A phase I dose escalation trial using intensity-modulated radiotherapy with simultaneous integrated boost in pelvic chemoradiotherapy for metastatic rectal cancer

Background In unresectable metastatic rectal cancers, the surgery of the primitive tumor remains highly debated. Chemoradiotherapy (CRT) of the primitive could allow sucient local control in order to avoid major and sometimes mutilating surgery. Dose escalation during CRT could increase this local control. The aim of this study was to evaluate the feasibility and tolerance of a CRT with radiation dose escalation delivered in intensity modulated radiotherapy (IMRT) with simultaneous integrated boost (SIB), in metastatic low and middle rectal cancers. Methods This multicenter phase I study included six patients treated for unresectable synchronous metastatic low and middle rectal adenocarcinoma in two dose levels. Radiotherapy was delivered using IMRT with SIB. The dose escalation was 52.5 Gy (level 1) and 56.25 Gy (level 2) in the primary tumor, in 25 fractions of 2.1 Gy and 2.25 Gy, respectively. High-risk clinical target volume (CTV) and low-risk CTV received respectively 50 Gy and 45 Gy in 25 fractions in the two levels. Concomitant chemotherapy was oral capecitabine and CRT was performed after four cycles of mFOLOX6 chemotherapy. The dose-limiting toxicity (DLT) was dened by a toxicity requiring the interruption of radiotherapy for more than ve consecutive fractions. All six patients received the full course of treatment at scheduled doses. No patients had acute toxicity requiring interruption of radiotherapy therefore no DLT has been reported. No patients had acute toxicity ≥ 3. Concerning late toxicity, three patients experienced grade 3. After CRT, four patients had a partial response and one patient had a complete clinical response. Two patients were considered in local progression at 9.4 months and 20.4 months of inclusion. Dose escalation at 56.25 Gy in the tumor lesion was possible with good acute tolerance. It needs to be evaluated in a larger study. It could allow sucient local control in order to avoid mutilating surgery in these metastatic patients.

A phase I dose escalation trial using intensity-modulated radiotherapy with simultaneous integrated boost in pelvic chemoradiotherapy for metastatic rectal cancer The maximum recommended dose (MRD) was de ned as the level immediately below the level at which two toxic limiting doses (DLTs) occurred or the last level if two DLTs did not occur.

Patients
Patients included in the study had histologically con rmed lower or middle rectum adenocarcinoma, with synchronous metastases deemed unresectable. They must be over 18 years old, have an estimated life expectancy of more than three months, a performance status according to WHO from 0 to 2. They should not have received previous treatment with pelvic radiotherapy or chemotherapy, have a complete de ciency of dihydropyrimidine dehydrogenase (DPD), have a severe or unstable disease, or have diarrhea or neuropathy grade ≥ 2 at baseline.

Intervention
Chemotherapy Before CRT, patients received four cycles of mFOLFOX6 chemotherapy administered every two weeks. This induction chemotherapy could be combined with targeted therapy (bevacizumab, cetuximab, panitumumab) based on KRAS / NRAS status. CRT started within two to four weeks after these four cycles of mFOLFOX6. Concomitant chemotherapy consisted of capecitabine at a dose of 800 mg / m² twice daily, ve days a week. Targeted therapies were not allowed during radiation therapy. First-line metastatic chemotherapy was resumed after the end of CRT.

Radiation therapy
Irradiation was delivered in intensity modulated radiotherapy (IMRT) with simultaneous integrated boost (SIB) (Fig. 3). The treatment was delivered in 25 fractions, ve per week, and one per day, over ve weeks. The gross tumor volume (GTV) was de ned using pre-chemotherapy 18FDG positron emission tomography (PET), ensuring concordance with clinical examination, rectoscopy, endorectal ultrasound, computed tomography (CT) and pelvic magnetic resonance imaging (MRI).
The dose escalation involved increasing the dose in GTV from 52.5 Gy to 60 Gy in fractions of 2.1 to 2.4 Gy, respectively, over the three dose levels (Table 1). The high-risk clinical target volume (CTV-HR) received a dose of 50 Gy (2 Gy per fraction). It corresponded to GTV with a margin of 10 mm excluding unaffected organs (prostate, uterus, vagina, bladder, and sacrum).
The low risk CTV (CTV-LR) received a dose of 45 Gy in 25 fractions (1.8 Gy per fraction). It included mesorectum, internal iliac ganglionic areas and high risk CTV, with the following limitations (Appendix 1): cranial: junction S1 / S2 (possibly S2 / S3 for a lesion of the lower rectum N0); dorsal: concavity of the sacrum (not including foramen); ventral: prostate in men, uterus and rectovaginal septum in women, bladder, ureters; lateral: pelvic wall ; caudal: 4 cm under the GTV (pre-chemotherapy). For low-rectal tumors, CTV-LR was extended to ischiorectal fossae (entire anal canal with a margin of at least 10 mm) to the coccyx back. The anal canal was excluded for a lesion of the middle rectum. The anal margin was excluded, except in case of massive invasion of the anal canal.
The planning target volume (PTV) was de ned by an automatic margin of 5 to 7 mm around previous volumes: PTV 1 = CTV-LR + 5-7 mm; PTV 2 = CTV-HR + 5-7 mm; PTV 3 = GTV + 5-7 mm. PTV 3 was therefore the volume receiving the highest dose by SIB (from 52.5 Gy to 60 Gy at 2.1 to 2.4 Gy per fraction depending on the level).
Dosimetry followed the recommendations of ICRU 83. Image-guided radiation therapy was mandatory with daily image guidance. 3D imaging was recommended, at least for the rst fractions.

Evaluation
The primary objective of the study was to determine the maximum recommended dose of radiotherapy delivered in IMRT with SIB during CRT. The primary endpoint was dose-limiting toxicity (DLT), de ned as the occurrence of toxicity requiring cessation of radiation therapy for more than ve consecutive fractions.
The secondary objectives were: acute (up to 3 months after the end of CRT) and late toxicity according to NCI-CTCAE V4.0; the local response; local progression-free survival at 12 months; 2-year overall survival; local surgery; the quality of life assessed by the QLQ-C30 and QLQ-CR29 questionnaires at inclusion, at the end of treatment and at follow-up.
Patients were followed clinically and biologically every two weeks during induction chemotherapy and weekly during CRT. An interim assessment was performed in the four weeks prior to CRT, including CT and 18FDG PET-CT. An end-of-treatment assessment was made within six weeks of the end of CRT, including clinical and biological evaluation, rectal echo-endoscopy (or rectoscopy), CT, pelvic MRI, 18FDG PET-CT, quality of life questionnaires. Then follow-up was done at 12 weeks and then every eight weeks for two years with at least a TAP CT scan, a biological assessment and quality of life questionnaires.

Ethics
This study received a favorable opinion from the committee for the protection of persons (CPP) (17 February 2014) and was authorized by the national agency for the safety of medicines (ANSM) ( rst April 2014). An independent committee has been appointed. All patients received from oral and written information, and signed a consent. The study was retrospectively registered on ClinicalTrials.gov (NCT03634202) (https://www.clinicaltrials.gov/ct2/show/NCT03634202) (16 August 2018).

Primary objective
A total of seven patients were included in the trial between May 2015 and February 2017 in our institution. One patient was wrongly included and excluded from the study because he did not meet the inclusion criteria (grade 2 diarrheas at baseline) and was progressing before CRT. Finally, three patients were included in level 1 and three in level 2. No patients were included in level 3. The characteristics of the population and initial disease are detailed in Table 2. The median follow-up was 27.4 months.  Acute tolerance of treatment was good with no acute toxicity ≥ 3 (Table 4). Of the six patients, four showed an acute gastrointestinal toxicity grade 1, and one patient a grade 2. On the urinary side, two patients had acute grade 1 toxicity and one patient grade 2. Note also grade 1 hematological toxicities in ve patients and grade 2 in one patient, and also grade 1 neurotoxicity in two patients, to be related to chemotherapy prior to CRT.  Regarding the late toxicity, 3 patients presented a grade 3. The rst patient, included in level 1, developed pelvic pain and pre-occlusive syndrome 3.9 months after the end of CRT, requiring a colostomy. The second patient, included in level 2, presented pelvic pain 4.1 months after the end of RT-CT, management resulted in a colostomy and later posterior pelvectomy. The third patient, enrolled in level 2, presented a recto-vesical stula 12.9 months after the end of CRT, requiring a suprapubic catheter. Note also grade 3 neurotoxicity in one patient to relate to oxaliplatin.
On the endoscopic evaluation after CRT, four patients had a partial response, one patient had a complete response, and one patient was not evaluable.
Of the six patients, two patients were classi ed in local progression. The rst, included in level 1, has been classi ed as progressive on PET-CT at 20.4 months of inclusion. However, the biopsies did not show any carcinomatous proliferation but only brous and cicatricial changes. During the follow-up, he had no local complications and no need for surgical management. The second patient, included in level 2 and considered in local progression at 9.8 months of inclusion on PET-CT, was nally operated by posterior pelvectomy. It was probably a false positive because the surgical specimen did not show any tumor in ltration but only cicatricial changes. The other four patients did not show any local progression during their follow-up. For the two patients classi ed in local progression, CRT allowed to increase local control compared with metastatic disease. Indeed, local progression-free survival was increased to 20.4 and 9.8 months compared with distant progression-free survival at 14.1 and 4.3 months, respectively.
Regarding overall survival, ve out of six patients were alive at 2 years of inclusion. At the time of the analysis, three patients had died with an overall survival ranging from 14.6 to 32.8 months.
Quality of life scores provided a qualitative assessment of several dimensions. The functional score for physical activity was improved between inclusion and the end of CRT for four out of six patients, a patient was stable, and one patient had a decrease just after CRT before increasing to a higher level than baseline. Pain appeared to decrease after CRT compared to baseline (Fig. 4). Patients with digestive symptoms at inclusion such as diarrhea, atulence, faecal incontinence, were improved at the end of CRT (Fig. 5). Two patients with no diarrhea at baseline experienced this symptom increase at the end of CRT before normalizing at the next assessment.

Discussion
This trial showed that a dose escalation with a SIB-IMRT technique for CRT of rectal cancer is possible at a dose of 56.25 Gy in the tumor lesion with good acute tolerance. Indeed, no interruption of radiotherapy and no grade ≥ 3 of acute toxicity was reported. The SIB technique made it possible to increase the dose without increasing the duration of the CRT. The good tolerability and the conserved duration of CRT allowed early resumption of rst-line metastatic chemotherapy after CRT. Nevertheless, three patients had late Grade 3 toxicity. Of these three patients, two had received anti-angiogenic treatment (bevacizumab) and one anti-EGFR treatment (panitumumab) with chemotherapy prior to CRT. The other three patients without reported late toxicity, had received chemotherapy by mFOLOX without targeted therapy associated before CRT. The other three patients without reported late toxicity had received chemotherapy by mFOLOX without targeted therapy associated before CRT. The involvement of these targeted therapies, administered just prior to CRT, needs to be assessed in terms of toxicity. The patient with recto-vesical stula had undergone colostomy surgery before inclusion in the trial. It should also be remembered that the included population is that of metastatic patients, with heavy treatments, and are potentially more fragile and prone to complications. Patients in the study also had advanced tumors (T3 or T4) with large volumes to irradiate. Patients with a less advanced disease could potentially bene t from higher dose escalation in a smaller volume, with good acute and late tolerance.
The dose escalation presented in the study provided good local control of the disease. Indeed, four out of six patients did not show any local progression and for the two patients classi ed in local progression, the histology was negative and the local progression was later than the distant progression. This doseescalation CRT strategy could provide su cient local control in metastatic patients to avoid heavy and mutilating surgery throughout their management.
This study has several limitations. First, it was a phase I feasibility study, the size of the population was therefore reduced. It is necessary to evaluate the dose escalation retained in a phase II study with a larger population. Then there was no inclusion in level 3. The trial was prematurely closed due to the di culty of including the target population, which can be explained by competitiveness with other clinical trials, and acceptable follow-up in included patients who had good tolerance and control in level 2. Regarding local control, the lines of chemotherapy can also participate, with a possible bias on the evaluation of control provided by the CRT. However, local control was superior to distant control, and the two treatment modalities are involved in the objective of surgical abstention in this population. Another limitation is the assessment of patients. The original schedule has not been applied in full. Indeed, an amendment has been made to lighten the monitoring schedule which was heavy in these metastatic patients treated by chemotherapy. Similarly, not all quality of life questionnaires were completed during the entire follow-up.
In the litterature, the principle of increasing the radiation dose in rectal cancer was already discussed in three-dimensional conformal radiotherapy (3D-CRT) 13 .
The development of IMRT in recent years has facilitated its development through the SIB 14 . An additional dose delivered by SIB-IMRT, for an equivalent total dose, appears less toxic than a complement by 3D-CRT, but without increasing the local response rate 15 .
Dose escalation appears to increase local response rate and pathologic complete response 11 . The dose escalation with a SIB-IMRT technique therefore seems promising on local control, while maintaining a good tolerance 16 . However, its bene t on the increase of the pathologic complete response is not yet provided. Delivering chemotherapy before the CRT has several interests. In our population of metastatic patients, it provides a global and rapid control of metastatic disease. It also has an effect on the local disease with a partial local response on the assessment before CRT in our population. Neo-adjuvant chemotherapy combined with CRT may increase the local response, also in non-metastatic patients. The study by Garcia-Aguilar et al. found an increase in the rate of pathologic complete response in locally advanced tumors to 18%, 25%, 30% and 38% by completing the CRT with zero, two, four, and six cycles of mFOLFOX6, respectively 18 .
In some patients, the tumor response may be longer after the CRT 10 . The study by Sloothaak et al. showed that out of 1,593 patients managed by preoperative CRT, the pathological complete response rate was maximal when surgery was performed at 14 weeks from the start of CRT, with a maximum plateau appearing to be reached at 17 weeks 19 . Shift surgery at 15 or 16 weeks after the start of the CRT (ie 10-11 weeks after the end of CRT) could increase the pathologic complete response rate. In a perspective of preservation, also in non-metastatic patients, this is an argument to wait for a complete clinical response up to 17 weeks from the beginning of the CRT before considering mutilating surgery. This study received a favorable opinion from the committee for the protection of persons (CPP) (17 February 2014) and was authorized by the national agency for the safety of medicines (ANSM) ( rst April 2014). An independent committee has been appointed. All patients received from oral and written Symptoms according to QLQ-C30 The rst assessment was performed at inclusion, the second after RT-CT.

Figure 5
Symptoms according to QLQ-CR29 The rst assessment was performed at inclusion, the second after chemoradiotherapy.