8. Adjuvant and Neoadjuvant Therapy – Colorectal Cancer

Therapy schemes

FOLFOX4: Folinic acid (FA) (200 mg/m2 as a 2-hour infusion, day 1 and 2) plus 5-FU (400mg/m2 as a bolus, then 600 mg/m2 as a 22-hour infusion; day 1 and 2) in combination with oxaliplatin (85 mg/m2 as a 2-hour infusion; day 1), repeated on day 15. 1 cycle lasts 2 weeks, 12 cycles total.

Modified FOLFOX6 scheme: Oxaliplatin (85 mg/m2 as a 2-hour infusion; day 1), folinic acid (400 mg/m2 as a 2-hour infusion, day 1) + 5-FU (400 mg/m2 as an IV bolus, day 1; then 2400 mg/m2 as a continuous 46-hour IV infusion). 1 cycle lasts 2 weeks, 12 cycles total.

XELOX: Oxaliplatin 130 mg/m2 on day 1; capecitabine 2x1,000mg/m2/day, day 1-14, repeated on day 22. 1 cycle lasts 3 weeks, 8 cycles total.

Background

Several randomised studies have demonstrated a significant reduction of the recurrence rate and an increase in overall survival when a combination of 5-FU and folinic acid was administered [834][836][O'Connell, M. J. et al. 1997].

The MOSAIC study (2,246 patients) compared adjuvant chemotherapy consisting of 5-FU/FA (LV5FU2) with a FOLFOX4 scheme (LV5FU2 + oxaliplatin 85 mg/m2) every 2 weeks for 12 cycles. In the study population as a whole, FOLFOX4 chemotherapy demonstrated a significant improvement in disease-free survival compared to LV5FU2 chemotherapy (73.3% vs. 67.4%, p=0.003) (10-year survival) [844][861]*. When focusing on stage III only, FOLFOX4 chemotherapy demonstrated a difference in disease-free survival of 7.5% (HR 0.78; 95% CI: 0.65-0.93; p=0.005). Overall survival was also significantly improved by FOLFOX4 chemotherapy in stage III, reflected by an increase of 4.4% (p=0.029) (250 deaths in the FOLFOX group (n=672) vs. 293 deaths in the 5- FU/FA group (n=675)). Four years after therapy, the rate of peripheral sensory neuropathy was 12% (grade I), 2.8% (grade II), and 0.7% (grade III) [844][861][863][864]*.

The NSABP C-07 study included 2,407 patients with stage II (28.6%) or stage III tumours who received either the Roswell-Park scheme with a weekly administration of 5-FU/FA as a bolus (3 cycles, 8 weeks each) or the same 5-FU/FA scheme with oxaliplatin 85 mg/m2 in weeks 1, 3, and 5 in an eight-week schedule (FLOX scheme). Patients in the FLOX group showed 20% fewer recurrences (p<0.04). Disease-free survival after four years was 73.2% for the FLOX group and 67.0% for the group of patients treated with 5-FU/FA [862]. When choosing between the different regimens, the side effects of the individual protocols should be considered. Due to the higher cumulative dose of oxaliplatin in the MOSAIC study, a slightly lower rate of level 3-4 neuropathies was observed in the NSABP study (12.4 vs. 8.4%). However, the rate of level 3 and 4 diarrhoea was three-fold higher in the bolus FLOX than in the infusional FOLFOX4 protocol (38% vs. 10.8%!). In the NSABP study, five patients (0.4%) died within the first 60 days after beginning chemotherapy due to chemotherapy-induced enteropathy [862]. While showing comparable efficacy, the toxicity of the FLOX protocol is not acceptable in comparison to that of the FOLFOX4 protocol. Hence, the FLOX protocol should not be used in adjuvant situations. Internationally, at this time preference is given to the modified FOLFOX6 scheme, which consists of a 46-hour continuous infusion of 5-FU after an initial 5-FU bolus on day 1. This way, the patient is spared the 5-FU bolus and the pump change on day 2 of the FOLFOX4 therapy.

After a follow-up of nearly 7 years, data of the NO16968 study (FU/LV vs. XELOX) show that the addition of oxaliplatin improves the 7-year OS from 67% to 73%. This long-term analysis thus confirms that the combination of the oral prodrug capecitabine in combination with oxaliplatin also plays a role in the adjuvant chemotherapy of stage III colon cancer [823]. A meta-analysis of 8,734 patients (NSABPC-08, XELOXA, X-ACT, and AVANT studies) showed no difference between DFS and OS relative to the administered fluoropyrimidine (5-FU/leucovorin vs. capecitabine). A multivariate analysis confirmed the results of the MOSAIC study, which showed that the addition of oxaliplatin is associated with a survival benefit [865].

Regarding the duration of adjuvant chemotherapy see 8.1.6.3.

Background

Even though 40% of all CRC patients are older than 75 years, only between 1-5% of randomised patients are aged ≥75 years in the adjuvant therapy studies. Whether adjuvant chemotherapy has a positive effect in patients aged ≥75 years was analysed on the basis of data from 5,489 patients from 4 data registers with UICC stage III CRC. The use of adjuvant chemotherapy in this group of patients led to an improvement in postoperative survival. The combination of 5-FU and oxaliplatin only led to an improved survival trend (HR 0.84; CI 0.69-1.04; p=NS), corresponding to an absolute improvement of survival of 5 percentage points after 3 years. An additional analysis of the data registers showed that more side effects were observed in elderly patients treated with oxaliplatin [866].

Retrospective analyses of the MOSAIC study [821] assessed the significance of supplementing the adjuvant therapy by oxaliplatin on the endpoints DFS (disease-free survival), TTR (time to recurrence) and OS (overall survival) in elderly patients. As a limiting factor it must be mentioned that patients aged >75 years could not be enrolled in the MOSAIC study. As a result, the subgroup of patients aged >69 years in the MOSAIC study accounted for only 315 of the total study population of 2,246 patients. This subgroup analysis did not show an advantage for oxaliplatin. With 1.10 (95% CI: 0.73- 1.56), the HR for OS was even to the advantage of the 5-FU/FA regimen alone. For DFS and TTR, the HR was 0.93 (95% CI: 0.64-1.35) and 0.72 (95% CI: 0.47-1.11), respectively. However, the significance level was not achieved, which may be due to the small sample size.

A similar analysis of the American NSABP C-07 study [867] showed no advantage for the FLOX therapy (5-FU bolus regimen plus oxaliplatin) compared to FU/FA therapy alone (HR 1.03; 95% CI: 0.77-1.36; p=0.87) in patients aged >69 years with regard to DFS. The same was found for overall survival (HR 1.18; 95% CI: 0.86-1.62; p=0.30). Here, the oxaliplatin/age interaction on OS was significant with a p-value of 0.039. 396 of the original 2,409 patients were older than 69 years. These data are supported by smaller, retrospective analyses. A Danish working group was able to retrospectively identify 191 patients aged >69 years who received adjuvant therapy with fluoropyrimidine (FP) or with FP plus oxaliplatin at a centre [868]. Here, too, the administration of oxaliplatin was associated with a shorter DFS (HR 0.58, p=0.016) and a shorter OS (HR 0.49; p=0.003) with an altogether higher toxicity (HR 3.69; p=0.001). A Canadian analysis of 90 patients aged >65 years showed that the benefit of adjuvant oxaliplatin therapy decreased with age [869].

The NO16968 study (FU/FA vs. XELOX) [823] was unable to demonstrate a negative effect of age on DFS or OS in the subgroup analysis. However, it is unclear to what extent this is related to capecitabine as a combination therapy drug.

A pooled analysis which assessed the effect of oxaliplatin in elderly patients [822] did not show a negative effect of age on the administration of oxaliplatin after adjusting for comorbidities. In the analysis of 904 patients from the NSABP C-08, XELOXA, X-ACT, and AVANT studies who received either fluoropyrimidine (FP) monotherapy or FP plus oxaliplatin, the interaction test for age and oxaliplatin was negative with regard to OS and DFS.

In summary, oxaliplatin should only be part in the adjuvant therapy in well-justified exceptions, for example, for patients without comorbidities.

Monotherapy with Fluoropyrimidines

Oral 5-FU Prodrugs:

Capecitabine 2 x 1,250 mg/m2 body surface area p.o. day 1-14, every 3 weeks. 1 cycle lasts 3 weeks, 8 cycles total.

Background

1,987 patients with stage III colon cancer were either randomised to the Mayo Clinic scheme (983 patients) or were given capecitabine monotherapy (1,004 patients) over a period of 24 weeks each (X-ACT study). The primary study objective was achieved by proving that capecitabine was at least equivalent to the Mayo scheme with regard to disease-free survival. The analysis showed a trend towards an improved disease-free survival with capecitabine (HR 0.87; 95% CI: 0.75-1.00; p=0.05). Furthermore, overall survival did not show a significant difference, but a trend towards the superiority of capecitabine was found (81.3% vs. 75.6%; p=0.05) [870]. A meta-analysis of 8,734 patients (NSABPC-08, XELOXA, X-ACT, and AVANT studies) did not show a difference in DFS and OS in connection with the administered fluoropyrimidine (5- FU/leucovorin vs. capecitabine) [865].

Even through a randomised study with UFT + folinic acid versus 5-FU/FA [876] did not detect a difference in overall survival and disease-free survival and Japanese meta- analysis of three studies even found a significant improvement of overall survival and DFS [877], UFT is currently not recommended, because it has not been approved for adjuvant chemotherapy of colon cancer in Germany.

Infusional 5-FU/Folinic Acid:

• LV5FU2 e.g. folinic acid (FA) (200 mg/m2 as a 2-hour infusion, day 1 and 2) plus 5-FU (400 mg/m2 as a bolus, then 600 mg/m2 as a 22-hour infusion; day 1 and 2) 1 cycle lasts 2 weeks, 12 cycles total
• 5-FU/folinic acid scheme e.g. folinic acid (FA) (500 mg/m2 as a 1-2-hour infusion) plus 5-FU (2,600 mg/m2 as a 24-hour infusion) 1x per week over a period of 6 weeks (day 1, 8, 15, 22, 29, 36). Repetition of therapy in week 9 (day 50).
• 2 cycles total.
• Protracted venous 5-FU infusion (PVI) e.g. 5-FU as a long-term infusion over 12 weeks total (300 mg/m2/day)

Background

Compared to the bolus schemes, several therapeutic studies with different types of infusional application show no difference to giving 5-FU/FA as a bolus in relation to disease-free and overall survival. However, the noticeably better toxicity profile obviously speaks in favour of the infusional route [871][872]* [873][874]. A comparison of a 12-week therapy with the protracted venous infusion (PVI) of 5-FU (300 mg/m2/day) versus a 6-month Mayo scheme showed no significant difference in recurrence-free survival (RFS) and in overall survival, while demonstrating lower toxicity for PVI 5-FU [875]. Beginning adjuvant chemotherapy within a period of 8 weeks after surgery showed a significant survival benefit [878]. The optimal duration of chemotherapy was 6 months [873][O'Connell, M. J. et al. 1998][880].

Background

Adjuvant therapy with an irinotecan-based protocol cannot be recommended on the basis of the data obtained from phase III studies [881], [882], [883].

Several studies have assessed the significance of monoclonal antibodies in the adjuvant situation. In the N0147 study, patients (N=1,863) with KRAS wild type tumours received adjuvant therapy with either mFOLFOX6 or mFOLFOX6 + cetuximab (12 cycles every 2 weeks). No positive effect of the adjuvant cetuximab therapy was demonstrated after a follow-up of 28 months: no difference was observed in either DFS (74.6% vs. 71.5%) or overall survival (87.3% vs. 85.6%). Significantly increased toxicity was observed in the cetuximab group. In the subgroup of patients aged over 70 years, the addition of cetuximab led to a marked decrease of DFS (86.2% vs. 72.5%). [884]. The addition of cetuximab to FOLFOX4 chemotherapy also did not improve the oncological outcomes in the PETACC-08 study [885].

The NASABP C-08 study compared the modified FOLFOX6 scheme (12 cycles every 2 weeks) with FOLFOX6 + bevacizumab. Bevacizumab was administered for 1 year in total. Patients with curatively resected colon cancer in stage II or III (75%) were enrolled in the study; patients aged over 70 years accounted for only 15.1% of the population. After a median follow-up of 5 years, this study did not show a significant improvement of DFS for the additional use of the antibody (75.1% vs. 77.9%), not even after a separate analysis of stage II and stage III. Overall survival was comparable in both groups [886]. The AVANT study (n=3,451) used a similar therapeutic design as the NSABP C-08 study. The primary endpoint was DFS. In this study, XELOX plus bevacizumab was tested as an additional third therapeutic arm. More serious side effects were reported in the groups treated with bevacizumab than in the group receiving FOLFOX therapy alone (25.5% vs. 20%). Within 60 days of starting the therapy, 2 patients died in the FOLFOX group, 4 patients in the FOLFOX4 plus bevacizumab group, and 6 patients in the XELOX plus bevacizumab group. The DFS was not improved by the addition of bevacizumab. A negative impact of the bevacizumab therapy on overall survival was even observed (HR for bevacizumab-FOLFOX4 vs. FOLFOX4 1.27 (p=0.02), bevacizumab-XELOX vs. FOLFOX4 1.15 (p=0.21)) [887]. The large QUASAR 2 study (Kerr) compared two therapeutic group with one another after curative therapy of stage II high-risk or stage III colon cancer (capecitabine vs. capecitabine plus bevacizumab). No difference in the 3-year DFS was established after a median follow-up of 4.92 years (75.4% vs. 78.4%). [888].

The intensification of postoperative chemotherapy by means of antibody therapy does not improve survival compared to combination chemotherapy alone. Monoclonal antibodies therefore have no role in the adjuvant therapy of colon cancer.

Background

At present there is no evidence to suggest that the addition of oxaliplatin leads to an improved efficacy of an adjuvant therapy in stage II (see section 8.1.4 UICC Stage II). Subgroup analyses of the MOSAIC study were conducted to assess the effectiveness of adjuvant chemotherapy with the FOLFOX4 scheme in stage II and in older patients (70- 75 years, N=315). 2,246 patients with a history of curative resection of stage II (N=899) or III colon cancer were enrolled in the MOSAIC study and received postoperative adjuvant therapy with either LV5FU2 or FOLFOX4 for 12 cycles. High-risk stage II was defined as: T4 tumour, tumour perforation, ileus, poorly differentiated tumour, blood vessel invasion or less than 10 lymph nodes examined. For the entire stage II population, the additional administration of oxaliplatin did not improve DFS or overall survival. The same was found for the subgroup of high-risk tumours and the group of older patients [821]. This outcome was also confirmed in the long-term analysis of the MOSAIC study for high-risk stage II: the addition of oxaliplatin did not lead to a prognostic improvement compared to LV5FU2 chemotherapy [844].

There are no randomised studies in patients with stage II colon cancer available for the oral 5-FU prodrug capecitabine. Stage II is characterised by a higher drug non- compliance than stage III [889].

Background

Rectal cancers in UICC stage I (T1/2N0) show a low rate of local recurrence and distant metastases when the treatment consists of radical surgery alone with adequate total mesorectal excision (TME) for tumours in the lower/middle third of the rectum or partial mesorectal excision (PME) for tumours in the upper third of the rectum (3% local recurrence and less than 10% distant metastases after 10 years [899]. For this reason, this early tumour stage has been excluded from modern, randomised studies assessing the role of neoadjuvant radiochemotherapy in locally advanced rectal cancer (cT3/4 and/or cN+) [901] [902] [903] [904] [905].

Nevertheless, randomised studies on preoperative short-term pre-radiation with 5x5 Gy versus surgery alone included tumour stage I. While long-term results of the Swedish study showed a significant benefit regarding local control for the additional radiation for UICC stage I as well (4.5% versus 14% after 13 years, p=0.009), the concept of TME had not yet been implemented in this study and the local recurrence rates in the study arm receiving surgery alone was unacceptably high [906]. The more recent Dutch TME study also showed a numerically significant reduction of the local recurrence rate for UICC stage I in the study arm with preoperative short-term radiotherapy (<1% vs. 3% after 10 years without preoperative radiotherapy, p=0.027) [899]; however, the local recurrence rates in this stage were low in both study arms. These low local recurrence rates in stage I were also confirmed in the most recent British study of short-term radiotherapy (MRC-CR07) (1.9% after 3 years with pre-radiation vs. 2.8% with immediately surgery, n.s.) [900].

For patients with deep-seated T1N0 high-risk cancer (G3/4, L1, V1, diameter larger than 3 cm, sm3) or with T2N0 tumours who refuse an extirpation, radio(chemo-)therapy followed by local excision/transanal endoscopic microsurgery or a wait-and-see strategy for clinically complete remission can be a treatment option [907][908][909][910][911]. This procedure, however, has not yet been validated and requires further prospective studies.

Background

A tumour is classified as rectal cancer if its aboral border is less than 16 cm away from the anocutaneous line, as measured by rigid rectoscopy [917]. Studies on the neoadjuvant therapy of rectal cancer have not used a standardised definition of the rectal third based on the distance of the lower border of the tumour to the anocutaneous line. In addition to the definition of ≤6 cm for the lower third, 6 to ≤12 cm for the middle third, and 12 to ≤16 cm for the upper third [902] [918] [919], the distances 0-5 cm, >5- 10 cm and >10-15 cm have also been used as limits [899] [900].

A German study on adjuvant and neoadjuvant radiochemotherapy (50.4 Gy in 28 fractions, 5-fluorouracil 1,000 mg/m2/day 1–5 in the first and fifth week of RT) of rectal cancer (up to 16 cm from the anocutaneous line) in UICC stage II and III (CAO/ARO/AIO-94) demonstrated a significant reduction of the local recurrence rate in the neoadjuvant arm (6% vs. 13% in the postoperative arm after 5 years, p=0.006) [901] [902]. The rate of postoperative complications was not increased for preoperative radiochemotherapy in comparison to immediate surgery; overall acute and chronic toxicity were significantly lower in the preoperative radiochemotherapy arm. For deep- seated tumours for which the surgeon had established an obligatory indication for extirpation prior to randomisation, the rate of sphincter-retaining surgical procedures was doubled by pretreatment in comparison to immediate surgery (19% for immediate surgery, 39% after neoadjuvant radiochemotherapy, p=0.004). In 2004, this study established neoadjuvant radiochemotherapy as a new standard for locally advanced rectal cancer in UICC stage II and III. Meta-analyses confirmed an improvement in local control by preoperative radiotherapy (with a biologically equivalent dose, BED, >30 Gy) compared to surgery alone or postoperative radiotherapy [915] [916]. For conventional fractionated preoperative radiotherapy (1.8-2 Gy single dose up to 45-50.4 Gy total dose), the simultaneous combination with 5-fluorouracil-containing chemotherapy was shown to be significantly superior with regard to local control in two randomised studies (HR 0.54, 95% CI 0.41-0.72) [903] [904] [905] [920]. This was confirmed in meta-analyses [912] [913] [914].

However, the improvement in local control did not translate into an improvement in disease-free survival and overall survival in any of these studies/meta-analyses. Long- term results in particular of the Swedish and Dutch studies on short-term preradiation showed a significantly decreased disease-related quality of life for a number of items after preradiation with 5x5 Gy compared to surgery alone (e.g. frequency of stool in patients without a stoma as per EORTC QLQ-CR-29: mean point score after surgery alone 19.4 vs. 26.3 after 5x5 Gy + surgery, p=0.006 [921] and a deterioration in functional results (sexual function: erectile dysfunction 30% after surgery alone vs. 50% after 5x5 Gy + surgery; faecal incontinence requiring pads: 37% after surgery alone vs. 56% after 5x5 Gy + surgery after a 14-year follow-up period [922] [923] [924] - however with an outdated radiation technique and radiation volumes).

For this reason, criteria for possible omission of preoperative radio(chemo)therapy in UICC stage II and III must be defined:

A problem of every neoadjuvant therapy is the potential “overstaging” and, thus, the resulting “overtreatment” of patients who have been misdiagnosed with lymph node positive tumour (cN+). The difficulty in imaging diagnostics for cN+ is due, among other reasons, to the occurrence of reactively enlarged lymph nodes and frequent micrometastases in healthy lymph nodes. Since the sensitivity and specificity for the evaluation of lymph node involvement are thus limited for all staging methods (MRI/endorectal ultrasound/pelvic CT: sensitivity 77%, 57%, 79%; specificity 76%, 80%, 76%) [925], primary surgery is considered an expedient option for cT1/2 tumours showing a questionable cN+ status (thus, formally UICC stage III) during imaging.

The extent of infiltration into perirectal adipose tissue can be used as further selection criterion for primary surgery in wall-penetrating cT3 tumours in the middle third of the rectum. For radial tumour infiltration below 5 mm (cT3a/b) determined in thin-layer MRI and an adequate distance to the mesorectal fascia, a risk of local recurrence comparable to that of stage I can be assumed for adequate TME [926]. Therefore, neoadjuvant radio(chemo)therapy can be omitted in this constellation if no additional risk factors (e.g. deep-seated tumour, confirmed lymph node involvement, EMVI) are present and the centre can provide high-quality MRI and TME surgery [927].

Background

A number of study groups have assessed the radial distance of the tumour (or involved lymph node) to the mesorectal fascia (= later circumferential resection margin, CRM) as a selection criterion for primary surgery without neoadjuvant therapy [928] [919]. If the distance is more than 1 mm (mrCRM-), a pathologically confirmed CRM-negative resection (pCRM >1 mm) is also achieved without pretreatment with adequate TME in more than 90-95% of cases. The Mercury Study Group reported a negative predictive value of MRI of 94% for the circumferential resection margin in primary surgery (pCRM- in 192 of 205 patients with mrCRM-). The absolute number of local recurrences after a median follow-up of 62 months in the group of 205 patients with negative mrCRM who underwent surgery alone was 7% (14/205) [928].

Both the mrCRM as determined by thin-layer MRI and the pCRM as determined by histopathology undoubtedly have a high prognostic value for local control as well as for disease-free survival and overall survival [928] [644]. However, subgroup analyses of large, randomised studies confirm that preoperative radiotherapy resulted in a further significant improvement of the local recurrence rate, especially when a pCRM-negative resection was achieved [899] [772]. A multivariate analysis within the scope of the British MRC-CR07 study on risk factors for locoregional recurrence identified the lymph node status, preradiation with 5x5 Gy, the position of the tumour relative to the anterior quadrants and the quality of surgical TME as independent prognostic factors, while pCRM was only confirmed as a significant factor in the univariate analysis [772]. An indication for neoadjuvant radio(chemo)therapy versus primary surgery based exclusively on the selection criterion mrCRM- therefore requires further quality- controlled, prospective studies.

The OCUM study group is assessing a risk-based indication for preoperative radiochemotherapy, which is carried out only for the mrCRM+ constellation within the scope of this non-randomised, prospective observational study for tumours >/=6 cm from the anocutaneous line irrespective of the T and N category, while patients with tumours in the lower third (<6 cm) from category T3 upwards receive neoadjuvant therapy. Of the 642 already enrolled patients, 389 (61%) have undergone primary surgery; the rate of pCRM- in these patients was 98% [919]. However, in the group with primary surgery, a total of 192 patients (49%) had early cT1-2 tumours and in 96 patients (25%) the tumour was located in the upper third of the rectum. Long-term results on local control and survival are not available yet.

Background

The following arguments speak in favour of treating the upper third of the rectum (defined variably as >10-15 cm, or >12-16 cm from the anocutaneous line, as measured by rigid rectoscopy) in the same way as colon cancer: Data from the American studies on adjuvant therapy which originally established postoperative radiochemotherapy for the treatment of locally advanced rectal cancer were based exclusively on rectal tumours with a distance of up to 12 cm between the distal edge of the tumour and the anocutaneous line [930] [931]. While the more modern studies on neoadjuvant radiotherapy/radiochemotherapy included tumours in the upper third, the Swedish and Dutch studies on short-term preradiation versus surgery alone showed no significant improvement of the local recurrence rate by additional radiotherapy for tumours located in the upper third of the rectum (defined here as >10-15 cm from the anocutaneous line) (8% vs. 12% for surgery alone after 10 years, p=0.3, in the Swedish study; 3.7% vs. 6.2% after 5 years, p=0.12, in the Dutch study) [906] [929]. In the British MRC CR07 study, however, the 3-year local recurrence rate for tumours >10-15 cm from the anocutaneous line after preradiation was 1.2% vs. 6.2% after primary surgery (HR 0.19; 95% CI: 0.07- 0.47) [900]. All in all, the local recurrence rate decreased with increasing distance to the anocutaneous line in all studies; it is thus justified to forgo general preradiation for tumours in the upper third of the rectum or to only establish a selective indication for this therapy in patients with risk factors for a local recurrence or with R1 resection (e.g. cT4, mrCRM+, cN2).

Background

In two randomised studies, pre-operative short-term radiation with 5x5 Gy over five consecutive days, immediately followed by surgery, was compared to preoperative, conventional fractionated 5-fluorouracil-based radiation radiochemotherapy (45 to 50.4 Gy in 25-28 fractions), followed by surgery after 4-8 weeks. A Polish study found a significantly superior result in relation to downsizing and downstaging (pCR rate 16% vs. 1%, p<0.001) as well as a significantly lower rate of CRM+ resections (4% vs. 16% after 5x5 Gy, p=0.02) after adjuvant radiochemotherapy; however, with an increased acute toxicity compared to short-term radiation (18% grade 3-4 toxicity vs. 3% after 5x5 Gy, p<0.001). The rate of sphincter-preserving surgical procedures (primary endpoint: 58% vs. 61% after 5x5 Gy), the local (16% local recurrences after 4 years vs. 11% after 5x5 Gy) and systemic tumour control (34.6% metastases after 4 years vs. 31.4% after 5x5 Gy), as well as late toxicity (grade 3-4 7% vs. 10% after 5x5 Gy) showed no significant difference in both arms [933] [934] [935]. These results were confirmed by the Australian Trans-Tasman study. Here, the primary endpoint, namely the rate of locoregional recurrences, showed no significant difference in both arms either (3-year results: 7.5% after 5x5 Gy vs. 4.4% after radiochemotherapy, p=0.24) [932]. The postoperative complications were similar in both study arms (53.2% after 5x5 Gy vs. 50.4% after radiochemotherapy); acute toxicity during radiochemotherapy was higher (e.g. grade 3-4 diarrhoea 1.3% vs. 14.2%, p<0.001) [Ansari, N. et al. 2017]. Within the first 12 months, no sigificant differences in the disease- related quality of life was established between the two study arms based on the EORTC questionnaires QLQ-C30 and OLC-C38 [936].

In principle, both fractionation schemes could thus be used for preoperative radiotherapy. In constellations in which tumour shrinkage prior to surgical procedures is desirable (e.g. T4 tumours, proximity of the tumour to the mesorectal fascia, distal tumours with an intention to preserve the sphincter), conventional fractionated radiochemotherapy with an interval before surgery should be preferred over short-term radiation immediately followed by surgery.

However, newer studies have shown that marked tumour regression is also observed with a longer interval between short-term radiotherapy and surgery. In the randomised Stockholm III study, surgery was performed either immediately after the administration of 5x5 Gy or after 4-8 weeks: the rate of ypT0 tumours was 2.1% following immediate surgery and 11.8% after a longer interval [937] [938]. A Polish phase III study compared conventional fractionated 5-fluorouracil-based radiochemotherapy (partly including oxaliplatin) with short-term radiotherapy followed by 3 cycles of neoadjuvant chemotherapy with FOLFOX-4 and surgery in week 12 in patients with clinically fixed T3 or T4 tumours. No significant difference was found in relation to the primary endpoint, the R0 resection rate, in both study arms (71% vs. 77%, p=0.07), nor in the local control and disease-free survival. The acute toxicity across all grades (1-4) was significantly lower in the arm with short-term RT and neoadjuvant chemotherapy (75% vs. 83%, p=0.006); the incidence of higher-grade 3-4 toxicity (diarrhoea, neutropenia) was similar in both arms overall (23% vs. 21%) [939]. Another randomised phase III study (RAPIDO) assessed short-term radiotherapy with 5x5 Gy, followed by six cycles of CAPOX and surgery in week 22-24 versus conventional radiochemotherapy with capecitabine and surgery in week 14-16 in high-risk patients as defined by MRI (T4, mrCRM+, N2, EMVI+) [Nilsson, P. J. et al. 2013]. Recruitment for this study was completed in June 2016; results are not available yet.

Background

For conventional fractionated preoperative radiotherapy (1.8-2 Gy single dose to 45- 50.4 Gy total dose), the simultaneous combination with 5-fluorouracil-containing chemotherapy was shown to be significantly superior to radiotherapy alone in terms of local control, but not in relation to disease-free survival and overall survival in two randomised studies (EORTC 22921, FFCD 9203) [903] [904] [905]. This has been confirmed in meta-analyses [913] [914]. In the EORTC 22921 and FFCD 9203 studies, patients in the preoperative combination arms received 5-fluorouracil as a bolus infusion at a dose of 350 mg/m2/day and folinic acid at a dose of 20 mg/m2/day in the first and fifth week of radiation over a period of 5 days each. In the German CAO/ARO/AIO-94 study, 5-fluorouracil without modulation was administered with folinic acid in the first and fifth week of radiation at a dose of 1,000 mg/m2/day as a 120-hour infusion [901].

Two subsequent, prospectively randomised studies assessed the replacement of infusional 5-fluorouracil by oral capecitabine [941] [O'Connell, M. J. et al. 2014] [Allegra, C. J. et al. 2015] (see evidence table in the Guideline Report). The German phase III study started as an adjuvant study in 2002 and was extended by a neoadjuvant stratum after publication of the CAO/ARO/AIO-94 study in 2004. In the adjuvant stratum, after having undergone resection, patients first received two cycles of capecitabine at a dose of 2,500 mg/m2/day over a period of 14 days (repeated on day 22), followed by radiochemotherapy (1.8 Gy to 50.4 Gy) with capecitabine 1,650 mg/m2/day throughout the period of radiotherapy and another three cycles of capecitabine. Patients in the 5-fluorouracil arm received two 5-FU bolus cycles after surgery (500 mg/m2/day 1-5, repeated on day 29), followed by radiotherapy with infusional 5-FU at a dose of 225 mg/m2/day throughout the period of radiotherapy and two additional 5-FU bolus cycles. In the neoadjuvant stratum, radiochemotherapy was carried out with capecitabine, followed by surgery and five adjuvant cycles of capecitabine at the same doses as in the adjuvant stratum. The neoadjuvant cohort in the infusional 5-FU arm underwent preoperative radiochemotherapy analogously to the CAO/ARO/AIO-94 study (1,000 mg/m2/day as a 120-hour infusion in the first and fifth week). Following surgery, four 5-FU bolus cycles were administered (500 mg/m2/day 1- 5, repeated on day 29). Overall survival was defined as the primary endpoint of the study (non-inferiority limit 12.5%). In 392 evaluable patients and after a median follow-up of 52 months, the 5-year survival was 76% in the capecitabine arm and 67% in the 5- fluorouracil arm (p=0.0004 for non-inferiority). In the neoadjuvant stratum (n=161 patients), signs of an increased anti-tumour effect were observed in the capecitabine group: pathological complete remission (pCR) was more frequent in the capecitabine arm (14% vs. 5%, p=0.09) and more patients had node-negative tumours in the resection (71% vs. 57%, p=0.08) [941]. The results of the American NSABP R04 study, which assessed the use of oxaliplatin (see below) in addition to neoadjuvant radiochemotherapy with infusional 5-fluorouracil (225 mg/m2/day during RT) versus capecitabine (1,650 mg/m2/day during RT) in a 2x2 factorial design, was confirmed by the German study with regard to surgical results, pCR rates and the tolerability of the therapy [O'Connell, M. J. et al. 2014]. The rate of locoregional recurrences after 3 years (4.0% vs. 3.9%) as well as disease-free survival (66.4% vs. 67.7%) and overall survival (79.9% vs. 80.8%) after 5 years was nearly identical in the arms with infusional 5-FU and capecitabine [Allegra, C. J. et al. 2015].

A large number of phase II studies have shown pCR rates of up to 30% for neoadjuvant 5-fluorouracil-based or capecitabine-based radiochemotherapy including oxaliplatin (summary: [Rodel, C. et al. 2012]). The significance of these combination therapies was subsequently assessed in a total of seven randomised phase III studies [O'Connell, M. J. et al. 2014] [Allegra, C. J. et al. 2015] [Aschele, C. et al. 2011] [C, Aschele et al. 2016] [Gerard, J. P. et al. 2012] [Francois, E. et al. 2016] [965] [918] [Jiao, D. et al. 2015] [953]. These studies differed mainly in the administration route and dosage of 5-fluorouracil/capecitabine and oxaliplatin during simultaneous radiochemotherapy. Five of the studies assessed the significance of oxaliplatin only during neoadjuvant radiochemotherapy; in two studies, randomisation also included adjuvant chemotherapy with or without oxaliplatin [918] [965]. The studies also differed with regard to the chosen primary endpoint (pCR rate [Gerard, J. P. et al. 2012], local control [Allegra, C. J. et al. 2015], disease-free survival [918] [865] [953], overall survival [C, Aschele et al. 2016] [Jiao, D. et al. 2015]), and the resulting sample size calculations and their follow-up periods. The German CAO/ARO/AIO-04 [918] and Chinese FOWARC [953] studies determined a significant improvement of the pCR rate as the secondary endpoint after neoadjuvant radiochemotherapy including oxaliplatin, while five additional individual studies showed no significant difference in the pCR rates. A meta-analysis including the ACCORD 12, STAR-01, NSAPB R-04 and CAO/ARO/AIO-04 studies described a significant increase in the pCR rate (HR 1.20; 95% CI, 1.01-1.42; p=0.04), but also a significant increase in grade 3-4 acute toxicity (HR 2.29; 95% CI, 1.31-4.00, p=0.004) after neoadjuvant radiochemotherapy with oxaliplatin [An, X. et al. 2013].

The German CAO/ARO/AIO-04 study demonstrated a significant improvement of disease-free survival (as the primary endpoint). A meta-analysis of four studies with details on the disease-free survival described a marginally significant effect on this endpoint (HR 0.89; 95% CI: 0.78-1.00, p=0.05) [Yang, Y. J. et al. 2016]. However, five of the seven individual studies showed no significant improvement of disease-free survival by including oxaliplatin; long-term results on the FOWARC study are still pending regarding this. All in all, an advantage of using oxaliplatin has not been confirmed.

Background

In the prospective studies to establish preoperative radiochemotherapy, surgery was performed within 3-10 weeks (median: 5.4 weeks; EORTC 22921, FFCD 9203 [903] [905]) or 6 weeks (CAO/ARO/AIO-94 and CAO/ARO/AIO-04 [901] [918]) after completing neoadjuvant therapy. The recently published GRECCAR-6 study randomised 265 patients after neoadjuvant radiochemotherapy (45-50 Gy, infusional 5-FU or capecitabine) into two arms with surgery after 7 weeks vs. 11 weeks [942] (see evidence table in Guideline Report). There was no significant difference in the primary endpoint of this study, the pCR rate (15% after 7 weeks vs. 17.4% after 11 weeks, p=0.6); however, more postoperative complications occurred after surgery in week 11 (32% vs. 44.5%, p=0.04) and the quality of TME was worse (complete TME 90% vs. 78.7%, p=0.02). Long-term results on oncological endpoints have not yet been published for this phase III study. Two meta-analyses and a systematic review based on non-randomised data with secondary analyses of prospective studies or observational studies suggest that an interval exceeding the usual 6-8 weeks after radiochemotherapy appears to be associated with an increased pCR rate, with an otherwise consistent complication rate [943] [944] [945] (see evidence table in the Guideline Report). Differences in the local recurrence rate, disease-free survival or overall survival were not observed. There is thus no higher level of evidence to support an increase in the interval between the conclusion of radiochemotherapy and surgery beyond the standard 6-8 weeks.

In the prospective studies to establish short-term radiotherapy with 5x5 Gy, surgery should be performed within one week of completing radiotherapy. A secondary analysis of the Dutch TME study showed that elderly patients (=/>75 years) in particular exhibited poorer survival if they underwent surgery more than 3 days (i.e. surgery on day 4-7) after completing radiotherapy [946]. For this reason, a maximum interval of 10 days is recommended as the total treatment time (start of RT until surgery) for this group.

In the randomised Stockholm III study, surgery was performed either immediately after the administration of 5x5 Gy or after 4-8 weeks: the rate of ypT0 tumours was 2.1% after immediate surgery and 11.8% after a longer interval; there was, however, no difference in the rate of pCRM+ resection and abdominoperineal extirpation [938]. Postoperative complications were lower after 5x5 Gy and a longer interval than after immediate surgery, and there were no differences in the long-term results on oncological endpoints (local recurrences, distant metastases) [Erlandsson, J. et al. 2017] (see evidence table in the Guideline Report). Another randomised, monocentric study from Poland with a total of 154 patients also revealed a significant improvement of tumour downstaging after a longer interval (4-5 weeks after 5x5 Gy), however, with an identical R0 resection rate, rate of sphincter-preserving surgical procedures and no significant difference in overall survival (primary endpoint) [Pach, R. et al. 2012]. The results of the systematic review by Bujko et al. do not allow for a definite conclusion owing to the heterogeneity of the included studies [Bujko, K. et al. 2014]. On the whole, a total treatment time of 10 days should be pursued for sole preoperative radiotherapy with 5x5 Gy and surgery. However, according to the publication of the Stockholm III study, a longer interval until surgery of 4-8 weeks can also be pursued and is a particularly appropriate alternative to conventional fractionated radiochemotherapy in patients requiring downsizing who are unsuitable for or refuse chemotherapy. Regarding the addition of neoadjuvant chemotherapy after 5x5 Gy and a longer interval to surgery, see the background text in recommendation 8.24). For patients showing clinically complete remission after neoadjuvant therapy and prior to planned surgery based on the digital rectal examination, the rectoscopy and thin-layer pelvic MRI, a “wait-and-see” strategy without radical surgery is being increasingly discussed. This treatment concept is currently being assessed in prospective studies and has not yet been validated. If the patient refuses to have surgery (at least in terms of an extirpation) in spite of being informed about the unclear data situation, close surveillance involving a digital rectal examination, rectoscopy and thin-layer MRI every three months during the first two years is to be recommended to facilitate early recognition and treatment of a local tumour growth (“re-growth”).

Comment on performing an MRI after completion of neoadjuvant therapy: Due to insufficient data, this issue was not discussed by the working group. A case-by-case decision should be made as required.

Background

Until now, only few studies have tested neoadjuvant chemotherapy before or after radiochemotherapy, or as sole therapy without RT/RCT. A small, randomised phase II study from Spain (n=108) compared 4 cycles of induction chemotherapy with capecitabine/oxaliplatin followed by radiochemotherapy with capecitabine/oxaliplatin and surgery, with the “classical” sequence (radiochemotherapy, surgery, adjuvant chemotherapy) with otherwise identical substances and dosages (evidence table 3) [947] [948]. There was no difference in the primary endpoint, the pCR rate and oncological long-term endpoints (for which, however, this randomised phase II study was not sufficiently powered), but the study did find a significant improvement in the tolerability and practicability of induction chemotherapy compared to adjuvant chemotherapy. Phase III studies on this question have not been conducted.

A prospective cohort study of the “Timing of Rectal Cancer Response to Chemoradiation Consortium” in the USA assessed the administration of zero, two, four and six cycles of FOLFOX chemotherapy, followed by surgery 6, 11, 15 and 19 weeks after completion of radiochemotherapy (see evidence table in the Guideline Report) after neoadjuvant radiochemotherapy with infusional 5-fluorouracil. The pCR rate increased successively in the cohorts (18%, 25%, 30%, 38%) without a simultaneous increase in surgical complications [949]. Two randomised phase II studies are currently (date of information: 04/2017) recruiting patients with the aim of assessing which sequence (neoadjuvant chemotherapy before or after radiochemotherapy) is superior in terms of efficacy, practicability and toxicity [950] and CAO/ARO/AIO-12. Results are expected in 2018.

Phase II studies [951] [952] and the randomised FORWARC study [953] have yielded R0 resection, downstaging and pCR rates for sole neoadjuvant chemotherapy without radio(chemo)therapy. A randomised phase III study (PROSPECT) in the US is currently assessing sole neoadjuvant chemotherapy (6 cycles of FOLFOX) versus standard radiochemotherapy with 5-FU/capecitabine in selected patients (cT1/2N1; cT3N0/N1 with mrCRM-, sphincter retention possible). Data on this study are expected in 2018/2019.

Regarding neoadjuvant chemotherapy after 5x5 Gy for primary rectal cancer without synchronous distant metastases, see the background text on recommendation 8.15. In case of a primary rectal tumour with synchronous, (potentially) operable metastases, radiotherapy with 5x5 Gy, followed by early effective systemic therapy and surgical treatment of the primary tumour/metastases within a reasonable time period, can be an expedient option [954].

Background

In all randomised studies on the adjuvant therapy of rectal cancer, patients with UICC stage I following R0 resection were excluded due to an altogether low rate of local recurrences and distant metastases (3% or less than 10% after 10 years [899]).

Background

Phase III studies performed in the 1970s and 1980s on the adjuvant, multimodal therapy of UICC stage II and III rectal cancer (pT3-4 and/or pN+, up to 12 cm from the anocutaneous line) showed a reduction of the local recurrence rate (absolute difference: 10-15%) as well as an improvement of overall survival through the addition of 5- fluorouracil-based simultaneous and adjuvant chemotherapy to postoperative radiotherapy compared to surgery alone or unimodal adjuvant therapy (absolute difference: 5-15%) [summary: [955]]. However, the surgical principles of total mesorectal excision were not implemented in these outdated studies, and the local recurrence rates after surgery alone were high. A pooled subgroup analysis of five of these early North American studies also suggests that especially patients with pT1-2N1 and pT3N0 tumours following R0 resection do not benefit from the additional radiotherapy [956].

Regarding the value of adjuvant radiochemotherapy in patients with UICC stage II and III cancer, no modern studies with quality-controlled TME surgery have been conducted since the establishment of neoadjuvant radio(chemo)therapy. Since postoperative radiochemotherapy is less effective and is associated with considerably more side effects than preoperative radiotherapy (see background text on 8.18)) [901], the benefit/risk ratio of postoperative radiochemotherapy when strictly applied according to the TME principles has not been sufficiently clarified [957]. A general recommendation in favour of postoperative radiochemotherapy for all patients with UICC stage II and III tumours can therefore not be given. This therapy should be reserved for patients with histopathologically confirmed risk factors for an increased risk of local recurrence (e.g. R1 resection, intraoperative tumour tear, pCRM+, inadequate TME quality, pT4, pT3c/d, pN2, extranodal tumours in the mesorectum, pT3 in the lower third of the rectum).

Background

A systematic review of postoperative chemotherapy versus observation after curative resection of rectal cancer (including 21 randomised studies between 1975-2011 and a total of 9,785 patients) confirmed a significant reduction in the mortality rate (HR 0.83, 95% CI: 0.76-0.91) and the recurrence rate (HR 0.75; 95% CI: 0.68-0.83) after 5- fluorouracil-based adjuvant chemotherapy [Petersen, S. H. et al. 2012]. However, the studies included in the review are very inhomogeneous and also involve studies with preoperative and postoperative radiotherapy, and do not allow for differential statements according to UICC stages.

The randomised QUASAR study on adjuvant chemotherapy (5-FU + folinic acid or levamisole) versus observation in patients with colorectal cancer and an unclear indication for chemotherapy (usually stage II) showed a significantly reduced recurrence rate (HR 0.68, 95% CI: 0.52-0.88) and an improvement of overall survival (HR 0.77, 95% CI: 0.54-1.00) [842] after a median follow-up of 5.5 years for the subgroup of patients with rectal cancer (n=948). However, 203 of these patients had also received preoperative radiotherapy and 264 patients had received additional postoperative radiotherapy. QoL surveys with regard to typical side effects of chemotherapy (diarrhoea, nausea, fatigue, anorexia, dry mouth) showed a significant deterioration in the therapy versus the observation arm for all categories, but were limited to the time of chemotherapy. Two randomised Japanese studies showed a significant advantage of postoperative chemotherapy (with tegafur-uracil for 1 year) versus observation with regard to disease-free and overall survival for patients with UICC stage III rectal cancer after TME and selective, extended lateral lymph node resection [Akasu, T. et al. 2006] [Hamaguchi, T. et al. 2011].

The evidence for adjuvant chemotherapy in patients with rectal cancer (without neoadjuvant therapy and without an indication for postoperative radiochemotherapy) is thus considerably lower than for colon cancer. Even though no phase III studies involving the use of modern adjuvant combination chemotherapies as sole adjuvant therapy after primary resection of rectal cancer without neoadjuvant therapy are available, expert consensus recommends proceeding according to the indication criteria and schemes of colon cancer in this constellation.