliplatin.

Table 1 Baseline demographic and clinical characteristics of the FAS

Parameter	Arm A (mFOLFOX6) n = 39	Arm B (mFOLFOX6 + aflibercept) n = 80	
Age (years)	
 Median (range)	64 (33-83)	63 (33-80)	
Sex, n (%)	
 Male	31 (80)	57 (71)	
 Female	8 (21)	23 (29)	
ECOG performance status, n (%)	
 0	30 (77)	73 (91)	
 1	9 (23)	7 (9)	
Clinical T category, n (%)	
 cT2	0	2 (3)	
 cT3	39 (100)	77 (96)	
 cT4	0	1 (1)	
If cT3, infiltration into perirectal fat, n (%)	
 T3 ≤ 5 mm	21 (54)	40 (50)	
 T3 > 5 mm	12 (31)	23 (29)	
 Unknown	6 (15)	14 (17)	
Clinical N status, n (%)	
 cN0	10 (26)	20 (25)	
 cN+	29 (74)	60 (75)	
Distance from anal verge (cm)	
 Median (range)	8 (3-14)	9 (1.5-15)	
Circumferential resection margin (CRM) status, n (%)	
 >2 mm	37 (95)	78 (98)	
 ≤2 mm	2 (5)	0	
 Unknown	0	2 (2)	
ECOG, Eastern Cooperative Oncology Group; FAS, full analysis set; mFOLFOX6, infusional modified 5-fluorouracil, leucovorin, and oxaliplatin.

Treatment adherence and efficacy

All patients (100%) in arm A and 73 of 80 patients (91%) in arm B completed all six planned cycles of mFOLFOX6, while 61 patients (76%) in arm B received all five scheduled aflibercept administrations. The median number of oxaliplatin and 5-FU cycles was six for both arms (arm A: range 5-6 and arm B: range 2-6), and the median number of aflibercept cycles in arm B was five (range 2-5). The median duration of treatment was 72 days in arm A (range 69-99 days) and 71 days in arm B (range 15-99 days). Dose modifications occurred at similar rates for oxaliplatin and 5-FU in both treatment arms, while 21% of patients had a dose modification for aflibercept (Supplementary Table S1, available at https://doi.org/10.1016/j.esmoop.2024.103703). The main reason for dose modifications was toxicity.

Following neoadjuvant treatment, one patient in arm A withdrew consent before surgery, resulting in 38 of 39 patients who underwent surgery in this arm. In arm B, 73 of 80 treated patients underwent surgery within the study (Table 2). Two patients in arm B did not undergo surgery: one due to death before and one because no residual tumor was detected after neoadjuvant therapy and the patient denied surgery. Five other patients in arm B received surgery at other centers (Figure 1).Table 2 Surgical morbidity and mortality of the FAS

Parameter	Arm A (mFOLFOX6) n = 39	Arm B (mFOLFOX6 + aflibercept) n = 80	
Time since randomization until surgery (months)	
 Median (range)	3.7 (3.1-5.4)	3.8 (3.1-4.8)	
Resection carried out, n (%)	38 (97)	73 (91)	
Mesorectal excision, n (% of resected)	
 Total mesorectal excision (TME)	38 (100)	69 (95)	
 Partial mesorectal excision (PME)	0	4 (5)	
Type of surgery, n (% of resected)	
 Open	11 (29)	24 (33)	
 Laparoscopic	22 (58)	38 (52)	
 Robotic	5 (13)	11 (15)	
Sphincter preservation, n (% of resected)	37 (97)	72 (99)	
Completeness of resection, n (% of resected)	
 R1	0	1 (1)	
 R0	38 (100)	71 (97)	
 R0-wide	37 (97)	66 (90)	
 R0-narrow	0	3 (4)	
 Unknown	1 (3)	2 (3)	
Postoperative hospitalization (days)	
 Median (range)	9 (1-35)	9 (1-53)	
Postoperative complications, n (% of resected)	6 (16)	16 (22)	
Clavien-Dindo Classification, n (% of resected)	
 Grade I	1 (3)	6 (8)	
 Grade II	1 (3)	1 (1)	
 Grade IIIa	3 (8)	3 (4)	
 Grade IIIb	0	4 (5)	
 Grade IVa	0	1 (1)	
 Unknown	1 (3)	1 (1)	
Local complications, n (% of resected)	
 Abscess	0	1 (1)	
 Abscess with anastomotic leak	1 (3)	0	
 Anastomotic leak	2 (5)	4 (5)	
 Gastrointestinal perforation	0	1 (1)	
 Hemorrhage	0	1 (1)	
 High output stoma	0	1 (1)	
 Ileus (paralytic)	2 (5)	4 (5)	
 Stoma complication	0	1 (1)	
 Re-surgery	0	3 (4)	
30-day postoperative mortality, n (% of resected)	0	2 (3)	
Local recurrence, n (%)	2 (5)	1 (1)	
Distant metastasis, n (%)	3 (8)	7 (9)	
FAS, full analysis set; mFOLFOX6, infusional modified 5-fluorouracil, leucovorin, and oxaliplatin.

MRI evaluation before surgery was available for 102 of 119 patients. Of these, downsizing to ymrT0 or 1 was documented in 10.3% (arm A) and 11.3% (arm B) of patients. No clear differences were also seen when downsizing to ymrT0-2 was considered (arm A 35.9% versus arm B 38.8%).

All resected patients in arm A (100%) underwent a TME whereas 5% of resected patients in arm B received a partial mesorectal excision. Resection was pathologically complete (R0) in 100% and 97% of resected patients in arms A and B, respectively. Of these, 97% (arm A) and 90% (arm B) obtained an R0-wide resection (Tables 2 and 3).Table 3 Pathological characteristics of the FAS

Parameter	Arm A (mFOLFOX6) n = 39	Arm B (mFOLFOX6 + Aflibercept) n = 80	
Pathological T category, n (%)	
 ypT0	9 (23)	16 (20)	
 ypT1	0	10 (13)	
 ypT2	12 (31)	17 (21)	
 ypT3	16 (41)	33 (41)	
 ypT4	1 (3)	1 (1)	
 Unknown	1 (3)	3 (4)	
Pathological N category, n (%)	
 ypN0	30 (77)	53 (66)	
 ypN1	7 (18)	16 (20)	
 ypN2	1 (3)	8 (10)	
 Unknown	1 (3)	3 (4)	
Pathological M category, n (%)	
 ypM0	32 (82)	58 (73)	
 ypM1	0	1 (1)	
 Unknown	7 (18)	21 (26)	
Pathological complete response (= Dworak grade 4), n (%)	10 (26)	15 (19)	
 95% CI	13-43	11-30	
Dworak regression grade, n (%)	
 0	5 (13)	4 (5)	
 1	13 (33)	16 (20)	
 2	9 (23)	27 (34)	
 3	1 (3)	8 (10)	
 Unknown	1 (3)	10 (13)	
CI, confidence interval; FAS, full analysis set; mFOLFOX6, infusional modified 5-fluorouracil, leucovorin, and oxaliplatin.

Table 3 shows the pathological findings of both arms. In arm A, 26% of patients [95% confidence interval (CI) 13% to 43%] in both the FAS and the per-protocol population achieved a pCR compared with 19% (95% CI 11% to 30%) and 22% in arm B, respectively. This difference in the pCR rate was not statistically significant (P = 0.47). Within the FAS, 26% in arm A and 44% in arm B had Dworak grade 2-3 regression, while 46% of the patients in arm A and 25% in arm B showed no regression (Dworak grade 0) or still had dominant tumor mass (Dworak grade 1). Node positivity (ypN1/2) upon surgery was found to be higher in arm B (30% versus 21%) and one patient in arm B, but none in arm A, showed perioperative distant metastases (ypM1).

Median duration of hospitalization after surgery was 9 days in both arms. Some 16% of resected patients in arm A and 22% of those in arm B suffered from postoperative complications. Of these, 6% and 9%, respectively, had minor surgical complications (grade I and II according to Clavien–Dindo classification), while 8% and 9%, respectively, experienced more severe grade III complications and one patient in arm B, but none in arm A, had a life-threatening grade IV complication. Most frequent local complications were anastomotic leak and paralytic ileus affecting 5% of patients in each treatment arm. Three re-surgeries in arm B were carried out due to these postoperative local complications (anastomotic leak, gastrointestinal perforation, ileus). Two patients in arm B, but none in arm A, died within 30 days after surgery (Table 2).

The overall median follow-up time was 40.1 months with 38.5 months (range 6.6-49.1 months) in arm A and 40.1 months (range 3.6-56.2 months) in arm B. The 3-year DFS rate was 83% (95% CI 70% to 95%) in arm A and 87% (95% CI 80% to 95%) in arm B and at 4 years 83% (95% CI 70% to 95%) and 85% (95% CI 77% to 93%), respectively (P = 0.82). RFS rate at 3 years was 85% (95% CI 73% to 97%) in arm A and 88% (95% CI 80% to 96%) in arm B (P = 0.70). During follow-up, local recurrence was reported in two patients (5%) in arm A and one patient (1%) in arm B, while distant metastases occurred in three patients (8%) in arm A and seven patients (9%) in arm B. The OS rate at 3 years was 91% (95% CI 81% to 101%) in arm A and 94% (95% CI 88% to 99%) in arm B (P = 0.72) (Figure 2).Figure 2 Disease-free survival, overall survival, and relapse-free survival in the FAS population according to treatment. FAS, full analysis set.

Safety

Each patient in both study arms experienced at least one AE irrespective of relatedness to study treatment (Supplementary Table S2, available at https://doi.org/10.1016/j.esmoop.2024.103703). Most frequent AEs (with an incidence of >20%) in both arms of any grade and irrespective of a potential treatment-relatedness were nausea (54% arm A and 49% arm B), peripheral sensory neuropathy (46% arm A and arm B), fatigue (44% arm A and 43% arm B), oral mucositis (26% arm A and 34% arm B), and diarrhea (26% arm A and 25% arm B) as well as paresthesia in arm A (23%) and vomiting (24%) and pain (21%) in arm B, all mostly at severity grades 1 and 2. AEs classified as treatment related were observed in 35 of 39 patients (90%) in arm A and 79 of 80 patients (99%) in arm B. The incidence of treatment-related AEs of CTCAE severity grade ≥3 was substantially higher in arm B, affecting 55% of patients compared with 18% of patients in arm A. Notably, treatment-emergent hypertension grade 3 was observed in 33% of patients in arm B, whereas it was only observed in one patient (3%) with grade 3 in arm A, reported, however, as unrelated to study treatment. The most common treatment-related AE with grade 3-4 in both study arms was neutropenia occurring in ∼10% of patients in each treatment arm.

Overall, AEs leading to mFOLFOX6 dose reductions were observed in 18% and 19% of patients in arm A and arm B, respectively. An interruption of chemotherapy administration occurred in 26% of patients in arm A compared with 35% of patients in arm B, however, whereas a permanent discontinuation was reported in 8% and 9% of patients, respectively (Supplementary Table S2, available at https://doi.org/10.1016/j.esmoop.2024.103703).

Discussion

The rectal cancer treatment landscape has undergone substantial changes within the past years including, for instance, the introduction of total neoadjuvant therapy (TNT) to improve DFS as well as concepts to increase the chance for organ preservation (e.g. using escalation of radiotherapy doses, TNT, or local excision instead of radical surgery). Recently, the possibility of obviating radiotherapy using nCT was established as another treatment option in LARC with low risk for local recurrence by randomized trials.

Firstly, the PROSPECT trial compared nCT with FOLFOX and only selective use of CRT with neoadjuvant CRT in patients with cT2 N1, cT3 N0/1, and mesorectal fascia (MRF)-negative (>3 mm) LARC not threatening the rectal sphincter. In this trial including 1194 patients, local recurrence rates were <2% in each arm and comparable 5-year DFS (81% versus 79%) and OS (89.5% versus 90.2%) rates were reported.5 Secondly, final data from the CONVERT phase III trial including 589 patients from Chinese centers were presented at the ESMO 2023 meeting. This study compared nCT using 3 months of capecitabine-oxaliplatin (CAPOX) with capecitabine-based CRT in LARC <12 cm from anal verge stage II/III with uninvolved MRF.6 Although the primary endpoint, non-inferiority in terms of 3-year locoregional recurrence-free survival (LRRFS) was formally missed [CRT 97.4% versus nCT 96.3%, hazard ratio (HR): 1.08] the LRRFS Kaplan–Meier curves virtually overlap and 3-year DFS (CRT 87.9% versus nCT 89.2%; HR: 0.88) and OS (CRT 94.1% versus 95.0%; HR: 0.86) are comparable. Thus, data from the CONVERT trial confirm the results of the PROSPECT trial although patients with a higher local recurrence risk were included in CONVERT (cT4 tumors 27%, location in the lower third of the rectum 41%, extramural venous invasion-positive tumors 18%). Long-term toxicity was in favor of the nCT approach (grades 2-4: CRT 31.1% versus nCT 19.0%). Finally, the FOWARC trial compared nCT with FOLFOX and two CRT arms (one arm 5-FU- and one FOLFOX-based, respectively) including a total of 495 patients with a high-risk profile (cT4 30.9%, CRM-positivity 21%, lower third 49.1%).7 Although the primary endpoint (superiority of neoadjuvant oxaliplatin-based treatment in terms of DFS) was missed, long-term results as shown at the ASCO 2023 meeting found comparable 10-year locoregional recurrence (nCT 9.6 versus 5-FU-based CRT 10.8%) DFS (nCT 60.5% versus 5-FU CRT 52.5%) and OS (nCT 74.4% versus 65.9%).8 Thus, even in these high-risk patients, perioperative chemotherapy may render CRT unnecessary in defined subgroups. Cumulating evidence from three randomized trials establishes nCT as a well-tolerated alternative to CRT for patients with LARC at low risk for local failure, at least if organ preservation by selected omission of TME is not attempted.

In patients with LARC and MRI-defined low risk of local recurrence, the addition of aflibercept to nCT with mFOLFOX6, as carried out in the current trial, was feasible. A substantially higher rate of ≥grade 3 AEs was, however, observed in the experimental arm (55% versus 18%). This elevated incidence of higher-grade AEs in the mFOLFOX6 + aflibercept arm, was attributable to aflibercept-related hypertension, a well-known toxicity of the anti-angiogenic agent.9 Accordingly, the addition of aflibercept to mFOLFOX6 was shown to result in dose compromises compared with chemotherapy alone, with 35% of patients having an interruption of chemotherapy administration in the experimental arm versus 26% of patients in the mFOLFOX6 arm. In general, the neoadjuvant mFOLFOX6/aflibercept combination did not compromise surgery as R0 resection rates, duration of hospitalization, Clavien–Dindo complication rates, and also local complications such as anastomotic leakage rates were shown to be comparable to neoadjuvant mFOLFOX6 alone. Notably, two patients in the experimental arm died within 30 days after surgery, however, without being related to study treatment. Of note, while a comparable percentage of patients had clinical evidence of nodal involvement (cN+: 74% versus 75%) at baseline, higher numbers of node-positive tumors were observed in the mFOLFOX6/aflibercept group following surgery compared with mFOLFOX6 alone (ypN1+N2: 30% versus 21%).

In terms of the primary endpoint, the addition of aflibercept to neoadjuvant mFOLFOX6 failed to improve the pCR rate compared with chemotherapy alone (19% versus 26%). There were also no differences (29% versus 29%) between the arms in the Dworak grade 3 (near pCR) and 4 (pCR) regression grades, which are generally referred to as ‘good response’ when added together. When our study was designed, the pCR rates for neoadjuvant FOLFOX were not known from larger studies, so we set the pCR for the standard arm at 10% in analogy to what could be expected with neoadjuvant CRT. In the meantime, the data from PROSPECT have clearly shown that pCR rates can be much higher when using FOLFOX (21.9%), so the expectations we had in our study design regarding the improvement in pCR were too optimistic.5

Similar to the pathological results, there were no relevant differences in MRI re-staging between the arms. Taking into account the preoperative MRI results, a total of 10.3% (arm A) and 11.3% (arm B) of patients had ymrT0-1 status (ymrT0-2: 35.9% versus 38.8%).

Smaller previous phase II trials had explored nCT in combination with anti-angiogenic drugs for LARC. Data from a phase II pilot trial by Schrag et al.3 indicated that preoperative FOLFOX in combination with bevacizumab led to a high pCR rate of 25% in MRI-defined stage II to III patients with rectal cancer excluding cT4, MRF-positive or unresectable tumors.3 Another phase II study, however, exploring nCT with mFOLFOX6 + bevacizumab in stage II to III LARC patients who had KRAS-mutant tumors revealed a lower pCR rate of 15%.10 Considerably, 15% of patients in the latter cohort were at high risk of having a CRM-positive tumor. In line with that, a single-arm study with MRI-defined high-risk rectal cancer patients who received neoadjuvant CAPOX in combination with bevacizumab followed by TME surgery showed a median pCR rate of 13.3%.11 Thus, the range of pCR rates in these phase II studies using nCT is most likely explained by different inclusion criteria favoring trials with inclusion of less advanced tumors and underscoring the need for randomized studies such as the current trial. Using aflibercept as part of a TNT strategy in combination with a FOLFOX induction chemotherapy, however, led to an increased pCR rate in mrT3c-d or T4 and/or N2 tumors from 13.8% to 22.6% in the randomized phase II GEMCAD 1402 trial.12

In terms of DFS rates, our data compare adequately with PROSPECT and CONVERT. No DFS benefit by using aflibercept, however, was seen in our trial. Likewise, the GEMCAD 1402 study reported comparable 3-year DFS rates between arms (experimental arm with aflibercept 75.2% versus standard arm 81.5%; HR: 1.21) after a median follow-up of 38 months.13

In summary, the addition of aflibercept to nCT in the current trial neither improved pCR rates nor long-term efficacy. Thus, the addition of anti-angiogenic agents to nCT cannot be recommended. Furthermore, given the unfavorable long-term results of the GEMCAD 1402 study, it is unlikely that antiangiogenic drugs will be pursued in the context of TNT strategies. Moreover, the OCUM phase II study has recently reported very low recurrence rates of primary surgery in patients selected by quality-assured MRI with good TME surgery.14 Therefore, in the context of optimal MRI and surgical quality, the potential benefit of nCT in DFS must ultimately be defined in a randomized trial rather than intensifying neoadjuvant treatment with the addition of monoclonal antibodies or other drugs. The ongoing ACO/ARO/AIO 18.2 randomized phase III trial of the GRCSG (ClinicalTrials.gov NCT04495088) therefore compares primary TME surgery followed by stage-appropriate adjuvant chemotherapy with 3 months of neoadjuvant FOLFOX/CAPOX using the same inclusion criteria as in the current trial.

Supplementary data

Supplementary Appendix

Acknowledgements

This study was sponsored by AIO-Studien-gGmbH. The authors thank AIO Studien gGmbH and cermed GmbH for supporting medical writing.

Funding

This work was supported by a grant from Sanofi-Aventis (no grant number).

Disclosure

RDH: consulting or advisory role: AbbVie, Amgen, AstraZeneca, Bayer, Bristol Myers Squibb (BMS), Boehringer, Daiichi, GSK, Leo, Lilly, Merck, Merck Sharp & Dohme (MSD), Nordic, Onkowissen, Pierre Fabre, Roche, Saladax, Sanofi-Aventis, Servier, Takeda, WALA; honoraria: AbbVie, Amgen, AstraZeneca, Bayer, BMS, Boehringer, Daiichi, GSK, Leo, Lilly, medupdate, Merck, MSD, Nordic, Onkowissen, Piere Fabre, Roche, Saladax, Sanofi-Aventis, Servier, Synlab, Takeda, WALA; research funding: Deutsche Krebshilfe, Sanofi-Aventis. CR: consulting or advisory role: GSK; research funding: Deutsche Krebshilfe. TG: leadership: Reprognostics; honoraria: Roche, MSD, BMS, Boehringer, Astellas; consulting or advisory role: Roche, MSD, BMS, Boehringer, Astellas. All other authors have declared no conflicts of interest.
==== Refs
References

1 Glynne-Jones R. Wyrwicz L. Tiret E. Rectal cancer: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up Ann Oncol 28 2017 iv22 iv40 28881920
2 Schrag D. Shi Q. Weiser M.R. PROSPECT: a randomized phase III trial of neoadjuvant chemoradiation versus neoadjuvant FOLFOX chemotherapy with selective use of chemoradiation, followed by total mesorectal excision (TME) for treatment of locally advanced rectal cancer (LARC) (Alliance N1048) J Clin Oncol 1 2023 LBA2-LBA2
3 Schrag D. Weiser M.R. Goodman K.A. Neoadjuvant chemotherapy without routine use of radiation therapy for patients with locally advanced rectal cancer: a pilot trial J Clin Oncol 32 2014 513 518 24419115
4 Dworak O. Keilholz L. Hoffmann A. Pathological features of rectal cancer after preoperative radiochemotherapy Int J Colorectal Dis 2 1997 19 23
5 Schrag D. Shi Q. Weiser M.R. Preoperative treatment of locally advanced rectal cancer N Eng J Med 389 2023 322 334
6 Ding P.-R. Wang X.-Z. Li Y.-F. LBA26 neoadjuvant chemotherapy with CAPOX versus chemoradiation for locally advanced rectal cancer with uninvolved mesorectal fascia (CONVERT): final results of a phase III trial Ann Oncol 34 2023 S1267 S1268
7 Deng Y. Chi P. Lan P. Neoadjuvant modified FOLFOX6 with or without radiation versus fluorouracil plus radiation for locally advanced rectal cancer: final results of the Chinese FOWARC trial J Clin Oncol 37 2019 3223 3233 31557064
8 Zhang J. Chi P. Lan P. Long-term outcome of neoadjuvant mFOLFOX6 with or without radiation versus fluorouracil plus radiation for locally advanced rectal cancer: a multicenter, randomized phase III trial J Clin Oncol 41 2023 3505
9 Pastorino A. Di Bartolomeo M. Maiello E. Aflibercept plus FOLFIRI in the real-life setting: safety and quality of life data from the Italian Patient Cohort of the Aflibercept Safety and Quality-of-Life Program Study Clin Colorectal Cancer 17 2018 e457 e470 29605592
10 Hasegawa S. Goto S. Matsumoto T. A multicenter phase 2 study on the feasibility and efficacy of neoadjuvant chemotherapy without radiotherapy for locally advanced rectal Cancer Ann Surg Oncol 24 2017 3587 3595 28685354
11 Uehara K. Hiramatsu K. Maeda A. Neoadjuvant oxaliplatin and capecitabine and bevacizumab without radiotherapy for poor-risk rectal cancer: N-SOG 03 phase II trial Jpn J Clin Oncol 43 2013 964 971 23935207
12 Fernández-Martos C. Pericay C. Losa F. Effect of aflibercept plus modified FOLFOX6 induction chemotherapy before standard chemoradiotherapy and surgery in patients with high-risk rectal adenocarcinoma: the GEMCAD 1402 randomized clinical trial JAMA Oncol 5 2019 1566 1573 31465088
13 Pesántez D. Ten Hoorn S. Machado I. Total neoadjuvant therapy with or without aflibercept in rectal cancer: 3-year results of GEMCAD-1402 J Natl Cancer Inst 115 2023 1497 1505 37405857
14 Ruppert R. Junginger T. Kube R. Risk-adapted neoadjuvant chemoradiotherapy in rectal cancer: final report of the OCUM study J Clin Oncol 41 2023 4025 4034 37335957
