Preoperative intensity-modulated chemoradiotherapy with simultaneous integrated boost in rectal cancer: five-year follow-up results of a phase II study

In recent years, many different treatment strategies have been tested to improve outcomes for patients with locally advanced rectal cancer, with toxicity being the main obstacle for intensification of the standard treatment.1, 2, 3, 4 Changing the preoperative radiotherapy (RT) technique from 3D conformal (3D CRT) to intensity-modulated radiotherapy (IMRT) allowed better sparing of normal tissue in dosimetric analyses5, 6, 7, 8 and was used in several phase II studies to achieve dose escalation with simultaneous integrated boost (SIB) with or without another drug in addition to standard concomitant capecitabine.9, 10, 11, 12, 13, 14 The published reports showed encouraging results for pathologic complete response (pCR) and local control (LC)9,11, but with no impact on gastrointestinal toxicity with the addition of oxaliplatin15 and important late toxicity with dose escalation.10

Due to the promising impact on clinical outcome, but, conflicting toxicity results of treatment intensification with IMRT-SIB dose escalation in preoperative locally advanced rectal cancer (LARC) treatment, we conducted a prospective phase II study, where we kept the biologically effective dose (BED) of experimental IMRT SIB fractionation similar to the standard 3D CRT protocol of 45 Gy elective dose and boost of 4.5 Gy to T3 and 9 Gy to T4 tumour.

Our previously published results have shown that radiochemotherapy with IMRT-SIB without dose escalation, concomitantly with capecitabine, achieved a high rate of pCR (25.5%) and downstaging rate, with favourable acute toxicity profile and excellent compliance.16 In this paper, we report LC, disease-free survival (DFS), overall survival (OS), late toxicity and quality of life (QoL) after median follow-up of 70 months.

Between January 2014 and January 2015, 51 (N = 51) patients were included. The patients’ characteristics are described in detail elsewhere16, but briefly – median age was 66 years (range: 33–81 years) and nearly half of the tumours were located in the lower third of the rectum. The tumour invaded the mesorectal fascia in 20 patients and 15 patients had suspicious extramesorectal lymph nodes on MRI. Clinical stages were: T2N1M0 (n = 1), T3N0M0 (n = 6), T3N1M0 (n = 15), T3N2M0 (n = 22), T4N1M0 (n = 4), T4N2M0 (n = 2), and T3N1M118, with small lung lesion prior to inclusion revealed as lung metastasis on control CT after the treatment in the last patient.

Preoperative radiochemotherapy was completed by 50 patients and 1 received preoperative short-course radiotherapy due to ischaemic stroke.

Altogether, 48 patients underwent surgery (47 treated per protocol). In 3 patients, surgery was omitted due to patient refusal, synchronous pancreatic cancer and rectal varices haemorrhage. Low anterior resection (LAR) was performed in 40 patients, abdominoperineal resection (APE) in 7, and pelvic exenteration in 1. One patient had a positive circumferential margin. Extramesorectal nodes exploration was based upon surgeon discretion and nodes were removed in 4 patients.

The total downstaging rate was 87% (41/47 patients), with a decrease in T and N stage observed in 32 and 39 patients, respectively. Pathologic complete response was achieved in 12 patients.

In median follow-up of 70 months (range 11–80 m) we recorded 13 deaths, 7 due to rectal cancer. Among the 6 remaining patients, 3 died of cardiovascular disease and one each of pancreatic cancer, alcohol hepatic cirrhosis and grade 5 (G5) ileus. One isolated local relapse and 1 with synchronous distant metastasis occurred 41 and 42 months after LAR and APE, respectively. Time to distant relapses was 0 m and 18 m (lung), 6 and 11 m (liver), 41 m (adrenal gland) and 42 m (abdominal lymph nodes). At the latest date of follow-up on 30.1.2021, there were 37 patients alive without disease and one patient on systemic treatment for disseminated rectal cancer.

Survival

We performed an intention-to-treat analysis for all 51 patients and for 47 patients that were treated according to protocol (Table 1). For the entire cohort, cumulative 5-year OS, 5-year DFS and 5-year LC were 76.5% (95% CI 64.9–88.1), 72.4% (95% CI 60.4– 84.6), and 89.7% (95% CI 81.1–98.3), respectively. In the treatment per protocol group 5-year OS, 5-year DFS and 5-year LC were 80.9% (95% CI 69.7–92.1), 77.1% (95% CI 65.1–89.1) and 95.2% (95% CI 88.7– 100), respectively. Five-year colostomy-free survival was 76% (29/38).

Table 1

Number of events after median follow-up of 70 months (11–80 m) and 5-year survival

	Intention to treat (N = 51)	Per protocol (N = 47)
5-year OS*	76.5%	80.9%
5-year DFS	72.5%	76.5%
5-year LC	90.2%	95.7%
	Number of events (%)	Number of events (%)
OS status
Alive	38 (74.5)	37 (78.7) *
Dead	13 (25.5)	10 (21.3)
DFS status
Alive without disease	37 (72.5)	36 (76.5)
Local/distant relapse/death	14 (27.5)	11 (23.5)
LC status
Local relapse -	46 (90.2)	45 (95.7)
Local relapse +	5 (9.8)	2 (4.3)

* = Numbers differ from OS status due to one noncancer death > 5-year after surgery; DFS = disease-free survival; LC = local control; OS = overall survival; for non-operated patients and patient with M+ disease local or distant recurrence was calculated as 0 months.

The potential influence of prognostic factors on survival was determined by means of the log-rank test (Table 2). There was no association between age at diagnosis, performance status, tumour grade, positive mesorectal fascia or suspicious extramesorectal lymph nodes, removal of ekstramesorectal lymph nodes, clinical stage (cT, cN), decrease in T and N stage or pathologic T stage on survival. We found no predictive value for pCR, TRG prognostic group or NAR prognostic group as survival surrogates. Significantly better DFS was found in patients where total downstaging was achieved and in patients with pathologically negative lymph nodes (Figure 1). There was significantly better OS and DFS with LAR compared with APE or pelvic exenteration (p = 0.000 and 0.013, respectively). Gender was a predictive prognostic factor for OS and treatment per protocol was associated with better OS and DFS.

Table 2

Influence of probable prognostic factors on OS and DFS

	Intention to treat (N = 51)		Per protocol (N = 47)
Prognostic factor	OS	DFS	OS	DFS
Age at diagnosis (≥ 65 years vs. <65 years)	ns	ns	ns	ns
Gender (male vs. female)	p = 0.044	ns	p = 0.064	ns
PS WHO	ns	ns	ns	ns
Tumour grade	ns	ns	ns	ns
Tumour location (upper/ middle/lower rectal third)	ns	ns	ns	ns
MRI +	ns	ns	ns	ns
Extramesorectal lymph nodes (positive/negative)	ns	ns	ns	ns
Time to treatment(≤ 7w / > 7w)	p = 0.045	ns	ns	ns
Surgery procedure (APE and pelvic exenteration/LAR)	p = 0.000	p = 0.013	p = 0.020	p = 0.016
cT stagea	ns	ns	ns	ns
cN stagea	ns	ns	ns	ns
Decrease in T stage	ns	ns	ns	ns
Decrease in N stage	ns	ns	ns	ns
Total downstaging	ns	p = 0.029	ns	p = 0.029
pT stage (0-2 vs. 3-4)	ns	ns	ns	ns
pN stage (0 vs. +)	ns	p = 0.044	ns	p = 0.019
Ekstramesorectal lymp node removal	ns	ns	ns	ns
pCR	ns	ns	ns	ns
TRG prognostic group	ns	ns	ns	ns
NAR prognostic group	ns	ns	ns	ns
Adjuvant chemotherapyb (5-6 / ≤ 4 cycles)	ns	ns	ns	ns
Treatment per protocol	p = 0.006	p = 0.001	/	/

^a according to AJCC, 7th edition18; ^b calculated for 36 patients with indication for adjuvant chemotherapy; APE = abdominoperineal excision; DFS = disease free survival; LAR = low anterior resection; MRI+ = positive mesorectal fascia; N = nodal; NAR = neoadjuvant rectal cancer score32; ns = not specific (p > 0.05). OS = overall survival; pCR = pathologic complete response; PS WHO = WHO performance status; T = tumour; TRG = tumour regression grade19

Figure 1

We recorded no local or distant relapses in the group of patients with pCR, with one death due to G5 adverse event, leading to 91.7% 5-year OS and 100% 5-year DFS and 5-year LC for this group of patients.

Late toxicity

Late toxicity data are available for all patients and are listed in Table 3. Patients reported mean 5 late adverse events (range 0–19) at last follow up. Major adverse events (CTCAE version 5.0 G ≥ 3)20 occurred in 12 patients (23%), of which 7 patients had one, 3 patients two and 1 patient had three G3 toxicities. Following protective stoma closure, 1 patient died due to G5 ileus complication 29 months after LAR. Gastrointestinal toxicity (GI) ≥ 3 was recorded in 7 (15%) and genitourinary (GU) in 5 (10%) patients. Two men have erectile dysfunction and two women are reporting problems due to dyspareunia, vaginal dryness and vaginal stricture. Due to complete faecal incontinence, permanent stoma was required in two patients, 10 and 36 months after LAR. Urgent surgical intervention was required for anastomotic dehiscence and hernia incarceration in one case where the patient later developed enterocutaneous fistula. In the remaining two patients with anastomotic dehiscence, protective stoma closure was omitted in one patient and permanent stoma was placed 23 months after LAR in the other patient. Permanent stoma placement was also required due to rectoprostatic fistula in one patient 36m after LAR. Altogether, 6 patients with sphincter-preserving surgery had stoma closure omitted or later placed as permanent due to late toxicity (faecal incontinence, anastomotic dehiscence and fistula). The last recorded serious adverse event possibly related to treatment was recorded after 60m of follow up in a patient with bladder carcinoma (Figure 2).

Table 3

Late toxicity after preoperative radiochemotherapy, surgery and adjuvant chemotherapy according to CTCAE version 5.020

	G1, n (%)	G2, n (%)	G3, n (%)	G4, n (%)	G5, n (%)
Anastomotic dehiscence	-	1 (2.1)	3 (6.3)	-	-
Anastomotic stenosis	4 (10.0)	-	-	-	-
Ileus	-	-	-	-	1 (2.1)
Hernia	4 (8.3)	1(2.1)	1(2.1)	-	-
Abdominal or pelvic pain	11 (22.9)	3 (6.3)	-	-	-
Anal stenosis	5 (10.4)	-	-	-	-
Fistula	-	1 (2.1)	2 (4.2)	-	-
Bloating	21 (43.8)	3 (6.3)	-	-	-
Constipation	10 (20.8)	4 (8.3)	-	-	-
Diarrhoea	9 (18.8)	5 (10.4)	-	-	-
Faecal incontinence	6 (15.4)	12 (30.8)	3 (7.7)	-	-
Faecal urgency*	5 (13.2)	1(2.6)	-	-	-
Flatulence	25 (52.1)	6 (12.5)	-	-	-
Haemorrhoidal haemorrhage	1 (2.1)	-	-	-	-
Haemorrhoids	3 (6.3)	-	-	-	-
Proctitis	1 (2.1)	-	-	-	-
Intestinal stoma leak	2 (8.3)	-	-	-	-
Dysuria	1 (2.1)	-	-	-	-
Urinary frequency	13 (27.1)	-	-	-	-
Urinary incontinence	9 (18.8)	4 (8.3)	1 (2.1)	-	-
Urinary retention	1 (2.1)	1 (2.1)	-	-	-
Urinary urgency	21 (43.8)	1 (2.1)	-	-	-
Ejaculation disorder (n = 20)	5 (25)	1 (5)	-	-	-
Erectile disfunction (n=20)	2 (10.0)	6 (30.0)	2 (10.0)	-	-
Dyspareunia (n=18)	1 (5.6)	2 (11.1)	1 (5.6)	-	-
Vaginal dryness (n = 18)	1 (5.6)	3 (16.7)	1 (5.6)	-	-
Vaginal stricture (n = 18)	1 (5.6)	1 (5.6)	1 (5.6)	-	-
Treatment-related secondary malignancy	-	-	1 (2.0)	-	-
Other**	-	-	-	-	-

* = data not available for all patients; ** = other: anal pain, anal, rectal or colonic haemorrhage, d, anal ulcer, rectal obstruction or stenosis

CTCAE = Common Terminology Criteria for Adverse Events version 5.0.; G = grade

Figure 2

Preoperative use of IMRT for LARC is increasing rapidly with great variations in clinical practice among centres24 indicating a lack of quality clinical studies reporting treatment outcome and toxicity for different fractionation regimes. To date, only four prospective phase II studies have been published with short-term outcome data after preoperative IMRT concomitant with capecitabine in patients with LARC9,16,25,26, but none of them reported long-term results. Our study is the first to report a 5-year treatment outcome with late toxicity and QoL.

With shorter treatment time and no dose escalation with SIB to primary tumour only, our pCR rate improved from 10% to 25.5%.16,27 With dose escalation in a Chinese study (41 Gy elective; 56 Gy tumour/lymph node; 22 fractions) and a Spanish study (46 Gy elective; 57.5 Gy tumour/lymph nodes; 23 fraction) they reported 31% and 30.6% pCR, respectively.9,11 The higher pCR rate can reflect higher BED in these trials, but pCR of 32.6% recently reported by Simson et. al. in a prospective observational study with single target dose of 50.4 Gy in 28 fractions suggests other possible factors influencing treatment results, since there are important differences in target definition and treatment verification between studies.28 With no boost to pathologic lymph nodes, but with detailed contouring guidelines, added internal safety margin and image-guided radiation therapy (IGRT), we have achieved an equal or better N downstaging rate (83%) in comparison to the Chinese and Spanish trials (79.2% and 47.2%, respectively) where a 5 mm uniform margin was used.

The encouraging total downstaging rate of 87% in our study translated into excellent 5-year OS, DFS and LC of 80.9%, 77.1% and 95.2%, respectively. OS and DFS were significantly higher compared to our historic cohort29 (OS 61.4%, p = 0.03 and DFS 52.4%, p = 0.01). Studies with more intensified escalated IMRT regimes with added concurrent oxaliplatin12,30, are reporting 5-year OS of 63–82% and 5-year DFS of 60–66%, that are comparable to our study.

A reported death from cardiovascular disease in three male patients can probably explain why male gender is significant prognostic factor for OS but not for DFS. Surprisingly, the type of operation significantly affected OS and DFS. In concordance with other studies, we found significant association between total downstaging and pN0 with the improvement of DFS31, but we found no predictive value for pCR, TRG prognostic group19 or NAR prognostic group32 as survival surrogates. However, we found excellent prognosis for the group of patients with pCR who had 91.7% 5-year OS and 100% 5-year DFS and 100% 5-year LC, confirming observations from other studies.33

Previously, we reported low acute toxicity of preoperative treatment and postoperative complications (G ≥ 3 of 4% and 8% respectively). All patients with LAR had protective stoma placement, so we recorded no anastomotic leakage, but with time, 4 (8%) cases of anastomotic dehiscence were detected, as expected from the published literature.34 In median follow-up of 70 months, we recorded 17 G ≥ 3 late adverse events in 12 patients (25% of patients who underwent surgery), with 15% of GI, 2% urinary and 8% sexual late toxicity, significantly lower than to our historic cohort (40%, 19.2% and 51.7%, respectively)29 and less than 35% reported late surgical complications after 3D CRT concomitant with 5-Fu/oxaliplatin.35 The only comparable IMRT study reporting late adverse events is a Belgian study, with preoperative IMRT-SIB to 46/55.2 Gy in 23 fractions and median 54 months of follow-up.10 Their estimated G ≥ 3 late toxicity was 13% and is lower than the 25% observed in our study with much lower GI and GU toxicity rates (9% vs. 15% and 4% vs. 10%, respectively) compared to ours. With different recording of late effects in our study, the late toxicity could have been overestimated. Anastomotic dehiscence was discovered late, when protective stoma closure was planned and was not counted as a postoperative complication. Late events were recorded with the actuarial method, so faecal incontinence, although not present at the time of the last follow-up, occurred previously in two patients. Also, we had no data on sexual activity and GI disorder prior to treatment, so all GI events are counted as late sequelae, although no difference in sexual functioning before and 5 years after treatment was found on QoL analysis. Exclusion of anastomotic dehiscence and two cases of faecal incontinence decreases our rate of GI G ≥ 3 toxicity to 8%, which is in concordance with 9% in the Belgian study and 9% reported after 3D CRT.34 The occurrence of G ≥ 2 late diarrhoea (10%) is also within the range of reported rates in the literature, with 9.5% of patients from the EORTC 22921 trial who reported grade 2 diarrhoea or higher after 5-year follow-up.34 The Spanish and Chinese investigators reported 2 and 4 fistulas in median follow-up at 17 m and 22 m. The occurrence of fistula in our study is similar, with 3 fistulas that occurred 29, 34 and 36 months after surgery, showing the importance of longer follow-up and subsequently underreporting of long toxicity events in clinical studies. Since the first publication with a reported sphincter preservation rate of 62% in our series, 6 patients with LAR ended up with permanent stoma due to faecal incontinence, anastomotic dehiscence and fistula, but still, we report a high rate of 5 y colostomy-free survival of 76%, comparable to other studies.35 Regarding major toxicity, 1 patient (2%) died due to treatment-related toxicity, consistent with a 1.4% and 2% death rate in the EORTC 22921 and German CAO/ARO/ AIO-94 trial, respectively.34,36

Our patient cohort have significantly lower quality of life compared to the general Slovenian population37 according to EORTC QLQ-C30 scores comparison, with inferior global function and functional mean scales with problems with fatigue, constipation, diarrhoea and finance. However, comparison to EORTC reference scores38 for CRC patients shows no difference in any of the QLQ-C30 items with borderline lower cognitive function, reflecting advanced patient age (median 75 years) at 5-year data collection. Time analysis of EORTC QLQ-C30 and QLQ-C29 median scores showed that improvement or deterioration of function/ symptom appeared one year after treatment and remained stable with longer follow-up. We recorded improvement in body image and a drop of anxiety and as expected after surgery, there was less mucus/blood in stool, and after radiation patients reported hair loss. A significant rise in reported faecal incontinence/leakage, is in concordance with the reported late toxicity but late detection of urinary incontinence (significant after 5 but not 1 year) indicates the importance of long follow-up for reliable reporting of late toxicity. Contrary to 8% sexual late toxicity findings, there were no differences in sexual end points in the QoL analysis, reflecting the possibility of overestimation of these late events in our cohort.

Together with uncertainty in the reporting of late toxicity, the main limitation of our study design is the small sample size and lack of randomization. According to definition of “locally advanced rectal cancers” before new subgroup division consensus in 2013, we used the term locally advanced also for the intermediate/bad group, without additional subdivision of T3 tumours. Nevertheless, our results are comparable to above mentioned studies, for they used the same definition in that time. The advantages of our study are the very strict radiotherapy protocol and quality control with the precise recording of acute and late toxicity events. In comparison to other IMRT studies for preoperative LARC, we were the only one not to intensify treatment with dose escalation and/or novel drug addition.

By shortening the overall treatment time using SIB, we reported excellent 25.5% pCR and after 5-year follow-up, our OS and DFS (80.9% and 77.1%, respectively) are in the survival range of more intensified treatments12,30, suggesting possible overtreatment for certain patients with LARC. In the era of high local control, more effort should be directed to reducing acute and late toxicity. Our fractionation regime showed a very low acute toxicity profile with a non-negligible late events rate. More high-quality data with longer follow-up is needed to determine the true effect on QoL and possibly determine relevant tolerances of the organ at risk for late consequences to optimize treatment planning.

The results of this long-term study confirm that IMRT SIB is feasible for preoperative treatment of intermediate/locally advanced rectal cancer. By shortening the overall treatment time, without dose escalation, we achieved high pCR, five-year overall survival, disease-free survival and local control. Due to the favourable acute toxicity profile, our treatment regime is suitable for treatment intensification with another drug in addition to capecitabine. More long-term data is needed for late toxicity assessment.