Neoadjuvant Chemotherapy Versus Adjuvant Chemotherapy for Very Low-Lying Clinical T3 Rectal Cancer: The NAIR Phase 2/3 Randomized Clinical Trial

Abstract

Objective:

To determine whether neoadjuvant chemotherapy (NAC) followed by total mesorectal excision (TME) and adjuvant chemotherapy (AC) is superior to TME followed by AC for very low-lying clinical (c) T3 rectal cancer.

Background:

Preoperative radiation is widely used for preoperative treatment of cT3 rectal cancer; however, it worsens patient-reported outcomes (PROs). Preoperative treatment without radiation is expected to preserve PROs.

Methods:

Patients with cT3N-anyM0 rectal cancer located within 5 cm from the anal verge were randomly assigned (1:1) to the NAC group (3 months of NAC followed by TME and 3 months of AC) or AC group (TME followed by 6 months of AC). NAC and AC comprised mFOLFOX6 (oxaliplatin, l-folinic acid, and fluorouracil) or CAPOX (oxaliplatin and capecitabine). The primary endpoint was the 3-year recurrence-free survival (RFS). PROs were analyzed.

Results:

Between February 2013 and March 2019, 130 patients were randomly assigned to the NAC (n = 65) or AC (n = 65) groups; of these, 127 were evaluable (NAC, n = 65; AC, n = 62). At a median follow-up of 37.4 months, the 3-year RFS was 75.5% and 70.9% in NAC and AC groups, respectively [hazard ratio (HR) = 0.67, 60% confidence interval (CI) = 0.48–0.86, 95% CI = 0.34–1.32; P = 0.098 by log-rank test] and the primary endpoint was met. There was no significant intergroup difference in the local recurrence rate (LRR) or overall survival. Histologically, good responders to NAC showed a trend toward better RFS than poor responders. The study groups showed similar PROs.

Conclusions:

NAC for very low-lying cT3 rectal cancer improved RFS without worsening PROs although LRR remained high.

Trial Registration:

UMIN Clinical Trials Registry: UMIN000009510/Japan Registry of Clinical Trials: jRCTs031180278.

INTRODUCTION

Preoperative chemoradiotherapy (CRT) using radiation plus concurrent fluoropyrimidine-based chemotherapy, followed by total mesorectal excision (TME), constitutes the standard treatment for locally advanced rectal cancer (LARC) as it reduces local recurrence rate (LRR).^1–3 Adjuvant chemotherapy (AC) is generally administered postoperatively; however, whether AC for LARC improves recurrence-free survival (RFS) remains controversial partly because of poor treatment compliance after TME.^4–6 Consequently, total neoadjuvant therapy (TNT), including preoperative CRT and preoperative or neoadjuvant chemotherapy (NAC), has been developed as the new standard treatment for LARC.^7–10 In previous Phase 3 clinical trials that used TNT as an investigational treatment, the transfer of postoperative chemotherapy to a preoperative setting provided survival benefits; however, the total dose of chemotherapy in the standard and investigational arms differed.^7,8 Therefore, whether the sequence of chemotherapy alone improves prognosis in LARC similar to that in locally advanced colon cancer remains unclear.¹¹

Meanwhile, incorporating patient experience through patient-reported outcomes (PROs) is recommended for assessing cancer-treatment tolerability.^12,13 Radiotherapy (RT) negatively impacts anal and sexual functions in patients with LARC, particularly those with very low-lying tumors requiring intersphincteric resection (ISR) to preserve the anus; conversely, preoperative treatment without RT is expected to preserve these functions.^14–19 Recent studies reported the prognostic noninferiority and better PROs for NAC followed by TME, compared to preoperative CRT followed by TME; however, the difference in prognosis and PROs between NAC followed by TME and TME followed by AC is unclear.^20,21 Therefore, we investigated whether altering the sequence of chemotherapy alone improves prognosis in LARC and whether NAC deteriorates PROs of patients with very low-lying rectal cancer. We included only very low-lying rectal cancers requiring ISR, which are particularly susceptible to the negative effects of preoperative RT.

This study presents the oncological outcomes and PROs of a multicenter randomized controlled trial that compared NAC followed by TME with TME followed by AC for very low-lying clinical (c) T3 LARC, wherein ISR is needed to preserve the anus.

METHODS

Study Design and Patients

The NAIR trial is a prospective, randomized, open-label, controlled, multicenter Phase 2/3 study that was conducted at 13 hospitals in Japan in accordance with the principles of the Declaration of Helsinki. The study protocol (Supplemental Tables 1–3 and Supplemental Figures 1–7, https://links.lww.com/AOSO/A499) was approved by the local ethics committees of all participating hospitals. This study followed the Consolidated Standards of Reporting Trials reporting guidelines. Written informed consent was obtained from all the study participants.

The major eligibility criteria were as follows: treatment-naive with rectal histologically proven adenocarcinoma located within 5 cm from the anal verge or 3 cm from the dentate line; clinical stage T3N-anyM0; age 20–75 years; Eastern Cooperative Oncology Group performance status 0 or 1; and presence of sufficient organ function. A positive lymph node was defined as 1 measuring ≥10 mm in the short-axis diameter on imaging. The MRI-circumferential resection margin (mrCRM) was considered positive when the distance between the tumor and neighboring organs or muscles was ≤1 mm. An MRI-extramural vascular invasion (mrEMVI) score of 3 or 4 was considered positive.²² Mismatch repair status was assessed using immunostaining in participants of the biomarker study. Patients who requested abdominoperineal resection (APR) or for whom anus-preserving surgery was not feasible at baseline evaluation were excluded.

Random Assignment and Stratification

Patients were enrolled by the study investigators and randomly assigned (1:1) to either the NAC group [NAC with 6 cycles of modified fluorouracil, leucovorin, and oxaliplatin (mFOLFOX6) or 4 cycles of capecitabine and oxaliplatin (CAPOX) followed by TME; postoperative identical chemotherapy] or the AC group (TME followed by AC with 12 cycles of mFOLFOX6 or 8 cycles of CAPOX) using a minimization method. Random assignment was stratified by cN stage, center, and sex, with treatment allocation not masked.

Treatment

NAC, TME, and AC were scheduled within 3 weeks after random assignment, 4–8 weeks after completing NAC, and within 8 weeks after TME, respectively, for patients in the NAC group. Imaging evaluations were performed after 3 cycles of neoadjuvant mFOLFOX6 or 2 cycles of neoadjuvant CAPOX. In nonresponders, TME was performed if the tumor was still resectable; if not resectable, study treatment was discontinued (Supplemental Figure 1, https://links.lww.com/AOSO/A499). For patients in the AC group, TME was scheduled within 8 weeks after random assignment, and AC was scheduled within 8 weeks after TME (Supplemental Figure 1, https://links.lww.com/AOSO/A499). In both groups, AC was initiated regardless of the histological stage at resection. The investigator could select either mFOLFOX6 or CAPOX but had to use the same regimen both preoperatively and postoperatively.

The mFOLFOX6 regimen comprised oxaliplatin (85 mg/m²) plus l-folinic acid 200 mg 2-hour infusion, fluorouracil (400 mg/m²) bolus, 2400 mg/m² 46-hour infusion, repeated once every 2 weeks. In total, a 24-week treatment regimen was planned (6 NAC cycles, 6 AC cycles, or 12 AC cycles). The CAPOX regimen comprised oxaliplatin 130 mg/m² plus oral capecitabine 1000 mg/m²/d (or 750 mg/m²/d in patients with 30 mL/min ≤ CCr ≤ 50 mL/min or age ≥ 70 years) for 2 weeks, repeated once every 3 weeks. In total, a 24-week treatment was planned (4 NAC cycles + 4 AC cycles or 8 AC cycles). The protocol allowed dose reduction, treatment delays, and early cessation of toxicity.

ISR was performed as described previously.²³ Although patients scheduled for APR at the time of enrollment were excluded, surgeons could switch to APR if intraoperative findings indicated that anal preservation was not feasible. Surgeons could choose between open and laparoscopic surgery. Lateral pelvic lymph node dissection (LLND)²⁴ was performed bilaterally according to the Japanese Society for Cancer of the Colon and Rectum (JSCCR) guidelines 2019²⁵; however, to increase registration, the protocol was amended to allow the omission of LLND if no obvious lateral lymph node metastasis was suspected on preoperative imaging.

Clinical follow-up included regular assessments (including serum carcinoembryonic antigen, carbohydrate antigen 19–9 levels, and a thorax/abdominal/pelvis computed tomography scan) every 4 months in the first 2 years and every 6 months in the following year.

Outcome Measures

Phase 2 aimed to assess the safety of NAC in the first 40 patients. The criteria for Phase 2 were as follows (Supplemental Table 1, https://links.lww.com/AOSO/A499):

1. Grade 4 hematologic toxicity requiring hospitalization or Grade 3 or higher nonhematologic toxicity attributable to preoperative FOLFOX or CAPOX therapy in the study treatment group occurred in a maximum of 6 of 20 patients.

2. Fewer than 4 of 20 patients in the study treatment group required emergency surgery during preoperative FOLFOX or CAPOX therapy and in the perioperative period.

3. Fewer than 4 out of 20 cases of APR cannot be performed due to tumor progression.

4. Fewer than 2 deaths out of 20 cases during the protocol treatment period.

The primary endpoint of the Phase 3 study was RFS, defined as the time between random assignment and the occurrence of macroscopic nonradical surgery, locoregional recurrence or metastasis, or death due to any cause. Secondary endpoints included the LRR, overall survival (OS), protocol treatment-completion rate, percentage of radical resection, surgical safety assessment, percentage of anal preservation, anorectal function survey using the Wexner Incontinence Score,²⁶ overall quality of life (QOL) survey using the Short-Form 36 questionnaire (SF-36),^27,28 modified Fecal Incontinence QOL Score,²⁹ chemotherapy safety assessment, the response rate to NAC, and the presence and extent of peripheral neuropathy. The PROs were collected longitudinally. Anal function data were collected from patients without permanent stoma, and overall QOL data were collected from all participants. Bladder function data were collected from all patients using the International Prostate Symptom Score,³⁰ Overactive Bladder Symptom Score,³¹ and International Consultation on Incontinence Questionnaire-Short Form.³² Sexual function data were collected using the international index of erectile function³³ and a questionnaire for ejaculation from male patients, and a simplified questionnaire of the female sexual function index³⁴ and the QLQ-CR38³⁵ from female patients.

All resected specimens were processed following the standardized protocol, including TNM classification according to the American Joint Committee on Cancer/International Union for Cancer Control (seventh edition). The tumor regression grade after NAC was semiquantitatively evaluated based on the 2019 JSCCR guidelines: Grade 0, no effect (no tumor cell necrosis or degeneration in response to treatment was observed); Grade 1a, minimal effect (tumor cell necrosis or degeneration is present in less than one-third of the entire lesion); Grade 1b, mild effect (tumor cell necrosis, degeneration, and/or lytic change is present in more than one-third but less than two-thirds of the entire lesion); Grade 2, moderate effect (prominent tumor cell necrosis, degeneration, lytic change, and/or disappearance is present in more than two-thirds of the entire lesion, but viable tumor cells remain); and Grade 3, marked effect (pathological complete response).²⁵ Adverse events (AE) were assessed according to the National Cancer Institute Common Terminology Criteria for Adverse Events (version 4.0).

Statistical Analysis

The sample size was originally estimated based on the threshold, and the expected 3-year RFSs were set at 70% and 80%. Seventy events were required to ensure a 75% statistical power with a 1-sided α of 0.10. Assuming enrollment and follow-up periods of 3 years each (total study period: 6 years), 200 patients (100 patients in each arm) were estimated. Due to the slow enrollment speed, the protocol was amended to continue enrollment until March 2019 and conduct an analysis with a 1-sided α of 0.20 and 70% power for 128 enrolled patients (33 expected events).

Efficacy analyses were performed on the full analysis set (FAS; included all enrolled patients who received at least 1 treatment session and met the eligibility criteria). RFS, LRR, distant metastasis-free survival (DMFS), and OS were evaluated using a stratified log-rank test based on factors other than the center. Survival curves were estimated using the Kaplan–Meier method. After protocol amendment, a 1-sided P < 0.20 for the primary endpoint was considered to be met. In addition to the 60% confidence interval (60% CI) for the primary analysis, 95% confidence intervals (95% CIs) are provided for reference. The relationship between histological tumor regression grade and prognosis was analyzed post hoc. Two groups were assessed: Grades 0–1a, in which NAC was considered ineffective, and Grades 1b–3, in which NAC was considered effective. PROs were analyzed using a mixed-effects model for repeated measures.³⁶ All analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC).

RESULTS

Accrual and Clinical Characteristics

In the NAIR trial, between February 2013 and March 2019, 130 patients were enrolled and randomly assigned to the treatment groups (65 patients in each treatment group) (Fig. 1). One patient withdrew consent postenrollment, and 2 developed unresectable distant metastases after enrollment; therefore, data from 127 patients were evaluable (FAS population: 65 and 62 in the NAC and AC groups, respectively). The median follow-up duration for the endpoints was 37.4 months (range 12.1–59.4 months). Baseline characteristics were well-balanced across the groups (Table 1).

FIGURE 1.

CONSORT diagram showing patient disposition in the NAIR trial. NAIR, Neoadjuvant chemotherapy for intersphincteric resection in low rectal cancer.

TABLE 1.

Baseline Characteristics of 127 FAS Patients

Characteristics	NAC Group	AC Group	P
Total number of patients (FAS)	65	62
Age, years			0.095*
Median (range)	60 (27–75)	64 (35–75)
Sex			0.98†
Male	46 (70.8)	44 (71.0)
Female	19 (29.2)	18 (29.0)
ECOG score			0.93†
0	59 (90.8)	56 (90.3)
1	6 (9.2)	6 (9.7)
BMI, kg/m2			0.84*
Median (range)	23.0 (17.0–35.8)	23.2 (16.5–42.4)
T stage			N/A
cT3	65 (100)	62 (100)
N stage			0.47†
cN0	36 (55.4)	38 (61.3)
cN1	25 (38.5)	18 (29.0)
cN2	4 (6.2)	6 (9.7)
Suspected lateral lymphatic metastasis	10 (15.4)	12 (19.4)	0.55†
Clinical stage			0.39†
II	36 (55.4)	39 (62.9)
III	29 (44.6)	23 (37.1)
Distance to anal verge, cm			0.64*
Median (range)	4.5 (2.0–5.0)	4.5 (0.5–5.0)
mrCRM positive	22 (33.8)	15 (24.2)	0.23†
mrEMVI positive	11 (16.9)	10 (16.1)	0.90†
MMR status			0.93†
Proficient	44 (67.7)	40 (64.5)
Deficient	2 (3.1)	2 (3.2)
Unknown	19 (29.2)	20 (32.3)
Chosen chemotherapy			0.45†
mFOLFOX6	49 (75.4)	43 (69.4)
CAPOX	16 (24.6)	19 (30.6)

Data are the frequency (proportion) unless otherwise indicated. mrCRM and mrEMVI were evaluated in patients who underwent MRI at the time of enrollment (NAC group: n = 63; AC group: n = 61).

^*Wilcoxon P-value;

^†χ2 P-value.

AC indicates adjuvant chemotherapy; BMI, body mass index; c, clinical; CAPOX, capecitabine and oxaliplatin; ECOG, Eastern Cooperative Oncology Group; FAS, full analysis set; FU, fluorouracil; mFOLFOX6, modified fluorouracil, leucovorin, and oxaliplatin; MMR, mismatch repair; mrCRM, MRI-circumferential resection margin; mrEMVI, MRI-extramural vascular invasion; N/A. not applicable; NAC, neoadjuvant chemotherapy.

Treatment Delivery

Treatment delivery is shown in Figure 1. In the NAC group, all 65 patients started NAC, and 63 (96.9%) completed the full 12-week course. One patient developed a cerebral infarction during NAC, for whom mFOLFOX6 was discontinued after 3 courses. This patient underwent TME after the infarction was resolved. NAC was terminated in 1 patient after 6 weeks due to disease progression, proceeding to TME. Two patients showed disease progression after 12 weeks but underwent TME. All patients in the NAC group underwent TME, and 57 (87.7%) received AC. Of the 62 evaluable patients in the AC group, all underwent TME, and 57 (91.9%) received AC. All 127 patients who underwent TME underwent radical surgery. Overall, 55 (84.6%) and 45 (72.6%) participants in the NAC and AC groups, respectively, completed the planned treatment protocol. The relative dose intensities of perioperative chemotherapy in each arm were similar (Supplemental Table 2, https://links.lww.com/AOSO/A499).

Tolerability

Overall, 3, 1, 2, and 0 patients met criteria 1, 2, 3, and 4 of the Phase 2 study, respectively. They also met the safety criteria for the Phase 2 study (Supplemental Table 1, https://links.lww.com/AOSO/A499).

Table 2 outlines surgical details and operative complications. Anal-preserving surgery (APS) was performed more frequently in the NAC group compared with the AC group (65 vs 55, P = 0.0053). Patients in the TME group who did not undergo APS were scheduled for APS preoperatively; however, intraoperative findings indicated that APS was not feasible, therefore, APR was performed.

TABLE 2.

Surgical Details and Intraoperative and Postoperative Complications

	NAC Group	AC Group	P
No. of patients (FAS)	65	62
Type of surgery			0.0053
Anal-preserving surgery (with diverting stoma)	65 (100) (65)	55 (88.7) (54)
APR	0	7 (11.3)
Approach for TME			0.31
Open	11 (16.9)	15 (24.2)
Laparoscopic	54 (83.1)	47 (75.8)
Lateral lymph node dissection	62 (95.4)	57 (91.9)	0.42
Operative time, min			0.36
Median (range)	356.0 (166–744)	385.5 (102–619)
Blood loss, mL			0.56
Median (range)	133.0 (6–2648)	153.5 (0–9794)
Intraoperative complications
All	1 (1.5)	3 (4.8)	0.29
Gastrointestinal tract injury	1 (1.5)	1 (1.6)	0.97
Urinary tract injury	0	1 (1.6)	0.30
Obturator nerve injury	0	1 (1.6)	0.30
30-day postoperative complications (All grade/grade ≥ 3)
Anastomotic leakage	6 (9.2)/2 (3.1)	10 (16.1)/ 2 (3.2)	0.24/0.96
Pelvic abscess	6 (9.2)/6 (9.2)	5 (8.1)/ 5 (8.1)	0.82/0.82
Wound infection	5 (7.7)/2 (3.1)	3 (4.8)/ 1 (1.6)	0.51/0.59
Urinary tract infection	1 (1.5)/0	8 (12.9)/ 5 (8.1)	0.013/0.020
Urinary retention	11 (16.9)/0	13 (21.0)/ 0	0.34/1.0
Bowel obstruction	7 (10.8)/0	5 (8.1)/ 1 (1.6)	0.60/0.30
Thromboembolism	1 (1.5)/1 (1.5)	0/0	0.33/0.33
Stoma-related complications	1 (1.5)/1 (1.5)	0/0	0.33/0.33
Late postoperative complications (All grade/grade ≥ 3)
Anastomotic fistula	2 (3.1)/ 2 (3.1)	2 (3.2)/ 0	0.96/0.16
Anastomotic stenosis	2 (3.1)/ 2 (3.1)	1 (1.6)/ 0	0.59/0.16
Rectal prolapse	3 (4.6)/ 3 (4.6)	3 (4.8)/ 3 (4.8)	0.95/0.95
Pelvic abscess	1 (1.5)/ 1 (1.5)	2 (3.2)/ 2(3.2)	0.53/ 0.53
Perianal abscess	1 (1.5)/ 1 (1.5)	0/ 0	0.33/0.33
Urinary tract infection	1 (1.5)/ 0	2 (3.2)/ 2 (3.2)	0.53/0.14
Bowel obstruction	3 (4.6)/ 2 (3.1)	2 (3.2)/ 1 (1.6)	0.69/0.59
Stoma-related complications	2 (3.1)/ 2 (3.1)	1 (1.6)/ 0	0.59/0.16
Ureteral stenosis	1 (1.5)/ 1 (1.5)	0/ 0	0.33/0.33

Data are reported as frequency (proportion) unless otherwise indicated.

AC indicates adjuvant chemotherapy; APR, abdominoperineal resection; FAS, full analysis set; NAC, neoadjuvant chemotherapy; TME, total mesorectal excision.

Chemotherapy-associated AEs are detailed in Table 3, and the timing of nonhematological toxicity onset is shown in Supplemental Table 3, https://links.lww.com/AOSO/A499. The incidence of any-grade peripheral sensory neuropathy lasting for 3 years did not significantly differ between the groups (22.7% in the NAC group vs 38.5% in the AC group in patients treated with mFOLFOX6, P = 0.32; 7.1% in the NAC group vs 33.3% in the AC group in patients treated with CAPOX, P = 0.10) (Supplemental Figure 2, https://links.lww.com/AOSO/A499).

TABLE 3.

Acute Toxicity of Perioperative Chemotherapy

Characteristics	NAC Group	AC Group	P
No. of patients (FAS)	65	62
Chemotherapy regimen			0.45
mFOLFOX6	49 (75.4)	43 (69.4)
CAPOX	16 (24.6)	19 (30.6)
Chemotherapy-related adverse events (All grade/Grade ≥ 3)
Leukopenia	28 (43.1)/ 3 (4.6)	27 (43.5)/ 1 (1.6)	0.96/0.33
Neutropenia	30 (46.2)/ 6 (9.2)	31 (50)/ 5 (8.1)	0.66/0.82
Anemia	29 (44.6)/ 1 (1.5)	37 (59.7)/ 0	0.090/0.33
Thrombocytopenia	55 (84.6)/ 1 (1.5)	47 (75.8)/ 0	0.21/0.33
Allergy	10 (15.4)/ 0	1 (1.6)/ 0	0.0058/ 1.0
Fever	1 (1.5)/ 0	6 (9.7)/ 0	0.045/ 1.0
Mucositis (features/symptoms)	3/ (4.6) 0	11 (17.7)/ 0	0.018/ 1.0
Skin rash	4/ (6.2) 0	0/ 0	0.047/ 1.0
Anorexia	16 (24.6)/ 2 (3.1)	11 (17.7)/ 1 (1.6)	0.34/0.59
Constipation	16 (15.4)/ 0	1 (1.6)/ 0	0.0001/ 1.0
Diarrhea	6 (9.2)/ 0	18 (29.0)/ 2 (3.2)	0.0044/0.14
Nausea	19 (29.2)/ 3 (4.6)	22 (35.5)/ 2 (3.2)	0.45/0.69
Vomiting	4 (6.2)/ 1 (1.5)	3 (4.8)/ 2 (3.2)	0.75/0.55
Mucositis (examination findings)	6 (9.2)/ 0	7 (11.3)/ 0	0.70/ 1.0
Febrile neutropenia	0/ 0	1 (1.6)/ 1 (1.6)	0.30/0.30
Hand–foot syndrome (CAPOX- group only)	9 (13.8)/ 0	4 (6.5)/ 0	0.17/ 1.0
Elevated AST	52 (80)/ 0	47 (75.8)/ 0	0.57/ 1.0
Elevated ALT	56 (86.2)/ 4 (6.2)	52 (83.9)/ 2 (3.2)	0.72/0.44

Data are reported as frequency (proportion) unless otherwise indicated.

AC indicates adjuvant chemotherapy; ALT, alanine aminotransferase; AST, aspartate aminotransferase; CAPOX, capecitabine and oxaliplatin; FU, fluorouracil; mFOLFOX6, modified fluorouracil, leucovorin, and oxaliplatin; NAC, neoadjuvant chemotherapy.

No treatment-related deaths occurred in either group.

Histopathologic Outcomes

The histopathologic outcomes are shown in Table 4. Among the 127 patients who underwent TME, 63 (96.9%) and 58 (93.6%) in the NAC and AC groups, respectively, achieved a pathologically R0 resection (P = 0.37). Statistically significant reductions in T and N stages were observed in the NAC group, with 2 of 65 patients (3.1%) showing pathological complete response (Table 4).

TABLE 4.

Histopathologic Outcomes in the Subgroups

Outcome Variables	NAC Group	AC Group	P
No. of patients (FAS)	65	62
T stage			0.0042*
pT0 (pCR)	2 (3.1)	0
pT1	6 (9.2)	2 (3.2)
pT2	26 (40.0)	15 (24.2)
pT3	26 (40.0)	44 (71.0)
pT4	5 (7.7)	1 (1.6)
N stage			0.0010†
pN0	48 (73.9)	28 (45.2)
pN1	13 (20.0)	16 (25.8)
pN2	4 (6.2)	18 (29.0)
Pathologically involved LLNs	2 (3.1)	10 (16.1)	0.012
R0	63 (96.9)	58 (93.5)	0.37
Histological effect of NAC
Grade 3 (=pCR)	2 (3.1)	N/A
Grade 2	10 (15.4)	N/A
Grade 1b	12 (18.5)	N/A
Grade 1a	40 (61.5)	N/A
Grade 0	1 (1.5)	N/A

Data are the frequency (proportion) unless otherwise indicated.

^*pT0–pT1 –pT2 vs pT3–pT4.

^†pN0 vs pN1–pN2.

AC indicates adjuvant chemotherapy; CR, complete response; FAS, full analysis set; LLN, lateral lymph node; N/A, not applicable; NAC, neoadjuvant chemotherapy; p, pathological.

Prognosis

In the FAS population, locoregional recurrence, metastasis, or death due to any cause was observed in 34 patients (15 and 19 in the NAC and AC groups, respectively). At 3 years, the probability of RFS was 75.5% (95% CI = 62.5–84.5) and 70.9% (95% CI = 57.2–80.9) in the NAC and AC groups, respectively [hazard ratio (HR) = 0.67, 60% CI = 0.48–0.86, 95% CI = 0.34–1.32; P = 0.098; Fig. 2A] and the primary endpoint was met.

FIGURE 2.

Kaplan–Meier curves of (A) RFS, (B) DMFS, (C) LRR, and (D) OS in patients with cT3NXM0 rectal adenocarcinoma originating within 5 or 3 cm from the anal verge or dentate line, respectively. AC group: TME followed by AC with 12 cycles of mFOLFOX6 or 8 cycles of CAPOX. NAC group: NAC with 6 cycles of mFOLFOX6 or 4 cycles of CAPOX followed by TME; and postoperative identical chemotherapy. AC, adjuvant chemotherapy; CAPOX, capecitabine and oxaliplatin; DMFS, distant metastasis-free survival; LRR, local recurrence rate; mFOLFOX6, modified fluorouracil, leucovorin, and oxaliplatin; NAC, neoadjuvant chemotherapy; OS, overall survival; RFS, recurrence-free survival; TME, total mesorectal excision.

The probability of DMFS at 3 years was 77.0% (95% CI = 64.1–85.7) and 71.7% (95% CI = 57.7–81.7) in the NAC and AC groups, respectively (HR = 0.69, 95% CI = 0.34–1.41; P = 0.14; Fig. 2B). The probability of local recurrence at 3 years was 15.1% (95% CI = 8.1–27.1) in the NAC group vs 11.4% (95% CI = 5.2–24.0) in the AC group (HR = 1.22, 95% CI = 0.45–3.28; P = 0.61; Fig. 2C), respectively. Overall, 9 patients died during the study: 6 and 3 in the NAC and AC groups, respectively. The probability of OS at 3 years was 92.3% (95% CI = 82.5–96.7) and 94.7% (95% CI = 84.3–98.3) in the NAC and AC groups, respectively (HR = 1.74, 95% CI = 0.43–7.01; P = 0.76; Fig. 2D). Subgroup analysis showed that the treatment effects on RFS were similar, regardless of the clinical prognostic factors (Supplemental Figure 3, https://links.lww.com/AOSO/A499).

When the patients in the NAC group were divided into those who showed a histological tumor regression grade of 0–1a and 1b–3, trends of better RFS and DMFS were observed in the patients with grade 1b–3 disease compared to those with grade 0–1a (HR = 0.41, 95% CI = 0.12–1.44; P = 0.15 and HR = 0.44, 95% CI = 0.12–1.59; P = 0.20, respectively); however, both RFS and DMFS in the patients who showed grade 0–1a in the NAC group were similar to those in the AC group (Figs. 3A, B). The LRR was not related to the histological tumor regression grade (Fig. 3C).

FIGURE 3.

Kaplan–Meier curves of (A) RFS, (B) DMFS, and (C) LRR for histological tumor regression grade. AC group: TME followed by AC with 12 cycles of mFOLFOX6 or 8 cycles of CAPOX. NAC group: NAC with 6 cycles of mFOLFOX6 or 4 cycles of CAPOX followed by TME; and postoperative identical chemotherapy. AC, adjuvant chemotherapy; CAPOX, capecitabine and oxaliplatin; DMFS, distant metastasis-free survival; LRR, local recurrence rate; mFOLFOX6, modified fluorouracil, leucovorin, and oxaliplatin; NAC, neoadjuvant chemotherapy; RFS, recurrence-free survival; TME, total mesorectal excision.

PROs

Three years after randomization, 61 (93.8%) and 52 (83.9%) patients in the NAC and AC groups, respectively, could defecate anally (P = 0.092; Supplemental Figure 4, https://links.lww.com/AOSO/A499). The mean Wexner Incontinence Score at 3 years was 11.2 and 9.5 in the NAC and AC groups, respectively (P = 0.42; Supplemental Figure 5, https://links.lww.com/AOSO/A499). No significant intergroup difference was observed between the NAC and TME groups in terms of urinary function, male sexual function, or overall QOL (Supplemental Figures 5 and 6, https://links.lww.com/AOSO/A499). Female sexual function was difficult to analyze because of the low number of responses to the questionnaire.

DISCUSSION

This study evaluated the efficacy and safety of NAC followed by TME and AC for very low-lying cT3 LARC, and postoperative PROs. The RFS was better in the NAC group than the AC group, meeting the primary endpoint. Although postoperative PROs were preserved, the LRR remained high in the NAC group. Unlike other clinical trials that had different chemotherapy doses in both groups and demonstrated a decreased rate of recurrence with TNT, our study had identical chemotherapy doses in both groups.^7,8 However, the RFS was better in the NAC group than in the AC group, indicating that the chemotherapy sequence alone improved RFS. The better DMFS in the NAC group compared to the AC group might have contributed to this result.

Postoperative PROs are important for the treatment of rectal cancer. Although the PROSPECT trial reported better PROs in the NAC group than those in the CRT group,²¹ it was unclear whether NAC worsened the postoperative PROs compared to upfront TME. Our results revealed no worsening of PROs after NAC, providing little concern about PROs when NAC is added to the treatment for LARC.

In this study, the LRR in both groups was higher than that of the PROSPECT trial, which showed about 2% of LRR in both groups.²⁰ In the PROSPECT trial that included patients with cT2 node-positive, cT3 node-negative, or cT3 node-positive LARC who were candidates for APS, the proportion of the tumor located within 5 cm from the anal verge was around 15% in both the NAC and the CRT groups.²⁰ Although the subgroup analysis is not shown in the PROSPECT study, generally, cT3 tumors located within 5 cm from the anal verge show high LRR. During the accrual period of the NAIR trial, preoperative CRT was not the standard treatment for cT3 rectal cancer in Japan,^25,37,38 and we expected that NAC would improve the LRR and RFS; however, LRR was not improved by NAC. In the subgroup analysis, the LRR was high even in good responders to NAC. This implies that NAC using doublet chemotherapy for very low-lying cT3 rectal cancer may result in inadequate local control, although it can preserve PROs. The pCR rate of 3.1% in the NAC group in this study was considerably lower than that of CRT (around 15% in general),^1–3 which was believed to reflect the low tumor shrinkage effect and affect the high LRR. Therefore, RT is considered necessary for local control of low-lying T3 rectal cancer. Whereas NAC without RT may be an option to improve RFS for patients who are contraindicated to RT for various reasons, they should be informed about the high LRR after NAC without RT. Additionally, the use of more effective chemotherapy than doublet chemotherapy, such as triplet chemotherapy, or molecular-targeted agents as the NAC may improve local control of very low-lying cT3 rectal cancer; however, this should be investigated in further clinical trials.³⁹

In the subgroup analysis, histologically good responders to NAC showed a trend toward better RFS and DMFS compared to poor responders in the NAC group. Additionally, both the RFS and DMFS of the poor responders in the NAC group were comparable to those of the AC group. This implies that only good responders may benefit from NAC; therefore, it may be necessary to develop methods to predict good responders before initiating treatment^40,41 or implementing a new strategy, such as adding on RT if the clinical response to NAC is poor.²⁰ Moreover, moving chemotherapy to the preoperative setting was expected to improve compliance with chemotherapy and thereby improve patient prognosis. However, no significant intergroup difference was observed in the relative dose intensity in this study. Therefore, the better RFS in the NAC group might have been achieved by the early control of micrometastasis by NAC. Monitoring circulating tumor DNA may be useful in testing this hypothesis.⁴²

Notably, this study was conducted in 2012. At that time, mFOLFOX6 was the mainstay for perioperative chemotherapy in Japan, and a protocol amendment allowed the use of the CAPOX regimen. There was no difference in the RFS, OS, or LRR between the FOLFOX and CAPOX groups (Supplemental Figure 7, https://links.lww.com/AOSO/A499); therefore, both regimens are available. Postoperative oxaliplatin dose intensity in patients treated with CAPOX in the NAC group was low; however, the reason is unclear due to the limited sample size. No intergroup difference between the NAC group and the AC group was noted in OS; however, the number of events was small. Therefore, a long-term follow-up of the patients enrolled in this study would be needed to verify whether there is an intergroup difference in OS.

Limitations

This study has some limitations. First, the sample size was small and statistical power was low. The sample size had to be reduced because of slow enrollment. One reason for the slow enrollment is believed to be a limited number of surgeons capable of performing ISR in Japan during the study period. Therefore, enrollment should have been more strictly predicted before the study was initiated. Although this forced us to amend the statistical settings, it is significant that changing the order of perioperative chemotherapy from the postoperative setting to the preoperative setting alone improved RFS without worsening PROs. Second, the LRR was high in both groups; therefore, the local recurrence of very low-lying cT3 rectal cancer should be controlled by preoperative RT in exchange for worsening PROs or introducing more effective NAC than doublet chemotherapy such as triplet chemotherapy or molecular-targeted agents. Thus, our data indicates the need for optimal use of preoperative therapies for advanced low-lying rectal cancers to control both local and distant recurrence, since the treatments of both arms in this study were not the current standard.

CONCLUSIONS

NAC for very low-lying cT3 rectal cancer improved RFS without worsening PROs although LRR remained high.

ACKNOWLEDGMENTS

We express special thanks to all participating patients, their families, all participating investigators, members of the Data and Safety Monitoring Board [Taira Kinoshita (Aichi Cancer Center), Koichi Goto (National Cancer Center Hospital East), Tomonori Yano (National Cancer Center Hospital East), and Hiraku Fukushima (Seiwa Memorial Hospital)], and research assistants (Mihoko Matsumoto, Aki Shiraishi, Keiko Yao, and Yuko Saji).

Supplementary Material

as9-6-e579-s001.pdf ^{(1.3MB, pdf)}