The investigators propose to conduct a randomised study on cT2, cT3a-b tumours less than 5 cm using two different techniques of radiotherapy boost following neoadjuvant chemoradiotherapy (nCRT) (CAP45): EBRT (9 Gy/5 fractions) or CXB (90 Gy/3 fractions). The endpoint will be organ preservation at 3 years without non-salvageable local pelvic recurrence. The proof of this concept will be of most benefit for all patients but especially for the elderly who usually are not fit for or keen to undergo major surgery.

The hypothesis of this study is to determine whether the addition of an endocavitary boost with CXB after standard treatment with nCRT, increases the chance of rectum and anus preservation by 20%-unites in early rectal adenocarcinoma without locally progressive disease (organ preservation in control arm 20%, in experimental arm 40%).

Main objective To demonstrate that neoadjuvant chemoradiotherapy in combination with a boost given with CXB (Arm B) is superior to the same neoadjuvant therapy plus a boost with EBRT alone (Arm A) in terms of rectum (organ) preservation without non salvageable local disease at 3 years post treatment start, or permanent deviating stoma.

Study Design Open-label, phase III, prospective, multi-centre, international, randomised 1:1, 2 arm study designed to evaluate the efficacy of a CXB boost versus an EBRT boost.

Rationale - current state of knowledge Current guidelines for cT2-T3 (clinical stage T2 and T3) low and middle rectal cancer recommend radical total mesorectal excision (TME) surgery that may involves permanent stoma or a low anterior resection with sometimes poor bowel function.

Randomised trials have shown that neoadjuvant (chemo)radiotherapy (nCRT) reduces the risk of local recurrence by more than half. At 3 years, it is close to or below 5% with acceptable toxicity at many centers. On the other hand, it provides no definite improvement in overall survival and does not increase the chances of conservative surgery.

Rectal adenocarcinoma is rather radioresistant and the dose required to achieve 50% sterilization is close to 90 grays (Gy), which is a high dose causing toxicities when given with external beam radiation therapy (EBRT). Among the radiotherapy techniques able to achieve safely such a high dose, Contact X-Ray Brachytherapy 50 Kv (CXB) is an appealing method.

The Lyon R96-2 randomised trial using CXB showed an increased clinical complete response (cCR) rate from 2% to 29% and an improvement by 30% the chance of avoiding a permanent stoma (72% vs. 42%). In addition, some patients achieving cCR were able to preserve not only the sphincter but the whole rectum (organ preservation) either after local excision or using only a "watch and wait" strategy.

Such a conservative approach is receiving a growing interest all over the world among colorectal cancer specialists. The extensive and pioneering work of Prof. Habr Gama in Brazil is presently considered as a reference for the use chemoradiotherapy and an EBRT boost (total dose 54 Gy/30 f/6 weeks) followed by watch and wait in case of cCR to preserve organ, i.E. to avoid surgery.

Research Hypothesis The investigators propose to conduct a randomised study on cT2, cT3a-b tumours less than 5 cm using two different techniques of radiotherapy boost following neoadjuvant chemoradiotherapy (nCRT) (CAP45): EBRT (9 Gy/5 fractions) or CXB (90 Gy/3 fractions). The endpoint will be organ preservation at 3 years without non-salvageable local pelvic recurrence. The proof of this concept will be of most benefit for all patients but especially for the elderly who usually are not fit for or keen to undergo major surgery.

The hypothesis of this study is to determine whether the addition of an endocavitary boost with CXB after standard treatment with nCRT, increases the chance of rectum and anus preservation by 20%-unites in early rectal adenocarcinoma without locally progressive disease (organ preservation in control arm 20%, in experimental arm 40%).

Main objective To demonstrate that neoadjuvant chemoradiotherapy in combination with a boost given with CXB (Arm B) is superior to the same neoadjuvant therapy plus a boost with EBRT alone (Arm A) in terms of rectum (organ) preservation without non salvageable local disease at 3 years post treatment start, or permanent deviating stoma.

Study Design Open-label, phase III, prospective, multi-centre, international, randomised 1:1, 2 arm study designed to evaluate the efficacy of a CXB boost versus an EBRT boost.

Randomisation:

Arm A: 3D conformal EBRT 45 Gy (1.8 Gy/fraction/5 weeks) with concurrent chemotherapy using capecitabine (825 mg/m2 bid, on radiation days).

A cone down EBRT targeting the Gross Tumour Volume (GTV) will deliver a boost dose of 9 Gy in 5 fractions. On week 14 after the start of treatment, the tumour response evaluation will guide the final strategy: surgery (radical TME or local excision) or watch-and-wait (W-W).

Arm B divided in 2 subgroups depending on the tumour diameter:

B1: If the tumour is < 3 cm, a CXB boost dose (90Gy/3 fractions/4 weeks) will be initially delivered to the tumour. After 2 weeks rest, patients will receive 3D conformal EBRT 45 Gy (1.8 Gy/fraction/5 weeks) with concurrent chemotherapy using capecitabine (825 mg/m2 bid, on radiation days). Clinical evaluation will be performed 3 weeks after the end of irradiation (week 14) and will guide the final strategy (surgery or W-W) as in arm A.

B2: If the tumour is ≥ 3 cm, patients will receive EBRT first 45 Gy (1.8 Gy/fraction/5 weeks) with concurrent chemotherapy using capecitabine (825 mg/m2 bid, on radiation days).

A CXB boost dose (90 Gy/3 fractions/4 weeks) will be delivered to residual tumour, after a rest of 2 weeks. On week 14 after the start of treatment, the tumour response evaluation will guide the final strategy (surgery or W-W) as in arm A. Adjuvant chemotherapy will be left to institution choice.

Number of subjects: Taking alpha=5% and bêta=7.5% with 10% patients not evaluable after randomisation, 236 patients (118 patients/arm) must be enrolled to show a 50% increase in the main endpoint at 3 years (from 20% to 40%).This study will show a hazard ratio of 0.56. Stopping rule: non-salvageable local recurrence rate > 10% checked by the independent Data Monitoring Committee every 80 patients.

Inclusion Criteria:

1. Adenocarcinoma of the rectum classified clinically T2, T3a, T3b (penetration in the mesorectal fat between 1 to 5 mm) by TNM classification (Tumour Node Metastase), < 5 cm largest diameter, < half rectal circumference (by MRI staging), N0-N1 (any node < 8 mm diameter on MRI), M0
2. Operable patient
3. Tumour accessible to endocavitary contact X-Ray Brachytherapy with a distance from the lower tumour border to the anal verge ≤ 10cm
4. 18 years or above
5. No comorbidity preventing treatment
6. Adequate birth control
7. Patient having read the information note and having signed the informed consent
8. Health care insurance available
9. Follow-up possible

Exclusion Criteria:

1. Inoperable patient
2. T1, T3cd, T4, T≥ 5cm, T≥ ½ circumference
3. Patient N2 at diagnosis or N1 with any node > 8 mm diameter
4. Patient presenting metastasis at diagnosis
5. Previous pelvic irradiation
6. Tumour with extramural vascular invasion
7. Simultaneous progressive cancer
8. Tumour invading external anal sphincter and within 1 mm, and the levator muscle
9. Patient unable to receive CXB or CRT
10. Tumour with poor differentiation (G3)
11. People particularly vulnerable as defined in Articles L.1121-5 to -8 of the French Healthcare Code, including: person deprived of freedom by an administrative or judicial decision, adult being the object of a legal protection measure or outside state to express their consent, pregnant or breastfeeding women
12. Any significant concurrent medical illness that in the opinion of the investigator would preclude protocol therapy
13. Patient with history of poor compliance or current or past psychiatric conditions or severe acute or chronic medical conditions that would interfere with the ability to comply with the study protocol
14. Concurrent enrolment in another clinical trial using an investigational anti-cancer treatment within 28 days prior to the first dose of study treatment