1 INTRODUCTION

Several randomized control trials (RCTs) of patients with locally advanced rectal cancer (LARC) have demonstrated the superiority of neoadjuvant chemoradiotherapy (nCRT) over postoperative CRT, because it provides better tumor control and reduces treatment toxicity.1-3 Thus, the National Comprehensive Cancer Network (NCCN) guideline now recommends nCRT plus total mesorectal excision (TME) as the standard treatment scheme for LARC patients.4 Research indicated that nCRT reduced the 5-year local recurrence (LR) rate from more than 25% to approximately 5%–10%, although it had no appreciable effect on the risk of distant metastasis (DM, about 30%), which remains the leading cause of cancer-related death in LARC patients.1-3 For ypTNM stage II or III rectal cancer (RC) patients who received nCRT and surgery, adjuvant 5-fluorouracil (5-FU)-based postoperative chemotherapy (POCT) did not affect overall survival (OS), disease-free survival (DFS), or DM, and only 43%–73.6% of patients received 95% or more of the planned 5-FU dose.5 Another study showed that adjuvant POCT did not affect recurrence-free survival (RFS) in stages II–III RC patients after nCRT and TME.6

Several Phase III RCTs showed that addition of oxaliplatin to 5-FU/capecitabine-based nCRT increased acute toxicity, but did not improve the pathological complete response (pCR), DFS, or OS.7-9 However, the CAO/ARO/AIO-04 trial reported a significant improvement of pCR (13% vs. 17%, p = 0.04) and 3-year DFS (71.2% vs. 75.9%, p = 0.03) in patients who received 5-FU-based nCRT without or with oxaliplatin.10 Moreover, 85% of patients in this trial received the full dose of neoadjuvant oxaliplatin. A meta-analysis by Zheng et al.11 also showed that adding oxaliplatin to preoperative nCRT significantly improved the pCR and DFS and decreased postoperative metastasis and LR, although it also led to increased CRT-related toxicities. Another meta-analysis of 10 RCTs (5599 LARC patients) showed that platinum-based drugs did not improve OS, DFS, or LR, but led to a statistically significant increase of pCR and a reduction of distant recurrence.12 However, these benefits were offset by the higher rate of grade 3/4 (G3/4) toxicities.

One of the largest meta-analyses to assess pCR-associated outcomes examined 14 datasets (3105 patients) and found that pCR (rather than residual histopathological tumor) was associated with a significantly lower incidence of LR (2.8% vs. 9.7%, p < 0.001) and improvements in 5-year distant relapse-free survival (88.4% vs. 74.9%, p < 0.001), 5-year DFS (83% vs. 66%, p < 0.001), and 5-year OS (87.6% vs. 76.4%, p < 0.001).13 Park et al.14 also reported that tumor response (complete vs. intermediate vs. poor) for LARC patients after nCRT was associated with 5-year RFS (90.5% vs. 78.7% vs. 58.5%; p < 0.001), 5-year DM rate (7.0% vs. 10.1% vs. 26.5%; p = 0.001), and 5-year LR rate (0% vs. 1.4% vs. 4.4%; p = 0.002). A Phase III, single-center, randomized trial by Akgun et al.15 showed that if the interval between nCRT and surgery exceeded 8 weeks in LARC patients, this significantly increased the pCR rate (10.0% vs. 18.6%, p = 0.027), but did not affect the quality of surgery. Another prospective, randomized, multicenter trial showed that when surgery was delayed to 12 weeks (rather than the usual 6 weeks) after CRT, it significantly increased magnetic resonance (MR) T downstaging (58% vs. 43%) and pCR (20% vs. 9%) and improved MR tumor regression grade (TRG) (52% vs. 34%).16 A multicenter retrospective analysis of LARC patients that compared an early surgery arm (≤7 weeks) and a delayed surgery arm (>7 weeks) reported an increased ypCR rate (8.6% vs. 12.3%, p = 0.011) and a higher sphincter preservation rate (89.9% vs. 92.4%, p = 0.078) after nCRT in the delayed group; however, the two groups had similar downstaging rate (37.0% vs. 36.9%), RFS (70.3% vs. 71.8%), and OS (83.3% vs. 86.5%).17 Kerr et al. also reported that delaying surgery beyond 8 weeks after nCRT may reduce postoperative morbidity without compromising prognosis.18 The Lyon prospective randomized trial of LARC patients, which had 17 years of follow-up, indicated that the LR and OS rates were similar when the interval between nCRT and surgery was long (6–8 weeks) or short (2 weeks).19 The Phase III randomized GRECCAR-6 trial of RC patients showed that there was no significant difference in the pCR rate of the standard group (7 weeks, 15%) and delayed group (11 weeks, 17%), but significantly more medical complications in the delayed group (32.8% vs. 19.2%; p = 0.01).20 However, 3-year OS, DFS, distant recurrence, and local recurrence were not influenced by the waiting period.21 Furthermore, an analysis of the National Cancer Database (NCDB) of 6397 LARC patients from 2004 to 2006 showed that radiation-to-surgery intervals greater than 60 days were associated with higher rates of positive surgical margins (6.7% vs. 4.8%, p = 0.009), lower rates of sphincter-preserving surgery (64.9% vs. 68.9%, p = 0.007), and shorter survival (HR = 1.314, 95% CI = 1.191–1.449; p < 0.001).22

Numerous studies have reported benefits of induction or consolidation chemotherapy (CONCT) with nCRT prior to surgery for LARC patients, a regimen referred to as total neoadjuvant therapy (TNT).23-31 These benefits include improved delivery of planned therapy, increased downstaging and pCR, earlier introduction of optimal systemic CT to address possible micrometastases, and availability of in vivo assessment of chemosensitivity. In addition, delivery of all CT preoperatively obviates the need for postoperative therapy, thus reducing the duration of a diverting ileostomy and alleviating the need for CT with a stoma, and the adverse events (AEs) following TNT were generally tolerable. These results led the NCCN to categorize TNT as an acceptable treatment strategy for patients with stage II/III LARC in 2018.4

Presently, there are three TNT schemes: nCRT followed by CONCT; induction chemotherapy (INCT) followed by nCRT; and INCT with nCRT followed by CONCT and then surgery. Here, we review the clinical evidence for the role of these three TNT schemes for LARC patients, with a focus on treatment compliance and toxicities, pCR rate, organ-preservation strategies, disease-related outcomes, and brief descriptions of ongoing trials.

2 NCRT FOLLOWED BY CONCT

To address concerns regarding the risk of systemic progression due to the longer wait time after nCRT and to further enhance neoadjuvant treatment response, several studies evaluated the effect of preoperative CONCT after nCRT for LARC patients (Table 1).

TABLE 1. Studies of nCRT followed by consolidation chemotherapy for LARC patients Study [year] Stage Design Pts (N) Neoadjuvant therapy regimen Adjuvant therapy pCR, % SR/R0/SP rate, % DFS/OS 3 y, % Garcia-Aguilar [2015]23 cII-III Nonrandomized 259 CRT (45 Gy [+5.4–9 Gy]+5-FU)+POCT mFOLFOX6 (8 cycles) 18* NR/98/77 50*/79 (5 y) Phase II four-arm CRT+mFOLFOX6 (2 cycles) mFOLFOX6 (6 cycles) 25 NR/100/75 81/92 (5 y) [2018]32 CRT+mFOLFOX6 (4 cycles) mFOLFOX6 (4 cycles) 30 NR/96/75 86/88 (5 y) CRT+mFOLFOX6 (6 cycles) mFOLFOX6 (2 cycles) 38 NR/100/69 76/84 (5 y) Lee [2016]33 cII-III 168 CRT (45 Gy+LF [5-FU+LV])+LF (1 cycle) LF (4 cycles) 19* 100/NR/93.5* 79.5*/86.9 160 CRT (45 Gy+LF) NR 10 100/NR/80 67.5/NR Zampino [2009]34 cT3–4/N0–2 Prospective single-arm 51 CRT (45 [+5.4] Gy+CAP]+CAP (2 cycles) NR 18 98/100/80 85/NR (5 y) Gao [2014]35 cII-III Prospective single-arm 36 CRT (46 [+4]Gy +XELOX)+XELOX (1 cycle) NR 36.1 100/100/75 NR/NR Tang [2018]36 cII-III Retrospective 45 CRT (46 [+4]Gy+XELOX)+XELOX (1 cycle) XELOX (5/3) + CAP (×2) 37.8 100/NR/NR 75.5/88.6 Zhu [2013]37 cII-III Phase II single-arm 42 CRT (44 Gy [IMRT]+XELOX)+CAP (1 cycle) XELOX (6–8 cycles) 16 90.1/92/78.9 57/66 Zhu [2014]38 cT3-4/cN+ Phase II single-arm 78 CRT (50/55 Gy+XELOX)+XELOX (1 cycle) XELOX (5 cycles) 23.7 97.4/100/36.8 63.8/77.4 CinClare [2020]24 cT3-4/N+ Phase III RCT two-arm 180 CRT (50 Gy+CAP)+XELOX (1 cycle) – 15* 88/96/- – 180 CRT (50 Gy+XELIRI)+XELIRI (1 cycle) – 30 87/97/- – Bujko [2016]41 cT3-4 Phase III RCT two-arm 261 5×5 Gy (for pelvis)+FOLFOX4 (3 cycles) NR 16 93/77/50 53/73* Polish II trial 254 CRT (50.4 Gy+FOLFOX4) NR 12 92/71/49 52/65 Markovina [2017]43 cT3–4/N0–2 Prospective Phase II 69 IMRT (4×5 [5×5] Gy)+mFOLFOX (4 cycles) FOLFOX (4 cycles) 28 100/95.7/75 85*/96 69 CRT (45 [+2.5–9]Gy+5-FU or CAP) FOLFOX/CAPEOX (×4–12) 16 100/88.6/72 68/88 Myerson [2014]44 cT3–4 Prospective single-arm 76 4×5 (5×5)Gy+FOLFOX (4 cycles) FOLFOX 25 100/95/75 NR RAPIDO [2020]46 MRI- Phase III RCT two-arm 462 CRT (5×5 Gy +CAPOX [×6] or FOLFOX4 [×9]) NR 28* 92/90/NR NR/89.1/NR [2021]47 high-risk 450 CRT (50.4/50 Gy+CAP) CAPOX (×8)/FOLFOX4 (×12) 14 89/90/NR NR/88.9/NR Abbreviations: 3DCRT, three-dimensional conformal radiotherapy; 3 y, 3 years; 5-FU, 5-fluorouracil; 5 y, 5 years; CAP, capecitabine; CAPEOX/XELOX, oxaliplatin + capecitabine; CRT, chemoradiotherapy; CT, chemotherapy; DFS, disease-free survival; dN, N-downstaging; dT, T-downstaging; dTN, T and/or N-downstaging; FOLFOX, oxaliplatin + leucovorin + 5-FU; Gy, Gray; h, hour; IMRT, intensity-modulated radiation therapy; IV, intravenous; LARC, locally advanced rectal cancer; LF, 5-FU + leucovorin; LV, leucovorin; mFOLFOX6, modified FOLFOX6; NAC, neoadjuvant chemotherapy; NR, not reported; OS, overall survival; pCR, pathologic complete response; R0, microscopically clear resection; RCT, randomized controlled trial; RT, radiotherapy; SP, sphincter preservation; TRG, tumor regression grade; Wk, week; XELIRI, capecitabine + irinotecan. * p < 0.05. 2.1 Long-course nCRT (5-FU-based LC-nCRT) followed by CONCT before surgery

Preoperative long-course nCRT (LC-nCRT) followed by surgery is a common scheme for LARC patients. A multicenter, Phase II trial by Garcia-Aguilar et al.23 randomized 259 LARC patients into four arms: standard LC-nCRT (RT with concurrent 5-FU), followed by zero, two, four, or six cycles of modified FOLFOX6 CONCT before TME. The compliance rate of CONCT was 77%–82%, much higher than the compliance of POCT (54%). The incidence of G3 or more AEs during CONCT was 3%, 18%, and 28% for two, four, and six cycles of mFOLFOX6, and the most common G3 AEs were diarrhea and lymphopenia. CONCT significantly improved the pCR rate (18% vs. 38%, p = 0.004) and DFS (50% vs. 86%, p = 0.004) but not OS (p = 0.37).32 However, CONCT did not influence the rate of sphincter-saving resection (p = 0.68) and resection with negative margins (R0, p = 0.089), the technical difficulty of the operation, and G3 or worse postoperative complications; however, CONCT significantly decreased pelvic fibrosis (17% vs. 2.4%, p < 0.001).23, 32 A pilot study by Lee et al.33 that used a historical control showed that compared with standard LC-nCRT and surgery (160 patients), the addition of one cycle of LF (5-FU + leucovorin) before standard LC-nCRT (45 Gy + two cycles of LF, 168 patients) and surgery for stage II/III RC patients significantly increased the sphincter preservation rate (80% vs. 93.5%, p < 0.001), the pCR rate (10% vs. 19%, p = 0.028), and DFS (79.5% vs. 67.5%, p = 0.042), and decreased the relapse rate (38.8% vs. 19.6%, p < 0.001) and deaths due to disease progression (31.9% vs. 10.1%, p < 0.001); however, T and N downstaging (p = 0.773 and 0.570), intraoperative and postoperative morbidity (p = 0.141) and OS (p = 0.572) were not significantly different. The compliance of chemotherapy was high (92.3%) and the AEs were tolerable, with 22.6% G3/4 neutropenia in the CONCT group.

These results showed that LC-nCRT followed by CONCT with mFOLFOX6 or LF increased the pCR and probably reduced the risk of disease progression, and that the AEs of CONCT were tolerable. However, Phase III randomized controlled trails (RCTs) are needed to prove the benefit of this regimen.

The ongoing KONCLUDE trial is investigating whether CONCT with mFOLFOX6 provides better pCR and 3-year DFS than adjuvant POCT alone with mFOLFOX6 for LARC patients who received nCRT before surgery (NCT02843191). Another ongoing RCT defined the primary endpoint as 3-year DFS for MRI-defined intermediate and high-risk rectal cancer patients who received standard LC-nCRT with continuous infusional 5-FU or oral capecitabine during RT (50.4 Gy in 1.8 Gy fractions) or standard LC-nCRT followed by three cycles of CONCT (mFOLFOX6) before TME (NCT04246684).

2.2 LC-nCRT (capecitabine-based) followed by CONCT before surgery

RT with concurrent capecitabine is an alternative standard nCRT regimen for LARC patients. Zampino et al.34 performed a single-arm Phase II study in which 51 LARC patients received capecitabine-based LC-nCRT, followed by two cycles of capecitabine CONCT before surgery. All patients finished the nCRT and 94% of them received two full cycles of CONCT. The sphincter preservation rate was 66% for patients with anal verges of 5 cm or less, the rate of R0 resection was 100%, and the rate of pCR was 18%. The rate of G3/4 AEs during nCRT or CONCT was 14% and the rate of postsurgical complications was 31%; the late AEs were mainly gastrointestinal and urogenital.

Another prospective single-arm study by Gao et al.35 examined high-risk LARC patients who received neoadjuvant RT with concomitant CT, CONCT, and POCT with XELOX (capecitabine + oxaliplatin). All patients completed the nCRT, 94% of them completed the CONCT, and 80% of them received POCT. However, the rate of pCR was 36.1% and the rate of postoperative complications was 19.4%. The estimated 3-year DFS and OS rates were 75.5% and 88.6%, respectively.36

Two single-arm studies by Zhu et al.37, 38 also showed that high-risk LARC patients who received nCRT and the CONCT with XELOX regimen had high compliance and good tolerance during the TNT, and the extent of postoperative complications was acceptable. Adding a dose of intensity-modulated radiation therapy (IMRT) had tendencies to improve the clinical complete response (cCR)/pCR rate, DFS, and OS; however, the pCR rate was lower than in study by Gao et al.35, 36 A retrospective study by Cui et al.39 showed that LARC patients who received nCRT followed by CONCT with two cycles of XELOX and TME had significantly improved pCR rate (19.3% vs. 4.9%, p = 0.01) and tumor downstaging (45.8% vs. 24.6%, p = 0.049) compared to those receiving standard nCRT and TME, but these groups had similar frequencies of AEs during and after treatment. Compared with LC-nCRT and CONCT with mFOLFOX, nCRT with synchronous and CONCT XELOX had a higher compliance rate and reduced AEs. Thus, Phase III RCTs should be performed to confirm the benefits of this regimen.

The multicenter, randomized, Phase III trial (CinClare study) was the first RCT to compare two regimens with nCRT plus CONCT for cT3-4/N+ LARC patients.24 The results indicated that compared with capecitabine-based nCRT followed by CONCT with one cycle of XELOX before TME, XELIRI (capecitabine + irinotecan)-based nCRT and one cycle of CONCT significantly improved pCR (15% vs. 30%, p = 0.001), but increased G3 AEs (6% vs. 38%, p < 0.001); although the two arms had similar rates of surgery (87% vs. 88%), R0 resection (96% vs. 97%) and G3 or worse postoperative complications (11% vs. 15%, p = 0.268). They also found that the pCR rate was 21.5% in the group that completed one to three cycles of weekly irinotecan and 33.6% in the group that completed four to five cycles.40

Three ongoing prospective studies are examining the advantage of CONCT with XELOX or mFOLFOX after nCRT in LARC patients. The EXPLORE study is comparing the pCR of LARC patients with MRI-defined high-risk patients (deeper infiltration, or extramural vessel invasion, or circumferential resection margin involvement, or surrounding organs and structures invaded, etc.) who receive capecitabine-based LC-nCRT followed by four cycles XELOX CONCT and then TME with those receiving LC-nCRT followed by TME (NCT03038256). Another trial is comparing the pCR and 3-year DFS in LARC patients receiving nCRT followed by CONCT (four cycles of mFOLFOX or three cycles of XELOX) with those receiving standard nCRT (NCT03957733). The CAO/ARO/AIO-16 trial is examining the effect of nCRT followed by CONCT on organ preservation in LARC patients (NCT03561142).

2.3 Neoadjuvant short-course RT (nSC-RT) followed by CONCT before surgery

nSC-RT is another scheme for LARC patients. The Poland II trial randomized 515 LARC patients to receive nSC-RT plus CONCT with three cycles of FOLFOX4 (Arm-A) or LC-nCRT synchronously with two cycles of FOLFOX4 (Arm-B). Arm-A had better RT compliance (100% vs. 92%, p < 0.001), lower acute AEs (any toxicity: 75% vs. 83%, p = 0.006), and better R0 resection (77% vs. 71%, p = 0.07); however, the two arms had similar RT and/or CT dose reductions and/or delays because of toxicity (37% vs. 34%, p = 0.40), G3-4 AEs (23% vs. 21%), rate of pCR (16% vs. 12%, p = 0.17), and postoperative complications. At 3 years, Arm-A had a better OS (73% vs. 65%, p = 0.046) and a lower cumulative incidence of death in patients with local or distant failure (23% vs. 31%, p = 0.049); however, the two arms had a similar 3-year rate of DFS (53% vs. 52%, p = 0.85), cumulative incidence of LR (22% vs. 21%, p = 0.82), and DM (30% vs. 27%, p = 0.25).41 After a median follow-up of 7.0 years, the two arms had no significant differences in 8-year rate of OS (49% vs. 49%, p = 0.38), DFS (43% vs. 41%, p = 0.85), LR (35% vs. 32%, p = 0.60), DM (36% vs. 36%, p = 0.54), and late complications (G3+: 11% vs. 9%, p = 0.66).42

Another matched pair analysis by Markovina et al.43 showed that LARC patients who were given TNT (SC-RT followed by four cycles of mFOLFOX-6) received significantly more total cycles of chemotherapy and more total cycles of FOLFOX than matched patients who received standard LC-nCRT with 5-FU-based concurrent CT, TME, and four or more cycles of adjuvant FOLFOX. The TNT significantly increased the rate of pT downstaging (75% vs. 41%, p < 0.001), the number of G3/4 hematologic AEs (22% vs. 0%, p < 0.001), 3-year DM-free survival (DMFS, 88% vs. 70%, p = 0.028), and DFS (85% vs. 68%, p = 0.032); however, the two groups had no significant differences in acute GI (p = 0.24), all nonhematologic toxicities (p = 0.23), postoperative and late toxicities, rate of sphincter preserving surgery (75% vs. 72%, p = 0.85), R0 resection (95.7% vs. 88.6%), pCR (28% vs. 16%, p = 0.17), and 3-year OS (96% vs. 88%, p = 0.67). A prospective single-arm Phase II trial by Myerson et al.44 reported similar rates of sphincter preservation (75%), R0 resection (95%), and pCR (25%). At their last follow-up, the LR rate was 5% and the freedom-from-disease relapse rate was 87%.

A retrospective study by Jia et al.45 showed that nSC-RT followed by CONCT was well tolerated (27% G3 radiation-related proctitis, 30% postoperative Clavien–Dindo grade 3 complications within 30 days of surgery) and associated with improved cCR (47%, nine of 19 patients) and pCR (35%, seven of 20).

A recent multicenter randomized Phase III trial (RAPIDO) found that compared with standard nCRT and POCT for stage cT3-4/N2 patients with LARC, nSC-RT followed by CONCT with CAPOX (six cycles) or FOLFOX (nine cycles) had higher compliance (75% vs. 58%) and more preoperative G3 AEs (48% vs. 25%), but these groups had no differences in surgical procedures or postoperative complications (50% vs. 47%).46 nSC-RT followed by CONCT significantly reduced disease-related treatment failure events (23.7% vs. 30.4%, p = 0.0019) and DM (20.0% vs. 26.8%, p = 0.0048); although it increased pCR (28% vs. 14%, p < 0.001), it did not significantly improve 3-year OS (89.1% vs. 88.8%, p = 0.59) or LR (8.3% vs. 6.0%, p = 0.12).47

The final report of another multicenter, randomized, Phase III trial (STELLAR) was presented at the 2021 ASCO meeting. This study compared SCRT (25 Gy/5 fractions/5 days) followed by four courses of CAPOX CONCT with LC-nCRT (50 Gy with concurrent capecitabine) before TME in 599 LARC patients with distal or middle third disease, T3–T4, and/or N+ RC. The SCRT + CONCT group had higher rates of completion (95.2% vs. 82.6%, p < 0.001), full-dose completion of preoperative treatment (93.2% vs. 74.8%, p < 0.001), pCR + cCR (22.5% vs. 12.6%, p = 0.001), and 3-year OS (86.5% vs. 75.1%, p = 0.036); however, these groups had similar pCR (16.6% vs. 11.8%, p = 0.134), 3-year DFS (64.5% vs. 62.3%), metastasis-free survival, and locoregional recurrence.48 Therefore, nSC-RT followed by CONCT before surgery may be a promising scheme for treatment of LARC patients.

A 2019 review by Kane et al. examined nSC-RT alone with NAC and concluded this regimen was probably beneficial for LARC patients.49 An ongoing trial of LARC patients is assessing the benefits of SCRT followed by CONCT consisting of two cycles over 6 weeks of CAPOX compared with standard LC-nCRT (ESCORT, NCT03676517).

3 INCT FOLLOWED BY NCRT

The second TNT regimen used for LARC patients is INCT followed by nCRT and then surgery (Table 2).

TABLE 2. Studies of induction chemotherapy followed by nCRT for LARC patients Study [year] cStage Design Pts (N) Neoadjuvant therapy regimen Adjuvant therapy pCR (dT/N), % SR/R0/SP, % Calvo [2006]50 cII–III Phase II RCT two-arm 52 FOLFOX4 (×2)+CRT(45–50.4 Gy+tegafur) LF (4 cycles) pT0:29 (dT:75; dN:40) 100/NR/68 62 CRT (45–50.4 Gy+tegafur) LF (4 cycles) pT0:8 (dT:58; dN:42) 100/NR/61 Marechal [2012]25 cT2–4/N+ Phase II RCT two-arm 28 mFOLFOX6 (×2 cycles)+CRT (45 Gy+5-FU) NR 25 (dT:46; dN:43) 96.4/96/NR 29 CRT NR 28 (dT:48; dN:55) 96.6/96/NR CONTRE [2017]51 cII–III Prospective single-arm 39 mFOLFOX6 (×8)+CRT(45[+5.4]Gy+CAP) None 33 (dTN:71) 97.4/98.5/NR Cercek [2014]52 cT3–4/N1–2 Prospective single-arm 61 mFOLFOX6 (×8)+CRT(45[+5/5.4 Gy]+XELOX) None 36 (dTN100) 100/NR/89.5 Cercek [2018]53 cII–III Retrospective cohort 320 CRT (45 Gy[+5/5.4 Gy]+5-FU or CAP) FOLFOX/CAPOX/FLOX (×4) 21* 90.3/100/NR 308 mFOLFOX6 (8)/CAPOX (5)/FLOX (×10)+CRT As above (4 cycles) 36 NR GCR-3 cT3–4 N0–2 Phase II RCT two-arm 56 XELOX (×4)+CRT (50.4 Gy+XELOX) None 14 (dT58) 100/86/NR [201054; 201555] 52 CRT (50.4 Gy+XELOX) XELOX (4 cycles) 13 (dT43) 100/87/NR EXPERT [2010]56 MRI high risk Prospective single-arm 105 XELOX (×4 cycles)+CRT (55 Gy+CAP) CAP (4 cycles) 21 (dTN76) 90.5/98/NR EXPERT-C [2012]57 MRI high risk